Treatment of Personality Disorders

TREATMENT OF
PERSONALITY DISORDERS
TREATMENT OF
Edited by
Jan Derksen
University of Nijmegen
Nijmegen, The Netherlands and
Free University of Brussels
Brussels, Belgium
Cesare Maffei
Vita-Salute San Raffaele University
and Scientific Institute San Raffaele
Milan, Italy
and
Herman Groen
Forensic Psychiatric Institute "Oldenkotte"
Rekker, The Netherlands
Springer Science+Business Media, LLC

Library of Congress Cataloging·in·PubHcation Data
Treatment of personality disorders/edited by Jan Derksen, Cesare Maffei, and Herman Groen.
p. cm.
"Including contributions from the First European Congress on Disorders of Personality,
held June 14-17, 1994, Nijmegen, The Netherlands, and the Second European Congress
on Personality Disorders, held June 26-29, 1996, Milan, Italy"-T.p. verso.
Includes bibliographical references and index.
ISBN 978-1-4419-3326-3 ISBN 978-1-4757-6876-3 (eBook)
DOI 10.1007/978-1-4757-6876-3
1. Personality disorders. I. Derksen, Jan, 1953- II. Maffei, Cesare, 1951- III. Groen,
Herman.
RC554.T72 1999
616.85'8-dc21
99-047601
Including contributions from the First European Congress on Disorders of Personality, held June 14-17, 1994,
Nijmegen, The Netherlands, and the Second European Congress on Personality Disorders, held June 26-29,
1996, Milan, Italy
ISBN 978-1-4419-3326·3
© 1999 Springer Science+Business Media New York

Originally published by Kluwer AcademicIPlenum Publishers, New York in 1999
http://www.wkap.nl
10987654321
A C.I.P. record for this book is available from the Library of Congress
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission
from the Publisher
PREFACE
It has been almost twenty years since DSM-III created a major shift in psychi-
atric classification procedures and in diagnostic and treatment practice by introducing
the multi-axial system and, for our patients specifically, the Axis II: Personality
Disorders.
Researchers and clinicians were forced to focus on many issues related to the field
of personality and its disorders. This meant an immense impetus for research, both
empirical and theoretical. Many recent developments are described in this book, as
reviews or as original articles. This book also covers developments in Europe as well
as in North America.
Important questions still remain unanswered, such as: What is the relationship
between the different clusters: A, B, & C? Are we talking about dimensions, categories,
or typologies? What can be done for patients who have more than one personality
disorder? Is a pro typical approach required? Consequently, is a multiconceptual
approach in treatment and research required? The authors contribute to this discus-
sion and provide guidelines for further thinking in research and treatment planning.
For clinicians, it is of major importance to know whether the disorder can be influ-
enced by treatment, and whether permanent change is really possible. A very impor-
tant question is whether a person indeed has a personality disorder, and how this
diagnosis affects clinical practice.
We have to differentiate between the genotypical and phenotypical, between bio-
logical and psychological, nature and nurture, trait and state. Differentiation implies
the use of diagnostic procedures, which as you will find, are promising. It also implies
different treatment strategies, depending on the phase of treatment for the same dis-
order, and the orientation of the therapist.
Accurate diagnoses can allow us to distinguish between different treatment
strategies for various disorders. Within the framework of a DSM personality disorder,
further differentiation then becomes possible. For example, good diagnostic procedures
will allow distinction between an impulsive and a dissociative borderline personality
disorder. Research is suggesting the first might better be influenced by biological!
psychopharmacological treatment, and the latter by focusing in a cognitive or psycho-
dynamic way on early childhood (sexual) trauma. It is also important to find out
whether the symptomatology we meet in clinical practice is caused by the personality
disorder itself, (an Axis I disorder) or whether this is the result of the interaction
between the two, complicated by problems and dysfunctioning on Axes IV and V.
v
vi Preface
Once we have established an accurate diagnosis, we can offer a range of treat-

ments, simultaneously or separately: psychopharmacological treatment together with
psychoanalytical and/or cognitive treatment.
Given the state of the art of the psychobiological foundations involved, we
analyze genetics, with a view to finding out how the carriers of our existence influence
our being. TIle genetic loading can explain some, but not all of the characteristics influ-
encing personality disorders. This has implications for biological research, especially
when foeuse d on treatment and in examining the psychobiology of personality disor-
ders. If we accept that phenotypical characteristics represent the same genetic pattern
as influenced by its environment, and if we accept that the relevant environment for
human species is not only biological but also psychological, then psychobiology gets
a far more existential meaning and therefore is of great interest to clinicians and
researchers interested in personality disorders from all perspectives.
Within this genotypical/phenotypical perspective, considerations on the psycho-
logical foundations of development and psychodynamic research will become dynamic
in itself. Attachment and the development of the self, vis-a-vis the disorders of per-
sonality, is of great importance, as is the occurrence of psychological, sexual, and phys-
ical trauma in early childhood. These issues are of clinical importance because human
beings grow through childhood and adolescence and beyond. As is shown by PET-
scanning even the brain can change functionally through psychological treatment. So
we are not a victim of our genes and the phenotypical outlook created in childhood.
Even with animals, this is not the case.
It is important for researchers and clinicians to decide on which position to take.
This position should be clear so that patients are aware of how their therapist views
their disorcer, and what outcomes for treatment are perceived. Consequently, our
therapeutic acts are affected by this position. Categorical thinking almost always intro-
duces the handicap model; dimensional and typological thinking creates room for opti-
mism with possible change from severe to less severe.
We have created an empirical and theoretical framework, aware that we were not
primarily interested in conceptual answers to questions such as: Is a personality disor-
der defined by 5, 9, 12 ... factors? Are personality disorders within the continuum of
Axis I disorders and therefore, by definition an Axis 1 disorder? among others. We
created a framework for clinical practice, trying to give clinicians a colorful background
on which basis they might be better able to give their work more depth using innova-
tive and creative therapies for these complex patients and their psychopathologies.
We think that with more alternative approaches, treatment will be less frustrat-
ing. One should and cannot adhere to one ideology, especially not with the very serious,
rigid, personality-disordered patient. Such a patient does, with respect to his psy-
chopathology, everything possible to preserve his homeostasis; not doing so is fearful,
as well as dangerous because of the lack of other possibilities in feelings and behavior.
Finding a way, creating a safe, holding environment, requires flexibility.
There is no final advice, nor concluding remarks, because we think that the
theoretical orientation and the way the therapist is trained, is too diverse throughout
Europe as well as throughout North America. Instead, the framework that we have
created, consisting of elements/modules for each individual patient is to be used by
every individual therapist according to his own frame of reference.
Jan Derksen
Cesare Maffei
Herman Groen
CONTENTS
Part I: Biological Foundations of Personality Disorders
1. Genetics of Personality Disorders ................................. 1

Svenn Torgersen and Marco Battaglia
2. Psychobiology of Personality Disorders: Implications for the Clinic 17

Larry 1. Siever
3. The Implications of Recent Research on the Etiology and Stability of

Personality and Personality Disorder for Treatment ............. 25
W. John Livesley
Part II: Psychological Foundations of Personality Disorders
4. Psychodynamic Research Can Help Us to Improve Diagnosis and

Therapy for Personality Disorders: The Case of Defense
Mechanisms ................................................ ' 39
1. Christopher Perry, Vittorio Lingiardi, and Floriana Ianni
5. Attachment, the Development of the Self, and Its Pathology in

Personality Disorders ....................................... 53
Peter Fonagy
6. Trauma and Personality .......................................... 69

Fabio Madeddu and Adolfo Pazzagli
7. Adolescence and Personality Disorders: Current Perspectives on a

Controversial Problem ...................................... 77
Enrico de Vito, Fran~ois Ladame, and Alvise Orlandini
Part III: Diagnostic Models of Personality Disorders
8. Integrative Perspectives on the Personality Disorders 97

Theodore Millon
vii
viii Contents
9. A Multidimensional Approach to Personality Disorders and Their

Treatment ................................................. 107
Joel Paris
10. The Structure of DSM-IV Borderline Personality Disorder and Its

Implications for Treatment ................................... 119
Cesare Maffei and Andrea Fossati
11. The Relationship between Anxiety Disorders and Personality Disorders:

Prevalence Rates and Comorbidity Models ..................... 129
Carol 1. M. Van Velzen and Paul M. G. Emmelkamp
Part IV: Treatment of Personality Disorders
12. Psychodiagnostics and Indications for Treatment in Cases of Personality

Disorder: Some Pitfalls ...................................... 155
1. Derksen and H. Sloore
13. The Psychotherapeutic Treatment of Borderline Patients 167

Otto F. Kernberg
14. Functional Analysis of Borderline Personality Disorder Behavioral

Criterion Patterns: Links to Treatment ......................... 183
Jennifer Waltz and Marsha M. Linehan
15. Psychopharmacological Treatment of Personality Disorders: A Review 207

Peter Moleman, Karin van Dam, and Veron Dings
16. New Drugs in the Treatment of Borderline Personality Disorder ....... 229
F. Benedetti, C. Colombo, L. Sforzini, C. Maffei, and E. Smeraldi
17. The Narcissistic Personality Disorder and Addiction. . . . . . . . . . . . . . . . . . 241

PerVaglum
18. Pharmacotherapy for Patients with Personality Disorders: Experiences

from a Group Analytic Treatment Program . . . . . . . . . . . . . . . . . . . . . 255
S. Friis, T. Wilberg, T. Dammen, and 0. Urnes
19. A New Interpersonal Theory and the Treatment of Dependent

Personality Disorder ........................................ 269
John Birtchnell and Giuseppe Borgherini
20. HIV Infection, Personality Structure, and Psychotherapeutic

Treatment ................................................. 289
R. Visintini, E. Campanini, A. Ama, R. Alcorn, S. Corbella, S. Gessler,
D. Miller, L. Nilsson Schonnesson, and F. Staracel
Index 305
1
GENETICS OF PERSONALITY DISORDERS
Svenn Torgersen 1 and Marco Battaglia2
1 University of Oslo
Department of Psychology
Blindern Oslo Norway
2 University of Milano and the San Raffaele Hospital
Department of Neuropsychiatric Sciences

29 via Prinetti Milano Italy
1. INTRODUCTION
When we wonder about whether genetic factors influence the development of a

disorder, we may think about necessary and sufficient conditions for the manifestation
of the disorder.
- Do we need to have some genes to get the disorder, or is it possible to develop
the disorder without the genes?
- Are the genes the full explanation, or do some non-random environmental
factors also have to be present?
- Are genetic factors completely irrelevant, the whole explanation being envi-
ronmental?
- Do some genes increase the likelihood of developing the disorder, without
being either sufficient or necessary?
Basically, the genetic influence on the development of a disorder is of two kinds:
1) One or more genes exert an independent influence on the development of
the disorder. This is called non-additive inheritance. The genes may be dom-
inant or recessive. If the genes are dominant, a critical allele from either the
mother or the father is sufficient. Recessive genetic transmission claims that
both the allele from the mother and from the father needs to carry the seed
of the disorder. It can be generally stated that while rare, mendelian disor-
ders whose inheritance is based on a single gene have traditionally been the
focus of medical genetics, psychiatric genetics, and behavioral genetics are
examining much more common illnesses and behaviors in which an inherited
Treatment of Personality Disorders, edited by Derksen et al.
Kluwer Academic / Plenum Publishers. Ncw York, 1999. 1
2 S. Torgersen and M. Battaglia
liability-or vulnerability-appears to interact with environmental factors in

a complex way, so that the methods of psychiatric genetics and behavioral
genetics are usually closer to those of genetic epidemiology.
2) A number of genes, being recessive or dominant, are adding up. The more
effective genes a person has, the higher liability for the disorder. This is
the basic model for additive inheritance. A multifactorial additive influence
means that a number of physical, psychological, and social factors contribute
in an additional fashion. The implication may be that the more traumatic the
environment, the fewer critical alleles are necessary for the development of
the disorder (Goldsmith and Gottesman, 1996). In principle, it is possible to
develop the disorder even when all at environmental conditions have been
favourable if the genetic load is massive. Conversely, a certain amount of trau-
matic experiences, physical, psychological or social, may imply a development
of the disorder even if the genetic basis is weak or absent. Given an average
genetic make-up, a certain amount of negative experiences are necessary. On
the other hand, with an ordinary environment, a lot of critical alleles are nec-
essary for the disorder to develop.
In general, it should be remembered that when data show a continuous distribu-

tion for a certain trait (a given behavior in our case) in the popUlation, then the trait
is likely to be influenced by several genes, each having a small effect, and environmental
factors, whereas non additive mechanisms are more likely to be involved when a trait
distribution tends to be bimodal, instead of normal (Plomin, DeFries, and McLean,
1990).
Even though understanding precisely how multiple genes with small effects influ-
ence behavior can be a very complex task, we have to remember that each of such
many genes is nonetheless submitted to the laws of heredity first discovered by Mendel.
Moreover, with the advent of molecular genetics techniques it has become pos-
sible to precisely isolate, identify, and quantify the effect of even a single gene of rela-
tively small effect that is able to influence behavior and personality, provided that the
behavioral scales employed to describe the phenotype are reliable and accurately
shaped (Cloninger,Adolfsson, and Svrakic, 1996). Very recent advancements in the field
showing significant associations between some polymorphisms of genes implicated in
the mechanisms of neurotransmission and temperamental, and personologic variants
in the human (Ebstein et aI., 1996; Benjamin et aI., 1996; Lesch et aI., 1996) are inter-
esting examples of such possibilities.
Many complications to these simple descriptions exist. Even if a person has seem-
ingly sufficient and necessary genes, the disorder does not need to develop. Some
arbitrarily or unknown factors have to be there in addition. This is called reduced pen-
etrance. Sometimes also so-called phenocopies exist::a disorder phenomenologically
similar to a normally strongly genetically determined disorder may develop out of
purely environmental reasons. Another important and common aspect in behavioral
genetics is pleiotropism, or the multiple behavioral effects that may be exerted by a
single gene. An example in animal behavior is the finding that the higher emotional
reactivity of the albino mice is partially due to the same gene that influences the coat
color (DeFries and Hegmann, 1970).
More important-and even more complex-is heterogeneity. That means that the
disorder may have many independent causes. Sometimes one gene or one pattern of
non-additive genes are responsible for the development of the disorder, other times
Genetics of Personality Disorders 3
completely different genes, non-additive or additive. Sometimes genes are not involved
at all.
For the same disorders, maybe a combination of non-additive genes and multi-
factorial additive genes are necessary.
The possibility also exists that only in interaction with some environmental factors
do the genes have any effect. As we study the phenotypical result of an eventual geno-
type and environment, the interaction may be difficult to trace. In fact there is growing
consciousness that a core issue for the understanding of the etiology of mental
ilnnesses-including the personality disorders-is the complex interaction between
genes and environment. It may be too simplicistic to think that one's vulnerability to
a mental illness is merely the sum of the genetic plus the non-genetic risk factors: genes
and environment may well be correlated, in that there can be a genetic predisposition
to selecting some specific environments, given a certain genetic background (see e.g.
Kendler, 1995). In this perspective a simple additive model that predicts the indepen-
dence between genotype and the impact of non-genetic factors would be inadequate.
Indeed, recent findings in the field of genetic epidemiology show that the relationships
between genes and environment in shaping normal and abnormal personality variants
need a multivariate, complex approach. Two very important concepts are clearly emerg-
ing from research in the gene-environment interaction. First, there may be a genetic
influence in leading people to select high-risk environments: this is what is meant when
we say that genes and environment are correlated. Second, while in a relatively neutral
environment people can be relatively homogeneous in ther levels of psychological dis-
tress, when an environmental stressor becomes active, new genes that influence the indi-
vidual differences in coping with adverse conditions would "come on line" (Kendler,
1995), so that interindividual differences and symptomatological variability would
become more evident.
2. METHODS
The most straightforward way to prove the genetic influence on the development
of a disorder would be to identify the critical allele of a gene in individuals with a spe-
cific disorder. Up to now, the technique of molecular genetics has not developed to a
point where this can be the standard procedure. Moreover, the more genes are involved,
the more the task can be complex.
The linkage method is therefore an approximation. By studying the process of
recombination of alleles in families of more generations, one is not only able to prove
the genetic influence, but also to track down the location of the gene. The technique is
that once a part of the human genome is known, genes in the neighbourhood can be
localised by the fact that they follow each other through generations of recom-
binations. However, once more, this method is only realistic when few genes are of
importance.
The more realistic procedure is to apply the twin or the adoption method. The
twin method capitalises on the fact that monozygotic (MZ) twins are genetically iden-
tical, while dizygotic (DZ) twins are not more similar in their genetic make-up than
sibs generally.
When the adoption method is applied, the features of the adoptees are compared
to the characteristics of the biological relatives. The social relatives of the adoptees, or
the biological relatives of adoptees without these features, are controls.
The best method is to combine the twin and the adoption method. In this case,
the relatively rare pairs of twins reared apart are studied. This procedure makes it pos-
sible to separate the genes from the environment. MZ twins reared together have iden-
tical genes and very similar environment, dizygotic (DZ) twins reared together have
somewhat different genes and similar environment, MZ twins reared apart have iden-
tical genes and different environment while DZ twins reared apart have both some-
what different genes and different environment.
The twin and adoption methods also make it possible in addition to proving
genetic influence, to study the effect of shared-in-families environment, and not-shared-
in-families environment. The shared-in-families environment, or common family envi-
ronment, consists among others of social class, place of living, nutrition, the upbringing
methods applied on all the children. The non-shared, or unique, environment, comprises
the different ways the parents treat the different children and their individual experi-
ences not shared by sibs. Also, the less-than-perfect reliability of the methods will be
part of the non-shared environment.
Family studies can not either prove or disprove the genetic influence. The reason
is that family members share both genes and environment. An aggregation of disor-
ders in families may thus either be due to heritability or shared-in-families environ-
ment, or both. However, given that genetic influence is proven, family studies may
disclose whether the genetic transmission is additive or non-additive, dominant or
recessive, or sex-linked. Furthermore, an etiological relationship between different dis-
orders may be revealed through family studies. This is the basis for the formulation of
the concept of spectrum of common liability between disorders that may appear phe-
notypically separated at different extents.
3. COMMON PERSONALITY DIMENSIONS
Before proceeding to the personality dimensions, some words may be said about
the genetics of common personality dimensions.
The two broad dimensions of personality, neuroticism, and extraversion, have
been very popular since they were introduced in the scientific world by Eysenck
(Eysenck, 1967). Later he introduced a third dimension, psychoticism (Eysenck and
Eysenck, 1976) (The Big Three). Later McCrae and Costa (1987) promoted a five factor
model consisting of openness to experience, agreeableness, conscientiousness in addi-
tion to neuroticism, and extroversion (The Big Five). Agreeableness and conscien-
tiousness maybe considered as a spin-off of psychotiscism, with opposite signs. Tellegen
(1985), applies three concepts very close to these two models. Negative emotionality is
close to neuroticism, positive emotionality is similar to extroversion, while constrain is
not far from conscientiousness.
A number of twin and adoption studies have been performed with these dimen-
sions. Twin studies have yielded high heritability, while adoption studies, not surpris-
ingly, have given lower genetic estimates. Luckily, also some studies comprising both
twin pairs reared together and twin pairs reared apart have been conducted.
Table 1 shows the results from two projects, the Swedish adoption/twin study of
ageing (SATSA) (Pedersen et aI., Bergeman et aI., 1993) and the Minnesota study of
twins reared apart (MSTRA) (Tellegen et aI., 1988).
SATSA shows lower heritability and shared environment, and higher non-shared
environment, compared to MSTRA. The reason is very likely due to the lower relia-
Table 1. The genetic variance (G), the shared-in-families

environmental variance (ES), and the non-shared-in-families
environmental variance (EN) according to two studies of twins
reared apart
G ES EN
SATSA 1)
Neuroticism 0.31 (0.44) 0.10 (0.14) 0.58 (0.42)
Extraversion 0.41 (0.59) 0.07 (0.10) 0.52 (0.31)
Openness to experience 0.40 (0.57) 0.06 (0.09) 0.54 (0.34)
Agreeableness 0.12 (0.17) 0.21 (0.30) 0.57 (0.53)
Conscientiousness 0.29 (0.41) 0.11 (0.16) 0.60 (0.43)
MSTRA2)
Negative affectivity 0.55 (0.61) 0.02 (0.02) 0.43 (0.37)
Positive affectivity 0.40 (0.44) 0.22 (0.24) 0.38 (0.32)
Constraint 0.58 (0.64) 0.00 (0.00) 0.43 (0.42)
1) Adjusted from Pedersen et aI., 1991; Bergeman et aI., 1993.

2) Adjusted from Tellegen et aI., 1988.
The numbers in parentheses are corrected for reliability deficiency.
bility of the very short scales they applied. According to the information about relia-
bility in the publications, a correction was made, based on a reliability of 0.70 in SATSA
and 0.90 in MSTRA. The numbers in parenthesis are the corrected numbers. The
genetic variance G is a combination of additive and non-additive variance, as such split-
ting is little robust in limited samples.
As it appears, the heritability (G) is between DAD and 0.60 for neuroticism (neg-
ative emotionality), extraversion (positive emotionality), conscientiousness (con-
straint), and openness to experience. The shared-in-families environment (ES) is
around 0.10, and the non-shared environment (EN) is around DAD. Agreeableness,
however, seems to have a very low heritability, while the shared and non-shared expe-
rience variance is high. . .
To the extent that the personality disorders are extremes of these broad person-
ality dimensions, one would perhaps also expect that personality disorders have a rel-
atively high heritability. Some studies have examined whether personality disorders
share variance with personality dimensions.
Table 2 gives an overview of the relationship between The Big Five and person-
ality disorders measured by questionnaires. The range of and median correlations are
noted. We observe that The Big Five are highly correlated to one or more of the per-
sonality disorders measured by interview, mostly so for the antisocial, narcissistic, and
dependant personality disorders, and least for the paranoid and schizoid personality
disorders.
If we look at the relationship of the The Big Five and personality disorders mea-
sured by interview, we find more modest correlations. One reason may be the common
disclosure tendency variance in questionnaires which measures either broad personal-
ity dimensions or personality disorders. Another, not completely independent reason
may be the fact that questionnaires and interviews have different measurement errors.
A third possibility is that personality disorder questionnaires simply do not measure
anything else than the broad personality dimensions studied by questionnaires.
Anyway, it is to be expected that personality disorders measured by question-
naires also show heritability, just as personality dimensions.
Table 2a. The correlations between The Big Five and personality disorders measured
by questionnaires
N E 0 A C R
Paranoid
MMPI Costa & McCrae (1990) 0.36 -Q.02 -Q.09 -0.31 -0.13
MMPI Thull (1992) 0.46 -Q.06 -Q.04 -0.48 -0.05 0.61
MCMI-I Costa & McCrae (1990) -0.08 -Q.02 -0.04 -0.27 0.15
MCMI-II Costa & McCrae (1992) 0.04 0.24 0.12 -0.07 0.02
MCMI-II Soldz et al. (1993) 0.30 0.11 0.08 -Q.13 0.06 0.44
PDQ-R Trull (1992) 0.45 -Q.10 -0.06 -Q.35 0.16 059
Median 0.33 -Q.02 0.00 -0.29 0.04 059
Schizoid
MMPI Costa & McCrae (1990) 0.16 -Q.62 0.06 -0.12 0.14
MMPI Trull (1992) 0.00 -0.7.2 -Q.27 -0.19 0.02 0.73
MCMI-I Costa & McCrae (1990) 0.04 -0.64 -Q.08 -0.04 -0.07
MCMI-II Costa & McCrae (1990) -0.14 -Q.49 0.04 0.10 0.14
MCMI-II Soldz et al. (1993) 0.26 -0.69 -Q.32 -Q.19 0.08 0.69
PDQ-R Trull (1992) -0.17 -Q.34 -Q.45 -0.34 0.02 0.61
Median 0.00 -Q.62 -0.08 -0.12 0.02 0.69
Schizotypal
MMPI Costa & McCrae (1990) 0.46 -Q.48 0.00 -0.15 0.04
MMPI Trull (1992) 0.45 -0.49 -Q.21 -0.40 -0.12 0.66
MCMI-I Costa & McCrae (1990) 0.43 -0.46 -0.19 0.11 -0.14
MCMI-II Costa & McCrae (1990) 0.39 -0.34 -Q.07 0.06 0.01
MCMI-II Soldz et aI. (1993) 052 -Q.49 -0.30 -0.38 -0.14 059
PDQ-R 1hJ1l (1992) 0.41 0.06 0.09 -0.33 -0.22 0.49
Median 0.44 -Q.47 -0.13 -0.24 -Q.13 0.59
Antisocial
MMPI Costa & McCrae (1990) 0.13 0.07 0.18 -0.35 -0.42
MMPI Trull (1992) 0.29 -0.22 -0.02 -0.42 -0.27 050
MCMI-I Costa & McCrae (1990) -0.27 0.12 0.22 -0.49 0.17
MCMI-II Costa & McCrae (1990) 0.15 0.21 0.08 -0.42 -Q.40
MCMI-II Soldz et al. (1993) 0.25 0.27 0.19 -0.18 -0.22 052
PDQ-R Trull (1992) 0.25 -0.16 0.14 -0.38 -0.37 055
Median 0.20 0.10 0.16 -0.40 -0.32 052
Borderline
MMPI Costa & McCrae (1990) 0.47 0.19 0.09 -0.21 -0.32
MMPI Trull (1992) 0.61 0.13 0.18 -0.45 -0.24 0.75
MCMI-I Costa & McCrae (1990) 0.52 -0.22 -Q.22 0.14 -0.10
MCMI-II Costa & McCrae (1990) 0.46 -0.09 -0.16 -Q.22 -0.22
MCMI-II Soldz et al. (1993) 0.56 0.04 -Q.02 -0.26 -0.34 0.69
PDQ-R Trull (1992) 0.60 0.19 0.28 -0.39 -0.17 0.76
Median 0.54 0.09 0.04 -0.24 -0.23 0.75
Not all personality dimensions embodied in the different available questionnaires

are equally homogeneous and structurally simple from a genetic vantage point;
however. For instance, the components of impulsivity and sociability of Eysenck's extra-
version are genetically independent, and share environmental determinants (Eaves
and Eysenck, 1975). Temperamental dimensions (Novelty Seeking, Harm Avoidance,
Reward Dependence, and Persistence) embodied in Cloninger's Temperament and
Character Inventory (TCI-Cloninger et aI., 1994) have the noteworthy advantage of
having been designed to be genetically homogeneous and independent of each other:
Table 2b. The correlations between The Big Five and personality disorders
measured by questionnaires
N E 0 A C R
Histrionic
MMPI Costa & McCrae (1990) -0.17 0.65 0.15 0.00 -0.22
MMPI Trull (1992) -0.01 0.61 0.39 --0.14 --0.14 0.71
MCMI-I Costa & McCrae (1990) -0.28 0.47 0.27 --0.30 0.10
MCMI-II Costa & McCrae (1990) --0.22 0.42 0.17 --0.31 -0.24
MCMI-II Soldz et al. (1993) -0.06 0.67 0.37 0.20 -0.26 0.77
PDQ-R Trull (1992) 0.54 0.39 0.23 -0.19 -0.05 0.76
Median -0.12 0.54 0.25 -0.17 --0.18 0.76
Narcissistic
MMPI Costa & McCrae (1990) -0.28 0.56 0.07 -0.18 0.01
MMPI Trull (1992) -0.30 0.59 0.43 -0.06 0.20 0.73
MCMI-I Costa & McCrae (1990) -0.28 0.47 0.27 -0.30 0.10
MCMI-II Costa & McCrae (1990) -0.22 0.42 0.17 --0.31 --0.24
MCMI-II Soldz et al. (1993) 0.03 0.49 0.45 --0.05 --0.06 0.66
PDQ-R Trull (1992) 0.69 0.02 0.22 -0.39 -0.28 0.77
Median -0.25 0.48 0.25 -0.24 --0.03 0.73
Avoidant
MMPI Costa & McCrae (1990) 0.52 --0.54 --0.03 -0.02 --0.02
MMPI Trull (1992) 0.55 --0.63 --0.27 --0.16 --0.19 0.78
MCMI-I Costa & McCrae (1990) 0.44 --0.53 --0.11 0.03 --0.07
MCMI-II Costa & McCrae (1990) 0.36 --0.32 --0.11 0.05 0.03
MCMI-II Soldz et al. (1993) 0.54 -0.59 -0.30 -0.23 -0.08 0.66
PDQ-R Trull (1992) 0.36 -0.39 --0.20 -0.20 .:.0.21 0.51
Median 0.48 -0.54 --0.16 --0.09 -0.08 0.66
Dependent
MMPI Costa & McCrae (1990) 0.50 -0.30 -0.10 0.22 -0.22
MMPI Trull (1992) 0.64 --0.21 -0.01 -0.08 -0.33 0.67
MCMI-I Costa & McCrae (1990) 0.37 --0.06 --0.36 0.38 --0.10
MCMI-II Costa & McCrae (1990) 0.20 0.09 --0.26 0.34 --0.04
MCMI-II Soldz et al. (1993) 0.13 -0.23 --0.29 -0.13 --0.06 0.39
PDQ-R Trull (1992) 0.51 --0.04 0.14 0.02 -0.27 0.57
Median 0.44 --0.14 --0.18 0.12 --0.16 0.57
Obsessive-Compulsive
MMPI Costa & McCrae (1990) 0.50 0.16 --0.07 -0.15 --0.06
MMPI Trull (1992) 0.52 --0.29 0.03 -0.27 --0.14 0.58
MCMI-I Costa & McCrae (1990) --0.39 -0.09 --0.19 0.09 0.38
MCMI-II Costa & McCrae (1990) -0.05 --0.03 --0.11 0.15 0.52
MCMI-II Soldz et al. (1993) 0.02 --0.35 --0.12 --0.02 0.43 0.55
PDQ-R Trull (1992) 0.38 -0.05 0.26 -0.34 -0.12 0.54
Median 0.20 -0.07 -0.09 -0.09 0.16 0.55
this may explain why genetic investigation-both at the formal and at the molecular
level-with the temperamental dimensions of the TCI appear to provide more clearcut
results as compared with those obtained with the NED (Costa and McCrae, 1990)
dimensions or other scales (Stallings et aI., 1994) in the same samples (Ebstein et aI.,
1996; Cloninger et aI., 1996).
The temperamental dimensions of Novelty Seeking, Harm Avoidance, Reward
Dependence and Persistence also show good power to explain comorbidity of psychi-
atric disorders (both on Axis I and II), thus supporting the hypothesis that this complex
Table 2c. The correlations between The Big Five and Personality Disorders
measured by questionnaires
N E 0 A C R
Passive-Aggressive
MMPI Costa & McCrae (1990) 0.39 -0.17 -0.02 -0.16 -0.33
MMPI Trull (1992) 0.56 -0.19 0.11 -0.45 -0.37 0.68
MCMI-I Costa & McCrae (1990) 0.50 -0.07 0.12 -0.04 -0.17
MCMI-Il Costa & McCrae (1990) 0.53 0.D1 -0.14 -0.20 -0.23
MCMI-II Soldz et al. (1993) 0.54 -0.08 -0.08 -0.32 -0.22 0.57
PDQ-R Trull (1992) 0.47 -0.12 0.06 -0.30 -0.40 0.60
Median 0.52 -0.10 0.D2 --0.25 -0.28 0.60
Sadistic
MCMI-ll Costa & McCrae (1990) 0.02 0.07 -0.08 -0.46 -0.16
MCMI-II Soldz et al. (1993) 0.27 0.18 0.22 -0.28 -0.02 0.56
Median 0.15 0.13 0.07 -0.37 -0.09 0.56
Self-defeating
MCMI-U Costa & McCrae (1990) 0.45 -0.08 -0.17 0.05 -0.05
MCMI-U Soldz et al. (1993) 0.61 -0.33 -0.18 -0.26 -0.33 0.61
Median 0.53 -0.21 -0.18 -0.11 -0.19 0.61
Explanation to Table 2a, 2b, and 2c:
N: Neuroticism
E: Extraversion
0: Openness to experience
A: Agreeableness
C: Conscientiousness
phenomenon may be understood in terms of interactions between independently inher-

ited temperamental dimensions during development (Battaglia et aI., 1996).
4. PERSONALITY QUESTIONNAIRES
Livesley et ai. (1993) have measured by means of a questionnaire personality

deviances which are similar to the personality disorders in DSM-III. They observed
that most of their types of personality deviances showed heritability. The heritability,
when additive and non-additive are combined, ranged from 0.64 to 0 with a median of
0.49. Narcissism, Identity problems (analogous to borderline) and Social avoidance
(similar to avoidant/schizoid), Callousness (antisocial), and Oppositionality (passive-
aggressive) were most heritable. Conduct problems were not at all. Submissiveness
(dependent), Self harm (borderline ?), Insecure attachment (dependent), and Intimacy
problems (avoidant) showed a relatively low heritability. Correspondingly, Conduct
problems showed a very high shared environment variance, also Submissiveness had a
high such variance. The non-shared environment variance was highest for Self harm,
followed by Intimacy problems. As is the case for most twin studies, only seven of the
eighteen traits displayed any shared-in-families variance. These estimates are probably
too low, due to more similar environment for MZ twin partners compared to DZ twin
partners, and possibly gene/environment interaction.
The pattern of heritability is not following the PD's in the study of Livesley et ai.
(1993). Personality deviance close to borderline, antisocial, and avoidant show both
high and low heritability. The reason may be the heterogeneity of the PD's defined by
Table 3a. The correlations between The Big Five and Personality Disorders
measured by interviews
N E 0 A C R
Paranoid
SIDP-R Trull (1992) 0.24 -0.01 0.02 -0.44 0.05 0.49
PDE Soldz et al. (1993) 0.34 0.01 0.15 0.00 0.11 0.42
Mean 0.28 0.00 0.09 -0.22 0.08 0.46
Schizoid
SIDP-R Trull (1992) 0.05 -0.47 --0.37 -0.34 0.07 0.58
PDE Soldz et al. (1993) 0.15 -0.28 --0.14 -0.05 0.15 0.28
Mean 0.10 --0.38 -0.26 -0.20 0.11 0.43
Schizotypal
SIDP-R Trull (1992) 0.26 --0.30 -0.07 --0.35 -0.13 0.44
PDE Soldz et aI. (1993) 0.21 -0.25 --0.11 0.04 0.15 0.32
Mean 0.24 -0.28 --0.09 -0.16 0.01 0.38
Obsessive-Compulsive
SIDP-R Trull (1992) 0.29 -0.28 0.Q1 -0.53 0.02 0.61
PDE Soldz et al. (1993) 0.25 -0.29 --0.09 -0.20 -0.29 0.36
Mean 0.27 -0.29 --0.04 -0.37 --0.14 0.49
Histrionic
SIDP-R Trull (1992) 0.37 0.19 0.10 -0.13 -0.25 0.49
PDE Soldz et al. (1993) 0.25 0.26 0.14 0.05 --0.12 0.45
Mean 0.31 0.23 0.12 --0.04 --0.19 0.47
Dependent
SIDP-R Trull (1992) 0.36 -0.09 0.02 --0.05 --0.02 0.38
PDE Soldz et al. (1993) 0.26 -0.05 --0.05 -0.05 --0.16 0.20
Mean 0.31 -0.07 --0.02 -0.05 --0.09 0.29
AntisoCial
SIDP-R Trull (1992) 0.08 --0.10 -0.06 --0.36 -0.32 0.46
PDE Soldz et aI. (1993) 0.01 0.19 0.00 -0.01 -0.10 0.00
Mean 0.05 0.05 --0.03 -0.19 --0.21 0.23
Narcissistic
SIDP-R Trull (1992) 0.32 0.04 0.16 -0.27 --0.10 0.42
PDE Soldz et al. (1993) -0.23 0.08 0.28 -0.10 --0.14 0.40
Mean 0.05 0.06 0.22 -0.19 --0.12 0.41
Avoidant
SIDP-R Trull (1992) 0.25 --0.36 -0.15 -0.19 0.03 0.43
PDE Soldz et al. (1993) 0.23 -0.57 --0.39 -0.17 0.03 0.58
Mean 0.24 -0.47 --0.27 -0.18 0.03 0.51
DSM (Torgersen et aI., 1993a). Another reason may be that Livesley et aI. (1993)
applied a questionnaire and thus other delineation of disorders may appear than pro-
posed in the clinical DSM approach.
Very few other twin studies exist. Kendler et aI. (1987) applied four items from
Eysenck's personality inventory, intending to measure suspiciousness in a large Aus-
tralian sample. They obtained a heritability of 0.41, no shared environment variance, as
common in twin studies, so the rest, 0.59, was non-shared-in-families variance.
More recently, Kendler et aI. (1992) have studied the heritability of schizotypal
features in nine scales intended to measure schizotypy. They found that for 7 of the
Table 3b. The correlations between The Big Five and Personality Disorders
measured by interviews
N E 0 A C R
Borderline
SIDP-R Trull (1992) 0.48 0.04 -D.08 -D.46 -0.31 0.65
PDE Soldz et al. (1993) 0.42 0.06 0.20 -0.13 -0.10 0.53
Mean 0.45 0.05 0.06 -D.30 -0.21 0.59
Passive-Aggressive
SIDP-R Trull (1992) 0.25 -0.13 0.02 -0.46 -D.19 0.49
PDE Soldz et al. (1993) 0.31 -0.05 0.06 -0.32 -D.49 0.55
Mean 0.28 -0.09 0.04 -0.39 -0.34 0.52
Sadistic
PDE Soldz et al. (1993) 0.11 0.12 0.19 -0.02 -0.06 0.17
Self-defeating
PDE Soldz et al. (1993) 0.27 0.07 0.16 -0.04 -0.08 0.32
scales, the heritability was relatively high, from 0.40 to 0.68, and with no shared envi-
ronmental variance. The highest heritability was found for anhedonia, picturing the so-
called "negative" features of schizotypy. On the other hand, for two scales measuring
perceptual aberration, "positive" schizotypy, no hereditary component was observed, a
high shared environmental variance (0.25-0.29) and a very high non-shared variance
(0.71-0.75).
These studies suggest, as expected from the high correlations between The Big
Five and PD measured by questionnaires, that heritability contribute a lot to develop-
ment of PD, such defined. However, as the correlations between PD's measured by
questionnaires and measured by interviews are moderate to low (Zimmermann, 1994),
we do not know whether interview PD's are also genetically transmitted. Furthermore,
the correlations between The Big Five and the PD's measured by interview are mod-
erate (Table 3).
5. PERSONALITY DISORDERS EVALUATED THROUGH

INTERVIEW AND RECORDS
This section is discussing studies where information about PD's are obtained
through records and/or interview. Almost all we know so far concerns antisocial and
schizotypal personality disorders.
6. ANTISOCIAL AND SCHIZOTYPAL

In 1976, Dalgard and Kringlen published a twin study of criminality. With a very
broad definition of criminality, they only observed a slightly higher concordance for
MZ twin pairs compared to DZ pairs. A more strict concept of crime yielded a con-
cordance of 26% for MZ and 15% for DZ twin pairs. Some will consider this differ-
ence as impressive, others will be more skeptical, taking into account the fact that MZ
partners are more together than DZ twins. To study the effect of similarity in environ-
ment, the authors analysed separately twin· pairs which were close and distant. They
then made the surprising discovery that the more close MZ twins showed lower con-
cordance and the more close DZ twins showed higher concordance. Thus, the concor-
dance difference only appeared among twin pairs which were distant. In disagreement
with the authors, we do not think that this result disproved genetic influence. However,
the results may show that an environmental factor as the twin relationship modifies the
effect of genes in criminality.
McGuffin and Gottesman (1984) reviewed a number of relatively systematic
ascertained twin studies of crime. They concluded that a fairly high difference in
concordance was found between MZ and DZ twin pairs. However, the same was not
true for juvenile delinquency. An American study of discharges for dishonesty in
the American army also showed a clearly higher concordance for MZ twin partners
compared to DZ partners. (Centerwall and Robinette, 1989). All these studies share
the problem that MZ twins often make offences together. Consequently, some
data from The Minnesota Study of Thins Reared Apart are important. By applying the
Diagnostic Interview Survey (DIS) to twins reared apart, they observed a heritability
for child antisocial features of 0.41 and adult antisocial features of 0.28 (Grove et aI.,
1990). Thus, this study did not find that the genetic influence is higher for antisociality
in older age. A recent study (Lyons et aI., 1995) found that resemblance for juvenile
antisocial traits was largely due to the familiar environment, while when adult traits
were analysed twin resemblance was almost completely explained by genetic factors.
This seems to provide further evidence that the action of genes is not constant during
development, and that genetic similarities may become more evident when twins
become able to slect their own environment, given the correlation between genes and
environment.
Not so few adoption studies of antisocial features and criminality have been per-
formed. Crowe (1974) studied offspring of female offenders which were given up for
adoption in infancy. The offsprings of the offenders had more often antisocial person-
ality, but not other personality deviations or psychiatric disorders compared to control
adopted-away offsprings. The length of time spent in temporary care prior to final
placement was important for the development of antisocial personality, pointing to the
interaction between genetic factors and environment.
Cadoret has published a number of articles from his adoption study of antisocial
personality. In a more recent article (Cadoret and Stewart, 1991), it is shown that not
only antisocial personality, but also attention deficit/hyperactivity were found among
adopted-away offspring of criminals. However, this was only true when the offspring
had been placed in lower socio-economic status home. Psychiatric problems in the
adoptive home were related to aggressivity in the offspring, and this aggressivity syn-
drome in its tum predicted antisocial adult personality. The study does show important
interaction between environment and genetics in the development of antisocial per-
sonality. In addition, ADIHD seems to be an alternative outcome of genetic factors
influencing the development of antisociality.
Other adoption studies (Cloninger et aI., 1982; Sigvardson et aI., 1982) have
also shown that prolonged institutional care before adoption and the socio-economic
status of the adoption home influences the likelihood of criminality in the adopted-
away offspring of criminals. Aopt In addition to ADIHD, an adoption study suggests
that somatoform disorder is genetically linked to antisociality (Cadoret, 1978). The
conclusion seems to be that genetic factors influence antisocial features, at least
criminality. This genetic influence may be relatively unspecific, and also of impor-
tance for other psychiatric disorders. Furthermore, environmental factors from early
childhood on seems to interact with the genetic endowment in the development of
antisociality. Specifically, it appears that individuals at high genetic risk are more sen-
sitive to the pathogenetic effects of an adverse adoptive home environment (Kendler,
1995).
In addition to antisocial, schizotypal is the most studied personality disorder in
the realm of genetics. An early twin study of 25 MZ and 34 DZ twin pairs showed a
concordance of28% for the MZ twins and 3% for the DZ twin partners. Genetic factors
thus seem to play a part in the development of schizotypal personality disorder
(Torgersen, 1984). A more recent twin study has demonstrated the heterogeneity
of the schizotypal personality disorder. Only the odd, eccentric, and affect-constricted
features of the schizotypal personality disorder seem to be genetically influenced
(Torgersen et aI., 1993a).
Kendler et ai. (1994) have recently updated the famous Danish Adoption Study
of Schizophrenia. Among their adopted-away pro bands were also some (13) they diag-
nosed, based on the interview material, as having a schizo typal personality disorder. It
turned out that 5 (21.7%) of their first degree biological relatives and 2 (8.3%) of their
second degree biological relatives also had schizotypal personality disorder. The fre-
quencies were statistically significantly higher than biological relatives of control
adoptees (respectively 3.7% and 1.6%).
Thus, also an adoption study confirms the genetic influence on the development
of schizotypal personality disorder.
More discussed has been an eventual genetic relationship between schizotypal
personality disorder and schizophrenia. The aforementioned Danish Adoption Study
observed a frequency of 13.2% of Schizotypal personality disorder among the first
degree relatives and 4.7% among the second degree relatives of schizophrenics. These
frequencies, although impressive, were not statistically significantly different from bio-
logical relatives of control adoptives. Turning the other way, no biological relatives of
index adoptees with schizo typal personality disorder had schizophrenia.
The Danish Adoption Study may thus suggest a tendency to a relationship
between schizotypal personality disorder when we look at relatives of schizophrenics,
but not when individuals with personality disorder is the point of departure.
The twin method is more powerful in detecting the etiological relationship
between two disorders. A twin study found a frequency of schizo typal personality dis-
order of 20.0% among MZ co-twins of schizophrenics, 14.8% among DZ co-twins, and
7.5% among other first degree relatives. These frequencies were clearly higher than
among co-twins and relatives of individuals with major depression (Torgersen et aI.,
1993b). More important, odd speech, inappropriate affects excessive social anxiety were
the schizotypal features which were frequent among co-twins and relatives of schizo-
phrenics. This is in accordance with the earlier mentioned twin study of personality dis-
order features (Torgersen, 1993a). The more dramatic "positive" psychotic-like features
of schizotypal personality disorder is neither genetically transmitted nor related to
schizophrenia. Now we can understand why several studies starting with schizotypal
probands failed to find a relationship with schizophrenia. When we study a very het-
erogeneous disorder, any relationship to another disorder is difficult to detect. Many
of the problems are "false" schizotypals in the meaning of being outside the spectrum
of schizophrenia, and the group of "true" schizotypals will often be too small to dis-
close a relationship to a rare disorder like schizophrenia.
However, several recent family studies that employed direct structured interviews
of relatives of index probands with schizotypal disorder consistently found a signifi-
cantly higher risk for schizophrenia compared to the risk found in families of control
subjects who were healthy or with other personality disorders (Battaglia et al., 1991;
Kendler et al., 1993; Battaglia et al., 1995). Again, the more "odd" features of schizo-
typy appeared to be the more "true" indicators of genetic liability to schizophrenia in
these studies.
Moreover, recent molecular genetic studies show that including schizotypal dis-
order in the spectrum of phenotypes genetically connected to schizophrenia signifi-
cantly improves the power of analyses and the chance of finding a significant linkage
(Straub et al., 1995).
As to other types of personality disorders, very little genetic research has been
performed. A twin study did not find any concordance for borderline personality dis-
order among 7 MZ pairs (Torgersen, 1984). On the other hand, 2 of 18 DZpairs were
concordant, pointing to some shared-in-families environmental variance.
A just finished Norwegian twin study suggests that borderline personality disor-
der does seem to be genetically influenced. The same seems to be true for narcissistic,
obsessive-compulsive, self-defeating, and avoidant personality disorders. Histrionic per-
sonality disorder may also be genetically influenced. Shared-in~families environmental
variance seems to be important for the transmission of dependent, paranoid, histrionic,
and maybe borderline personality disorder. Also the "positive, false, psychotic-like"
schizotypal syndrome seems to be influenced by shared-in-families environmental
variance.
For personality disorders at large, shared-in-families environmental variance
seems to be almost as important as heredity.
To conclude, genetic factors seem to have some influence on the development of
the broad common personality dimensions. Except for agreeableness, the variance
explained by genetic factors may be up to 50%, with 10% accounted for by shared-
in-families environmental variance, and the rest non-shared-in-families environmental
variance.
Also measured by questionnaire, personality disorder features may be similarly
influenced by genes. The reason may be that with questionnaires we simply
measure the broad personality dimensions, even if we intend to measure personality
disorders.
As to personality disorders proper, antisocial personality disorder seems
to be genetically influenced, perhaps sharing genetic variance with attention deficit,
hyperactivity disorders, and somatoform disorders. Also the eccentric, affect-
constricted features of schizotypal personality disorder may be genetically influenced,
sharing genetic etiology with schizophrenia. As to the other personality disorders,
maybe the narcissistic, obsessive, borderline, avoidant self-defeating, avoidant, and
histrionic personality disorders are genetically influenced. Shared-in-families environ-
mental variance seems unusually important for the development of personality disor-
ders. Any speculation about the mode of genetic transmission is premature. Interaction
with environment seems important in the development of antisocial personality. The
genetic relationship between schizophrenia and schizotypal personality disorder seems
complex, and maybe only part of genetic transmission is common (Torgersen et al.,
1993b; Kendler et al., 1994). In the near future, it is reason to believe that more will be
known also about the genetic influence on the development of other personality dis-
orders. However, it is the authors belief that this influence may be modest, especially
for the borderline, histrionic, and avoidant personality disorders as defined by DSM-
III-R and DSM-IY. More or less arbitrarily definitions of personality disorders, without
taking etiological research into account, may create heterogeneous syndromes
with questionable validity. A reciprocal interaction between clinical observations, treat-
ment studies, and etiological research may promote the progress in the understanding
of personality disorders.
REFERENCES
Battaglia, M., Bernardeschi, L., Franchini, L., Bellodi, L., and Smeraldi, E. (1995). A family study of Schizo-
typal disorder. Schizophrenia Bulletin, 21, 33-45.
Battaglia, M., Gasperini, G., Sciuto, G., Scherillo, P., and Bellodi, L. (1991). Psychiatric disorders in the fam-
ilies of schizotypal subjects. Schizophrenia Bulletin, 17, 659-668.
Battaglia, M., Przybeck, T.R, Bellodi, L., and Cloninger, CR (1996). Temperament dimensions explain the
comorbidity of psychiatric disorders. Comprehensive Psychiatry, 37, 292-298.
Benjamin, I, Li, L., Patterson, C, Greenberg, B.D., Murphy, D.L., and Hamer, D.H. (1996). Population and
familial association between D4 dopamine receptor gene and measures of Novelty Seeking. Nature
Genetics, 12, 81-84.
Bergeman, C.S., Chipuer, H.M., Plomin, R, Pedersen, N.L., McClearn, G.E., Nesselroade, IR, Costa, P.T. Jr.,
and McCrae, RR (1993). Genetic and environmental effects on openness to experience, agreeable-
ness, and conscientousness: An adoption/twin study. Journal of Personality, 61, 59-178.
Cadoret, RJ. (1978). Psychopathology in adopted-away offspring of biologic parents with anti-social behav-
ior. Archives of General Psychiatry, 35, 176-184.
Cadoret, RI and Stewart, M.A. (1991). An adoption study of attention deficit/hyperactivity/aggression and
their relationship to adult antisocial personality. Comprehensive Psychiatry, 32, 73-82.
Centerwall, B.S. and Robinette, CD. (1989). Twin concordance for dishonorable discharge from the military:
With a review of the genetics of antisocial behavior. Comprehensive Psychiatry, 30, 442-446.
Cloninger, CR,Adolfsson, R, and Svrakic, N.M. (1996). Mapping genes for human personality. Nature Genet-
ics, 12, 3-4.
Cloninger, CR, Przybeck, T.R, Svrakic, D.M., and Wetzel, RD. (1994). The Temperament and Character
Inventory: a guide to its development and use. St Louis MO, Washington University.
Cloninger, CR, Sigvardson, S., and Bohman, M. (1982). Predisposition to petty criminality in Swedish
adoptees. III Cross-fostering analysis of gene-environment interaction. Archives of General Psychia-
try, 39,1242-1253.
Costa, P.T. Jr. and McCrae, RR (1990). Personality disorders and the five factor model of personality. Journal
of Personality Disorders, 4, 362-37l.
Crowe, RR (1974). An adoption study of anti-social personality. Archives of General Psychiatry, 31, 785-79l.
Dalgard, O.S. and Kringlen, E. (1976). A Norwegian twin study of criminality. The British Journal of Crimi-
nology, 16, 213-232.
DeFries, IC and Hegmann, IP. (1970). Genetic analysis of open field behavior. In Contributions to behavior
genetic analysis: the mouse as a prototype. Eds G. Lindzey and D.C Thiessen, pp. 23-56. New York.
Appleton Century Crofts.
Eaves, L.I and Eysenck, H.I (1975). The nature of extraversion: a genetic analysis. Journal of Personality and
Social Psychology, 32, 102-112.
Ebstein, RP., Novick, 0., Umansky, R, Priel, B., Osher, Y, Blaine, D., Bennett, E.R, Nemanov, L., Katz. M.,
and Belmaker, RH. (1996). Dopamine D4 receptor exon III polymorphism associated with the human
personality trait of Novelty Seeking. Nature Genetics, 12,78-80.
Eysenck, H.I and Eysenck, S.B.a. (1976). Psychoticism as a dimension of personality. New York: Crane,
Russak & Co.
Eysenck, H.I (1967). The biological basis of personality. Springfield, Ill.: Charles C Thomas.
Goldsmith, H.H. and Gottesman, 1.1. (1996). Heritable variability and variable heritability in developmental
psychopathology. In M.F. Lenzenweger and II Haugaard (eds): Frontiers of developmental psy-
chopathology. New York, Oxford University Press.
Grove, W.M., Eckert, E.D., Heston, L., Bouchard, T.I Jr., Segal, N., and Lykken, D.T. (1990). Heritability of
substance abuse and antisocial behavior: A study of monozygotic twins reared apart. Biological Psy-
chiatry, 27,1293-1304.
Kendler, K.S. (1995). Genetic epidemiology in psychiatry. Taking both genes and environment seriously.
Archives of General Psychiatry, 52, 895-899.
Kendler, K.S., McGuire, M., Gruenberg, AM., O'Hare, A, Spellman, M., and Walsh, D. (1993). The Roscom-
mon family study. I Methods, Diagnosis of probands and risks for schizophrenia in relatives. Archives
of General Psychiatry, 50, 527-540.
Kendler, H.S., Gruenberg, AM., and Kinney, D.K. (1994). Independent diagnoses of adoptees and relatives
as defined by DSM-III in the Provincial and National sample of the Danish Adoption Study of Schiz-
ophrenia. Archives of General Psychiatry, 51, 456-468.
Kendler, K.S. and Hewitt,IK. (1992). The structure of self-report schizotypi in twins. Journal of Personality
Disorders, 6, 1-17.
Kendler, K.S., Heath, A, and Martin, NG. (1987). A genetic epidemiologic study of self-report suspicious-
ness. Comprehensive Psychiatry, 28,187-196.
Lesch, K.P., Bengel, D., Heils, A., Sabol, S.Z., Greenberg, B.D., Petri, S., Benjamin, I, Muller, CR, Hamer,
D.H., and Murphy, D.L. (1996). Association of anxiety-related traits with a polymorphism in the sero-
tonin transporter gene regulatory region. Science, 274, 1527-1531.
Livesley, W.I, Jang, K.L., Jackson, D.N, and Vernon, P.A (1993). Genetic and environmental contributions
to dimensions of personality disorder. American Journal of Psychiatry, 150, 1826--1831.
Lyons, M.I, True, W.R, Eisen, S.A., Goldberg, I, Meyer, 1M., Faraone, S.M., Eaves, L.I, and Tsuang, M.T.
(1995). Differential heritability of adult and juvenile antisocial traits. Archives of General Psychiatry,
52, 906--915.
McGuffin, P. and Gottesman, 1.1. (1984). Genetic influence on normal and abnormal development. In: M.
Rutter and L. Hersou (Eds.). Child Psychiatry: Modern Approaches (2nd ed.) London: Blackwell.
McCrae, RR and Costa, P.T. Jr. (1987). Validation of the five-factor model of personality across instruments
and observers. Journal of Personality and Social Psychology, 52, 81-90.
Pedersen, NL., McClearn, G.E., Plomin, R, Nesselroade, IR, Berg, S., and De Faire, U. (1991). The Swedish
Adoption/Twin Study of Ageing: An update. Acta Geneticae Medicae et Gemellologiae, 40, 7-20.
Perry, IC (1992). Problems and considerations in the valid assessment of personality disorders. American
Journal of Psychiatry, 149, 1645-1653.
Plomin, R, De Fries, IC, and McClearn, G.E. (1990). Behavioral Genetics. A primer. Second Edition WH
Freeman & Co., New York.
Sigvardson, S., Cloninger, CR, and Bohman, M. (1982). Predisposition to petty criminality in Swedish
adoptees III. Sex differences and validation of the male typology. Archives of General Psychiatry, 39,
1248-1253.
Soldz, S., Budman, S., Demby, A, and Merry, I (1993). Representation of personality disorders in circumplex
and five-factor space: Exploration with a clinical sample. Psychological Assessment, 5, 41-52.
Stallings, M.C, Hewitt, IK., Cloninger, CR, Heath, AC, and Eaves, L.I (1994). The structure of tempera-
ment: a joint analysis of the models of Cloninger, Eysenck, and Gray. Behavior Genetics 24 (6),532.
Straub, RE., MacLean, CI, O'Neill, F.A., Burke, I, Murphy, B., Duke, F., Shinkwin, R, Webb, B.T., Zhang,
I, Walsh, D., and Kendler, K.S. (1995). A potential vulnerability locus for schizophrenia on chromo-
some 6p24-22: evidence for genetic heterogeneity. Nature Genetics 11, 287-293.
Tellegen, A (1983). Structure of mood and personality, and their relevance to assessing anxiety, with an
emphasis on self-report. In: AH. Thma and ID. Maser (Eds.). Anxiety and the anxiety disorders (pp.
681-706). Hillsdale, NJ: Erlbaum.
Tellegen, A, Lykken, T.D., Bouchard, T.I, Wilcox, K.I, Segal, N.L., and Rich, S. (1988). Personality similarity
in twins reared apart and together. Journal of Personality and Social Psychology, 54, 1031-1039.
Torgersen, S. (1984). Genetic and nosological aspects of schizotypal and borderline personality disorders. A
tWin study. Archives of General Psychiatry, 41, 546--554.
Torgersen, S., Skre, I, Onstad, S., Edvardsen, I., and Kringlen, E. (1993b). "True" schizotypal personality dis-
order: A study of co-twins and relatives of schizophrenic probands. American Journal of Psychiatry,
150,1661-1667.
Torgersen, S., Skre, I., Onstad, S., Edvardsen, I, and Kringlen, E. (1993a). The psychometric-genetic structure
of DSM-III-R personality disorder criteria. Journal of Personality Disorders, 7,196-213.
Trull, T.I (1992). DSM-III-R personality disorders and the five-factor model of personality: An empirical
comparison. Journal of Abnormal Psychology, 101,553-560.
Zimmermann, M. (1994). Diagnosing personality disorders. A review of issues and research methods.
2
PSYCHOBIOLOGY OF PERSONALITY
DISORDERS
Implications for the Clinic
Larry 1. Siever
INTRODUCTION
While it has been traditional to seek out environmentally based or psychologi-

cally rooted explanations for the personality disorders, it is becoming increasingly clear
that our understanding of these disorders must also include an appreciation of its bio-
logic substrates. Brain system observed to regulate affective expression, cognitive orga-
nization, anxiety thresholds, and impulse control may playa crucial role in determining
an individual's "set point" which can become the basis around which the personality is
organized. The basis of this set point may be partially genetic, and indeed, recent studies
suggest that activity of enzymes involved in the synthesis and reuptake of the
monoamines may play an important role in determining these system's activities with
important behavioral implications. However, salient environmental events, particularly
traumas have also been shown to have lasting effects on brain modulatory systems. It
becomes important to understand the convergences between these biologic systems
and the psychologic constructs we use to understand personality disorder. It may be
useful to identify these convergences in terms of core dimensions such as the regula-
tory domains as discussed above: cognitive organization, affective stability, anxiety, and
impulse control and examine their underlying biologic substrates.
IMPULSE CONTROL
The serotonergic system seems to playa key role in modulating the suppression
of behaviors that later meet with punishment, particularly aggressive behaviors, and
individual differences in serotonergic activity may contribute to the susceptibilities
toward impulsive aggressive behaviors. Animal studies suggest that lesions of seroton-
ergic neurons lead to unrestrained killing of mice placed in the lesioned rodents vicin-
ity. Furthermore, these lesioned rodents have difficulty suppressing learned behavior
Kluwer Academic / Plenum Publishers, New York, 1999. 17
18 L.J. Siever
such as lever pressing that was previously coupled to a reward but is replaced by
a painful consequence. While it is easy to "anthropomorphize" these results to suggest
that humans who are unable to suppress aggressive behaviors or "learned" the con-
sequence of these behaviors have deficits in serotonergic system, it is now possible to
more directly test these possibilities rather than merely extrapolating from animal
data.
There are a number of techniques by which one can measure serotonergic activ-
ity, including measurement of serotonergic metabolite, measurement of hormonal
responses to agents which enhance serotonergic activity, and measurement of meta-
bolic responses to such agents, all of which suggest reduced serotonergic activity asso-
ciated with aggression. Diminished activity of the serotonergic system, whether
reflected in low concentrations of metabolites of serotonin or in blunted responses to
agents that challenge this system, may result in the failure to suppress punished or
aggressive behaviors resulting in "disinhibited" aggression. This model is supported by
animal studies, suggesting that lesions of the serotonin system in rats will result in their
showing marked increases in muricidal behavior (that is, killing of mice), and failure to
extinguish behavior that is met with punishment. In primates, serotonergic deficits are
also associated with impulsive aggression and differences in serotonergic activity seem
to be largely genetically based, although they may be modified by environmental
manipulations. For example, changing the high dominance status of an individual in a
primate troop may change their serotonergic activity. Nonetheless, genetic contribu-
tions seem to outweigh environmental ones, at least, in controlled settings. Genetic
factors may contribute to individual differences in serotonergic system activity through
affecting the synthesis, or breakdown, of serotonin, and the sensitivity of receptors that
mediate its effect on other neurons. Initial studies suggest that one allelic variation of
the tryptophan hydroxylase gene may be associated with increased suicide attempts in
criminal offenders and, in a pilot study from our laboratory, with increased impulsive
aggression in personality disorder patients. While these results require extension and
replication in other laboratories, it is likely that individual differences in serotonergic
activity based on genetic and environmental factors will modify the "setpoint" at which
impulsive aggression occurs in response to stress or frustration.
These temperamental predispositions may have important contributions to the
development of personality. For example, infants who can not easily suppress impul-
sive or aggressive behaviors may be more prone to temper-tantrums that will punc-
tuate their early development. These tantrums may interfere with the smooth
development of appropriate attachment behaviors and may result in disturbed or inter-
rupted attachment patterns. Caretakers may attempt to cope with these tantrums by
either become alternatively overindulgent or deprivational, inadvertently increasing
the intermittent reinforcement properties of these tantrum-like behaviors. The result
maybe a less firm and well developed attachment system, which will affect how that
individual negotiates later stages of personality development. For example, separa-
tion/individuation may become more problematic without the benefit a solid, secure
attachment. In addition, as this phase involves mastery of aggression in the service of
seeking autonomy and separation from a secure base, impulsive aggression will impact
directly on this phase of development as well. Finally, the impulsive aggression may
make more problematic sustained, relatively mature identifications with appropriate
adults to ensure transition to a self-directed developmental trajectory. In this way,
impulsivity/aggression may not only impact on these various developmental phases and
their requisite tasks, but also, by contributing to the resolution of previous phases and
Psychobiology of Personality Disorders 19
the psychological structures associated with them, have profound influences on later
development. Thus, it would be difficult to disentangle a predisposition to impulsive
aggression from the integrated development trajectory of personality.
An increased understanding of the brain mechanisms underlying this impulsive
aggression may now be furthered by imaging studies that actually visualize brain
response to agents, such as fenfluramine, which cause release of serotonin, and
brain metabolism in regions such as orbital frontal cortex, which playa major role in
inhibiting and regulating aggression. Thus, we may be able to visualize the brain
metabolic correlates of diminished suppression of aggression. More accessible tests of
hormone responses to agents such as fenfluramine can also indicate blunted of the
serotonin system and may help us understand better how psychopharmacologic inter-
ventions may ameliorate impulsivity in severe cases. Preliminary studies already
suggest that reduced serotonergic responsiveness may diminish the effectiveness of
selective serotonin reuptake inhibitors and/or delay their onset of action. Furthermore,
they suggest that SSRI's may be useful in the control of impulsive aggression in indi-
viduals with impulse dysregulation. In some studies of SSRI's in the treatment of
borderline patients, it has been the anger dyscontrol rather than depression that is alle-
viated by these agents. Thus, an understanding of these biologic mechanisms may
improve our psychopharmalogic treatment, as well as increase our empathic under-
standing of the developmental challenges that an impulsive/aggressive temperament
brings about.
It is essential to recognize that the lens through which we view the world is going
to be affected by our temperament and therefore environmental circumstances cannot
be differentiated from the biologic apparatus through which they are filtered. These
considerations can then inform psychotherapeutic treatment, whether psychodynamic
or cognitive/behavioral oriented. I suggest a value for dimensional approaches that
consider multifactorial contributions at both the biologic and psychosocial levels.
For example, alterations in noradrenergic system may effect irritability and eng-
agement with the environment and thus importantly impact on the expression of
serotonergic deficits. Without considering multiple biologic and environmental factors,
it would be difficult to appreciate the complexity of variability in the personality
disorders. However, fortunately for the investigator, it may be that a more limited
number of these factors at both the biologic and psychosocial levels can be identified
so that plausible mechanistic schema and longitudinal studies could be ultimately
undertaken.
Similar considerations hold for alterations in brain structure and organization,
that may be a result of both heritable and/or environmental factors such as viral illness
or nutritional factors, hypothesized to contribute to the symptoms of the schizophre-
nia spectrum disorders. These disorders range from severe chronic schizophrenia to
schizo typal and possibly schizoid personality disorder, personality disorders marked by
eccentricity, social withdrawal, and suspiciousness. It is quite possible the differences in
central nervous system development that may be more subtle than gross neurologic
lesions, yet ultimately detectable by imaging or histopathologic methods, may influence
how attention is deployed, information is internalized and organized, and coping strate-
gies are generated to respond to the environment. Subtle shifts in these organizational
capacities might have a profound impact on how an individual is able to interact with
others in their environment and perform satisfactorily in their occupation or, more
particularly, form satisfying interpersonal relationships. Accumulating evidence from
such studies suggest that alterations in brain metabolism, structure, and secondary
20 L. J. Siever
cognitive/attentional dysfunction are not only associated with chronic schizophrenia

but with the schizophrenia spectrum personality disorders as well. These studies suggest
an important role for brain development for understanding individual differences in
personality disorder.
Our knowledge of biologic validators of personality disorders is only in its infancy.
As this understanding deepens, knowledge of biologic correlates of the personality
disorders may indeed enable a better identification of psychosocial factors that inter-
act with genetic factors, the impact of trauma and other formative environmental
influences, and a more fine grained understanding of the interaction between tem-
perament and caretaking environment in the development of personality disorder
by enabling the selection of more appropriate samples in which to understand these
interactions.
COGNITIVE ORGANIZATION
The capacity of an individual to organize their perceptions of the world around

them, formulate appropriate responses and behavioral strategies and correct these pro-
grams on the basis of their experience requires effective cognitive executive function
to coordinate representations of the environment and planning of actions on the envi-
ronment. Individuals with deficits in cognitive organization appear to have difficulties
in such executive functions and as a result appear "out of step", unable to integrate
their experiences with ongoing maps of the world around them, and appear to have
difficulty in interacting smoothly with individuals around them at work or socially. It
appears that at least some forms of cognitive disorganization may be rooted in altered
organizations of underlying brain structures. These relationships can be identified
clearly in patients with schizophrenia spectrum personality disorder whose oddities of
behavior and dress, peculiar perceptions and socially isolated behaviors may actually
be a function of anomalous brain organization. The prototypic disorder of cognitive
disorganization and schizotypal personality disorder. People with schizotypal person-
ality disorder appear eccentric, detached, have few close friends, may use words pecu-
liarly and may experience cognitive/perceptual distortions. Because they may have
difficulty in decoding the complex signals associated with both verbal and non-verbal
communications, they may appear to be "out of sync" with others in social conversa-
tion and describe themselves as finding social interactions fraught with difficulty and
frustration. As a result, they often pursue a solitary life with occupations that don't
make many interpersonal demands and few ongoing personal relationships. The social
deficits, however, may be understood in part as related to an underlying cognitive
disorganization.
Cognitive testing does suggest underlyingneuropsychologic deficits in patients
with schizotypal personality disorder. Tests of executive functioning, which are sensi-
tive to damaged frontal lobe, are performed more poorly by schizotypal patients. SPD
patients have been demonstrated to make more perseverative errors on the Wisconsin
Card Sort Test which is a reflection of the capacity to abstract, utilize working memory,
and shift behavioral strategies according to incoming feedback. They also perform more
poorly on the Trails B Test and the Stroop Test, two tests also sensitive to frontal dys-
function. SPD patients also demonstrate deficits in verbal learning. Their performance
on the California Verbal Learning Test (CVLT) suggest deficits both in coding and
more prominently in retrieval of learned words. This capacity is particularly sensitive
to temporal dysfunction. They also show deficits in a broad range of tasks designed to
tap into involuntary attention. Their performance on the Continuous Performance Test
(CPT) suggests that schizotypal patients, like schizophrenic patients, demonstrate
deficits in sustained attention pairing. SPD patients also perform more poorly in eye
tracking tasks with low pursuit gain and longer periods of saccadic tracking. Finally,
some, but not all, reports suggest abnormalities in the backward masking tests
which evaluate shorter term visual processing. Cumulatively, these tests point to an
underlying dysfunction of attention that may be mediated by subcortical-frontal and
subcortical-temporal networks.
SPD patients also demonstrate structural alterations in the brain that may in fact
be associated with these cognitive impairments. Ventricular enlargements have been
recorded using both CT scanning and MRI methodology in our samples. Increased ven-
tricular size may be associated with greater cognitive impairment. The preliminary data
suggests reduction in temporal lobe volume in these patients, although surprisingly,
pilot data suggests increases in frontal volumes, possibly consistent with greater frontal
reserves which may buffer the cognitive impairment of the schizo typal in contrast to
the schizophrenic patient.
Imaging studies visualizing cerebral blood flow or glucose metabolism also
suggest anomalous brain activation patterns in patients with SPD. Asymmetric tempo-
rallobe activation and inefficient activation of frontal cortex may be observed during
performance of tasks such as the WCST or CVLT in our laboratory. Some of these acti-
vation patterns may represent attempts to compensate for dysfunction of brain regions
which are primarily affected by the schizophrenia spectrum pathophysiology. It may be
that SPD patients are spared the severity of schizophrenic psychopathology because
of their adaptive strategies to compensate for this underlying dysfunction.
The neurotransmitter dopamine, may also playa role in both the modulation of
the cognitive deficit as well as the psychotic-like symptoms of SPD. For example,
decreases in dopaminergic activity have been associated with impairments in working
memory and other cognitive functions in animal studies and decreases in dopaminer-
gic indices have been associated with reduced cognitive performance in both schizo-
phrenia and schizotypal personality disorder. Increased dopaminergic activity, on the
other hand, particularly in subcortical regions, may be related to the psychotic-like
symptoms of SPD as reflected in correlations between the number of these symptoms
and increases in CSF in plasma homovanillic acid (HVA) , a major metabolite of
dopamine. Thus, in summary, alterations in the modulation and/or activity of brain
structures in a schizophrenia spectrum personality disorder may contribute to their
cognitive/perceptual distortions of social deficits.
AFFECTIVE INSTABILITY
A hallmark feature of many of the Cluster B or dramatic cluster personality

disorders, particularly borderline personality disorders, is their emotional reactivity or
affective instability. Their affects may rapidly shift from despondent to irritable or angry
to excited, usually in response to interpersonally charged environmental events.
Their difficulty in regulating their affects makes them more sensitive to disappoint-
ment, separation, or loss. As individuals with these personality disorders lack the ego
mechanisms or coping strategies to construct or re-alter their intra-structural milieu,
they tend to act impulsively by ingesting substances, engaging in promiscuity, or getting
22 L. J. Siever
involved in other risky or self-destructive behaviors such as binge-eating, gambling, or

even self-mutilating behavior. Thus, the affective instability coupled with impulsivity
may provide the substrates for the ongoing maladaptive acting out of the borderline
patient.
The neurobiologic basis of affective instability is as of yet unknown. However,
personality disorder patients with affective instability, particularly borderline person-
ality disorder patients, may respond to pharmacologic challenge with greater negative
affective responses than comparison groups. For example, in our laboratory, patients
with borderline personality disorder demonstrated greater dysphoric or depressive
responses to the cholinesterase inhibitor physostigmine, than did other personality dis-
order patients or normal controls. The dysphoric response to physostigmine was par-
ticularly associated with those criteria of borderline personality disorder related to the
affective instability including the specific criterion of affective instability, fears of aban-
donment, and identity disturbance. Borderline patients also showed greater affective
responses to serotonergic challenge in a study in our laboratory as well as in other
studies. While these studies may be consistent with supersensitivity of cholinergic
systems, for example, they may also suggest that post receptor amplification mecha-
nisms may be at the levels, for example, possibly if second messengers may be enhanced
in patients with borderline personality disorder and affective instability accounting in
part for their excessive reactivity affectively to the environment.
It is now possible to image brain metabolism in specific regions in response to
affective provocation. For example, activation of the amygdala in related areas can
be observed in CAT studies following emotionally charged stimulation and these
responses are abnormal in mood disorder patients. This experimental paradigm may
offer an option to study the heightened emotional reactivity of the borderline patient.
Thus, it may be possible to map out both these pharmacologic pathways and brain
regions implicated in the excessive affective instability and sensitivity of the patient
with borderline personality disorder.
ANXIETY
The cluster C personality disorder diagnoses are marked by persistent maladap-

tive strategies to reduce anxiety, particularly around prospects of future punishment or
disappointment. Thus the avoidant individual seeks to forestall the experience of rejec-
tion by staying away from social or public settings where he might be subject to nega-
tive judgments or in his mind even ridicule. A dependent personality disordered
individual structures their life around submission to a dominant other avoiding the
ambiguities, conflicts, and potential rejections inherent in making their own decisions.
The obsessive compulsive personality disordered patient is obsessively rule bound
diminishing possibilities of spontaneity and flexibility that for them may be an occa-
sion for uncertainty or conflict. While these disorders bear a relationship to the Axis I
anxiety disorders conceptually and to some degree in comorbidity, this relationship is
probably least uniform or empirically supported of the Axis I-Axis II relationships.
Furthermore, less is understood about the biology of the anxious cluster personality
disorder.
There are some indications that social phobia, which overlaps a great deal with
avoidant personality disorder, may be characterized by reductions of noradrenergic
activity and increases in serotonergic activity. However, results of studies in this circa
have been mixed and there have been no studies of identified avoidant personality dis-
order patients. There have been virtually no biologic studies of dependent personality
disorder. One study of obsessive compulsive personality disorder in our laboratory
suggested blunted prolactin responses to fenfluramine in the obsessive compulsive
personality disorder patients. However, these decreases could be accounted for by vari-
ability and impulsive aggression which coexisted with the obsessive compulsive
features of these patients.
REFERENCES
Coccaro E.P., Siever L.J., Klar H., Maurer G., Cochrane K., Cooper T.B., Mohs R.C., and Davis K.L.: Sero-
tonergic studies in patients with affective and personality disorders: Correlates with suicidal and
impulsive aggressive behavior. Arch Gen Psychiat 46:587-599, 1989.
Siever L.J. and Davis K.L.: A psychobiologic perspective on the personality disorders. Am J Psychiatry
148:1647-1658,1991.
Siever L.J., Kalus 0., and Keefe R.: The boundaries of schizophrenia. Psychiatric Clinics of North America
(Vol 16) 2:217-244, 1993.
3
THE IMPLICATIONS OF RECENT RESEARCH

ON THE ETIOLOGY AND STABILITY OF
PERSONALITY AND PERSONALITY
DISORDER FOR TREATMENT
W. John Livesley
Department of Psychiatry
University of British Columbia
Recent empirical research on normal and disordered personality suggests the

need to re-appraise some of the assumptions underlying traditional approaches to
treating personality disorder. Over the last decade, diverse disciplines including
clinical psychiatry, personality psychology, cognitive psychology, behavior-genetics, and
evolutionary psychology have contributed to our understanding the condition. This
work has shed new light on the structure of personality disorder, classification, and
diagnosis, the relationship between normal and disordered personality, the nature
of the dysfunction associated with the diagnosis, the stability of personality, and the
etiology of personality problems. Although much of this research is not immedi-
ately relevant to clinical practice, work on the causes of personality disorder and the
stability of personality in particular, suggests a conception of personality disorder
that appears to conflict with the models and ideas that guide some commonly used
treatments.
Most psychotherapeutic approaches including psychoanalytically based therapies,
interpersonal therapy, and cognitive and cognitive-behavior therapy share certain
assumptions about the nature of personality disorder and personality change. Two
assumptions seem especially important. First, the causes of personality disorder are
thought to lie largely in adverse psychosocial experiences. Many approaches to treat-
ment acknowledge the importance of constitutional factors, but rarely incorporate
these factors into theories of therapeutic change or take them into account when
planning and implementing therapy. It is as if theorists need to acknowledge constitu-
tional factors before getting on with the real business of therapy - and the real
business is understanding the psychosocial etiology of personality disorder and using
this understanding as the model for treatment. Thus, much of therapy focuses on
26 W. J. LivesJey
memories, conflicts, affects, and cognitions associated with traumatic experiences. This
is appropriate given the etiological importance of the these events, but they are not the
only etiological agents involved in the development of personality pathology. Biologi-
cal factors are also important and they also need to be taken into account when
theorizing about the treatment and planning the management of individual cases.
Second, most therapies seem to assume that personality is relatively plastic and
open to change. In one sense this is a paradoxical idea; personality disorder is defined
in terms of enduring traits. According to the DSM-IV, personality disorder is "an endur-
ing pattern of inner experience and behavior" (p. 630, 1994). Yet, we seek to treat per-
sonality disorder, and treatment implies change. In the early days of psychoanalysis,
there was considerable optimism about the extent to which personality could be
changed. Subsequently, this optimism was tempered by recognition that personality is
highly stable. Nevertheless, the idea persists that personality and personality disorder
are malleable and that personality can be changed with the appropriate psychological
interventions. Thus, it is common to encounter ideas about "changing the structure of
personality". It is not clear, however, what such statements really mean. Nor is it clear
what the term "structure" means in this context. Taken literally, they suggest that pro-
found changes are feasible. This idea, however, potentially conflicts with substantial evi-
dence that some important components of personality are extremely stable and show
little change during the adult lifespan.
The assumptions underlying contemporary approaches to therapy give rise to the
two major theoretical models of personality disorder that are used in treatment. The
developmental-conflict model which, in its simplest terms, proposes that the major traits
of personality disorder originate in developmental frustrations and conflicts, and that
defenses against these conflicts become translated into the traits of character structure.
These ideas are expressed most clearly in Abraham's (1921/1927,1925/1927) typology
which relates personality structure to problems encountered at specific psychosexual
stages. Thus, obsessional traits are assumed to originate in anal conflicts, and depen-
dency traits are assumed to arise from oral conflicts. The second model suggests that
personality disorder involves deficits due to the failure to develop certain important
aspects of personality and that regression to earlier levels of functioning occurs under
stress. With both models the goal of therapy is to bring about change in personality
structure. But, is change in personality structure a feasible goal? And, are the assump-
tions underlying traditional these models consistent with current empirical knowledge?
Etiological studies and research on the stability of personality suggest that these
assumptions need to be reconsidered and that therapeutic strategies need to be mod-
ified to accommodate these new findings. The problem is not that therapies based on
these assumptions are wrong, rather that they deal with only a limited aspect of the
problem.
1. THE CAUSES OF PERSONALITY DISORDER

The theory that personality disorder is a psychosocial condition caused by adverse
developmental experiences conceptualizes personality as the product of learning and
experience. Within this framework, biological factors are assigned a relatively unim-
portant role. Etiological studies, however, suggest a different and more complex picture.
There is little doubt about the important role that traumatic and abusive experiences
play in the development of personality problems. This role is documented by many
The Etiology and Stability of Personality and Personality Disorder 27
studies, especially for borderline personality disorder (Perry and Herman, 1993).
Although deprivation and trauma are important etiological agents they are not neces-
sary factors. Nor are they the only factors involved; genetic and other biological factors
are also important.
Several years ago, Siever and Davis (1991) reviewed some of the biological
correlates of major traits underlying personality disorder. Their work emphasized the
biological basis of personality disorder and encouraged exploration of the relationship
between Axis I and Axis II. These studies do not, however, demonstrate a biological
etiology to personality disorder. More convincing evidence for this comes from
genetic studies. Behavior-genetic studies comparing identical and fraternal twins indi-
cate that 40-60% of the variability in normal personality traits is genetic in origin
(Plomin, Chiuper, and Loehlin, 1990). Although much of this work was conducted
on normal personality traits, similar results have been reported for traits delineating
personality disorders (Livesley, Jang, Jackson, and Vernon, 1993; Jang, Livesley, Vernon,
and Jackson, 1996). In these studies, a self-report questionnaire-the Dimensional
Assessment of Personality Pathology (Livesley and Jackson, in press) was used to
assess four higher-order dimensions of personality disorder: lability, antagonism, inter-
personal unresponsiveness, and compulsivity. These empirically derived components
show some resemblance to DSM-IV diagnostic concepts. Lability captures the unsta-
ble component of many DSM-IV Cluster B diagnoses especially borderline personal-
ity disorder. Antagonism resembles antisocial personality disorder. Interpersonal
unresponsiveness describes the socially withdrawn component of Cluster A diagnoses.
Finally, compulsivity resembles obsessive compulsive personality disorder. This higher
order structure also resembles four of the five factors used to describe normal
personality structure namely, neuroticism, agreeableness, extroversion, and conscien-
tiousness. Only the openness to experience factor described by Costa and McCrae
(1985) is not represented (Schroeder, Wormsworth, and Livesley, 1992). The higher
order dimensions are sub-divided into a variable number of basic dimensions. A total
of 18 dimensions are used to describe the more specific components of personality dis-
order. Table 1 shows the basic dimensional traits that define the four higher-order
dimensions. Estimates of the heritability of the higher-order dimensions based on a
sample of 483 twin pairs were: lability 53%, antagonism 50%, interpersonal unrespon-
siveness 52%, and compulsivity 38%. In the case of the 18 basic dimensions, heritabil-
ity ranged from 56% for callousness and conduct problems to 35% for rejection, with
a median value of 47% (Jang, Livesley, Vernon, and Jackson, 1996). Thus, there is a sub-
stantial genetic basis to the traits composing personality disorder. The evidence also
Table 1. Higher-order and basic dimensions of personality disorder

Higher-order Dimension Basic Dimensions
Lability anxiousness, identity problems, submissiveness, affective lability,
cognitive distortion, social avoidance, insecure attachment,
oppositionality, self-harm, suspiciousness, narcissism
Antagonism callousness, rejection, stimulus seeking, conduct problems, narcissism,
suspiciousness
Interpersonal unresponsiveness social avoidance, intimacy problems, restricted expression (of feelings and
restricted self-disclosure)
Compulsivity compulsivity (this component is defined by a single basic trait)
28 W. J. Livesley
suggests that genetic influences are important throughout the lifespan and that the
genetic contribution to personality traits increases a little with age (Jang, Livesley, and
Vernon, 1996).
The picture emerging from etiological studies is that personality disorders are
psychobiological entities with a complex etiology involving biological and psychosocial
factors. The traits and other characteristics that constitute personality disorder are not
simply the products of learning and experience. Rather, they are psychobiological struc-
tures that develop out of the interaction between genetic predisposition and envi-
ronmental influence. The probable sequence of events is that genetic predisposition
influences responsiveness to certain environments and these environmental factors in
turn influence the expression of the genetic predisposition. As this interaction unfolds,
experience becomes encoded in the mental apparatus as a set of cognitions and behav-
iors, that is, a set of beliefs and expectations that come to influence the way events are
interpreted and subsequent responses to them. In this way traits become incorporated
into the maladaptive patterns that form an important part of the clinical presentation
of personality disorder. The traits that constitute personality disorder therefore, are
complex structures consisting of a genetic predisposition and associated cognitions,
affects, and behaviors. They are not simply the products of experience.
This understanding of personality traits is not consistent with theories of person-
ality disorder that conceptualize traits only as enduring expressions of developmental
conflicts or associated ideas that traits are characterological consequences of defenses
against these conflicts. For example, the notion that narcissistic traits are simply the
products of defenses of against overwhelming feelings of inferiority and inadequacy is
not consistent with observations that narcissism is 53% heritable and that the heri-
tability of the specific traits of need for adulation, attention-seeking, grandiosity, and
need for approval that constitute narcissism ranges from 37% (grandiosity) to 50%
(need for approval). Developmental conflicts may contribute to the development of
these traits, and these traits may be used for defensive purposes, but psychosocial
factors alone do not provide an adequate explanation of their origin.
What are the implications of this understanding for treatment? The idea that
personality has a biological underpinning and the associated idea that traits consist
of interlocking cognitions, affects, and behaviors raises questions about the extent to
which personality can be changed using currently available techniques. The fact that a
trait has a genetic component does not necessarily mean that it cannot be changed. It
does, however, raise questions about the limits of change. Thus, it seems unlikely that
treatment can radically alter the structure of personality in the sense of modifying
major dispositional traits. This suggests that treatment approaches that manage traits
as if they were only the products of defenses are unlikely to be productive. Studies of
the stability of personality and the way personality changes during the lifespan point
to similar conclusions.
2. STABILITY AND CHANGE IN PERSONALITY
To find information on the stability of personality across the lifespan we need to

turn to the study of normal personality because there are few studies of the natural
history and long-term course of personality disorder. And, those that have been con-
ducted have rarely used the kinds of structured personality measures required to
provide information of stability and change. Studies of normal personality suggest that
there is considerable variability across the different components of personality in sta-

bility and potential for change (Heatherton and Weinberger, 1994).
Differences in stability are most readily understood if personality is described in
terms of a system of inter-related structures and processes (Costa and McCrae, 1994;
Vernon, 1963). The basic dispositional traits that are prominent in many theories of per-
sonality and the DSM-IV definition of personality disorder form a major part of the
system (see Fig. 1). Also important are cognitions used to understand the self, other
people, and the environment. These consist of personal constructs, beliefs, and expec-
tations that are used to organize experience and predict events. These cognitions are
shaped by life experiences under the influence of basic dispositional traits. Thus, highly
introverted persons are likely to use different constructs and hold different beliefs
about themselves and others from those who are more sociable. Constructs in turn,
shape experience and influence the way basic dispositional traits are expressed. For
descriptive purposes, the construct system may be divided into concepts of self and con-
cepts of the environment. Knowledge about the self and the processes that regulate
and maintain self-images and self esteem may be conveniently thought of as the self
system. These processes figure prominently in clinical accounts of personality disorder.
The cognitions used to understand the environment that are important in understand-
ing personality, are those used to understand other people. These will be referred to as
the interpersonal system. This is composed of beliefs and expectations, and associated
behavioral strategies and affects that influence our characteristic ways of relating to
others. The self and interpersonal sub-systems are not independent but overlap and
share many features. Both probably develop in parallel, influenced by similar factors
especially basic dispositions and the nature and quality of early relationships. The psy-
choanalytic concept of object relationships captures the close connection between the
self and person sub-systems.
In addition to basic dispositional traits, self system, and interpersonal system, the
personality system also comprises behavior. The behavior that is significant for per-
sonality is behavior that is regular and consistent including consistent forms of experi-
ence. These products of the functioning of the personality system will be referred to as
characteristic expressions. Finally, any systematic account of personality should also
include the environmental context in which behavior occurs. This important component
of the personality system is often neglected when treating personality disorder due to
the emphasis that most clinical explanations of disorder place on internal personality
Construct System
Self System
Figure 1. Personality System.

30 w. J. Livesley
structure and dynamics. The social environment, however, plays an important role in
shaping behavior and in maintaining maladaptive behavioral patterns. Finally, in the
case of personality disorder, the personality system also includes symptoms such as
dysphoria, psychological distress, and impulsive and self-harming behaviors. The
components or sub-systems of personality are not separate entities but interacting com-
ponents of an organized and integrated system. Thus, processes that integrate and
regulate affects, impulses, and behavior are an integral part of the system.
The division of personality into sub-systems is a heuristic that makes it easier
to describe the different aspects of personality and personality pathology while at
the same time capturing the complexity of personality. Personality disorder is associ-
ated with dysfunction in all components of the system. For this reason, the idea of
the personality system is a useful way to organize information on the stability of
personality.
The evidence suggests that the different components of personality differ in
the extent to which they are stable across the life span and the extent to which they
can be changed through therapeutic intervention. Intellectual traits appear to be the
most stable, followed closely by dispositional traits, with self esteem and some
self attitudes being the most subject to change (Conley, 1984a, b). Basic dispositional
traits appear to change little from the late twenties onwards, and even earlier change is
limited (Caspi and Herbener, 1990; Costa and McCrae, 1994). Costa and McCrae (1992)
estimate that approximately 60% of the variance in personality trait scores is stable
over the full adult lifespan. If the mean scores on large groups drawn from different age
groups are compared the means are remarkably similar. Neuroticism, extroversion, and
openness show a slight decrease during the adult life span, while agreeableness and
conscientiousness increase a little (Costa and McCrae, 1994). These differences, however,
are small.
The conclusions drawn from these studies are based on the average scores of dif-
ferent groups. It is possible for the mean score to remain the same and individual scores
to vary considerably. The evidence, however, suggests that this is not the case. The rank
order of a group of individuals remains very stable when followed for as long as thirty
years (Costa and McCrae, 1994). This finding appears to be robust. The accumulation
of evidence points to impressive stability in basic dispositions once they have been
formed. Personality traits are a little less stable during the period from late adolescence
to early adulthood but even during this period there is impressive evidence for stabil-
ity. These results have lead several authors to suggest that personality crystallizes some-
where in the late twenties and changes little thereafter (Costa and McCrae, 1994;
Helson and Moan, 1987). As William James put it "personality is set like plaster". The
characteristic behaviors through which these dispositions are expressed, however, seem
to be more malleable.
Empirical evidence of the stability of self and interpersonal system characteris-
tics is limited. There is some evidence, however, that some aspects of self and identity,
especially self-attitudes and aspects of self-esteem are less stable than basic disposi-
tions (Conley, 1984a, b). Clinical evidence suggests that other aspects of the self, espe-
cially those aspects usually referred to a self-identity (Kernberg, 1975) and a stable and
cohesive sense of self are highly stable and resistant to change, as are core beliefs about
the self or self schemas (Beck, Freeman, and Associates, 1990). Similarly, some com-
ponents of the interpersonal system change under the influence of new experiences
encountered in the course changing life circumstances. Other interpersonal behaviors,
attitudes, and cognitions, like core aspects of the self, appear to be more persistent.
Perhaps the most variable and most malleable aspect of personality pathology are the
symptoms observed in personality disorder. These seem to wax and wane under the
influence of internal and external events. Some of the more affective traits also appear
to vary over time perhaps due to the effects of state factors.
The stability in personality observed throughout adulthood merely indicates that
change does not occur under normal circumstances. It should not be taken to indicate
that personality cannot be changed. Nor do changes in traits that occur until the
late 20s indicate that personality can be changed deliberately with therapy. They
merely indicate that change occurs, perhaps as a result of maturation. At present,
there is relatively little information about whether change can be deliberately brought
about during the early adult years, or later life. Nevertheless, given the accumula-
tion of evidence on stability it is reasonable to conclude that the dispositional
traits that form an integral part of personality structure are highly stable and that
current treatments of personality disorder are unlikely to lead to major changes in
these traits.
This conclusion has important implications for treating personality disorder and
planning treatment programs. Personality is not totally malleable and that there are
limits to the extent to which some components of personality can be modified. Thus,
therapeutic effort should be directed with these ideas in mind. Little is gained from
attempting to change basic structures that are highly stable. Although information on
the stability and potential for change of personality pathology is incomplete, evidence
from studies of normal personality and clinical observation suggest an approximate
hierarchy of stability. The least stable and most amenable to change are symptoms and
some situational factors. More stable, and therefore a less amenable to change, are the
characteristic expressions through which basic dispositional traits are expressed, some
self attitudes including elements of self-esteem, and some interpersonal behaviors and
problems. Finally, basic dispositional traits, core aspects of the self and associated inter-
personal behaviors are the most stable and resistant to change. This suggests that in
the shorter-term therapies and in the earlier phases of longer-term therapy, attention
should be directed toward symptomatic components, interpersonal situations that
trigger symptoms and crises, and related personality characteristics that are open to
change. The core problems of self and interpersonal systems are more appropriately
the subject of medium to longer-term therapy.
To understand more fully the implications of research on personality change for
treatment we need to consider why personality is so stable. Three factors seem impor-
tant. First, there is an adaptive value to stability. Social exchange depends upon people
remaining the same from day to day. Interpersonal relationships would be unpre-
dictable and chaotic if the characteristics of the participants varied considerably from
one occasion to the next. Second, genetic influences contribute to stability. Third, sta-
bility arises from the organization of personality-the different components of the per-
sonality system are dynamically inter-related to form an interlocking and mutually
supportive system that is resistant to change. Basic dispositional traits are integral to
this stability. From the perspective of treating personality disorder, the stability of traits
is a problem because traits act as anchors or fixed points that help to stabilize the system
and the maladaptive expressions of these traits are a key component of most person-
ality disorders. For this reason we need to consider the stability of traits in more detail
to identify ways to promote more adaptive expressions of these traits.
32 w. J. Livesley
3. CHANGE AND THE STRUCTURE OF TRAITS

The clinical relevance of this structure is illustrated by the following clinical
vignette. The patient, a woman in her mid-twenties with a labile personality pattern
would probably be diagnosed as having borderline personality disorder by those who
prefer categorical diagnoses. She had a long psychiatric history involving frequent para-
suicidal behaviors and severe cutting that led to frequent visits to the emergency room.
Her other problems included affective lability and dysphoria, intense feelings of anger
and rage, difficulty in maintaining close relationships, various kinds of impulsive behav-
ior, and self pathology. After several months of twice-weekly therapy the cutting and
visits to the emergency room decreased sharply whereupon the patient complained bit-
terly that her life was extremely dull and boring. She wondered whether she could tol-
erate the boredom much longer. If this new complaint is examined from a traditional
perspective, the parasuicidal behaviors could be considered to met various needs, and
to fill the void created by the inner sense of emptiness and the lack of a coherent sense
of self. The implication is that the sense of boredom would decrease as a more coher-
ent sense of self emerged. Undoubtedly these behaviors served this purpose, but this
is only a partial explanation of their occurrence and persistence. It was apparent that
the self-harming behaviors satisfied additional needs including attention, care, and
dependency. They also provided considerable satisfaction because they were part of
intensely exciting episodes. Invariably the cutting occurred in social situations and
evoked considerable attention and activity on the part of others. On most occasions
the police and paramedical services were called and each episode was dramatic and
exciting. It was apparent that the patient enjoyed the excitement and sense urgency
and crisis. She commented that one of the things that she missed was a the thrill that
the incidents provided. This behavior seemed to be part of a more general pattern of
sensation or stimulus seeking behaviors. Viewed from the perspective of trait psychol-
ogy, this patient occupied an extreme position on the dimension of stimulus or sensa-
tion seeking (Livesley, Jackson, and Schroeder, 1989; Zuckerman, 1971). The self
harming and impulsive behaviors were maladaptive ways of expressing this trait that
also served other functions. The boredom could be understood to be the consequence
of the inability to satisfy the need for stimulation.
Stimulus seeking, like other personality traits, has a substantial genetic compo-
nent. In addition, this trait involves various cognitive schemas that initiate and main-
tain thrill-seeking and sensation seeking acts. For example, the patient had a sense of
invulnerability and believed that she would not be harmed by the self harming behav-
iors or other exciting but high risk actions. Other schemata supported this behavior
including the belief that life was relatively dull, that normality was to be avoided. Self-
harming acts were rewarding. They produced pleasure and excitement and a relief from
emotional distress. These satisfactions in turn reinforced the actions. Thus, the cogni-
tive, affective, and behavioral components probably built upon a genetic predisposition
supported and maintained each other.
Undoubtedly these behaviors may also have served defense and coping functions
including helping the patient to cope with disturbing feelings and self-experiences. But
this is not the only reason for the persistence of the these behaviors or the intense
boredom and need for excitement that followed when these acts were discontinued. If
this problem is understood in terms of the ideas derived from studies of etiology and
personality change it seems unlikely that therapy would lead to major changes in
sensation seeking or that the problem would be totally solved by continued therapy
designed to resolve the structural problems believed to account for borderline pathol-
ogy. Instead, the task of therapy is to help the patient to acquire more adaptive ways
of expressing sensation seeking and the general goal of therapy is to promote adapta-
tion. In this case, the patient began to pursue various high risk sports that were excit-
ing which met her needs for sensation and thrill. These seemed to provide a useful
temporary substitute for the excitement derived from the more maladaptive behaviors
of cutting and self harm.
This understanding of the nature of basic dispositions form the basis for a rational
approach to therapeutic change. From a clinical perspective, it may be useful to distin-
guish between underlying dispositions and the way that these dispositions are expressed.
Although basic dispositions are relatively fixed, characteristic expressions may be more
open to change because they are determined by a variety of other factors in addition to
the basic disposition. Thus, it is important to differentiate between underlying mecha-
nisms and the behavioral expression of these mechanisms (Buss, 1994). This provides a
model to conceptualize change. Therapeutic effort should be directed toward helping
patients to develop more adaptive ways of expressing basic dispositions rather than
attempting to change the dispositions themselves. Expressing this more simply, it is diffi-
cult to believe that it is possible to help someone who is highly introverted to become an
extrovert, or someone who enjoys seeking excitement to tolerate a dull existence. Intro-
version and sensation seeking appear to be relatively fixed at the level of basic disposi-
tions. There is, however, an opportunity to help patients find more adaptive ways to
express their basic personality traits. This approach requires a slight shift in emphasis and
attitude. Rather that seeking to change personality traits the task is to align with these
traits and to work with them to establish a more effective adaptation.
4. STRATEGIES FOR PROMOTING MORE ADAPTIVE

EXPRESSIONS OF BASIC DISPOSITIONAL TRAITS
Traits may be expressed through characteristic expressions that are maladaptive

for several reasons. First, the genetic predisposition to develop certain traits may be
sufficiently strong that it leads to problems even in relatively benign environments.
Second, maladaptive responses may be learned because they serve adaptive needs.
Third, environmental factors may amplify the expression of a given genotype. Fourth,
mechanisms that regulate and control trait expression may be inadequate. Finally, prob-
lems may also arise when the person is intensely intolerant of their own personality
traits and become excessively self-critical.
The conception of traits as psychobiological structures consisting of behaviors,
cognitions, and affects with a genetic predisposition suggests three strategies for
working with traits and promoting more adaptive expressions: (1) progressive substi-
tution of more adaptive expressions; (2) attenuating trait expression through the acqui-
sition of strategies to regulate and control trait based responses; and (3) increasing
acceptance and tolerance of basic dispositional traits. Each of these strategies involves
the use of cognitive interventions to modify the beliefs and expectations associated with
maladaptive trait expression. It is not sufficient in most cases, however, to rely only on
cognitive interventions. The structural model of traits also draws attention to the other
components. It is often necessary to use behavioral strategies to promote behavioral
change and to ensure that changes made in therapy are generalized to everyday
situations and that they are maintained when therapy ends.
34 w. J. Livesley
4.1. Progressive Substitution of More Adaptive Expressions

With this strategy, maladaptive expressions are gradually replaced by more adap-
tive behaviors. The brief case vignette illustrates this strategy. Initially, the cutting and
other self harming acts seemed to the patient to be almost normal and natural ways of
dealing with problems and distress that had the additional benefit of eliciting care and
providing excitement. By exploring the ideas and beliefs associated with these acts it
was possible to identify alternative ways of coping. The patient identified alternative
ways of meeting her need for excitement and stimulation through high risk sports. This
provided an intermediate solution that was more adaptive. Subsequently, alternatives
ways of expressing this trait were identified that involved changes in life style and the
decision to pursue career opportunities that were satisfying and exciting. Many of the
maladaptive behaviors of personality disorder appear amenable to this approach of
gradual substitution of more adaptive behaviors that resembles the graded exposure
method for treating phobias.
4.2. Attenuating Trait Expression through the Acquisition of Strategies to

Regulate and Control Trait Based Responses
Some dysfunctional expressions of traits arise due to problems in regulating
and controlling behavior. This problem is most apparent with affective traits such as
affective lability and neuroticism. Difficulties arise when high levels of these traits
are associated with difficulties in controlling affective responses. Many patients with
personality disorder seem to have acquired maladaptive strategies for coping with
emotional arousal. Instead of using coping strategies such as distraction, problem
solving, or simple acceptance of mood changes to modulate affect and mood change,
they use strategies that amplify the perturbations in mood. This may occur when
patients ruminate about their moods or feelings thereby intensifying them. It can also
occur when patients react fearfully to mood changes and affects telling themselves that
these moods are intolerable and that they are unable to handle them, responses that
usually lead to a further increase in anxiety and distress. A variety of interventions
including anxiety management and cognitive restructuring may be helpful in increas-
ing control of the expression of these traits and especially in modulated degree of affect
arousal.
4.3. Increasing the Acceptance and Tolerance of Basic

Dispositional Traits
Many of the traits of observed in patients with personality disorder are egosyn-
tonic in the sense that the patient feels that they are an integral part of their person-
ality and hence they are not clearly recognized. This does not, however, apply to all
traits. Many patients express considerable dissatisfaction and distress about the per-
sonal qualities. For example, it is not unusual for patients who are highly introverted
to express distress at their inability to relate to others. Similar problems may also arise
during the course of therapy as patients become increasingly aware of personal quali-
ties only to feel discontented with themselves because they possess qualities that they
dislike or find undesirable. Under these circumstances it may be useful to combine
the above strategies for modifying trait expression with interventions designed to help
patients to accept of their basic traits. This may be achieved by helping the patient
recognize and use the adaptive potential of these qualities. For example, one patient
with high levels of affective lability and trait anxiety learned to manage affects more
effectively while at the same time learning to recognize that experiences of intense
affect could have positive aspects. She was in a creative profession learned to use her
strong feelings to increase her creativity. Thus, instead of fearing affect arousal she
began to recognize the positive benefits of some mood changes which in turn helped
to modulate these changes.
Although these strategies form only a small component of the interventions
required to provide comprehensive treatment of personality disorder, the management
of traits is an important component of the process. The clinical course of personality
disorder is often episodic with exacerbation of symptoms and problems triggered by
situational factors, usually interpersonal circumstances that are often echoes of earlier
conflict ridden relationships. Basic dispositional traits play an important part in these
events by influencing the way the person seeks out and responds to situations. They
also influence the nature and intensity of responses to these triggering events. For this
reason, the management of traits is important not as an alternative to exploration of
conflicts, traumatic experiences, affects and other dynamic issues that are the traditional
foci of attention in the treatment of personality disorder but as an additional set of
issues that often need to be considered for therapy to be effective.
5. CONCLUSION
The results of both etiological studies and investigations of personality change

appear to point to the same conclusions; personality is not totally malleable and there
are limits to the extent to which some components of personality can be changed. These
conclusions make an important contribution to the foundation for an empirically based
approach to treatment. It is important, however, that we do not allow the conclusion
that personality disorder has a biological and environmental etiology and that person-
ality is highly stable and resistant to change to lead us to adopt a pessimistic view
of the treatment of personality disorder. A growing number of studies have demon-
strated that treatment is effective and that patients with personality disorder can be
helped to lead more satisfying and adaptive lives (see for example, Piper, Rosie, Joyce,
and Azim, 1996). Instead of pessimism, we should view these results as providing a
rational basis for re-thinking what we hope to achieve when treating personality dis-
order. In essence, empirical research provides a rational basis for planning clinical work.
It also acts as a guide that helps us to direct therapeutic effort where it is likely to be
most effective.
The application of these findings would involve the development of a more tai-
lored approach to the management of individual cases. This requires a greater focus
on assessment and the use of structured assessment techniques to identify the targets
for therapeutic change. This in turn indicates the need for a clinically relevant system
to describe the major traits of personality disorder. Current classifications are not
appropriate because of their reliance on global diagnostic concepts that are not suffi-
ciently specific to serve as the basis for therapeutic interventions. Most interventions
are directed toward particular behaviors, attitudes, or traits rather than diagnoses
(Sanderson and Clarkin, 1994). The systematic application of such an approach to
36 w. J. Livesley
treatment would involve the integration of diverse therapeutic interventions selected

from different treatment modalities and theoretical positions according to patient need.
The application of this approach will require the development of detailed treatment
plans based on in-depth description of personality pathology derived from a combina-
tion of clinical and structured assessment.
REFERENCES
Abraham, K, 1927, Contributions to the theory of the anal character. In Selected papers on psychoanalysis.
London: Hogarth. (Original work published in 1921).
Abraham, K, 1927, Character formation on the genital level of the libido. In Selected papers on
psychoanalysis. London: Hogarth. (Original work published in 1925).
American Psychiatric Association, 1994, Diagnostic and statistical manual of mental disorders (DSM-IV).
Washington, DC: American Psychiatric Association.
Beck, AT., Freeman, A., and Associates, 1990, Cognitive therapy of personality disorders. New York:
Guilford Press.
Buss, D.M., 1994, Personality evoked: The evolutionary psychology of stability and change. In T.F.
Heatherton and IL. Weinberger, Can personality change? Washington DC, American Psychological
Association.
Caspi, A. and Herbener, E.S., 1990, Consistency and change: Assortative marriage and the consistency of
personality in adulthood. Journal of Personality and Social Psychology, 58, 250-258.
Conley, 11, 1984a, Longitudinal consistency of adult personality: Self-reported psychological characteristics
across 45 years. Journal of Personality and Social Psychology, 47,1325-1333.
Conley, 11, 1984b, The hierarchy of consistency: A review and model of longitudinal findings on adult indi-
vidual differences in intelligence, personality, and self-opinion. Personality and Individual Differences,
5,11-26.
Costa, P.T. and McCrae, RR, 1985, The NEO Personality Inventory manual. Odessa, FL: Psychological
Assessment Resources.
Costa, P.T. and McCrae, RR, 1992, Trait psychology comes of age. In T.B. Sonderegger (Ed.), Nebraska sym-
posium on motivation: Psychology and aging. Lincoln: University of Nebraska Press.
Costa, P.T. and McCrae, RR, 1994, Set like plaster? Evidence for the stability of adult personality. In T.F.
Heatherton and IL. Weinberger, Can personality change? Washington DC, American Psychological
Association.
Helson, Rand Moane, G., 1987, Personality Change in Women from College to Mid-Life. Journal of
Personality and Social Psychology, 53, 176--186.
Heatherton, T.F. and Weinberger, IL., 1994, Can personality change? Washington DC, American Psycholog-
ical Association.
Jang, KL., Livesley, W.I, Vernon, P.A, and Jackson, D.N., 1996, Heritability of personality disorder traits: A
twin study. Acta Psychiatrica Scandinavica, 94, 438--444.
Jang, K.L., Livesley, W.I, and Vernon, P.A, 1996, The genetic basis of personality at different ages: A
cross-sectional twin study Journal of Personality and Individual Differences, 21, 299-30l.
Kernberg, O.F., 1975, Borderline conditions and pathological narcissism. New York: Aronson.
Livesley W.I and Jackson, D.N., Manual for the Dimensional Assessment of Personality Pathology. Port
Huron: Sigma Press, (in press).
Livesley, W.I, Jackson, D.N., and Schroeder, M.L., 1989, A study of the factorial structure of personality
pathology. Journal of Personality Disorders, 3, 292-306.
Livesley, W.I, Jang, KL., Jackson, D.N., and Vernon, P.A, 1993, Genetic and environmental contributions to
dimensions of personality disorder. American Journal of Psychiatry, 150, 1826--183l.
Perry, Ie. and Herman, IL., 1993, Trauma and defense in the etiology of personality disorder. In 1 Paris
(Ed.), Borderline personality disorder. Washington DC, American Psychiatric Press.
Piper, W.E., Rosie, IS., Joyce, AS., and Azirn, H.F.A, 1996, Time-limited day treatment for personality disor-
ders. Washington, DC: American Psychological Association Press.
Plomin, R, Chiuper, H.M., and Loehlin, Ie., 1990, Behavior genetics and personality. In L.A Pervin (Ed.),
Handbook of personality theory and research. New York, Guilford Press.
Sanderson, C. and Clarkin, IF., 1994, Use of the NEO-PI personality dimensions in differential treatment
planning. In P.T. Costa and T.A. Widiger (Eds.), Personality disorders and the five factor model of
personality. Washington, DC: American Psychological Association Press.
Siever, L.l and Davis, K.L., 1991, A psychobiological perspective on the personality disorders. American
Journal of Psychiatry, 148,1647-1658.
Schroeder, M.L., Wormsworth, lA., and Livesley, W.l, 1992, Dimensions of personality disorder and their
relationship to the big five dimensions of personality. Psychological Assessment, 4, 47-53.
Vernon, P.E., 1964, Personality assessment: A critical survey. London: Metheun.
Zuckerman, M., 1971, Dimensions of sensation seeking. Journal of Consulting and Clinical Psychology, 36,
45-52.
4
PSYCHODYNAMIC RESEARCH CAN HELP

US TO IMPROVE DIAGNOSIS AND THERAPY
FOR PERSONALITY DISORDERS
The Case of Defense Mechanisms
J. Christopher Perry,! Vittorio Lingiardi,2 and Floriana Iannil
1Institute of Community & Family Psychiatry

Hopital General Juif Sir MortimerB. Davis
Jewish General Hospital
McGill University
Montreal, Quebec, Canada
2Department of Psychology
Facolta' di Psicologia
Universita' "La Sapienza"
Roma, Italy
Personality Disorders (PDs) are found in up to 10% of adults in the general

population, and even higher percentages in psychiatric clinics and hospital settings
(Perry and Vaillant, 1989). Individuals with PDs experience moderate to severe impair-
ment in social and/or occupational functioning as well as significant distress much of
the time. Research has consistently shown that PDs are also commonly associated with
many other psychiatric problems, such as mood disorders, anxiety disorders, suicide
attempts, and completed suicide, substance abuse, etc. (Oldham et aI., 1995). These indi-
viduals also repeatedly use both in- and out-patients psychiatric services (Perry, Lavori,
and Hoke, 1987; Perry, 1993). Despite their suffering, they are widely misunderstood
by others. The concept itself of PD has been for a long time unrecognized by psychia-
try, so that Hirschfeld (1993) described them as "the stepchildren of psychiatry". Only
in the last decades, PDs have obtained more attention by clinicians and they have
gained a more specific diagnostic and therapeutic dimension.
From a public health point of view, these conditions have significant impact on
costs to society and to the individuals themselves and their families. We review below
some revealing findings about improvement in PDs without and with psychotherapy,
Kluwer Academic I Plenum Publishers, New York, 1999. 39
40 J .. C. Perry et al.
and then explore the important issues which need to be addressed in future research
endeavors. We conclude this chapter presenting a research project centered on the role
of Defense Mechanisms (DMs) in PDs evaluation with particular attention to the
relationship between psychotherapy referral and cost-benefit criteria.
1. DO PERSONALITY DISORDERS RECOVER?
From the vantage point of most clinicians, many individuals with PDs cycle in and
out of crises, receiving short-term treatments that result in little substantial change as
they return to life as usual. Despite this short-term view, surprisingly, the long-term
natural history of PDs is for slow recovery. Natural history refers to observing what
individuals do on their own over time, which means some will choose therapy, others
will try it but drop out of it prematurely, some will cycle in and out of therapy, and
some won't try at all, but all do so of their own accord. In the review of natural history
studies of PDs, there were only a sufficient number of studies of only one major type,
Borderline Personality Disorder (BPD), to be able to view its long-term course. There
are four published large-scale follow-ups of borderline patients 15 years after their
initial assessment (McGlashan, 1986; Plakun et aI., 1986; Paris et aI., 1987, 1988, 1989;
Stone, 1990). BPD is characterized by instability in mood, sense of identity and close
relationships, and by self-destructive behaviors. Despite a substantial proportion of
completed suicides (up to 9% over about 15 years), all studies consistently reported
improvements at similar rates over long-term follow-up (Perry, 1993). Taking these
studies together we could estimate that 3.7% of cases recovered per year, meaning that
they would improve to the point where they no longer met the criteria for BPD. The
recovered persons might still have troubles, but not as serious and pervasive as before.
Using the same data, we could estimate that it would require about 24 years for 100%
of BPD subjects to recover just by natural course alone. This is important because many
people fear that a diagnosis of BPD means a lifetime of suffering. Nonetheless it is the
goal of psychotherapy to shorten the time until improvement.
2. DOES PSYCHOTHERAPY HASTEN THE RECOVERY OF

PERSONALITY DISORDERS?
The field of psychotherapy research has fairly clearly demonstrated that psy-
chotherapy is efficacious for many disorders. In his review of psychotherapy studies in
general, Lambert (1994, 180) stated: "psychotherapy facilitates the remission of
symptoms. It not only speeds up the natural healing process but also often provides
additional coping strategies and methods for dealing with future problems".
Left unanswered by this and other reviews is whether psychotherapy helps indi-
viduals with PDs. We recently reviewed three studies of dynamic psychotherapy for
PDs which reported the recovery rate at follow-up sometime after treatment ended
(Banon, Perry, and Ianni, 1995). The studies had a high percentage of patients with BPD
along with other PDs types as well. Like the natural history studies, the three studies
were in surprisingly close agreement, and we were able to estimate that, on average,
11.6% of individuals with PDs recovered each year. Extrapolating into the future, we
estimated that at about 8.33 years 100% of treated cases would be recovered. However,
Psychodynamic Research and Improved Diagnosis and Therapy for Personality Disorders 41
this estimate combined both time in treatment with some subsequent follow-up after
treatment was completed, and therefore might over-estimate the time needed until
recovery. So we reexamined the three studies to determine how many therapy sessions
the subjects received, and then we based our estimates on that. Our model found that
the individuals with PDs recovered at a rate of 0.191 % per session. The model esti-
mated that 50% of individuals would recover after receiving 123 sessions, while 100%
recovery was predicted to take 384 sessions which would require approximately 5.1
years of active treatment.
While the small number of studies make it difficult to know how accurate or stable
the above estimates are, nonetheless, if true, it has highly meaningful implications.
Long-term dynamic psychotherapy may speed up natural improvement by a factor of
4.7 times (i.e. 5.1 years with psychotherapy versus 24 years with natural history only,
for 100% recovery). While we need further studies to validate that these findings are
accurate overall and to determine whether there are variations among the different
PDs types, the existing evidence is sufficient for us to believe that it should be very
fruitful to study those factors in psychotherapy which hasten improvement.
Other studies suggest that after even one or two years in treatment, psychother-
apy can decrease the need for hospitalization and emergency visits (Waldinger and
Gunderson, 1984; Linehan et aI., 1993) even prior to more complete recovery. Thus we
can think in terms of a dollar cost for therapy versus the cumulative costs of continu-
ing with a PD. If one psychotherapy session costs an average of about 100 US$, then
123 sessions would cost 12, 300 US$ to bring 50% of PD individuals in treatment
to recovery from the PD. This should result in significant decrease in costs to society
attributable to decreases in hospitalizations, emergency visits, the aftermath of suicides,
automobile crashes, and other impulsive actions, as well as lost earnings and other
indirect costs of illness. It should also diminish the hidden negative effects upon other
family members, especially the children, which are hard to estimate accurately.
3. WHAT DO WE NEED TO STUDY TO IMPROVE THE

PSYCHOTHERAPY OF PERSONALITY DISORDERS?
3.1. Not Efficacy, but Improved Efficacy

We no longer need to do studies just to prove the efficacy of psychotherapy. The
results are consistent that most psychotherapies which have been studied are more
effective than doing nothing or engaging in a so-called placebo-treatment, where no
real active treatment is conducted (Luborsky et aI., 1993; Crits-Christoph, 1992). Thus
we need to move on to more focused questions which will improve the specificity of
our treatments for specific problems, improve the conduct of the treatment as offered
by most clinicians, and shorten the duration of time until substantial improvement or
full recovery. Addressing these issues will than make psychotherapy both more helpful
and cost-effective for patient and society alike.
3.2. Designer Treatments

We need to develop "designer treatments", that is modifications of a basic treat-
ment modality, such as dynamic or cognitive-behavioral, fitted to the individual patient.
42 J. C. Perry et al.
Most good clinicians already share this goal, but research has generally not addressed
how to make treatment specific for an individual patient's problems; rather treatments
are designed and studied for classes of individuals. Treatment manuals are often written
as if "one size fits all", which everyone knows can lead to some uncomfortably fitting
clothes! To further the development of specific treatments for specific problems, we
need to consider what research is needed.
3.3. More Valid Measurement

We need more research to improve the validity of our diagnostic assessments,
our assessment of the underlying mechanisms of psychopathology, and our measure-
ment of improvement in everyday life functioning, or what is often called patient
outcome. Previous generations of researches have been very successful in measuring
problems which are close to the surface, such as symptoms that result from deeper,
more inferred mechanisms. For example, rating the symptoms of depression to deter-
mine a person's level of depression is done quite reliably and validly with current
interviews and questionnaires. By contrast, we still have a great need for valid ways
to measure the vulnerability to recurrence of depression, or the mechanisms which
maintain depression.
The measurement situation is worse all around for the PDs. When standardized
instruments for assessing PDs have been compared, their level of diagnostic agreement
is only a little better than chance (Perry, 1992). This means that it is hard to compare
the results of two studies and be sure that the diagnoses were made in an equivalent
way. Apart from diagnosis, there are only a few areas where we have begun to study
the underlying mechanisms of PDs. This includes research on defense mechanisms,
dynamic formulations, and cognitive schemata. We know that scientific study of psy-
chodynamics is a difficult challenge for researchers, especially because the assessment
of dynamic phenomena requires inference. Therefore we need to study variables that
are capable of guiding the observer in arriving at each inference and/or justifying that
an inference is warranted. The usual alternative is between a measure that is easily
learned and reliably applied, but not truly dynamic, versus a measure that attempts to
assess dynamic phenomena at too deep a level, but fails to do so reliably. As we'll see,
defense mechanisms represent an area of investigation more accessible than other
aspects of dynamics, such as intrapsychic conflict or transference. This is because pat-
terns of affect, behavior, and cognition indicating an underlying defense are less
complex or require less inference than other dynamic phenomena.
For example, we now know that individuals with BPD use specific defenses like
splitting, acting out, repression, and dissociation. With a few years of systematic study,
we could have a good characterization of the mechanisms of all PDs using existing
instruments. These instruments are also capable of yielding a unique description of
the individual person, like a dynamic formulation. This should facilitate research and
clinical work. Later in this article we'll discuss some aspects of empirical research on
DMs in PDs.
3.4. The Links between Etiology and Current Mechanisms

We need more studies of the links between original causal factors that pushed the
individual toward developing a PD and the current mechanisms that maintain it. For
example, there is a growing evidence that a number of traumata, such as physical and
sexual abuse or severe emotional neglect, are found in the early childhood of many
individuals with BPD, but not all (Fossati et aI., in press). When past traumata and
current mechanisms are studied together it is probable that we will delineate more than
one pathway to becoming BPD. Furthermore, treatment studies might find that each
pathway also influences the type of treatment which is shown to be most successful.
Thus our most effective treatments will ultimately be driven by proven theories about
the etiology and maintenance of a disorder. Until our research links etiology with
current mechanisms, we will still be relying on treatments designed by good clinicians
which are a mix of effective elements, personal preferences which may work in
some hands and not others, harmless but ineffective ideas, and techniques which are
actually harmful at times. Progress will occur when we link a better understanding of
the underlying psychopathology to a specific treatment for it.
3.5. Specificity in the Treatment Process

Assuming that we can accurately diagnose and assess the underlying mechanisms
of an individual's PD, we need to specify which psychotherapy techniques to use. How
do we promote change? Psychodynamic therapists often report that therapy goes
through phases of providing emotional support and developing a common under-
standing of the patient's problems as part of building a positive therapeutic alliance.
The therapeutic alliance (TA) is a prerequisite for change, providing the "glue" which
holds the patient in treatment despite the urge to run away from problems, and pro-
vides a buffer that mitigates the distress over seeing one's problems clearly. As therapy
progresses, the emphasis shifts more toward understanding maladaptive patterns, their
origin, and what makes it difficult to change (resistance). Accurate and empathic inter-
pretations then playa special role. Research is increasingly capable of providing a sys-
tematic picture of the past events, current patterns and their underlying mechanisms
which clinicians can incorporate in the process of developing their interpretations. Here
the goals of good clinical treatment and research converge.
Research should also be capable of delineating the pathways by which an indi-
vidual improves. For example individuals who act out may first substitute reaction for-
mation and eventually use self-assertion as they improve in dealing with stress and
conflicts. Psychotherapy research eventually should provide a scientific description of
the patient's dynamics and what to expect in treatment, knowledge which the therapist
can use to help guide the elements of therapy. It will incorporate knowledge developed
by studying patients with specific profiles of personality traits (PTs) or PDs, but be
capable of describing what is unique to each individual.
Of course psychotherapy will always require more than applying a bag of tech-
niques based on an accurate picture of the patient. We need a better understanding of
the role of so-called non-specific factors which works alongside these specific tech-
niques. All clinicians know that there is much about the patient-therapist relationship
that facilitates therapy. For instance, whenever validating the meaning of present events
vis a vis earlier formative experiences, the therapist paradoxically also promotes revis-
ing those meanings and holds out the possibility of choosing a more adaptive way to
cope in the present. This is at least as tricky as coaxing a toddler to try walking again
after soothing him or her following a painful spill! Thus we will never do without indi-
viduals who naturally show care, concern and a genuine wish to be helpful while also
having the capacity to endure someone else's distress as they learn to surmount their
problems.
It is clear that the research and clinical agendas to improve the psychotherapy of
PDs are convergent. We now have many of the tools needed to design and conduct the
studies that will make tomorrow's treatments more specific and better than today's. We
will have to conduct long-term follow-up of our patients to document that change is
stable and permanent rather than temporary. Finally, we will need to take our improved
treatments and demonstrate that they are cost-effective to society. With the mounting dif-
ficulty of paying for health services worldwide, cost-effectiveness analyses will become an
increasingly important aspect of evaluative research of psychotherapy.
3.6. Improving Clinical Care: Not Just a Manual

Currently a great deal of emphasis has been placed on developing manual-based
treatments for psychiatric disorders, and on assessing the adherence and competence of
therapists conducting those treatments (Barber and Crits-Christoph, in press). Despite
the accomplishments of manual-based treatments, we are concerned that their use in the
heterogeneous population of PDs may prove to be too inflexible or static for both
patients and therapists. In this case, therapists might learn a manual-based treatment, but
then never apply it in their practices outside of a treatment study.
We look to a growing trend in which researchers help clinicians design treatments.
We need a new generation of approaches that will marry individual patient assessment
with specific treatment guidelines in order to make clinicians' therapies better targeted
to a patient's problems, less likely to provoke negative therapeutic reactions, and more
efficient. It will be crucial to have clinicians inform researchers of their experiences in
applying the treatment guidelines in order continually to adjust the guidelines based
on real experience. A working partnership or collaboration between researchers and
clinicians will facilitate improvement in clinical care as well as in the training of future
therapists.
4. DEFENSE MECHANISMS EVALUATION MAKES THE

TREATMENT OF PERSONALITY DISORDERS MORE
SPECIFIC AND EFFECTIVE
The main aim of researchers should has to help clinicians to focus on mechanisms
of personality functioning when managing or treating the patients and to determine
more precisely which patients will benefit from a psychotherapy. As we have seen, there
is more than a good reason to consider empirical evaluation of DMs an indispensable
requisite for: a) choosing patients indicated for psychotherapy in the contest of a
logic of cost-benefit; b) evaluating, on the basis of the defensive style variations, the
treatment efficacy tailored on every single patient.
Notwithstanding the major debates coursing through modern psychoanalysis con-
cerning the structure and nature of the psychic apparatus and the causes of therapeu-
tic change, the overall concept of ego defenses has proved to be comparatively free of
controversy and an integral part of disparate theoretical persuasions. In the final analy-
sis, this attests to the central role an understanding and recognition of ego defenses has
in helping to make sense out of the often bewildering complexity of the individual
clinical situation.
According to Paris (1996), therapeutic alliance, functional level, and defensive
style are the main parameters we consider when we assess treatability in personality
Psychodynamic Research and Improved Diagnosis and Therapy for Personality Disorders 4S
disordered patients. The aim would be to carry out a triage among those who are
untreatable, those who require primarily support and crisis intervention, and those who
can benefit from more extensive treatment methods.
Measures of the alliance are robust predictors of the outcome in psychotherapy
(Luborsky et aI., 1988). For example we know that the alliances of patients with BPD
are very fragile (Frank, 1992) and over half of these patients become early dropouts
from psychotherapy (Skodol et aI., 1983; Gunderson et aI., 1989). Even if TA strength
probably represents the most effective outcome index in psychotherapy (Marmar and
Gaston, 1988) there are no ways to calculate this index before the treatment has been
started!
A number of researchers have consistently demonstrated a significant correlation
between pre-treatment functional levels, measured with GAF (Global Assessment
Functioning, DSM) or HSRS (Health-Sickness Rating Scale; Luborsky, and Bachrach,
1975), and therapy outcome (Kernberg et aI., 1972; Luborsky et aI., 1988; Propst et aI.,
1994). The significant variability and the poor correspondence between functioning
levels and diagnosis make this index not significant by itself.
Examining the association between PDs-PTs and defense mechanisms is very
helpful in conducting psychotherapy and in predicting TA strength and dropout behav-
iors. The defensive style is predictive of treatment response and gives us important
informations about patient's functional level. This approach, following the leaning
emerged in late Sixties, testifies of a shift from the high level of inference of psycho-
analysis to more phenomenological models, such as the DSM approach and the
research oriented approach.
The course of DMs over the years of psychotherapy is a very empirical way to
evaluate therapy efficacy. In a qualitative-few cases study, we have followed the course
of DMs over 1 to 2 years of psychotherapy. Defenses were measured at the beginning
of the therapy, at 1, and at 2 years into the therapy, respectively.
Defenses were rated using DMRS-Defense Mechanism Rating Scale (Perry,
1991,1992; Perry and Cooper, 1989; Perry et aI., 1993; Perry and Kardos, 1995).
The DMRS is an observer-based method which measures the use of DMs from
clinical interviews or therapy sessions transcripted or video and/or audio recorded. It
comprehends 28 individual defense mechanisms, hierarchically ordered in 7 clusters
from the less mature defenses (Action Defenses) to the most mature ones (Table 1).
Table 1. The DMRS hierarchy of adaptation: defense and defense levels

Defense Levels Defenses
Level 7 = Mature (High adaptive level) Affiliation, altruism, anticipation, humor, self-assertion, self-
observation, sublimation,suppression
Level 6 = Obsessional Isolation, intellectualization, undoing
LevelS = Other Neurotic a) Repression, Dissociation,
b) Reaction formation, displacement
Level 4 = Minor Image-Distorting Omnipotence, idealization, devaluation
Level 3 = Disavowal Denial, projection, rationalization, (fantasy is scored at this
level)
Level 2 = Major Image-Distorting Splitting of other's images, splitting of self-images, projective
identification
Levell = Action Acting out, passive aggression, help reject complaining
(hypochondriasis)
46 J. C. Perry et aL
A qualitative and quantitative scoring yield a final profile which classes the subject on
a scale (range 0 to 7), measuring the Overall Defensive Functioning.
One ofthe questions posed by this pilot study were whether and how DMs change
overtime.
Figure 1 shows the DMs profile of the patient MB at three different times of her
psychotherapy. MB is a white, single, female patient, 30 years old, with no Axis I diag-
nosis, and histrionic (HPD) and narcissistic (NPD) personality disorders on Axis II,
involved in a twice a week expressive psychotherapy. At the beginning of the therapy
her GAF was around 50. T1 is the DMRS evaluation during the clinical assessment, T2
is the evaluation after 1 year and T3 after 2 years of psychotherapy. The X Axis reports
the 7 levels of the hierarchy of defenses. The Y Axis reports the percentage of defenses
used by the patient during the DMRS evaluations. Low level defenses decrease and
disappear overtime, living their place to more adaptive defenses.
Commenting a similar ongoing study of a small sample of patients, Perry notes
that "defenses have a fortunate characteristic from a research point of view. Because
they are a basic building block of psychodynamic psychopathology, they function as
markers of the patient's functioning in psychotherapy, which is useful for process
research. However, when summed up over a session or series of sessions, they also serve
as an outcome measure representing how the person is doing over that same time
period" (Perry et ai., 1993).
In this kind of psychotherapy pilot studies, researchers can begun to examine the
relationship between DMs, motives (wishes and fears), conflicts, therapeutic alliance,
and outcome assessed by other measures of psychopathology. It is also possible to
examine the relationship between the patient's defenses and the therapist's interven-
54
52
50
48
46
44
42
40
38
36
34
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
o
Action Borderline Disavowal Narcissistic Neurotic Obsessional Mature
Figure 1. DMs profile of the patient MB at three different times of her psychotherapy.
tions and subsequent outcomes. This data will help us develop hypotheses about some
old questions (Perry and Cardos, 1995):
- Is there a differential relationship between the patient's defenses used at one
point in a session and is it more effective for the therapist to address the
defense or the underlying motive?
- To what extent does defensive function change in a successful therapy?
- As the use of less adaptive defenses decreases, are they supplanted by spe-
cific defenses that are more adaptive?
- Is there a differential relationship between the prominent constellation of
defenses (such as the major or minor image distorting defenses) and the mix
of therapeutic techniques that predict successful outcomes?
- To what extent is the maturity of defenses at the start of therapy related to
the patient's ability to form and maintain a therapeutic alliance?
Although questions such as these will require a generation of research, the advent
of methods such as the DMRS makes it possible to address these clinically important
issues. Whereas recent papers describe how clinicians should deal with patient's spe-
cific defenses (Perry and Cooper, 1987; Vaillant, 1992), research on defenses in psy-
chotherapy also should provide an empirical foundation for what is now left to clinical
acumen.
5. DEFENSE MECHANISMS AND PERSONALITY DISORDERS:

AN EMPIRICAL RESEARCH
In conclusion of this chapter we present some preliminary data of an ongoing

research project aimed to evaluate DMs (and other indexes) role in the decision-
making process about psychotherapy referral. We are also studying these data in rela-
tion to dropout phenomenon and TA indexes. This research project is part of an
international multicentric project directed by Christopher Perry in collaboration with
Jewish General Hospital-McGill University (Montreal, Canada); Austin Riggs Center
(Massachusetts, USA); Medical Psychology and Psychotherapy Unit, Neuropsychiatric
Science Department, San Raffaele Hospital, University of Milan (Italy)*; Mother-
Infant Treatment Program (Geneva, Switzerland); Department Universitaire Psichia-
trique B (Losanne, Switzerland). We are evaluating a sample of 100 adult outpatients
consecutively admitted to clinical assessment and seeking psychotherapeutic treatment
at the Psychotherapy Unit of S. Raffaele Hospital, University of Milan. We want to
study the association between: Axis II diagnoses (DSM-IV, SCID-II), defense mecha-
nisms (DMRS), psychopathological symptoms (SCL-90), level of social functioning
(GAF). In a second step, these indexes are compared with dropout phenomenon fre-
quency and TA characteristics (CALPAS).
* The Italian research group is coordinated by Vittorio Lingiardi, MD and Laura Vanzulli, MD, and com-
posed by: Clizia Lonati, MD, Daniela Croce, PhD, Francesca Beretta, PhD, Massimiliano Simula, PhD, Chiara
Pozzi, PhD, Stefania Roberti, PhD, Francesca Delucchi, medical school student. The Psychotherapy Depart-
ment of the San Raffaele Hospital, University of Milan, is directed by Cesare Maffei, MD. We thank Andrea
Fossati, MD, for statistical consultation.
As we have already mentioned DMs have been assessed by a group of clinical

psychologists and psychiatrists trained in using DMRS. IRR coefficient of individual
overall score (which represents the overall defensive functioning of the individual) was
0.87. IRR coefficients of subtotal scores (relating to the 7 defenses clusters) ranged
from 0.61 to 0.95.
TA has been evaluated using the California Psychotherapy Alliance Scale, both
Patient and Therapist versions (CALPAS, Marmar and Gaston, 1988). The CALPAS is
a 24 items-7 point scale, filled in at 5th, 10th, and 15th therapy sessions, assessing 4
relatively independent alliance dimensions: Patient Commitment (PC); Patient
Working Capacity (PWC); Working Strategy Consensus (WSC); Therapist Under-
standing and Involvement (TUI).
In a further contribution we will publish the final results of our research. At the
present time we evaluated 26 adult outpatients (gender = 60% female, 40% male; age
= 20 to 45 years old). They all have a secondary school license. GAF range from 40 to
82 (mean 63.07). In collecting the sample we excluded subjects < 18 > 45 years old, with
presence of organic diseases or a major psychotic syndrome, and IQ < or = 75. The
assessment of these patients has been completed before starting psychotherapy. DMs
evaluation has been conducted blind to DMS-IV diagnoses. All the patients have signed
a consensus form.
Tables 2 and 3 show the diagnostic and the DMs distribution of our sample. Of
the sample, 54% has an Axis I diagnosis, mostly Anxiety Disorders and Eating Disor-
ders, and around 80% has an Axis II diagnosis. Borderline personality disorder is the
most represented Axis II "pure" diagnosis, but around 23% of the sample has a cluster
B codiagnosis. The sample size precludes an extensive examination of the relationship
between DMs and Axis I1Axis II disorders and of the effects of comorbidity. However,
as we can see in Table 3, defensive functioning tends to be healthier in patients with
only PTs than in patients with full diagnoses of PDs. In both cases the medium level
of defenses is the most represented, but the general trend moves to an opposite direc-
Table 2. Diagnostic distribution

Axis I % N (26)
No diagnosis 46 12
Anxiety Disorder 27 7
Eating Disorder 11.5 3
Mood Disorder 7.6 2
Anxiety + Mood Disorders 3.8 1
Substance Abuse 3.8
Axis II % N (26)
No diagnosis (only PTs) 19.2 5
BPD 19.2 5
NPD 7.7 2
AVPD 7.7 2
OCPD 3.8
DPD 3.8
NePD 3.8
NASPD 3.8
Codiagnosis: 30.8 8
Mainly Cluster B 23 6
Table 3. Defense mechanisms distribution

PDs (full diagnoses)
DMs %±s.d. PTs
N26 N21 N5
High level (7) 0.72 2.86
Mature 0.72-± 1.62 2.86 ± 6.39
Medium level (4-6) 56.15 64.93
Obsessive 22.02 ± 12.64 33.04 ± 7.49
Neurotic 13.72 ± 8.01 15.71 ± 14.12
Narcissistic 20.42 ± 10.07 16.17 ± 6.76
Low level (1-3) 43.14 32.26
Disavowal 25.55 ± 9.3 21.82 ± 10.45
Borderline 7.56 ± 4.81 5.41 ± 6.4
Action 10.02 ± 7.71 5.02 ± 5.21
Overall Defensive Functioning 3.89 ±0.4 4.42 ± 0.57
range 3.18 to 4.57 range 4.07 to 5.43
Number of Defenses per Session 35.50 ± 13.56 33.60 ± 10.21
tion for the two diagnostic classes. Mature defenses rate is higher in PTs and low level
defenses rate is higher PDs.
We tested also if specific defenses were significantly associated with certain PTs.
The findings are consistent with the idea that specific defenses may function as mech-
anisms underlying the surface behaviors and personality traits which clinicians find so
difficult to treat. Generally, action defenses are positively correlated with all cluster B
disorders. These findings partially replicates Perry and Cooper (1989) and Perry et aI.
(1993, 1995) ones, confirming that the cluster B patients use the lowest and the most
unstable defenses. An other interesting result is the correlation between HPT and affil-
iation, whose meaning is probably due to the seductive capacity of the histrionic patient
in asking for help, bringing the rater to mistake the patient attitude for affiliation. The
positive correlation between OCPT (also significant in using the defense of humor) and
narcissistic defenses (in particular devaluation) shows the underlying connection
between obsessive and narcissistic traits. Generally we can say there is some relation
between certain defensive styles and certain Axis II disorders, yet this is not a rigid
association. The two categories remain independent: DMs cannot be considered as the
dynamic version of PDs. Findings however suggest that changes in defenses will be
useful indicators of improvement with treatment.
We also found a correlation between DMs and psychopathological acute
symptoms as referred using SCL-90. In particular, action defenses have a positive
correlation with almost all SCL-90 symptom categories. Developing such an acute
symptomathology in presence of a PT-PD could be the result of the employment of
maladaptive defenses. This hypothesis is confirmed by the observation that the trait-
symptom correlation falls if we don't consider low-level defenses in our statistical
elaboration.
We finally considered our sample form the point of view of the dropout
behavior and the TA. We already knew (Skodol et aI., 1983; Waldinger and Gunderson,
1984; Gunderson et aI., 1989; Yeomans et aI., 1992) dropout is a relevant and early
phenomenon in PDs psychotherapy. This part of our research project was designed to
50 J. C. Perry et aL
study the aSSOCiatIOn between TA measures (evaluated using CALPAS-P, and

CALPAS-T), DMs and dropout behavior in adult outpatients with Personality Traits
(PTs) or Personality Disorders (PDs), involved in a long term expressive/supportive
psychotherapy. Comparing CALPAS-P and CALPAS-T, we found that therapists'
alliance rates are commonly lower than patients' ones. This disagreement increases in
patients dropping out very early. WSC-T (Working Strategy Consensus evaluated by
the Therapist) is the best predictor of early dropout, but also TUI-T (Therapist Under-
standing and Involvement evaluated by the Therapist) and TUI-P (evaluated by the
Patient) seem to be significant in discriminating between remaining in therapy vs drop-
ping out patients. Cluster A disorders are associated with lower PWC-T (Patient
Working Capacity evaluated by the therapist), whereas cluster B disorders correlate
with higher PWC-T and WSC-T rates. When cluster C disorders are present, all patients
alliance indexes, PWC-T and PC-T (Patient Committment) are higher. A good DMs
level influences higher PWC-T and TUI-T. Obsessive and Neurotic Defenses are asso-
ciated with higher PWC-T, WSC-P, and WSC-T, but with lower PC-P and TUI-P. Nar-
cissistic Defenses are associated with higher PWC-P and Borderline Defenses correlate
with higher WSC-P and TUI-P.
Our findings support the hypothesis that TA is a good predictor index of dropout
behaviors in the very early phases of therapy with PDs patients. In particular, the level
of agreement on tasks and goals of therapy evaluated by the therapist and the thera-
pist understanding and involvement capacity are the dimensions that more easily dis-
criminate treatment outcome. Therapist evaluations seem more "prudent" than patient
evaluations and they have more predictive power. This discrepancy in the rates is even
more evident in dropout cases. Axis I and Axis II diagnoses don't explain dropout phe-
nomenon, but the underlying presence of certain PTs and DMs has a specific impact
on the quality of TA.
6. CONCLUSIONS
Recent years have brought a mess of new studies on DMs. Despite the different
approaches, all the methods seem to produce similar findings and the different samples
suggest a robust relationship between adult personality functioning, defensive style and
psychopathology. Considering defenses along with the other traditional features of a
phenomenological diagnosis could be the most useful way of integrating a dynamically
oriented clinical approach with the support of empirical research. There is empirical
evidence that individual defenses have meaningful relationship to psychopathology and
adaptation. In addition, specific defenses are associated with specific symptoms and
with psychosocial role impairment.
Our evidence suggests that mature defenses exert some important role in
healthy functioning and that immature and image-distorting defenses are associated
overall with higher general levels of symptoms and impairment in psychosocial
functioning.
The consideration of individual defenses and global defensive functioning of
our patients appears to be very useful in the devising of more effective treatment
interventions and in planning psychotherapies more tailored on the individual
patient.
A new impulse for the scientific study of defense mechanisms has derived from the
challenge that PDs poses both to clinicians and theorists. To some extent, the consider a-
tion of the psychic life of borderline and narcissistic patients has shifted the researcher's
focus from the mere intrapsychic reality of defenses to a more interpersonal context. For
example, unsatisfactory or traumatic environmental situations have been considered
important factors in generating a pathological use of defenses. Finally, DSM phenome-
nological approach and its request for a diagnostic standardization has influenced the
empirical study of defense mechanisms. Some recent contributions to the empirical
research in PDs assessment and psychotherapy seem to show a correlation between
defensive functioning, PTs and PDs, and dropout in psychotherapy. This approach, fol-
lowing the leaning emerged in late Sixties, testifies of a shift from the high level of infer-
ence of psychoanalysis to more phenomenological models, such as the DSM approach
and the research oriented approach. Defense mechanisms and therapeutic alliance
indexes warrant further attention as we design treatments for PDs.
REFERENCES
Banon, E., Perry, IC., and Ianni, F., 1995, The Effectiveness of Psychotherapy for Personality Disorders:
Recommendations for Future Study, Poster presented at the 4th International Congress on Personality
Disorders, Trinity College, Dublin Ireland, June 22, 1995.
Barber, IP., Crits-Christoph, P., and Luborsky, L., 1996, The effects of therapist adherence and competence
on patient outcome in brief dynamic therapy. J Consult and Clin Psychology. In press.
Crits-Christoph, P., 1992, The efficacy of brief dynamic psychotherapy: a meta-analysis. Am J Psychiatry,
149:151-158.
Fossati, A, Madeddu, F., and Maffei, C., Borderline Personality Disorder and Childhood Sexual Abuse: A
Meta-analytic Study. J Pers Dis, in press.
Frank, AF., 1992, The therapeutic alliances of borderline patients. In: Clarkin, IF., Marziali, E., and
Munroe-Blum, H., (eds), Border Personality Disorder: Clinical and Empirical Perspectives. Guilford,
New York.
Gunderson, IG., Franck, AF., Ronningstam, E.F., Wahter, S., Lynch, V.l, and Wolf, P.I, 1989, Early discon-
tinuance of borderline patients from psychotherapy. J of Nervous Mental Diseases, 177:38-42.
Hirschfeld, R.M.A, 1993, Personality disorders: Definition and diagnosis. J of Personality Disorders, supp!.
7:9-18.
Kernberg, O.F., Coyne, L., Appelbaum, A, Horwitz, L., and Voth, H., 1972, Final report of the Menninger
Psychotherapy Research Project. Bulletin of the Menninger Clinic, 36:1-275.
Lambert, M., 1994, The effectiveness of psychotherapy. In: Bergin, AE. and Garfield, S.L., (eds), Handbook
of Psychotherapy and Behavior Change, Fourth ed. John Wiley and Sons, New York.
Linehan, M., Heard, H., and Armstrong, H.E., 1993, Naturalistic follow-up of a behavioral treatment for
chronically parasuicidal borderline patients. Arch Gen Psychiatry, 50:971-974.
Lingiardi, v., Vanzulli, L., Madeddu, F., Croce, D., and Maffei, c., 1993, Borderline Personality Disorder Psy-
chotherapy and Very Early Drop Out Rate: a preliminary report. 3rd International Congress on Per-
sonality Disorders. September 8-111993, Cambridge, USA
Luborsky, L., Crits-Christoph, P., Mintz, I, and Auerbach, A, 1988, Who Will Benefit from Psychotherapy?
Basic Books, New York.
Luborsky, L. and Bachrach, H., 1974, Factors influencing clinicians judgments of mental health: Eighteen
experiences with the Heath-Sickness Rating Scale. Arch Gen Psychiatry, 31:292-299.
Luborsky, L., Diguer, L., Luborsky, E., Singer, B., Dickter, D., and Schmidt, K., 1993, The efficacy of dynamic
psychotherapies: Is it true that everyone has won all must have prizes? In: Miller, N., Luborsky, L.,
Barber, I, and Docherty, I, (eds), Psychodynamic Treatment Research: A Handbook for Clinical
Practice. Basics Books, New York, pp. 497-516.
Marmar, G. and Gaston, L., 1988, Manual for the California Psychotherapy Alliance Scale CALPAS. Unpub-
lished manuscript. Department of Psychiatry, U of California, S. Francisco.
McGlashan, T.H., 1986, The Chestnut Lodge Follow-up Study: Long term outcomes of borderline personal-
ities. Archives of General Psychiatry, 43:20-30.
Oldham, 1M., Skodol, AE., Kellman, H.D., Hyler, S.E., Doidge, N., Rosnick, L., and Gallaher, P.E., 1995,
Comorbidity of Axis I and Axis II disorders. Am J Psychiatry, 152:571-578.
Paris, 1,1996, Social Factors in the Personality Disorders. Cambridge University Press.
Paris, 1, Brown, R., and Nowlis, D., 1987, Long term follow-up of borderline patients in a general hospital.
Comprehensive Psychiatry, 28:530-535.
Paris, 1, Nowlis, D., and Brown, R, 1988, Developmental factors in the outcome of borderline personality
disorder. Comprehensive Psychiatry, 29:147-150.
Paris, 1, Nowlis, D., and Brown, R, 1989, Predictors of suicide in borderline personality disorder. Canadian
J of Psychiatry, 34:8-9.
Perry, Ie., 1991, Defense Mechanisms Rating Scale. Cambridge Hospital-Harvard Medical School, Boston.
Perry, Ie., 1992, Problems and considerations in the valid assessment of Personality Disorders. Am J
Psychiatry, 149:1645-1653.
Perry, Ie., 1993, Longitudinal studies of Personality Disorders. J Pers Dis, 7 suppl:63-85.
Perry, Ie. and Cooper, S.H., 1987, Empirical studies of psychological defense mechanisms. In: Michels, R.
and Cavenar, 10., (eds.), Psychiatry 1 (30). New York, JB Lippincot and Basic Books, 1-19.
Perry, Ie. and Cooper, S.H., 1989, An empirical study of defense mechanism. Clinical interview and life
vignette ratings. Arch Gen Psychiatry, 46:444-452.
Perry, IC. and Kardos, M.E., 1995, A Review of the Defense Mechanism Rating Scales. In: Conte, H.R
and Plutchick, R, Ego Defenses. Theory and Measurement. John Wiley and Sons, New York,
pp.283-299.
Perry, Ie., Kardos, M.E., and Pagano, e.I, 1993, The study of defenses in psychotherapy using the Defense
Mechanism Rating Scales DMRS. In: Hentschel, U. and Ehlers, W., (eds), The concept of defense mech-
anisms in contemporary psychology: Theoretical, research, and clinical perspectives. Springer, New
York.
Perry, Ie., Lavori, P.H., and Hoke, L., 1987, A Markov model for predicting levels of psychiatric services use
in Borderline and Antisocial Personality Disorders and bipolar type II affective disorder. J Psychiatr
Res, 21:215-232.
Perry, Ie. and Vaillant, G.E., 1989, Personality Disorders. In: Kaplan, H.I. and Sadock, B.I, (eds), Com-
prehensive Textbook of Psychiatry, Fifth Edition, vol. II. Baltimore: Williams and Wilkins Co.,
pp.1352-1387.
Plakun, E.M., Burkhardt, P.E., and Muller, IP., 1986, 14 year follow-up of borderline and schizotypal per-
sonality disorders. Comprehensive Psychiatry, 27:448-455.
Propst, A., Paris, 1, and Rosberger, Z., 1994, Do therapist experience, diagnosis and functional level predict
outcome in short-term psychotherapy? Canadian J of Psychiatry, 39:178-183.
Skodol, AE., Buckley, P., and Charles, E., 1983, Is there a characteristic pattern in the treatment history
of clinic outpatients with borderline personality? J of Nervous and Mental Diseases, 171:405-410.
Stone, M., 1990, The Fate of Borderline Patients. Guilford, New York.
Vaillant, G., 1992, The beginning of wisdom is never calling a patient a borderline; or, The clinical manage-
ment of immature defenses in the treatment of individuals with personality disorder. J of
Psychotherapy Practice and Research, 1 (2):117-134.
Waldinger, R. and Gunderson, IG., 1984, Completed psychotherapies with borderline patients. Am J
Psychotherapy, 38:190-202.
5
ATTACHMENT, THE DEVELOPMENT OF THE

SELF, AND ITS PATHOLOGY IN
Peter Fonagy
Freud Memorial Professor of Psychoanalysis

Director, Sub-Department of Clinical Health Psychology
University College London
Director of Research, Anna Freud Centre
Co-ordinating Director, Child and Family Centre and
Centre for Outcomes Research and Effectiveness
Menninger Foundation
Topeka, Kansas
1. INTRODUCTION
Attachment theory concerns the nature of early experiences of children and the
impact of these experiences on aspects of later functioning of particular relevance to
personality disorder. The question we attempt to address here is how deprivation, in
particular early trauma, comes to affect the individual's propensity to personality dis-
order. As part of this question we are naturally also concerned to understand how
such adverse consequences may be avoided. The key assumption made by the invok-
ing of attachment theory is that individual social behaviour may be understood in
terms of generic mental models of social relationships constructed by the individual.
These models, although constantly evolving and subject to modification, are strongly
influenced by the child's experiences with the primary caregivers. Let us now turn to
the details of the theory.
Address for correspondence: Sub-Department of Clinical Health Psychology, University College of London,
Gower Street, London WCIE 6BT, Tel: +44 171 391 1791, Fax: +44 171 916 8502, e-mail: [email protected].

54 P.Fonagy
2. THE NATURE OF THE ATTACHMENT SYSTEM
Attachment theory, developed by John Bowlby (1969, 1973, 1980), postulates a

universal human need to form close affectional bonds. It is a normative theory of how
the "attachment system" functions in all humans. Bowlby described attachment as a
special type of social relationship, paradigmatically between infant and caregiver,
involving an affective bond. More significantly, it may also be seen as the context within
which the human infant learns to regulate emotion (Sroufe, 1990).
The stability of early childhood attachment patterns is well demonstrated. Mary
Ainsworth and her colleagues (Ainsworth, 1985; Ainsworth and Wittig, 1969; Ainsworth
et aI., 1978) developed a procedure commonly known as the Strange Situation, which
classifies infants and toddlers into one of four attachment categories. Secure infants
explore readily in the presence of the primary caregiver, are anxious in the presence
of the stranger, are distressed by their caregivers departure and brief absence, rapidly
seek contact with the caregiver following a brief period of separation, and are reas-
sured by renewed contact. The recovery from an over-aroused disorganised state is
smooth and carried to completion in the sense that the infant returns to exploration
and play.
Some infants, who are usually made less anxious by separation, do not automat-
ically seek proximity with the caregiver on her return following separation and may
show no preference for the caregiver over the stranger; these infants are designated
"Anxious/Avoidant". A third category, the "Anxious/Resistant" infant manifest im-
poverished exploration and play, tend to be highly distressed by separation from the
caregiver, but have great difficulty in settling after reunion showing struggling, stiffness,
or continued crying, or fuss in a passive way. The caregiver's presence or attempts at
comforting fails to offer reassurance and their anxiety and anger appears to interfere
with their attempts to derive comfort through proximity. Both these insecure groups
appear to be coping with arousal and ambivalence through a precautious over-control
of affect because they appear to be uncertain in their expectation that the caregiver
will do his or her part to modulate their emotional arousal (Main and Weston, 1981;
Sroufe, 1990).
It is generally held that the patterning of attachment related behaviour is under-
pinned by different strategies adopted by children to regulate their emotional reac-
tions. As affect regulation is acquired with the help of the child's primary caregiver, the
child's strategy will be inevitably a reflection of the caregiver's behaviour towards
him/her. Secure infants' behaviour is based on the experience of well co-ordinated, pos-
itive interactions where the caregiver is rarely over-arousing and is able to restabilise
the child's spontaneously emerging disorganising emotional responses. Therefore, they
remain relatively organised in stressful situations. Negative emotions are not seen
as threatening in and of themselves but are regarded by the infant as serving a
communicative function (Grossman et aI., 1986; Sroufe, 1979, 1996).
By contrast, Anxious/Avoidantly attached children are presumed to have experi-
ences when their emotional arousal was not restabilised by the caregiver because of
personal or social pressures on the caregiver and an associated mild neglect or even
resentment of the child. The same expectations may arise in children who were over
aroused through intrusive parenting, therefore they over-regulate their affect and steer
away from situations that are likely to be emotionally arousing. Anxious/Resistantly
attached children under-regulate, heightening their expression of distress possibly in
an effort to elicit the expectable response of the caregiver. These children have low
Attachment and Its Pathology in Personality Disorders ss
thresholds for threat and may become preoccupied with having contact with the care-
giver, but show signs of frustration regarding this contact even when it is available
(Sroufe, 1996).
A fourth group of infants appear to exhibit a range of seemingly undirected behav-
ioural responses giving the impression of disorganisation and disorientation (Main and
Solomon, 1990). Infants who manifest freezing, handclapping, headbanging, the wish to
escape from the situation even in the presence of the caregiver, are referred to as "Disor-
ganised/Disoriented". It is generally held that for such infants the caregiver has served
both as a source of fear and as a source of reassurance, thus the arousal of the attachment
behavioural system produces strong conflicting motivations. Not surprisingly, a history of
severe neglect or physical or sexual abuse is often associated with the manifestation of
this pattern (Cicchetti and Beeghly, 1987; Main and Hesse, 1990).
2.1. Determinants of Attachment Security

It is beyond the scope of this paper to consider in detail the rich literature on
determinants of infant security.1 Clearly genetic transmission may account for some
component of the prediction from parental attachment status to the child's security of
attachment (van Ijzendoorn, 1992). The influence of temperament on attachment secu-
rity is controversial, but the balance of the evidence is now against the appropriateness
of a temperamental account (Kagan, 1982; Lamb et aI., 1984).2
The quality of maternal care has been repeatedly shown to predict infant secu-
rity. The sensitive responsiveness of the parent is traditionally regarded as the most
important determinant of attachment security in the infant (Isabella, 1993; Isabella and
Belsky, 1991). The parameters assessed include: ratings of maternal sensitivity (e.g. Cox
et aI., 1992; Isabella, 1993), prompt responsiveness to distress (Del Carmen et aI., 1993),
moderate stimulation (Belsky et aI., 1984), non-intrusiveness (Malatesta et aI., 1986),
interactional synchrony (Isabella et aI., 1989), warmth, involvement, and responsive-
ness (O'Connor et aI., 1992). These associations have been strengthened by findings
from experimental studies, where the enhancement of maternal sensitivity has been
shown to increase the proportion of secure infants in high-risk populations (van den
Boom, 1995). Similar parameters have been predictive for fathers (Cox et aI., 1992) and
for professional caregivers (Goosens and van Ijzendoorn, 1990).
Negative parental personality traits are associated with insecurity in many studies,
although by no means all (Zeanah et aI., 1993). This has been shown for anxiety (Del
Carmen et aI., 1993), aggression (Maslin and Bates, 1983) and suspicion (Egeland and
Farber, 1984). Parental psychopathology is also found to be a risk factor in some studies
(Campbell et aI., 1993). Of the contextual factors, support from the partner (Goldberg
and Easterbrooks, 1984) and from others in the mother's environment (Crnic et aI.,
1983) appear important. The strength of these associations is reinforced by experi-
mental studies where social support was systematically manipulated (Lyons-Ruth et aI.,
1990; Jacobson and Frye, 1991; Lieberman et aI., 1991).
IThere are many excellent reviews available, notably by Belsky (Belsky et aI., 1995).
2There is little evidence that distress-prone infants become anxious-resistant babies (van den Boom, 1990).
Temperament changes in the first year of life (Belsky et aI., 1991) and the attachment pattern of a child to
his two parents is often inconsistent (Fox et aI., 1991) and appears to be dependent on the internal working
model of each parent (Steele et aI., 1996).
56 P.Fonagy
These predictors of infant security are correlated to one another and are all
likely to be unequally distributed across socio-economic groups. It is known that
socio-economic status and other indicators of social deprivation are linked to both
infant and adult classifications (e.g. Ward and Carlson, 1995; van Ijzendoorn and
Kroonenberg, 1988; Crittenden et aI., 1991; Zeanah et aI., 1993). Poor parenting skills
and the maltreatment of children are more common in families suffering economic
hardship (Gabarino, 1992). Insecure classification is more common in deprived groups.
Maltreatment of children, strongly associated with economic deprivation (Belsky, 1993)
is most likely to be associated with the disorganised/disoriented pattern of infant
attachment.
2.2. The Continuity of Patterns of Attachment

Bowlby proposed that the quality of childhood relationships with the caregivers
results in internal representations or working models of the self and others that provide
prototypes for later social relations. Internal working models are mental schemata,
where expectations about the behaviour of a particular individual toward the self are
aggregated. The expectations are themselves abstractions based on repeated interac-
tions of specific types with that individual. If the child's physical injury is quickly dealt
with, sources of unhappiness are rapidly addressed, the child will develop the legiti-
mate expectation that, with that person at least, his distress is likely to be met by reas-
surance and comforting. The internal working model is the result of a natural process
of abstraction of the invariant features from diverse social situations with a particular
individual (Stern, 1994).
Such internal models of attachment remain relatively stable across the lifespan
(Collins and Read, 1994). Secure children, with the benefit of well-regulated caregiver-
infant relationships behind them, are expected to evolve positive expectations con-
cerning their exploratory competence, to achieve a reliable capacity for modulation of
arousal, a good capacity for communication within relationships and, above all, confi-
dence in the ongoing availability of the caregiver. Early experiences of flexible access
to feelings is regarded as formative by attachment theorists, enabling secure children
both to maximise the opportunities presented to them by the environment and draw
on socially supportive relationships. The autonomous sense of self emerges fully from
secure parent-infant relationships (Emde and Buchsbaum, 1990; Lieberman and Pawl,
1990; Fonagy et aI., 1995). The increased control of the secure child permits it to move
towards the ownership of inner experience and come to recognise the self as com-
petent in eliciting regulatory assistance, to develop metacognitive control and to
achieve an understanding of self and others as intentional agents whose behaviour is
organised by mental states, thoughts, feelings, beliefs, and desires (Sroufe, 1990; Fonagy
et aI., 1995).
2.3. Prediction from Adult Attachment Measures

The stability of these attachment assessments are dramatically illustrated by lon-
gitudinal studies of infants assessed with the Strange Situation and followed up in ado-
lescence or young adulthood with the Adult Attachment Interview (AAI) (George
et aI., 1985). The AAI asks subjects about childhood attachment relationships and the
meaning which an individual currently gives to attachment experiences. The instrument
is rated according to the scoring system developed by Main and Goldwyn (1994) which
classifies individuals into Secure/Autonomous, Insecure/Dismissing or Avoidant,
Attachment and Its Pathology in Personality Disorders 57
Insecure/Preoccupied or Resistant, or Unresolved/Disorganised with respect to loss or

trauma, categories according to the structural qualities of their narratives of early ex-
periences. (While autonomous individuals clearly value attachment relationships and
regard these as formative, insecure individuals are poor at integrating memories of
experience with their assessment of the meaning of that experience. Those dismissing
of attachment deny or devalue early relationships. Preoccupied individuals tend to be
confused, and angry or passive in their current relationships with their parents and
others). Two studies (Hamilton, 1994; Waters, et aI., 1995) have shown a 68-75% cor-
respondence between attachment classifications in infancy and classifications in adult-
hood. This work speaks to the remarkable stability of attachment classifications
across the lifespan. Similar findings are beginning to emerge using other measures of
attachment in adults (Hazan and Zeifman, 1994).
3. THE TRANS GENERATIONAL TRANSMISSION OF

ATTACHMENT PATTERNS
There is further important evidence that attachment relationships may playa key
role in the transgenerational transmission of hardship and deprivation. Individuals cat-
egorised as secure are 3 or 4 times more likely to have children who are securely
attached to them (van Ijzendoorn, 1995). This turns out to be true even in prospective
studies where parental attachment is assessed before the birth of the child (Benoit and
Parker, 1994; Fonagy et aI., 1991; Radojevic, 1992; Steele et aI., 1996; Ward and Carlson,
1995). These findings also emphasise the importance of quality of parenting in deter-
mining the child's attachment classification. The findings from our lab suggest that
parental attachment patterns predict variance in addition to temperament measures
or contextual factors, such as experience, social support, marital relationship, psy-
chopathology, and personality (Steele, Steele, and Fonagy, in preparation). If attach-
ment is linked to personality disorder we may anticipate a substantial overlap between
determinants of infant security and long-term predictors of criminality.
3.1. Attachment and Mentalizing

A compelling model for the transmission of secure attachment, which has moved
the field beyond a simple view of caregiver sensitivity, was suggested by Mary Main
(1991) in her seminal chapter on metacognitive monitoring and singular versus multi-
ple models of attachment. Main (1991) showed that the absence of metacognitive
capacity, the inability to "understand the merely representational nature of their own
(and others') thinking" (p. 128), makes infants and toddlers vulnerable to the incon-
sistency of the caregiver's behaviour. They are unable to step beyond the immediate
reality of experience and grasp the distinction between immediate experience and the
mental state which might underpin it. Main drew our attention to the development in
the child of the mental state that Dennett (1987) called "the intentional stance".
Dennett stressed that human beings are perhaps unique in trying to understand each
other in terms of mental states: thoughts, feelings, desires, beliefs, in order to make sense
of and, even more important, to anticipate each other's actions. It is self-evident that
by attributing an emotional or cognitive state to others we make their behaviour explic-
able to ourselves. If the child is able to attribute a withdrawing, non-responsive mother's
apparently rejecting behaviour to her emotional state of depression, rather than to
himself as bad and unstimulating, the child is protected from, perhaps permanent,
58 P.Fonagy
narcissistic Injury. Perhaps even more central is the child's capacity to develop
representations of the mental states, emotional and cognitive, which organise his/her
behaviour toward the caregiver.
We attempted to operationalise individual differences in adults' metacognitive
capacities which we believe might help to fill the "transmission gap". We were curious
to know if the extent of self-reflective observations about the mental states of self and
others in Adult Attachment Interview (AAI) narratives could predict infant security.
We chose the term "reflective self-scale" (as opposed to self-reflection scale) to under-
score that we were concerned about the clarity of the individual's representation of the
mental states of others as well as the representation of their own mental state.
Consistent with our expectation, reflective-self ratings were reliable (intraclass
r = 0.8 and above) and provided a good pre-natal prediction of the Strange Situation
behaviour of the child. Both fathers and mothers who were rated to be high in this
capacity were three or four times more likely to have secure children than parents whose
reflective capacity was poor. (Fonagy et aI., 1991).
The capacity for metacognitive control may be particularly important when the
child is exposed to unfavourable interaction patterns, in the extreme, abuse or trauma.
For example, in the absence of the capacity to represent ideas as ideas, the child
is forced to accept the implication of parental rejection, and adopt a negative view
of himself. A child who has the capacity to conceive of the mental state of the other
can also conceive of the possibility that the parent's rejection of him or her may be
based on false beliefs, and therefore is able to moderate the impact of negative
experience.
We examined this issue by administering a brief structured interview to parents
in our sample, 18 months after they had completed the Adult Attachment Interview,
concerning a number of simple indicators of family stress and deprivation which had
been reported in past studies to increase dramatically the probability of adverse
outcome, including, in a recent study, the likelihood of insecure infant attachment.
These indicators included: single parent families residing separately, overcrowding,
paternal unemployment, etc. We divided our sample into those who had reported sig-
nificant experience of deprivation (more than 2 items) and those who had not. Our
prediction was that mothers in the deprived group would be far more likely to have
children securely attached to them if their reflective-self rating (metacognitive
capacity) was high.
10 out of 10 of the mothers in the deprived group with high reflective-self ratings
had children who were secure with them, whereas only 1 out of 17 of deprived mothers
with low ratings did so. Reflective-self function seemed to be a far less important pre-
dictor for the non-deprived group. Our findings imply that the intergenerational repli-
cation of early negative experiences may be aborted, the cycle of disadvantage
interrupted, if the caregiver acquires a capacity to fully represent and reflect on mental
experience (Fonagy et aI., 1994).
3.2. Metacognitive Monitoring and the Development of the Self

Metacognitive monitoring completes one aspect of the intergenerational cycle.
Not only are parents high in reflective capacity mQre likely to promote secure attach-
ment in the child, particularly if their own childhood experiences were adverse, but
also secure attachment may be a key precursor of robust reflective capacity (Fonagy
et aI., 1995).
In London, we have collected cross-sectional data from 3-5 year olds which
appears to indicate a strong correlation between security on a projective measure of
attachment (the SAT) and the early development of a theory of mind, using the belief-
desire reasoning task. We found the children who were rated Secure on the Separa-
tion/Anxiety test were more likely to pass the theory of mind tasks (Fonagy et aI., 1997).
In a longitudinal study, we found that of 92 children, the 59 who had passed the task
at 5 years, 66% were secure at one year with their mother. Of the 29 who failed, only
31 % had been secure. Attachment security to father was less significantly associated
with greater competence at this task. There was clear indication that the reflective self
function of mother was associated with the child's success. 80% of children whose
mothers were above the median in reflective self function passed, whereas only 56%
of those whose mothers were below did so.
These results suggest that the parents' capacity to observe the child's mind facil-
itates the child's general understanding of minds mediated by secure attachment. The
availability of a reflective caregiver increases the likelihood of the child's secure attach-
ment which, in turn, facilitates the development of theory of mind. Throughout these
studies we assume that a secure attachment relationship provides a congenial context
for the child to explore the mind of the caregiver, and, as the philosopher Hegel (1807)
taught us, it is only through getting to know the mind of the other that the child devel-
ops full appreciation of the nature of mental states. The process is intersubjective: the
child gets to know the caregiver's mind as the caregiver endeavours to understand and
contain the mental state of the child.
The child perceives in the caregiver'S behaviour not only her stance of reflec-
tiveness which he infers in order to account for her behaviour, but also he perceives in
the caregiver'S stance an image of himself as mentalizing, desiring, and believing. He
sees that the caregiver represents him as an intentional being. It is this representation
which is internalised to form the self. "I think therefore I am" will not do as a psycho-
dynamic model of the birth of the self; "She thinks of me as thinking and therefore I
exist as a thinker" comes perhaps closer to the truth.
If the caregiver's reflective capacity enabled her accurately to picture the infant's
intentional stance, the infant will have the opportunity to "find itself in the other" as a
mentalizing individual. If the caregiver's capacity is lacking in this regard, the version
of itself that the infant will encounter will be an individual conceived of as thinking in
terms of physical reality rather than mental states.
4. SOME SPECULATIONS ABOUT PATHOLOGICAL

DEVELOPMENT BASED ON THE DIALECTIC MODEL
The fundamental need of every infant is to find his mind, his intentional state, in
the mind of the object. For the infant, internalisation of this image performs the
function of "containment", which Winnicott has written of as "giving back to the baby
the baby's own self" (Winnicott, 1967, p. 33). Failure of this function leads to a desper-
ate search for alternative ways of containing thoughts and the intense feelings they
engender.
The search for alternative ways of mental containment may, we suggest, give rise
to many pathological solutions, including taking the mind of the other, with its distorted,
absent or malign picture of the child, as part of the child's own sense of identity.
Winnicott (1967) wrote: "What does the baby see when he or she looks at the mother's
60 P.Fonagy
face? ... ordinarily, the mother is looking at the baby and what she looks like is related
to what she sees there ... [but what of] the baby whose mother reflects her own
mood or, worse still, the rigidity of her own defences ... They look and they do not see
themselves ... what is seen is the mother's face" (p. 27).
This picture then becomes the germ of a potentially persecutory object which is
lodged in the self, but is alien and unassimilable. There will be a desperate wish for
separation in the hope of establishing an autonomous identity or existence. However,
tragically, this identity is centred around a mental state which cannot reflect the chang-
ing emotional and cognitive states of the individual, because it is based on an archaic
representation of the other, rather than the thinking and feeling self as seen by the
other.
Paradoxically, where the child's search for mirroring or containment has failed,
the later striving for separation will only produce a movement towards fusion. The
more the person attempts to become himself, the closer he moves towards becoming
his object, because the latter is part of the self-structure. This in our view accounts
for the familiar oscillation of borderline patients, between the struggle for indepen-
dence and the terrifying wish for extreme closeness and fantasised union. Develop-
mentally, a crisis arises when the external demand for separateness becomes irresistible,
in late adolescence and early adulthood. At this time, self-destructive and (in the
extreme) suicidal behaviour is perceived as the only feasible solution to an insoluble
dilemma: the freeing of the self from the other through the destruction of the other
within the self.
In some individuals, for whom separateness is a chronic problem, we assume that
the experience of self-hood can only be achieved through finding a physical other onto
whom the other within the self can be projected. Naturally, this increases the individ-
ual's need for the physical presence of the object. Thus, many such individuals experi-
ence considerable difficulty in leaving home and if they finally achieve physical
separation, they can only do so by finding an alternative and comparable figure onto
whom the other within the self may be projected. If the other dies, or abandons the
individual, a pathological mourning process may be initiated whereby the person
feels compelled to maintain a live picture of the other, in order to retain the integrity
of the self.
Another possible outcome of poor development of the psychological self, with
consequent conflicts over separation, is that the body may be used to contain and enact
mental states. In these cases the child's own body comes to serve the function of
metarepresentation of feelings, ideas and wishes. Violence towards the body of the self
(e.g. self cutting) or that of the other (apparently unprovoked aggression or "mindless
violence" may be a ways of "controlling" mental states which are invested in bodily
states (e.g. the mother seen as part of ones own body) or destroying "ideas" experi-
enced as within the body of the other. In other young children the search for the psy-
chological self in the other may lead to the physical image of the object being
internalised as part of the child's identity. In extreme cases, this may result in gender
identity disorder (Fonagy and Target, 1995).
If the child finds no alternative interpersonal context where he is conceived of
as mentalizing his potential in this regard will not be fulfilled. In cases of abusive, hostile
or simply totally vacuous relationship with the caregiver, the infant may deliberately
turn away from the mentalizing object because the contemplation of the object's mind
is overwhelming as it harbours frankly hostile intentions toward the infant's self. This
may lead to a widespread avoidance of mental states which further reduces the chance
of identifying and establishing intimate links with an understanding object.
As studies of resilient children suggest, even a single secure/understanding rela-

tionship may be sufficient for the development of reflective processes and may "save"
the child. Metacognitive monitoring is biologically prepared and will spontaneously
emerge unless it's development is inhibited by the dual disadvantage of the absence of
a safe relationship and the experience of maltreatment in the context of an intimate
relationship. We do not anticipate that trauma outside of the context of an attachment
bond would have pervasive inhibitory effects on mentalizing. It is because the theory
of mind or, more broadly, and reflective self function evolve in the context of intense
interpersonal relationships, that the fear of the mind of another can have such devas-
tating consequences on the emergence of social understanding. To illustrate the clini-
cal relevance of this model, it may be helpful to consider borderline personality
disorder from the point of view of attachment theory.
4.1. A Transgenerational Model of Borderline Personality Disorder

Although accurate figures are hard to come by and vary across studies, consider-
able evidence has accumulated to support the contention that child abuse is transmit-
ted across generations. Oliver (1993), in his recent review of 60 studies, mainly from
the United States and the UK, concluded that approximately one third of child victims
of abuse grow up to continue a pattern of seriously inept, neglectful or abusive child-
rearing as parents. Research has documented that a specific link exists between the
history of childhood maltreatment and borderline personality disorder and sexual
abuse is especially implicated. In brief, as infants and children, borderline individuals
frequently have caretakers who are themselves within the so called "borderline spec-
trum" of severely personality disordered individuals. The social inheritance aspect
of BPD may be an important clue in our understanding of the disorder.
George Moran, Mary Target, and I (Fonagy et aI., 1993) have put forward an
attachment theory formulation of severe narcissistic and borderline states based on epi-
demiological findings of the association of severe personality disorder and a history of
childhood maltreatment and sexual abuse. We proposed that borderline individuals are
those victims of childhood (sexual) abuse who coped by refusing to conceive of the
contents of their caregiver's mind and thus successfully avoided having to think about
their caregiver's wish to harm them. They go on to defensively disrupt their capacity
to depict feelings and thoughts in themselves and in others. This leaves them to operate
upon inaccurate and schematic impressions of thoughts and feelings and they are thus
immensely vulnerable in all intimate relationships.
. Many of the symptoms of BPD individuals may be understood in terms of a
defensive strategy of disabling mentalizing or metacognitive capacity.
1. Their failure to take into consideration the listener's current mental state
makes their associations hard to follow.
2. The absence of concern for the other which may manifest as extreme vio-
lence and cruelty, arises because of the lack of a compelling representation
of suffering in the mind of the other. A key moderator of aggression is there-
fore absent. The lack of reflective capacity in conjunction with a hostile world
view may predispose individuals to child maltreatment but such inhibition
may be a necessary component of all violence against persons. Military train-
ing has the apparent and explicit aim of fashioning men into machines and
the enemy into an inanimate or sub-human object. Seeing the other as imbued
with thought and feeling is very likely imposes a break.
62 P.Fonagy
3. Their fragile sense of self (identity diffusion to use Kemberg's term), may be
a consequence of their failure to represent their own feelings, beliefs, and
desires with sufficient clarity to provide them with a core sense of themselves
as a functioning mental entity. This leaves them with overwhelming fears of
mental disintegration and a desperately fragile sense of self.
4. Such patient's mental image of object remains at the immediate context
dependent level of primary representations-he/she will need the object as
they are and will experience substantial difficulties when confronted with
change.
5. Absence of prominence "as if" in the transference requires meta representa-
tions, the capacity to entertain a belief whilst at the same time knowing it to
be false. Psychotherapy requires such pretence and it's absence manifests as
so called "acting out" of the transference.
4.2. The Association of Attachment Status and Borderline States

In an ongoing study (Fonagy et aI., 1996) we administered AAI's to a sample of
85 consecutively admitted non-psychotic inpatients at the Cassell Hospital in London,
which is run along the principles of a psychoanalytic therapeutic community. About
40% of the patients met diagnostic criteria for borderline personality disorder (BPD)
on the basis of a structured interview (SCID-II).
The distribution of AAI classifications arrived at totally independently of the
diagnostic process, did not distinguish well Borderline Personality Disorder (BPD)
from other personality disorder diagnoses but the number of entangled (particularly
E3) classifications were well above the number which would be expected by chance
(75%).
Borderline patients' interviews were, however, differentiated by a combination of
3 characteristics:
(1) higher prevalence of sexual abuse reported in the AAI narratives,
(2) significantly lower ratings on the reflective self-function scale,
(3) a significantly higher rating on the lack of resolution of abuse, but not loss
scale of the AAI.
Further, there was a significant interaction between abuse and RSF: individuals
with experience of abuse who had low RSF were very likely to have a diagnosis of
BPD.
These findings are consistent with our assumption that individuals with experi-
ence of severe maltreatment in childhood who respond to this experience by an inhibi-
tion of reflective self junction are less likely to resolve this abuse, and are more likely
to manifest borderline psychopathology.
Childhood maltreatment mayor may not have long term sequelae and the deter-
minants of the outcome are only partially understood. Here we propose that if chil-
dren are maltreated but they have access to a meaningful attachment relationship
which provides the intersubjective bases for the development of mentalising capacity,
they will be able to resolve (work through) their experience and the outcome of the
abuse will not be one of severe personality disorder. We do not expect that their reflec-
tive processes will protect them from episodic psychiatric disorder, such as depression,
and epidemiological data suggests that victims of childhood maltreatment are at an
elevated risk for many forms of [Axis-I] disorder.
However, if the maltreated child has no social support of sufficient strength and
intensity for an attachment bond to develop which could provide the context for the
acquisition of a reliable capacity to envisage the psychological state of the other in
intense interpersonal relationships, then the experience of abuse will not be reflected
on or resolved. Naturally, the unresolved experience of abuse diminishes the likelihood
of meaningful relationships which, in a self-perpetuating way, further reduces the like-
lihood of a satisfactory resolution of the disturbing experience through the use of reflec-
tive processes. In fact a pattern may be established whereby suspicion and distrust
generalises and leads to a turning away from the mental state of most significant
objects and an apparent "decoupling" of the "mentalizing module" leaving the person
bereft of human contact. This may account for the "neediness" of borderline person-
ality disordered individuals; yet no sooner do they become involved with another then
the malfunctioning of their inhibited mentalising capacity leads them into terrifying
interpersonal confusion and chaos. Within intense relationships their inadequate men-
talising function rapidly fails them, they regress to the intersubjective state of the devel-
opment of mental representation and they are no longer able to differentiate their own
mental representations from those of others and both of these from actuality. These
processes combine and they become terrorised by their own thoughts about the other
experienced (via projection) in the other, particularly their aggressive impulses and
fantasies; these become crippling and most commonly they reject or arrange to be
rejected by their object. Psychoanalysis or psychotherapy can break the ,:"icious cycle
by reinforcing reflective capacity.
5. CRIME, VIOLENCE, AND ATTACHMENT
As with borderline patients, a history of maltreatment is present in 80--90% of

juvenile offenders and approximately a quarter of those with histories of severe mal-
treatment are likely to have criminal convictions (e.g. Taylor, 1986). We have suggested
that attachment to individuals as well as social institutions may be critical in reducing
the risk of delinquency and adjustment processes are severely disrupted by childhood
maltreatment. More specifically, if attachment to the primary caregiver is intimately
linked to the acquisition of reflective capacity (see section on moral development,
above), the latter may be a key mediator in predisposing an individual to criminality,
particularly to violent offences. We may suppose that those individuals, who were never
exposed to interpersonal relationships where the acquisition of a reflective capacity
would have been facilitated, or who were exposed to care giving environments where
their only route to adaptation was the inhibition of mentalizing, are most likely to
develop insecure attachments and manifest low reflective capacities, thus removing
essential inhibitions on criminal activities. The capacity to envision the mental state of
the potential victim may be essential in preventing us from deliberately harming other
members of our social group (or species).
To put these ideas to a test, Levinson and Fonagy (in preparation) collected AAI's
from 22 prisoners (convicted or on remand with diagnosable psychiatric disorder) and
matched them with 2 control groups on age, gender, social class, and IQ: 1) a psychi-
atric inpatient control group matched for diagnoses (Axis 1111) and 2) a normal
control group recruited from a medical outpatient department. The findings may be
summarised as follows:
(1) There was significantly more secure attachments in the normal control group
64 P.Fonagy
but the two clinical groups did not differ in terms of overall level of security. (2) 36%
of the prison group vs. 14% of the psychiatric group were classified as "Dismissing"
with normal controls in between (23%). (3) 45% of the prisoners vs.64% of psychi-
atric controls were classified as "Preoccupied" with only 14% of non-criminal
controls receiving this classification. (4) 82% of psychiatric patients but only 36% of
prisoners and 0% of non-clinical controls received "Unresolved" classifications. (5)
82% of prisoners and only 36% of psychiatric patients were rated as having been
abused with only 4 % of normal controls. (2/3 of abuse was physical, 1/3 sexual in both
clinical groups.) (6) Neglect was more prevalent in the prison group but rejection
was more frequently reported by psychiatric patients. (7) Current anger with attach-
ment figures was dominant in psychiatric patients but relatively more among prison-
ers. (8) Prisoners had significantly lower ratings on the reflective function scale (RSF)
than either psychiatric patients or those from the non-clinical group, but RSF ratings
of normals was still significantly higher than those of psychiatric patients. (9) When
the prison group was split into those with violent index offences (murder, malicious
wounding, GBH, armed robbery, indecent assault to child), vs. non violent ones (pos-
session, importation, obtaining property by deception, theft, handling stolen goods) the
rating on reflectiveness of the former group was found to be significantly lower than
the latter.
This pattern of results is consistent with our assumption that criminality arises in
the context of weak bonding with individuals and social institutions and the relatively
ready dismissal of attachment objects. Criminal behaviour may be seen as a socially
maladaptive form of resolving trauma and abuse (which was almost ubiquitous in our
small sample). Violent acts are committed in place of experienced anger concerning
neglect, rejection, and maltreatment. Committing antisocial acts is facilitated by a non-
reflective stance of the victim which may be of particular significance in cases where
the victim is clearly identifiable as in violent acts against another person.
This is only a pilot investigation, but the results are promising to the extent that
they link attachment related narratives to the nature of the offence committed.
Naturally, an important alternative account to the one proposed here may be that it
was these crimes which caused the disorganisation of the attachment system and it
was the psychological impact of crime which permeated the interviews of the violent
group. The less serious offences may have made less impact on the representation of
relationships.
6. PSYCHOTHERAPY AND MENTALIZING
Clinical psychoanalysis inevitably deals with individuals whose past experience

has left them vulnerable to current stress and the repetition of adverse early experi-
ences. The treatment imposes a non-pragmatic elaborative, mentalistic stance. This
enhances the development of reflective self function and may in the long run enhance
the psychic resilience of individuals in a generic way, providing them with improved
control over their system of representation of relationships. It equips them with a kind
of self-righting capacity where through being able to operate on their representational
models, the latter can become an object of review and change. Such gradual and con-
stant adjustments facilitate the development of an internal world where the behaviour
of others may be experienced as understandable, meaningful, predictable, and charac-
teristically human. This reduces the need for splitting of frightening and incoherent
mental representations of mental states, and new experiences of other minds can more
readily be integrated into the framework of past relationship representations.
The abused child, evading the mental world, never acquires adequate meta-
control over the representational world of internal working models. Unhelpful models
of relationship patterns emerge frequently and the internal world of the child and adult
comes to be dominated by negative affect. The individual's enhanced suspiciousness of
human motives reinforces hislher strategy to forego mentalizing, thus further distort-
ing the normal development of a reflective function. Caught in a vicious cycle of para-
noid anxiety and exaggerated defensive manoeuvres, the individual becomes
inextricably entangled into an internal world dominated by dangerous, evil and above
all mindless objects. He has abnegated the very process which could extract him from
his predicament, the capacity to reflect on mental states.
Psychotherapeutic treatment in general, and psychoanalytic treatment in partic-
ular, compels the patient's mind to focus on the mental state of a benevolent other,
that of the therapist. The frequent and consistent interpretation of the mental state of
both analyst and patient (i.e. the interpretation of the transference in the broadest
sense) is then desirable, if not essential, if the inhibition on this aspect of mental func-
tion is to be lifted. Over a prolonged time period, diverse interpretations concerning
the patient's perception of the analytic relationship would enable him to attempt to
create a mental representation both of himself and of his analyst, as thinking and
feeling. This could then form the core of a sense of himself with a capacity to
represent ideas and meanings, and create the basis for the bond that ultimately permits
independent existence.
7. SUMMARY
The paper puts forward an attachment theory model of severe personality dis-
turbance. It is suggested that (1) secure attachment is the basis of the acquisition of
metacognitive or mentalizing capacity; (2) the caregiver's capacity to mentalize may
foster the child's bonding with the parent; (3) maltreatment may undermine the acqui-
sition of a mentalizing capacity; (4) symptoms of borderline personality disorder may
arise as a consequence of inhibited mentalizing; (5) violent crime and anti-social per-
sonality disorder may be possible because the capacity to reflect upon the mental
states of the victim is compromised; (6) psychotherapeutic work may facilitate the
reactivation of this inhibited capacity.
REFERENCES
Ainsworth, M.D.S. (1985). Attachments across the lifespan. Bulletin of the New York Academy of Medicine,
61, 792-812.
Ainsworth, M.D.S. and Wittig, B.A. (1969). Attachment and exploratory behavior of one-year-olds in a
strange situation. In B.M. Foss (Ed.), Determinants of infant behavior (pp. 113-136). London: Methuen.
Ainsworth, M.D.S., Blehar, M.e., Waters, E., and Wall, S. (1978). Patterns of attachment: A psychological study
of the Strange Situation. Hillsdale, NJ: Erlbaum.
Belsky, J. (1993). Etiology of child maltreatment: A developmental-ecological analysis. Psychological
Bulletin, 114, 413-434.
Belsky, J., Rovine, M., and Taylor, D.G. (1984). The Pennsylvania Infant and Family Development Project,
III: The origins of individual differences in infant-mother attachment: Maternal and infant contribu-
tions. Child Development, 55, 718-728.
66 P.Fonagy
Belsky, I, Fish, M., and Isabella, R (1991). Continuity and discontinuity in infant negative and positive
emotionality: Family antecedent and attachment consequences. Developmental Psychology, 27,
421-431.
Belsky, I, Rosenberger, K., and Crnic, C. (1995). The origins of attachment security: "Classical" and contex-
tual determinants. In S. Goldberg, R Muir, and I Kerr (Eds.), John Bowlby's Attachment Theory:
Historical, Clinical and Social Significance (pp. 153-184). Hillsdale, NJ: Analytic Press.
Benoit, D. and Parker, K.C.H. (1994). Stability and transmission of attachment across three generations. Child
Development, 65,1444-1457.
Bowlby, I (1969). Attachment and loss, vol. 1: Attachment. London: Hogarth Press and the Institute of
Psycho-Analysis.
Bowlby, I (1973). Attachment and loss, vol. 2: Separation: Anxiety and anger. London: Hogarth Press and
Institute of Psycho-Analysis.
Bowlby, I (1980). Attachment and loss, vol. 3: Loss: Sadness and depression. London: Hogarth Press and
Institute of Psycho-Analysis.
Campbell, S.B., Cohn, IF., Meyers, T.A, Ross, S., and Flanagan, C. (1993). Chronicity of maternal depression
and mother-infant interaction. Presented at Biennial meeting of the Society for Research in Child
Development, New Orleans.
Cicchetti, D. and Beeghly, M. (1987). Symbolic development in maltreated youngsters: An organizational
perspective. In D. Cicchetti and M. Beeghly (Eds.), Atypical symbolic development. New Directions
for Child Development (Vol. 36, pp. 5-29). San Francisco: Jossey-Bass.
Collins, N.L. and Read, S.I (1994). Representations of attachment: The structure and function of working
models. In K. Bartholomew and D. Perlman (Eds.), Advances in Personal Relationships Vol 5:
Attachment Process in Adulthood (pp. 53-90). London: Jessica Kingsley Publishers.
Cox, M., Owen, T., Henderson, v., and Margand, N. (1992). Prediction of infant-father and infant-mother
attachment. Developmental Psychology, 28, 474-483.
Crittenden, P.M., Partridge, M.F., and Claussen, AH. (1991). Family patterns of relationship in normative and
dysfunctional families. Development and Psychopathology, 3, 491-449.
Crnic, K.A, Greenberg, M.T., Ragozin, AS., Robinson, N.M., and Basham, RB. (1983). Effects of stress and
social support on mothers and premature and full-term infants. Child Development, 54, 209-217.
Del Carmen, R, Pedersen, F., Huffman, L., and Bryan, Y. (1993). Dyadic distress management predicts
security of attachment. Infant Behavior and Development, 16, 131-147.
Dennett, D. (1987). The intentional stance. Cambridge, MA: MIT Press.
Egeland, B. and Farber, E.A (1984). Infant-mother attachment: Factors related to its development and change
over time. Child Development, 55, 753-771.
Emde, R and Buchsbaum, H. (1990). "Didn't you hear my mommy?" Autonomy with connectedness in moral
self-emergence. In D. Cicchetti and M. Beeghly (Eds.), The self in transition (pp. 35-60). Chicago:
University of Chicago Press.
Fonagy, P. and Target, M. (1995). Understanding the violent patient: The use of the body and the role of the
father. International Journal of Psycho-Analysis, 76, 487-502.
Fonagy, P., Steele, M., Moran, G., Steele, M., and Higgitt, AC. (1991). The capacity for understanding mental
states: The reflective self in parent and child and its significance for security of attachment. Infant
Mental Health Journal, 13, 200--216.
Fonagy, P., Moran, G.S., and Target, M. (1993). Aggression and the psychological self. International Journal
of Psycho-Analysis, 74, 471-485.
Fonagy, P., Steele, M., Steele, H., Higgitt, A, and Target, M. (1994). Theory and practice of resilience. Journal
of Child Psychology and Psychiatry, 35, 231-257.
Fonagy, P., Steele, M., Steele, H., Leigh, T., Kennedy, R, Mattoon, G., and Target, M. (1995). The predictive
validity of Mary Main's Adult Attachment Interview: A psychoanalytic and developmental perspec-
tive on the transgenerational transmission of attachment and borderline states. In S. Goldberg, R
Muir, and I Kerr (Eds.), Attachment Theory: Social, Developmental and Clinical Perspectives (pp.
233-278). Hillsdale, NJ: The Analytic Press.
Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R, Mattoon, G., Target, M., and Gerber, A (1996). The
relation of attachment status, psychiatric classification, and response to psychotherapy. Journal of
Consulting and Clinical Psychology, 64, 22-31.
Fonagy, P., Redfern, S., and Charman, T. (1997). The relationship between belief-desire reasoning and a
projective measure of attachment security (SAT). British Journal of Developmental Psychology, 15,
51-61.
Fox, N.A, Kimmerly, N.L., and Schafer, W.D. (1991). Attachment to mother/attachment to father: A
meta-analysis. Child Development, 62, 210--225.
Attachment and Its Pathology in Personality Disorders .67
Gabarino, 1 (1992). The meaning of poverty in the world of children. American Behavioral Scientist, 35,
.220-237.
George, c., Kaplan, N., and Main, M. (1985). The Adult Attachment Interview. Unpublished manuscript,
Department of Psychology, University of California at Berkeley.
Goldberg, W.A and Easterbrooks, M.A (1984). The role of marital quality in toddler development.
Developmental Psychololgy, 20, 504-514.
Goosens, F. and van Ijzendoorn, M. (1990). Quality of infants' attachment to professional caregivers. Child
Development, 61, 832--837.
Grossmann, K.E., Grossmann, K., and Schwan, A (1986). Capturing the wider view of attachment: A reanaly-
sis of Ainsworth's Strange Situation. In C.E. Izard and P.B. Read (Eds.), Measuring Emotions in Infants
and Children, volume 2 (pp.I24-171). New York: Cambridge University Press.
Hamilton, C. (1994). Continuity and discontinuity of attachment from infancy through adolescence. Unpub-
lished doctoral dissertation, UC-Los Angeles.
Hazan, C. and Zeifman, D. (1994). Sex and the psychological tether. In K. Bartholomew and D. Perlman
(Eds.), Advances in Personal Relationships, volume 5 (pp. 151-178). London: Jessica Kingsley
Publishers Limited.
Hegel, G. (1807). The Phenomenology of Spirit, translated by AV. Miller. Oxford: Oxford University Press,
1977.
Isabella, RA (1993). Origins of attachment: Maternal interactive behaviour across the first year. Child Devel-
opment, 64, 605-621.
Isabella, RA and Belsky, 1 (1991). Interactional synchrony and the origins of mother-infant attachment: A
replication study. Child Development, 62, 373-384.
Isabella, RA, Belsky, 1, and von Eye, A (1989). Origins of infant-mother attachment: An examination of
interactional synchrony during the infant's first year. Developmental Psychology, 25,12-21.
Jacobson, S.W. and Frye, K.F. (1991). Effect of maternal social support on attachment: Experimental evidence.
Child Development, 62, 572-582.
Kagan, 1 (1982). Psychological Research on the Human Infant: An Evaluative Summary. New York: Wiley.
Lamb, M.E., Thompson, RA, Gardner, w.P., Chamov, E.L., and Estes, D. (1984). Security of infantile attach-
ment as assessed in the "strange situation": Its study and biological interpretation. The Behavioral and
Brain Sciences, 7, 127-147.
Levinson, A and Fonagy, P. (submitted). Attachment classification in prisoners and psychiatric patients.
Lewis, M. and Feiring, C. (1989). Infant, mother, and mother-infant interaction behavior and subsequent
attachment. Child Development, 60, 831--837.
Lieberman, AF. and Pawl, 1H. (1990). Disorders of attachment and secure base behavior in the second
year of life: Conceptual issues and clinical intervention. In M.T. Greenberg, D. Cicchetti, and E.M.
Cummings (Eds.), Attachment in the Preschool Years (pp. 375-398). Chicago: University of Chicago
Press.
Lieberman, AF., Weston, D.R, and Pawl, 1H. (1991). Preventive intervention and outcome with anxiously
attached dyads. Child Development, 62, 199-209.
Lyons-Ruth, K., Connell, D.B., and Grunebaum, H.U. (1990). Infants at social risk: Maternal depression and
family support services as mediators of infant development and security of attachment. Child Devel-
opment, 61, 85-98.
Main, M. (1991). Metacognitive knowledge, metacognitive monitoring, and singular (coherent) vs. multiple
(incoherent) models of attachment: Findings and directions for future research. In P. Harris, 1 Steven-
son-Hinde, and C. Parkes (Eds.), Attachment Across the Lifecycle (pp. 127-159). New York: Routledge.
Main, M. and Goldwyn, R (1994). Adult Attachment Rating and Classification System, Manual in Draft:
Version 6.0. Unpublished manuscript, University of California at Berkeley.
Main, M. and Hesse, E. (1990). Parents' unresolved traumatic experiences are related to infant disorganized
attachment status: Is frightened and/or frightening parental behavior the linking mechanism? In
D.C.M. Greenberg and E.M. Cummings (Eds.), Attachment in the preschool years: Theory, research,
and intervention (pp. 161-182). Chicago: University of Chicago Press.
Main, M. and Solomon, E. (1990). Procedures for identifying infants as disorganized/disoriented during
the Ainsworth Strange Situation. In D.C.M. Greenberg and E.M. Cummings (Eds.), Attachment
in the preschool years: Theory, research, and intervention (pp. 121-160). Chicago; University of Chicago
Press.
Main, M. and Weston, D. (1981). The quality of the toddler'S relationship to mother and to father: Related
to conflict behavior and the readiness to establish new relationships. Child Development, 52, 932-940.
Malatesta, C.Z., Grigoryev, P., Lamb, c., Albin, M., and Culver, C. (1986). Emotional socialisation and
expressive development in pre-term and full-term infants. Child Development, 57, 316-330.
68 P.Fonagy
Maslin, C.A. and Bates, IE. (1983). Precursors of anxiolts and secure attachments: A multivariant model at
age 6 months. Paper presented at the biennial meeting of the Society for Research in Child
Development: Detroit, MI.
O'Connor, M., Sigman, M., and Kasari, C. (1992). Attachment behavior of infants exposed prenatally to
alcohol: Mediating effects of infant affect and mother-infant interaction. Development and
Psychopathology, 4, 243-256.
Oliver, IE. (1993). Intergenerational transmission of child abuse: Rates, research, and clinical implications.
American Journal of Psychiatry, 150, 1315-1324.
Radojevic, M. (1992). Predicting quality of infant attachment to father at 15 months from pre-natal paternal
representations of attachment: an Australian contribution. Paper presented at the 25th International
Congress of Psychology: Brussels, Belgium.
Sroufe, L.A. (1979). Socioemotional development. In I Osofsky (Ed.), Handbook of Infant Development (pp.
462-516). New York: Wiley.
Sroufe, L.A. (1990). An organizational perspective on the self. In D. Cicchetti and M. Beeghly (Eds.), The
selfin transition: Infancy to childhood (pp. 281-307). Chicago: University of Chicago Press.
Sroufe, L.A. (1996). Emotional development: The organization of emotional life in the early years. New York:
Cambridge University Press.
Steele, H., Steele, M., and Fonagy, P. (1996). Associations among attachment classifications of mothers, fathers,
and their infants: Evidence for a relationship-specific perspective. Child Development, 67, 541-555.
Steele, H., Steele, M., and Fonagy, P. (in preparation). A path-analytic model of determinants of infant-parent
attachment: limited rather than multiple pathways.
Stern, D.I (1994). One way to build a clinically relevant baby. Infant Mental Health Journal, 15, 36-54.
Taylor, P.I (1986). Psychiatric disorder in London's life sentenced prisoners. British Journal of Criminology,
26,63-78.
van den Boom, 0. (1990). Preventive intervention and the quality of mother-infant interaction and infant
exploration in irritable infants. In W. Koops et al. (Eds.), Developmental Psychology Behind the Dikes
(pp. 249-270). Amsterdam: Eburon.
van den Boom, D.C. (1995). Do first-year intervention effects endure? Follow-up during toddlerhood of a
sample of Dutch irritable infants. Child Development, 66, 1798-1816.
van Ijzendoorn, M.H. (1992). Intergenerational transmission of parenting: A review of studies in non-
clinical popUlations. Developmental Review, 12,76-99.
van Ijzendoorn, M. and Kroonenberg, P.M. (1988). Cross-cultural patterns of attachment. A meta-analysis of
the Strange Situation. Child Development, 59, 147-156.
van Ijzendoorn, M.H. (1995). Adult attachment representations, parental responsiveness, and infant attach-
ment: A meta-analysis on the predictive validity of the Adult Attachment Interview. Psychological
Bulletin, 117,387-403.
Ward, M.I and Carlson, E.A. (1995). Associations among Adult Attachment representations, maternal sen-
sitivity, and infant-mother attachment in a sample of adolescent mothers. Child Development, 66,
69-79.
Waters, E., Merrick, S., Albersheim, L., Treboux, D., and Crowell, I (1995). From the Strange Situation to the
Adult Attachment Interview: A 20-year Longitudinal Study ofAttachment Security in Infancy and Early
Adulthood. Presented at meeting of the Society for Research in Child Development: Indianapolis, IN.
Winnicott, D.W. (1967). Mirror-role of the mother and family in child development. In P. Lomas (Ed.), The
Predicament of the Family: A Psycho-Analytical Symposium (pp. 26-33). London: Hogarth.
Zeanah, C.H., Benoit, D., Barton, M., Regan, c., Hirshberg, L.M., and Lipsitt, L.P. (1993). Representations
of attachment in mothers and their one-year-old infants. Journal of the American Academy of Child
and Adolescent Psychiatry, 32, 278-286.
AUTOBIOGRAPHICAL NOTE
Peter Fonagy, PhD (London), Dip. Clin. Psy. (London), Freud Memorial Profes-
sor of Psychoanalysis, Director, Sub-department of Clinical Health Psychology, Depart-
ment of Psychology, University College London; Research Director, The Anna Freud
Centre, London; Director, Child and Family Center, Menninger Foundation, Topeka,
Kansas, Training, and Supervising Analyst, British Psycho-Analytical Society.
6
TRAUMA AND PERSONALITY
Fabio Madeddu 1 and Adolfo Pazzagli2
1 Medical Psychology and Psychotherapy Unit

Department of Neuropsychiatry Sciences
San Raffaele Hospital
University of Milan-School of Medicine
Via L. Prinetti 29-20127 Milan
2Department of Neuropsychiatric Science
University of Florence-School of Medicine
Viale Pieraccini 17-50139 Florence
1. INTRODUCTION
The typical human proclivity of attributing one's own afflictions to external events
is currently very widespread and is often a way of building a defensive barrier. This
tendency has always been connected with the genesis of permanent traits which
characterize individuals and form their personality. An example is the relationship
between humoral theories and astral influences. This .relationship becomes even
more explicit when considering behaviors currently identified as "personality disor-
ders": lunatic, martial, jovial are personality traits that can be related to our present
distinctions.
At the end of the nineteenth century, Sigmund Freud attributed sexual traumas
as being at the core of many neuroses, including personality neuroses. The traumatic
etiology of these neuroses was included in the complex complementary series system:
congenital and inherited traits contribute, in a complementary manner with early child-
hood experiences, to determine a "tendency" which is integrated with current experi-
ences in determining if an individual will respond in a sane or in a neurotic manner to
certain situations. Freud's scheme is still valid today, even if the relative weight of
certain traits may vary depending on the case-or on the prevailing theory at the time.
However, the problem related to the traumatic origin of neuroses has often been eluded
in such a complex scheme. It has been juxtaposed in an internal dilemma related to
the individual's conscious and unconscious fantasies. Thus, the origin of psychic trau-
70 E Madeddu and A. Pazzagli
mas have always been irreconcilably attributed either to traumatic experiences or to

their fantasies. It was Freud himself, probably in relation to the different stages of self-
analysis, who caused the sexual trauma hypothesis to overturn, radically and abruptly
blaming the child for his fantasies rather than the adult for the abuse.
In more recent years, Masson (1984) turned this dilemma over again, but without
substantially modifying its nature. However, the problem related to the relationship
between traumas and personality, and, even more importantly, between traumas and
personality disorders (PDs) has recently been given new consideration, probably due
to the tremendous and widespread traumas which characterize recent history such as
the consequences of the holocaust for its survivors and of the Vietnam war for its vet-
erans. Moreover, there has been a growing interest in studying the repercussions of
sexual childhood traumas on individuals. These recent studies did not deal with the
problem from a subjective or ideological point of view but through scientific research,
making it possible to establish assessments and to compare test data.
2. EMPIRICAL STUDIES
Most of the research done in this area involves the study of borderline personal-
ity disorders (BPD). Since this type of disorder involves the "borders" of different enti-
ties, it plays a leading role in understanding the factors which cause major personality
disorders and the psychopathological behaviors related to them. More specifically, from
Adolf Stern's (1938) first observations to the most recent concepts on borderline dis-
order (Perry and Hermann, 1993), the problem related to the role of traumas has been
dealt with both from a pathogenic point of view and, subsequently, in relation to treat-
ment (Gabbard et aI., 1994). As previously mentioned, the most important aspect of
these studies has been the introduction of empirical research which has obligated
researchers and clinicians to reexamine concepts and approaches which had been
locked in a clinical-speculative dimension. From this point of view, literature has
focused its attention on borderline disorders as coded in the most recent DSM editions,
thus giving considerable insight on the matter. The results of these studies have shown
that the childhood of these patients is often characterized by relational traumatic phe-
nomena such as parental neglect, lack of protection (Frank and Paris, 1981), prolonged
detachment from parents or their absence (Soloff and Milward, 1983) and, finally, phys-
ical, and sexual abuse, mostly perpetrated by parents (Stone, 1988; Perry and Hermann,
1993). Regarding abuse, many studies at first stressed the frequency of definite events,
as physical and sexual abuse, in a BPD's childhood (Perry and Hermann, 1993), build-
ing up the "traumatic" conceptual patterns that partially refer to the post traumatic-
stress disorder area (van der Kolk, McFarlane, and Weisaeth, 1996). In general, the
consequences of the trauma have great impact on controlling emotions: these victims
seem to constantly be living within the emotional milieu of the event, with increased
hyperalertness and hypersensibility (van der Kolk, 1987). The type of emotional "insta-
bility" of borderline patients looks similar to the syndromes described in patients who
have gone through traumas or catastrophes: consistent affective symptoms, inclination
to drug abuse, dissociative episodes, "clairvoyance" and, more in general, disorders in
interpersonal relationships and a chronic sense of loss of identity. These aspects of emo-
tional disregulation could be an interesting way of coping with borderline issues. In
particular, the tendency to emotionally exaggerate apparently neutral stimuli, the
impUlsive area, the short-circuit area, often described as a tendency to act-out, features
Trauma and Personality 71
related to certain symptoms (e.g. self-aggression) can be considered expressions of a

troubled manner of handling feelings/emotions. These aspects also find a symptomatic
expression in anger, hot-temper, hypersensibility, and aggressive behavior which are
very frequent in these patients. To simplify this discussion, we can say that most bor-
derline patients are characterized by the "inability to control impulses" (Zanarini,
1993), by "emotional disregulation" (Linehan and Koerner, 1993) and "overreaction"
(Stone, 1993). Patients with borderline disorders find it impossible to control or mod-
ulate emotions and the reverberation of this impossibility could be the origin of later
distortions: identity will obviously be impaired, as will be the possibility of building
enduring relationships since these patients are continuously hit by emotional storms.
The "flaws" in examining reality could be attributed to excessive affectivization as
emphasized by many projective tests (Singer, 1977).
Other hypotheses have correlated the borderline pathology to trauma starting
from the concept of primitive defense mechanisms. In fact, it is well known that
these patients resort to primitive defenses, induding splitting, and dissociation.
Such defenses may represent a way of dealing with trauma as an emotional event
which could not be handled in any other way. These defenses will then characterize the
patient's personality structure (and perhaps build it) in some sort of permanently
distorted style that from defensive becomes cognitive, repeating itself and then ris-
ing up in different relationships, induding the transference relationship (Perry and
Hermann, 1993). Regarding this latter issue, Gabbard et al. (1994) have written: "A
history of early trauma may also contribute to negative reactions to transference inter-
pretation. Many patients who experienced abuse during childhood, characteristically
externalize aggression and see others as "bad objects". If the therapist attempts to force
the aggression back onto the patient with transference interpretations, the patient may
feel retraumatized and humiliated, leading to a deterioration of the therapeutic
alliance" .
Aside from therapeutic correlatives, it can be said that there are a number of
empirical studies and conceptual frameworks which strengthen the hypothesis that
child abuse and borderline disorders are strictly related. However, these hypotheses
need more empirical development; data obtained from these studies are not univocal
and, more importantly, they do not seem to support a specific relationship between
event and pathology. For example, recently Laporte and Guttman (1996) have found
that even if 93 % of women with BPD had experienced some sort of separation or abuse
in their childhood, the same occurred to 74% of women with other personality disor-
ders. Furthermore, in this particular study, women with different types of personality
disorders had experienced more physical abuse than the general popUlation; as a ref-
erence we can recall that the percentage of all types of physical abuse amongst the
overall population is estimated at 11 % (Strauss and Gelles, 1986). So far we have no
data to help distinguish between women affected by personality disorders, with the
exception of BPD, and women affected by Axis I disorders, who appear to have suf-
fered-as did male patients-more physical abuse than the general population (Brown
and Anderson, 1991; Carmen, Ricker, and Mills, 1984). The frequency of sexual abuse
(22% in the above mentioned study) is not greater in women with personality disor-
ders with respect to the ones of the general population, for which frequency varies
between 6% and 30%, depending on the definition given to abuse and the research
methodology chosen (Bagley, 1984). In order to better understand why the results given
in the empirical literature were not always univocal regarding the relationship between
trauma and BPD, a recent meta-analytic evaluation has been carried out on the
72 F. Madeddu and A. pazzagli
empirical literature published between 1980 and 1995, relevant to the relationship
between child sexual abuse (CSA) and BPD (Fossati, Madeddu, and Maffei, 1996).
Twenty-one studies have been taken into consideration and they have confirmed that
there are several problems regarding methodology (Madeddu et aI., 1991). The results
have shown that a linear hypothesis which correlates sexual abuse, its variations and
BPD cannot be found by a meta-analytic review of the literature. Regarding the vari-
ations considered (use of force, family ties between abuser and abused, greater patho-
genic impact of earlier abuse), the results obtained are: (1) even if severity of CSA was
positively and moderately associated with BPD, a clear gradient in effect size signifi-
cance and relevance from fondling to intercourse as expected according, for example,
to Brown and Finkelhor (1986) observation, was not observed; (2) the findings on per-
petrators are somehow at variance with studies showing that abuse by a relative or
caretaker is more pathogenic that abuse by a non-relative. Data show that abuse by a
non-relative, rather than by caretakers, had the higher effect-size in BPD; (3) even the
data indicating that the younger the child's age at abuse the severer the long term
sequelae (Brown and Finkelhor, 1986) were not fully replicated in studies on BPD,
given that rank order (from larger to smaller) for the effect .size estimates was latency,
adolescence, early childhood, and not, as expected, early childhood, latency, adoles-
cence. This latter observation appears consistent with other data (Paris, 1994). The most
important result appears the most general one related to the existence of only an
overall moderate association between later reports of CSA and BPD diagnosis, rather
than a strong association between a documented CSA and later development of
BPD. As a whole, these findings concerning CSA parameters associated with BPD
development seem to support the need for future research on the relationship between
CSA and BPD, avoiding oversimplifications and emphasizing the importance of a
multifactorial etiologic model (Paris, 1994).
These findings seem to reduce the cause-effect hypothesis but surely do not elim-
inate the issue regarding the weight of environmental factors in general. This hypoth-
esis must be inserted in a larger context in which other parental and protective factors
playa leading role (Paris, 1994; Spaccarelli, 1994). As previously highlighted, other
studies have focused on the chronic, traumatic milieu, trying to give an empirical foun-
dation to the psychodynamic area tradition, in which the role of primary pathogenic
relationships in the development of borderline disorders are of great importance
(Adler, 1985; Mahler, 1971).
As a matter of fact, a persistent traumatic environment is probably involved in a
wider range of personality disorders pathogenesis. According to Akhtar (1992), the sig-
nificant variants of major PDs development include early, unmitigated, major traumatic
events. Such PDs, including disorders that were not taken into account by DSM classi-
fications, are: borderline, infantile (Histrionic), paranoid, hypomanic, antisocial, as if,
schizoid, schizotypal. An overall explanation might be given by the presence, as men-
tioned earlier, of a closer tie between "timely" traumas and the impulsive area (van der
Kolk, Perry, and Herman, 1991) with symptoms which might be connected to some BPD
aspects and, most likely to Antisocial disorders (Luntz and Spatz-Widom, 1994), while
wider dimensions of insufficient parental functions may lead to different personalities.
The key variants are temperamental and developmental factors which lead to a com-
plex bio-psycho-social model (Paris, 1994).
The matter regarding the consequences of traumatic factors experienced during
childhood, which include significant losses, early, and prolonged separations, and "even"
physical, sexual, and verbal abuse is still open to discussion as far as its specificity to
personality disorders is concerned, in particular to the ones different from BPD. It is

true that in psychiatric histories, the more specific traumatic events are more frequent
than in non psychiatric populations, but there is no tangible specificity relevant to
personality disorders nor is it possible to decide if they are causal factors or epiphe-
nomena of a widespread disorder in a family's ability to protect and contain the
child's needs.
3. THE ATTACHMENT THEORY AND

THE TRAUMATIC ENVIRONMENT
Another interesting area regarding these issues is represented by the studies

which stem from Bowlby's approach to the problem of attachment since these studies
focus their attention on the quality of the relationship between a child and the care-
giver, on the elements involved in building it or in its distortions. Bowlby defined attach-
ment as a psychological inner setting which allowed the individual to maintain or search
for contacts with the person seen as the attachment figure. The attachment to persons
or places and the fear of what is new and of strangers are considered elements of a
behavioral system whose purpose is that of maintaining a relatively stable condition
between the individual and his environment. The organization of this behavioral system
starts in the early stages of life and tends to be structured in a strategy organized around
the first year of life.
Individual differences, initially identified in terms of secure or insecure attach-
ment modes, in relation to a specific caregiver have been described and coded by
Ainsworth et al. (1978) through a laboratory technique called "Strange Situation"
which is still the most widely used systematic procedure to test the quality of a child's
attachment to the caregiver.
Studies in this area have been boosted by the processing of a semi-structured
interview which makes it possible to classify the adult mental representation of attach-
ments. This tool is known as the "Adult Attachment Interview" (AAI) (George, Kaplan,
and Main, 1985). It includes a set of questions which make it possible to completely
identify the history of experiences of child attachment in the individual and to evalu-
ate their effects on his current personality. The manner in which these experiences are
narrated, more than the nature of the experience itself, leads to a general classification
of the current mental status of the adult regarding attachment. With AAI it has been
possible to gather evidence on the association between the way in which a mother
recalls her own childhood experiences and the quality of her relationship with her own
child: intergenerational concordance. Recently, development psychologists who study
the sources of individual differences in child attachment patterns, have started to
explore the influence of the mother's childhood experiences, and of her own personal-
ity as well, on the mother-child relationship. It has also been supposed that the adult's
assessment of childhood experiences and their consequences on current personality
regarding attachment are structured in a relatively firm "state of mind" (Main, Kaplan,
and Cassidy, 1985). This "state of mind" or "mind representation" is defined as a set of
rules "for the organization of information relevant to attachment and for obtaining or
limiting access to that information" (Main, Kaplan, and Kassidy, 1985). This implies a
focus on the cognitive organization and on the reconstruction of childhood attachment
experiences which is consistent with the findings made by cognitive science on mental
representations (Mandler, 1985).
74 F. Madeddu and A. pazzagli
Clinical and retrospective data suggest that abused children, in turn, become
abusers in a number of cases greater than non abusers and that, in general, many emo-
tionally disturbed adults have had insecure or interrupted attachment relationships
during their childhood (van Ijzendoorn, 1995). However, estimates regarding the rela-
tionship between early experiences and later consequences on behavior are contradic-
tory (Kaufman and Zigler, 1987). Self-report measures, which are the ones most
frequently used in these studies, are based on a frankly optimistic point of view of
autobiographic memory capacities which describe the "objective" experiences of the
subjects (Wagenaar, 1987; van Ijzendoorn, 1992).
With AAI, on the other hand, it is possible to study the manner in which memo-
ries on the organization of mental representations emerge, thus avoiding a merely "sub-
jective" approach, based for instance on the narrative or historical reality of a trauma.
Many scales for assessment of childhood experiences and of the current mental con-
dition regarding attachment figures have been identified based on these findings. From
these scales it has been possible to classify mental representation patterns of early
attachment experiences.
A study of attachment relationships on a significant number of clinical and
general population samples (Pazzagli, Guerrini Degl'Innocenti, and Selvi, 1996) has
considered 50 subsequent AAIs which were divided into groups based on the presence
of: (1) at least one major grief due to the loss of a significant attachment figure; (2) sep-
aration from parents for at least one year during childhood; (3) physical and sexual
abuse. There was also a fourth control group in which none of the events described
above were reported.
In general, findings showed how the presence of grief due to the loss of a signi-
ficant figure or an early and prolonged detachment can cause a restrained disorgani-
zation of the psychic processing of events which can be revealed by the presence of
incongruous elements such as lapses, confusion or irrational statements. Trauma,
instead, intended as physical and sexual abuse results in many "indicators" which show
that the traumatic experience has not been resolved (with the exception of a resilient).
Furthermore, the second category is often the most insecure, or it not possible to find
a coherent and prevalent operational pattern.
In summary, we can confirm the harm caused on personality organization result-
ing from major grieves or early detachment, probably confirming the existence complex
"traumatic milieu" experience during the childhood (Madeddu, 1996); this milieu
often includes sexual and physical abuse. Even if research has failed to identify a spe-
cific cause-effect factor, it has highlighted that a "traumatic" relational aspect does
exist, especially for the most severe Pds, which is a useful indication for a correct
therapeutic approach.
REFERENCES
Adler G., 1985, Borderline psychopathology and its treatment, Aronson, Northvale.
Akhtar S., 1992, Broken Structures, Aronson, Northvale.
Ainsworth M.D., Blehar M.e., Waters E., and Wall S., 1978, Patterns of attachment: a psychological study of
the Strange Situation, Erlbaum, Hillsdale.
American Psychiatric Association, 1980, Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition,
American Psychiatric Press, Washington DC.
Bagley R., 1984, Sexual offenses against children, Ministry of Justice, Ottawa.
Bowlby 1., 1973, Attachment and loss, Penguin Books, London.
Brown G.R and Anderson B., 1991, Psychiatry morbidity in adult inpatients and childhood histories of sexual
and physical abuse, Am. J. Psychiat., 148:55-61.
Brown A. and Finkelhor D., 1986, Impact of childhood sexual abuse: A review of the literature, Psycholog.
Bull., 99:66-77.
Carmen E.H., Ricker P.P., and Mills T., 1984, Victims of violence and psychiatric illness, Am. J. Psychiat.,
141:378-383.
Fossati A, Madeddu F., and Maffei e., 1996, Childhood Sexual Abuse and Borderline Personality Disorder:
A Meta-Analysis, Journ. Pers. Dis., in press.
Frank H. and Paris J., 1981, Recollections of family experience in borderline patients, Arch. Gen. Psychiat.,
38:1031-1034.
Gabbard 0.0., Horwitz L., Allen J.G., Frieswyk S., Newsom G., Colson D.B., and Coyne L., 1994, Transfer-
ence Interpretation in the psychotherapy of Borderline Patients: A High-Risk, High-Gain Phenomenon,
Harvard Rev. Psychiat., July-August:59-69.
George e., Kaplan N., and Main M., 1985, Adult Attachment Interview, Unpublished Manuscript, University
of California, Berkeley.
Kaufman J. and Zigler E., 1987, Do abused children become abusive parents?, Am. Journ. Orthopsychiat.,
57:186-192.
Laporte L. and Guttman H., 1996, Traumatic Childhood experiences as risk factors for borderline and other
personality disorders, Journ. Pers. Dis., 10:247-259.
Linehan M.M. and Koerner K., 1993, A behavioral theory of borderline personality disorder, In Paris J. (Ed.),
Borderline Personality disorder. Etiology and Treatment, American Psychiatric Press, Washington De.
Luntz B.K. and Spetz Widom e., 1994, Antisocial Personality Disorder in Abused and Neglected Children
Grown up, Am. J. Psychiat., 151(5):670-674.
Madeddu F., Visintini R, Fossati A, and Maffei e., 1991, Borderline personality and family factors: an
overview of empirical literature. Exp. Clin. Psychiatry, 8(4):151-168.
Madeddu F., 1996, L'esperienza traumatica infantile nel disturbo borderline di personalita, In Barone L. and
Maffei e. (Eds), Emozione e Conoscenza nei disturbi di personalitil, Franco Angeli, Milano.
Mahler M.S., 1971, A study of the separation-individuation process and its possible application to borderline
phenomena in the psychoanalytic situation, Psychoanalytic study of the child, 26:402--424.
Main M., Kaplan N., and Kassidy J., 1985, Security in infancy, childhood, and adulthood: A move to the level
of representation, Monograph of the Society for Research in Child Development, 50, 1-2 (Serial 209).
Masson J.M., 1984, The assault on Truth, Farrar, Straus, and Giroux, New York.
Paris J., 1994, Borderline personality disorder. A multidimesional Approach, American Psychiatric Press,
Washington DC.
Pazzagli A., Guerrini Degl'Innocenti B., and Selvi A., 1996, Mourning, trauma and Internal Working Model,
II European Congress on Personality Disorders, June 26-29, Milan.
Perry J.e. and Hermann J.L., 1993, Trauma and defences in the etiology of borderline personality disorder, In
Paris J. (Ed.) Borderline Personality disorder. Etiology and Treatment, American Psychiatric Press,
Washington De.
Singer M., 1977, The borderline diagnosis and psychological tests: review and research, In Hartocollis P. (Ed.),
Borderline Personality Disorder: the concept, the syndrome, the patient., International Universities
Press, Cambridge.
Soloff P.H. and Millward J.w., 1983, Developmental histories of borderline patients, Compreh Psychiat,
23:574-588.
Spaccarelli S., 1994, Stress, appraisal, and coping in child sexual: a theoretical and empirical review, Psychol.
Bull., 116(2):340-362.
Stern A., 1938, Psychoanalytic investigation of and therapy in the borderline group of neuroses, Psychoanal.
Q.,7:467--489.
Stone M., 1988, Incest in female borderlines: its frequency and impact, Int. J. Fam. Psychiat., 9:277-293.
Stone M., 1993, Etiology of borderline personality disorder: psychobiological factors contributing to an under-
lying irritability, In Paris J. (Ed.), Borderline Personality disorder. Etiology and Treatment., American
Psychiatric Press, Washington De.
Strauss H.A and Gelles RJ., 1986, Societal change and change in family violence from 1975 to 1985 as revealed
by two national survays, Journ. Marriage and Fam., 48:465--479.
van der Kolk B.A., 1987, Psychological Trauma, American Psychiatric Press, Washington De.
van der Kolk B.A, Perry J.e., and Herman J.L., 1991, Childhood origins of Self-Destructive Behavior, Am. J.
Psychiat.,148(12):1665-1671.
van der Kolk B.A, Mc Farlane Ae., and Weisaeth L. (Eds), 1996, Traumatic Stress, Guilford Press, New York.
76 F. Madeddu and A. Pazzagli
van Ijzendoorn M.H., 1992, Intergenerational trasmission of parenting: a review of studies in nonclinical
population, Developmental Review, 12:76-99.
van Ijzendoorn M.H., 1995, Adult Attachment Representations, Parental Responsiviness, and Infant Attach-
ment: A meta-Analysis on the Predictive Validity of Adult Attachment, Inter. Psychological Bull.,
117:387-403.
Wagenaar W.A., 1987, My memory. A study of autobiographical memory over six years, Cognitive
Psychology, 18:225-252.
Zanarini M.C., 1993, Borderline personality disorder as an impulse spectrum disorder, In Paris J. (Ed.), Bor-
derline Personality disorder. Etiology and Treatment, American Psychiatric Press, Washington DC.
7
ADOLESCENCE AND PERSONALITY

DISORDERS
Current Perspectives on a Controversial Problem
Enrico de Vito,! Pran~ois Ladame,2 and Alvise Orlandini3
1Director of Progetto A
Centre for the Study of Adolescence
Azienda USSL 26
Via Cefalonia, 5-20097 San Donato Milanese
Milan, Italy
2Paculty of Medicine of the University of Geneva
Geneva University Hospitals
16-18 Bd. Saint-Georges
1205 Geneva Switzerland
3Medical Psychology and Psychotherapy Unit
San Raffaele Hospital
University of Milan School of Medicine
Via L. Prinetti, 29-20127 Milan, Italy
EDITOR'S INTRODUCTION
This chapter focuses on the controversial problem of the meaning and validity
of the diagnosis of Personality Disorders (PD) during adolescence. Is there anything
specific to PD during this life phase? What are the internal and external factors
producing such specificity? Is it possible to go beyond the difficulty of discriminating
between PD and other more transient fonns of psychological distress during
adolescence?
In the following three sections, the authors focus on these topics through differ-
ent perspectives to provide an articulate contribution. The first section presents a patho-
genetic model which considers PD inside a continuum of psychological suffering,
78 E. de Vito el al.
deriving from more or less severe disturbances of the development of Self-regulation

in a relation system. Normative stresses and other unpredictable events occurring
during adolescence can worsen the disorders of the Self-regulatory functions, and deter-
mine the onset of different symptomatological pictures, which can become stable. Such
pathogenetical model is empirically validated by recent studies based on the attach-
ment theory.
The second section focuses on the controversies of PD diagnosis during adoles-
cence, relying only on descriptive and behavioral criteria, ignoring the underlying
psychic functioning. The psychic functioning of borderline adolescents will be presented
from a psychoanalytic perspective, considering particularly the impossibility of defin-
ing and maintaining psychic space limits, because of binding activity failure. An effec-
tive treatment has to take into account specific characteristics of psychic functioning
of borderline adolescents (particularly splitting). An intensive and reliable psy-
chotherapeutic setting is highly recommended to achieve substantial and not superfi-
cial modifications.
The third section presents follow-up study data focusing on the stability of PD
diagnosis in adolescence according to DSM-IY. These data, coherent with the literature
briefly reviewed, show low levels of stability for many PD and a tendency towards
decrease of PD diagnosis from adolescence to adulthood. Future research should iden-
tify markers of stability of PD, together with risk and resilience factors specific of ado-
lescence for diagnosis, prognosis, and treatment.
1. VULNERABILITY AND PERSONALITY DISORDERS IN

ADOLESCENCE
Enrico de Vito, M.D.
1.1. Introduction
Most clinicians and researchers share the opinion that it is particularly difficult
to recognize and diagnose Personality Disorders (PD) in adolescence, due to the flu-
idity and lack of constancy that characterize this phase.
For instance, there has been considerable discussion on the difficulty of distin-
guishing between borderline syndromes and transient developmental disorders. These
clinical pictures more or less overlap with the concept of turmoil or identity crisis, in
which sometimes can be seen very intense increase in impulsiveness, anxiety and con-
flicts, relational difficulties, feelings of emptiness, inferiority and shame, and, more in
general, changing psychic and behavioural symptoms.
In a recent study (1994) conducted at Yale University, Garnet et al. noted
that stability of the diagnosis of PD in a population of hospitalised adolescents
(aged between 15 and 19) was very low (only 33% of the patients diagnosed at the
start as borderline according to the DSM III-R criteria maintained the same diagnosis
two years later). Also the specific nature of the diagnostic criteria (linked to certain
symptoms) was found to be very low. Differing from the theories maintained in
the past, for instance by Masterson (1972, 1980), the authors of this study have there-
fore reiterated the uncertainty of a diagnosis of borderline personality disorder in
adolescence-to be confirmed during the course of a long-term assessment-and the
usefulness of applying less specific diagnostic tools, which could retain their validity
better in time.
Adolescence and Personality Disorders 79
To model these less specific diagnostic tools, it is essential to refer to the field of
developmental psychopathology, with its main concepts rooted in the psychoanalytical
theories of object relations and with the recent acquisitions originating from infant
research and the attachment theory.
In this paper, the concept of vulnerability of the Self system in adolescence is con-
sidered as a diagnostic frame implying the recognition of a continuum of psychic suf-
fering with variable psychopathological expression, placing the adolescent in an area
in which there are risks of breakdown in the integrity of the Self (that is to sayan area
adjacent to the borderline and psychotic area).
To better define this area, after a brief review of theoretical prospects on the
arrested development underlying PD, it is useful to focus on the contributions provided
by studies on attachment to the understanding of vulnerability as a failure in the devel-
opment of self-regulation within a relation system during the first three years. Lastly,
we intend to show that the setting of the "Short Individuation Psychotherapy" can be
a specific and effective setting for a differential assessment in this area of adolescent
pathology.
1.2. Review of Literature

Using the developmental symbiosis and separation-individuation model, M.
Mahler (1971) has assumed that a developmental arrest between the second and third
year of age (during the so-called rapprochement sub-phase, when the child makes a
transition from exploration to asking hislher mother for new emotional supply) may
be at the root of the disorder of Self construction typical of borderline personality
disorders.
Starting out from this model, albeit from significantly different points of
view, authors such as Masterson and Rinsley (1975), Kernberg (1978), and Adler
(1985) all agree in recognizing the pathogenetic importance of a failure in the separa-
tion-individuation process as a basis for the borderline disorder, maintaining
partial object and primitive defence relations (in particular splitting). Most of the symp-
toms of borderline patients are apparently based on the incomplete and distorted
development of structures capable of adapting to the growing and diversified demands
coming from inside and outside, during the course of the subsequent phases of the
life-cycle.
Adolescence, as Bios (1967) has indicated, can be seen as a new version of the
separation-individuation process. The individual finds himself/herself swaying from the
urge to explore, a desire for independence, and moving away from the parental figures,
to the ever-present need for dependence, stressed by his/her bewilderment and fear
caused by the new demands coming from hislher body, from the intensity of personal
drives and from the external world. The imbalance on the affective level is at times
offset, but often complicated and made problematical by the changes at cognitive level,
that is to say by the transition from concrete operational thought to formal thought.
The psychological world becomes exciting and confusing, intriguing and threatening to
work through. In this framework, the adolescent also sways between on one hand
thrusts to progress towards a horizon that is rapidly expanding in a conceptual and
abstract sense in a wish for "others", and therefore for possible futures, and on the
other hand his/her feeling of being restricted, small and insignificant, with the conse-
quent need to regress back to concreteness and omnipotence and to the denial of a
reality which is seen as concretely terrifying. To explain this swaying condition, and the
80 E. de Vito et al.
often difficult or broken-off transition through this area of experience, it has become
essential to resort to Winnicott's concepts (1953) of transitional phenomena and objects
and of potential space, that is to sayan intermediate area between internal psychic
reality and external reality, in which it is possible to develop the process of dialectics
with others which leads to the creation of subjectivity, of the true Self, or in which, on
the contrary, this process may fail.
Masterson (1972, 1980) and Rinsley (1988) have suggested that the mothers of
borderline adolescents may, in the course of the second separation-individuation
process, send their children the message that growing up, that is to say asserting their
own ineluctable impulse towards separation and individuation, will cause the loss of
maternal support and that, conversely, remaining dependent, and therefore essentially
symbiotic, will guarantee her continued care and support.
This message is linked to the mother herself not being separated and individu-
ated by her own mother, and therefore being a borderline herself, who retains her bond
with her child due to her own need to project in the child her own unresolved separa-
tion dilemma. The effect for the child is the threat of abandonment and loss as pun-
ishment for his efforts to grow up. This leads to a pervasive form of anxiety and fear
of separation, which these authors have called abandonment depression. Going back
to Winnicott's concepts, one could say that a relationship with a not good enough
mother makes it difficult for the individual to build up one's Self (or causes this process
to fail), first as a child then as an adolescent (due to the impossibility of finding his/her
mind separate in the mind of the object). This therefore entails a tendency towards
regression and fusion and towards a collapse of potential space, which alters more or
less seriously the dialectics between reality and fantasy.
The attachment theory (Bowlby, 1969, 1973, 1980) has addressed the vicissitudes
of attachment to parental figures, of separation and loss, and has brought into focus
normal and pathological development of the Self and of object relations during the
various phases of life. According to this theoretical model, at the centre of personality
organisation there is the structuring of Internal Working Models (IWM), that is to say
of dynamic mental representations of the Self and of the attachment figures emerging
from the dyadic relationship and repeated interactive sequences. The characteristics of
cohesion of the Self, of affective regulation, and of relatedness, reflect the nature of the
Internal Working Models.
During adolescence, the IWM's undergo reorganization, which takes place on the
basis of the quality of the earlier and current attachment relationships, in other words,
of how a response was and is given (according to phase-specific procedures) to the need
for security, proximity, and emotional contact with parents or new figures (de Vito,
Muscetta, 1992).
1.3. The Continuum Vulnerability-Psychopathology

This model seems to provide a valid framework, sufficiently consistent with the
concepts mentioned above, for a continuum between normality and pathology. This
implies a continuum between the area of vulnerability characterised by affective insta-
bility which jeopardises the ability to develop both adequate relationships and per-
sonality disorders. The complex changes which take place during adolescence,
combined with contingent stresses and traumas of various origins and magnitudes, can
lead to disorders of Self image or to a syndrome more or less overlapping the border-
line syndrome, where a failure to regulate the relational attachment system lies. This
failure could be the parent's in providing affective attunement or could be found in
non-availability, intrusiveness and indulgence, seductivity or role-reversal, in the inabil-
ity to provide containment, set limits, and offer space for the other in one's mind. This
may also occur where there are shortcomings or distortions in the framework of poten-
tially supportive relationships other than in the family (peer groups, etc.).
The more this condition of emphasised vulnerability is grafted into a serious
developmental failure which took place in the early stages of attachment-in the first
two years of life-and/or into a situation affected by traumas or losses, and/or into a
predisposed constitution, the more likely it is that a borderline personality disorder will
develop and stabilise at the end of adolescence, with different degrees of severity (see
Fig. 1).
Moreover, the conceptualisation of attachment provides a trans-generational per-
spective for developmental disorders and psychopathology, validated empirically. This
DISREGULATION IN THE RELATIONAL TRAUMAS AND SEVERE DEVELOPMENTAL

SYSTEM DURING THE PREVIOUS FAILURES IN THE RELATIONAL SYSTEM
PHASES OF DEVELOPMENT DURING THE PREVIOUS PHASES OF
DEVELOPMENT
NORMATIVE STRESSES NORMATIVE STRESSES
UNPREDICTABLE UNPREDICTABLE
LIFE EVENTS LIFE EVENTS
\ I
\ DISREGULATION N THE RELATIONAL
SYSTEM IN ADOLESCENCE
VULNERABILITY IN SELF-REGULATION
I
•
I
BORDERLINE PERSONALITY DISORDER
IN ADOLESCENCE I IN ADOLESCENCE
IN THE
MODCLATION
IN THE
CONTROL
INTHE
MONITORING
I
OF AFFECT OF BEHAVIOR OFSELF •
1 1
SYMPTOMATOLOGY
Figure 1. Developmental model showing interaction of early relational experiences with other subsequent
contributing factors in producing vulnerability in self-regulation or Borderline Personality Disorder in
adolescence.
suggests that a parent's ability to respond appropriately to his child's needs and anxi-
eties, in particular those relating to separation (that is the parent's ability to convey
what Ainsworth (1978) defined as a "secure attachment"), depends on whether or not
he has structured "secure models" of attachment to his parents in his own mental life.
The "insecure" attachment cycle is perpetuated when parents with unhappy or trau-
matic experiences which have not been metabolised transfer the insecure attachment
pattern to their child, pressing him/her to develop or strengthen primitive defence
mechanisms to keep his/her affects within tolerable limits and to reduce the risks of
fragmentation.
A recent introduction to this field of research is the Adult Attachment Interview
(AAI) developed by Mary Main (1993), aimed at studying the mental state of the adult
and of the adolescent with regards to attachment.
In assessing the answers to questions on one's past experiences of attachment,
the examiner pays particular attention to the "coherence" with which the person relates
his own story, or in other words his autobiographical competence, which, as Fonagy
(1991) has remarked, is an expression of the spontaneous ability to observe one's own
mental functioning, or the validity of one's "self-reflective function" (SRF).
Fonagy himself (1995) is perfecting a trans-generational model of borderline
pathology, starting from the study of metacognitive monitoring and of self-reflective
function by means of the AAI method. According to Fonagy, "severe narcissistic and
borderline states can be understood in connection with a dysfunction of the self-
reflective functions. Individuals exposed to early traumas, for example, may face the
intolerable prospect of conceiving the mental state of their tormentors only by refus-
ing to recognise it, through a defensive destruction of their ability to define feelings
and thoughts about themselves or others. This forces them to work on the basis of
schematic and inaccurate impressions of thoughts and feelings, and makes them vul-
nerable in all close relationships".
In adolescence, the search for identity leads through more or less successful
attempts to focus one's representations of the Self, to link them to one another in a
synchronic and diachronic sense: in other words, to activate the individuation processes
in a functional manner. This entails a complex work of self-reflection which is closely
linked to and made possible by having at one's disposal a "theory of the mind", built
up during the development on a relational matrix, and updated on the basis of the new
conditions or relational systems peculiar to this phase. This construction of the theory
of mind must be linked both to the relationship between fantasy and reality and to the
functioning of memory, in order to allow mature and personal working through of the
person's time dimension.
The disorder of one's sense of identity in adolescence may therefore be seen as
a disorder of self-representation, relationship between fantasy and reality and record-
ing of mnestic traces, seen as a consequence of a failure in the relationship with
an object having psychic qualities, capable of reflecting the thinking mind of the
adolescent.
We have recently started a research project to acquire empirical data on the rela-
tionship between psychic vulnerability in adolescence, "insecure" Internal Working
Models and possible alterations in the construction of the Self representations. To do
so we have administered the Adult Attachment Interview to 80 adolescents from the
general popUlation. The preliminary results seem to show a correlation between ado-
lescents with severe disorders of Self image and Internal Working Models of "insecure"
attachment.
The possibility of providing a specification of the different categories of

these "insecure" patterns (for example, of classifying temporary collapses during the
monitoring of the conversation, as may happen while treating specific topics), and
thus of trying to correlate these different patterns with different expressions of vul-
nerability and of disorder, would suggest the clinical use of the A.A.I. in adolescence,
during the diagnostic phase and to check the changes occurring during the course of
therapy.
1.4. The Short Individuation Psychotherapy and

The Vulnerable Adolescent
Let us now return to the more strictly clinical framework to comment on the spe-
cific nature of the setting of the Short Individuation Psychotherapy, conceptualised by
Tommaso Senise (Aliprandi, Pelanda, and Senise, 1991) to assess the vulnerability of
the Self during adolescence.
In his paper "For adolescence: psychoanalysis or analysis of the Self?", Senise
(1981) wrote regarding the self-reflective functions discussed above: "During adoles-
cence, the individuation processes are severely tried. Still inadequate due to their imma-
ture investigation procedures, they are rendered even more ineffective by the loss
of validity of the images reflected by adults. They are therefore often suspended,
inhibited, or worked only sectorially, due to defensive activities and/or previous
shortcomings".
To probe the introspective capabilities, the activity and use of individuation
processes and, at the same time, where possible, to try to help the patient overcome
blocks or stiffening spontaneously, the Short Individuation Psychotherapy proposes
three basic operations:
1) Correct empathic and overall identification with the adolescent, but more
selective and precise in the framework of Self-monitoring processes.
2) Promotion in the adolescent of a counter-identification with the therapist
identified with him/her, specifically in the selective framework referred to
above, which causes a mirror effect.
3) The third operation is both a basis for and a consequence of the other two.
It consists of seeking at the end of a brief activity an early working alliance
to draw, together with the adolescent, a self portrait reflecting as realistically
as possible the adolescent's self image and that of his/her fundamental rela-
tionships, and of the investigation procedures by means of which he/she
reached this image.
It is clear that this pathway is followed in a very different way by adolescents with
different levels of psychological functioning. There are adolescents in a condition of
increased vulnerability, but still able to differentiate themselves from others and to
reflect themselves (and therefore to bring their self-monitoring activity, spontaneity,
and curiosity back into play fairly rapidly), and other adolescents who are at a serious
standstill in terms of development, only partially able, proceeding by trials, escapes, and
falls, to recognise the importance of the other and to use this for flashes of represen-
tations or of memories and for brief reconstructing segments.
For the latter, this intervention is basically the preliminary stage of a therapeutic
relationship which will allow for a significant psychological change only in the medium
or long term.
84 E. de Vito et aL
2. BORDERLINE PERSONALITY DISORDER IN ADOLESCENCE

Fran~ois Ladame, M.D.
2.1. Introdnction
Between DSM (either III -R or IV) and a psychodynamic or psychoanalytic
point of view there may be a possibility of agreement around the issue of "identity dis-
order". Nevertheless we have serious doubts regarding the relevancy, at least in ado-
lescence, of the DSM distinction between borderline personality disorder (BLPD) and
narcissistic personality disorder (NPD).According to our clinical and therapeutic expe-
rience, it is much more like a Janus-faced situation, mainly depending on the impor-
tance of the mood disorder. When depressive features are on the foreground, we
observe dependant as well as clinging attitudes and it is not at all difficult to collect
most criteria for BLPD diagnosis. On the contrary, whenever depressive features are
not clearly present, we are struck by omnipotence and self-sufficiency as well as by atti-
tudes of contempt, characteristic of a NPD. Hence, according to the current psychiatric
way of speaking, it seems that the difference is more of a dimensional nature than of
a categorial one.
We all know that a narcissistic breach characterises all pathologies which are not
clearly neurotic. This breach is also evident in psychosis, but in the latter the conse-
quences are much more dramatic, i.e. the absence of a subject but in a delirious neo-
reality, whereas the borderline patient is able to mask it while relying on the many and
various figures of the double. This means that borderline patients will adopt borrowed
identities. Hence the question: are not all adolescents borderline? The answer is defi-
nitely "no" for the reason just mentioned above: in the borderline patients, the bor-
rowed identities poorly and superficially succeed in masking the narcissistic breach
which evidently has to be considered as one of the "criterias" for diagnosing the
disorder.
2.2. Questioning the Specificity of the BLPD

Authors who question the specificity of BLPD underline the fact that this diag-
nosis is not clearly validated according to the criteria of demonstration of a consistent
course, response to treatment, biological markers, and familial aggregation. Tyrer, in
London, is one of the most violent attackers, pinpointing the vague and fuzzy borders
of borderline personality disorder (Tyrer, 1994). He argues that the borderline state is
a coaxial diagnosis of personality traits and current axis I disorder. He also stresses sim-
ilarities with unstructured and narcissistic aspects of adolescence and takes for granted
that borderline conditions are developmental disorders. Nurnberg et al. (1991), from
the Albert Einstein College of Medicine, also consider BLPD as a broad, heteroge-
neous category with unclear boundaries. There is extensive overlapping with neigh-
bouring personality disorder categories (82% against 41 % among those with other PD
diagnoses) .
Papers that strictly refer to adolescent patients conclude in the same way. Accord-
ing to Bernstein et al. (1993), a longitudinal follow-up of a randomly selected commu-
nity sample of 733 youths reveals that most PD do not persist over a 2-year period
(persistence of BLPD varies between 24 and 29% according to the severity of the dis-
order). Garnet et al. (1994), from Yale University School of Medicine, have contacted
21 adolescent inpatients with BLPD 2 years after index hospitalisation, and only 7 of
them met criteria for BLPD at follow-up. "For such patients, diagnoses of identity dis-
order along with depression or conduct disorder may best capture what is clinically
salient, both currently and prognostically" (p. 1382).
In comparison, those papers which attest to the specificity of BLPD are fewer,
and less convincing. For instance, Johnson et al. (1995), from the Western Psychiatric
Institute at Pittsburgh, studied the validity of PD diagnoses in adolescents according
to the criteria of familial aggregation. Their results support the validity of axis II diag-
noses, particularly avoidant and borderline disorders. But personally we have doubts
regarding a demonstration based only on this criteria. Consideration for the mecha-
nisms of identification would offer an alternative explanation.
This short review of the current psychiatric literature confirms the difficulty, not
to say impossibility, of establishing a diagnosis of borderline personality in adolescents
based only on semiological criteria, i.e. without taking into account the psychic func-
tioning behind symptoms. We think that the diagnosis of borderline functioning is reli-
able according to other diagnostic criteria stated below.
2.3. Borderline Disorder in Adolescence and Sexual Abuse in Childhood

We would have preferred to avoid the whole issue, but do not feel we have the
right to do so considering the current excesses, especially in today's American litera-
ture. To phrase it simply, the question is whether BLPD is the direct outcome of a sexual
abuse during childhood or, even, a particular form of PTSD. The answer cannot be any-
thing different from "no", if not for the simple reason that sexual abuse is quite common
within the general population whereas borderline disorder is a fairly rare condition
(according to a Geneva study among a randomly selected sample of more than one
thousand 13-16 year-old schooled adolescents, 20% of girls and 4 % of boys have been
forced to direct physical sexual contacts either by adults or older mates, Halperin, 1996).
This may not be evidence but it shows how necessary it is to take into account factors
of resilience which may allow children to resist even extreme conditions and on the
other hand, as far as etiopathogeny is concerned, to rely on more complex hypotheses
than it is unfortunately the case in many current papers. Most laymen can today attest
that the notion of child sexuality and its central role in the psychological develop-
ment-whether it follows a normal or a pathological course-are simply ignored in a
surprisingly high number of studies which try to make a causal relationship from a
simple correlation. Even when we consider the higher estimates of past sexual abuse
among groups of borderline patients, they range from 50 to 80% (see Pahud, 1996).
This means that sexual abuse is not necessary for the development of BLPD. Sexual
abuse may aggravate the condition, but we are not allowed to infer that it is the "cause"
of the disorder.
2.4. Psychoanalytic Point of View

Is a subject borderline or does he have something borderline about him? We are
in favour of the second hypothesis, starting from the idea that psychic life is a contin-
uum between psychosis and neurosis, between the paranoid-schizoid and depressive
positions. In an exemplary way the borderline patient shows us these two modalities
of function, and his capacity to pass from one to the other (sometimes during the course
of the same session) cannot but astonish us. Other opinions are put forward in the psy-
choanalytic literature, for instance Bergeret (1974), in France, as well as Kernberg
(1977) speak of a particular form of psychic organisation, suggesting the character of

a stable system, whereas Masterson (1986) defines borderline pathology as a person-
ality organisation type which is not psychotic.
One cannot speak of borderline states without considering the question of limits.
Whenever the limits of psychic space (representational space, internal and external,
conscious and unconscious) cannot play their role, one flows over into bodily space or
that of enactment. That is one of the principal characteristics of borderline pathology
around which most symptomatic productions are articulated. The binding activity is at
issue between the drive and its representations, as well as the work of transformation
which permits a passage from somatic to psychic activity, to differentiate a perception
from a representation and to establish reality testing. This binding activity is one of the
main functions of the preconscious which are attested through displacement and
repression as well as the work of dreams. For those who are familiar with the Laufers'
formulations, similarities should be obvious with the "developmental breakdown", the
condition when nothing prevents unconscious fantasies to overwhelm conscious ones
(Laufer and Laufer, 1984). The failure of binding activity is responsible for an exces-
sive condensation of drive activity and for enactment. The latter can be compared to
"acted out nightmares" in the situations when the work of dreaming has been switched
off as a binding activity.
All authors agree with the importance of splitting in the borderline disorder. We
personally would disregard the diagnosis of borderline pathology when there is no evi-
dence that splitting is one of the main defence mechanisms. It is the excess of splitting
that permits the borderline subject to function at the same time in the register of neu-
rosis and psychosis while avoiding, sometimes throughout life, a clinical psychotic
decompensation. A few symptoms attest in some way to the "failure" of splitting as an
airtight mechanism. According to our experience, it is the case with suicidal attempts
which should therefore be considered as a unique opportunity-still too often missed-
for proposing and beginning a treatment.
A question arises naturally. What prevents the balance to lean toward rebinding,
as it is the case in neurotic/normal adolescents? Probably borderline pathology depends
on diverse factors, before and during adolescence, but a constant is the failure of the
infantile neurosis, meaning the lack of a specific organiser permitting to maintain on
the level of castration everything that is susceptible to initiating the breakdown of nar-
cissistic homeostasis. The diverse and inevitable traumas of the pregenital phases do
not allow the achevement the integration made possible by the Oedipus complex or,
more properly, its decline. It is thus a failure of the after-effect, heavy with consequences
for the ego: absence of history and of insertion in time.
2.5. Therapeutic Perspectives

The particularities just described will influence the treatment. But, first of all
and before addressing the issue of treatment, it is necessary to be very clear regarding
the therapeutic goals. It is acceptable to propose minimal solutions when we only
aim at a symptomatic improvement. However, we must know that symptomatic treat-
ments cannot really modify the situation. If something is really to change in the psychic
functioning of the borderline adolescent (besides symptomatic modifications or
evolutions of an "as if" facade), it can only be within a fixed and intensive therapeutic
setting.
To illustrate the meaning of therapeutic goals, we will focus on a particular group
of borderline adolescents, those who have attempted suicide. It has gradually occurred
to us since the start of working on the issue of adolescent suicide and borderline pathol-
ogy that one of the main characteristics of the suicidal act is to be paradoxical, which
means to aim at two simultaneous but incompatible goals (Ladame and Ottino, 1993).
This aspect of the question, when treating these adolescents, taxes our own ability to
think because we are faced with the difficulties of a dual thinking pattern. Once more
it is the splitting which permits this strange state of things. For example: we all are very
familiar with the ever-present question whether suicidal adolescents really wish to die
when attempting suicide. This question is relevant only as part of a logical thought
pattern. Within the framework of a dual thought process, the answer is also double: the
adolescent wishes to die and at the same time he does not wish to die. Both sets of
thoughts are concomitantly present and expressed through the same gesture. When-
ever we aim at unmasking suicidal potentiality, undoing the underlying pathology, it is
necessary to also work through suicide as a paradoxical act. Therapists should there-
fore be very watchful about this specific "thought disorder" which mirrors at the cog-
nitive level the co-occurence of neurotic and psychotic functioning permitted by
splitting.
To come back to the therapeutic setting, it has to be intensive and as unvarying
as possible, i.e. potentially indestructible, because these patients habitually have a
fantasy of destroying the oedipal parents and of realising a fatal incestuous desire.
Usually this fantasy is no more unconscious, it can no more be repressed and it intrudes
into the conscious, as underlined it before. At those moments, the therapist becomes
identical with tyrannical, persecutory, internal objects, demanding tribute for present
and not past offences, for a disaster which has already occurred and still is not yet expe-
rienced, because it has not been registered. The failure of the infantile neurosis pre-
cludes as a primary objective the reconstruction of repressed ideas of the past
childhood. It also means a lack of containment, and the latter function has to be tem-
porarily endorsed by the therapist whose position is now clearly paradoxical. Hence
the necessity, once more, of a potentially indestructible setting.
Let us admit that most of these issues have been worked through over the years
in a satisfactory way. Yet we cannot be sure of the final outcome before the termina-
tion phase when all prior benefits can well be challenged again. The main difficulty is
not separation in itself, as usually stressed. The main difficulty is to accept that treat-
ment comes to an end even if the patient has not succeeded in seducing his/her thera-
pist. This implies frustration on two levels. Being able to accept it would mean that the
patient became successfully disentangled from the Oedipus complex and its agonies:
murder and castration-incest, parricide, and suicide. But still, borderline patients are
very well known for their rebelliousness; they are fundamentally insubordinate to
human laws ascribing to every of us one sex and one place within a genealogy (Ladame,
1991).
2.6. Conclusion
Beyond the notion of identity disorder, shared both by the DSM and the psy-
choanalytic point of view, we wonder whether these two approaches have anything else
in common, being opposed in their respective fundamentals: the first looks at descrip-
tive criteria statistically shared by a large group of individuals, the second stresses
psychic functioning and considers the uniqueness of every human subject.
We do not see how it might be possible to have a representation of a borderline
patient without relying on the specificities of the mental functioning, as the ones par-
ticularly underlined above, like splitting, which permits the co-occurence of neurotic
and psychotic functioning (as Freud put it in his short 1924 article on neurosis and psy-
chosis) as well as a dual paradoxical thought process.
Regarding the therapeutic approach, we argued in favour of an intensive and reli-
able setting insofar as we aim at a real modification of psychic functioning and also pin-
pointed the challenges of the termination phase.
3. THE VALIDITY OF THE DIAGNOSIS OF PERSONALITY

DISORDERS IN ADOLESCENCE THROUGH
FOLLOW-UP STUDIES
Alvise Orlandini, M.D.
3.1. Introduction
According to the Diagnostic and Statistic Manual of Mental Disorders~IV
Edition (DSM IV) "Personality Disorders (PD) categories may be applied to children
or adolescents in those unusual instances in which the individual's particular maladap-
tive personality traits appear to be pervasive, persistent and unlikely to be limited to a
particular developmental stage, or an episode of an Axis I disorder. It should be recog-
nised that the traits of Personality Disorders that appear in childhood will often not
persist unchanged into adult life. To diagnose a PD in an individual under age 18 years,
the features must have been present for at least 1 year". As clearly stated above, tempo-
ral stability of a diagnostic category over time is considered a relevant theoretical
aspect of PD diagnosis, particularly in adolescence; in fact it provides a basis on which
to predict course and outcome. Therefore, lack of stability of PD during adolescence
suggests lack of validity. For this reason the validity of PD diagnoses during adoles-
cence has always been controversial. Although Axis II diagnoses can quite precisely
describe disturbed adolescents, it is important to consider and distinguish them from
problems of a specific developmental stage. What is typically problematic of PD in ado-
lescence is the prognosis: how long will they last? Do they appear during adolescence
exactly as they appear during adult life? A long term follow up of PD is the only way
to study what really happens during adolescence, and particularly if it is possible to
predict which adolescents will present a PD diagnosis even in adult life. There is clini-
cal evidence of the controversies of PD diagnosis in adolescence. Long term stability
of PD in adults has been investigated more often. For example Reich (1989), in a test
re-test reliability study at eight weeks on 70 out-patients with Personality Diagnostic
Questionnaire-Revised (PDQ-R), found a good temporal stability for Paranoid,
Schizotypal, Borderline, Avoidant, and Dependent Po.
There are instead very few studies on long term stability of PD in adolescence
(Golombek, 1987; Bernstein, 1993; Mc Glashan, 1995) and such studies can be hardly
compared because of differences concerning sample characteristics (clinical or non clin-
ical sample, different age ranges), PD assessment (semi-structured interviews or self-
administered questionnaires) re-test interval, statistical and results (single PD or
Clusters; categorial or dimensional approach).
Korenblum (1987) studied a non clinical sample of 61 adolescents 13 years old
with a follow-up after three years. The diagnosis of PD according to DSM III, was made
by ratings of a semistructured interview focusing on interpersonal relationships. The
diagnoses were grouped in five clusters: "A" (paranoid, schizoid, schizotypal), "B"
(histrionic, borderline, narcissistic), "C" (antisocial), "D" (avoidant, dependent,
obsessive-compulsive, passive-aggressive), and "E" (atypical, mixed, other). The find-
ings of this study were: 1) a decrease in the prevalence of personality dysfunction from
13 years to 16 years (from 46% to 33%); 2) the features of personality dysfunction
change from a bimodal distribution of anxious-fearful and antisocial types in early ado-
lescence to a more diffuse, less well defined pattern in middle adolescence; 3) the anti-
social and the eccentric withdrawn groups were stable over time, while
"anxious-fearful" adolescents grew out of their difficulties; 4) 25% manifested person-
ality dysfunction in both early and middle adolescence and 28% with personality dys-
function at 13 years fully recovered at 16 years.
Bernstein (1993) studied prevalence and stability of DSM III-R Personality Dis-
orders in a randomly selected community sample of 733 youths ranging in age from 9
to 19 years with a two year follow-up. Both structured interviews (modified versions of
SCID II administered to adolescents and their mothers) and self administered ques-
tionnaires (PDQ-R) were used to diagnose PD. The prevalence of PD peaked at age
12 in boys and at age 13 in girls. 57% subjects at time 1 received a diagnosis of mod-
erate PD, while only 25% received a diagnosis of severe PD. After two years follow-
up, 31 % subjects received a diagnosis of moderate PD, while 17% a diagnosis of severe
PD. Obsessive-compulsive personality was the most prevalent moderate axis II disor-
der, narcissistic personality disorder the most prevalent severe disorder (32 % ). Schizo-
typal personality was the least prevalent axis II disorder. A two year follow-up revealed
that more than a half of PD diagnosis at time do not persist, but those subjects who
had a PD diagnosed were at higher risk for further PD diagnosis after two years. The
probability to receive a diagnosis of PD decreased from 11 to 21 years, reaching the
incidence expected in adult general population.
Mattanah (1995), examined the stability of DSM III-R PD in a sample of 65 ado-
lescent inpatients, contacted for follow up after two years from discharge. This is the
only study conducted on a clinical population. The age ranged from 12 and 18 years.
Semistructured interviews (PDE) were used for axis II disorders. Percentage of sub-
jects diagnosed at baseline (Tl) who met criteria for the same diagnosis at follow-up
(T2), and Cohen's simple K were used as stability statistics. Borderline PD was the most
frequent specific disorder, both at time 1 (48 % ), and at time 2 (23 % ). Schizoid PD was
the less frequent (Tl = 2 % ,; T2 = 0%). Cluster B presented the highest percentage of
stable cases (48%) while cluster A (11 %) the lowest.
The research on PD during adolescence focused also on childhood antecedents.
Bernstein (1996) studied 641 youths with a 10 year follow up. The antecedents
considered were conduct problems, depressive symptoms, anxiety fear and immaturity.
Childhood conduct problems resulted as a predictor of PD in all DSM III-R
clusters, depressive problems predicted cluster "A" PD in boys, and immaturity
predicted cluster "B" in girls. It is interesting that the authors considered PD diagnoses
only if lasting for at least two years and 43.5% subjects did not maintain the diagnosis
for such period.
Some studies focused on stability of single PD and not on all axis II. For example,
two studies suggest stability of schizoid and borderline symptoms from childhood
through adolescence and early adulthood (Aarkrog, 1981; Wolf, 1980). Garnet (1994)
found stability for borderline PD in 33% (7) subjects in a two years follow-up study.
Other studies focused on Antisocial PD which proved to have adequate stability
(Robins, 1966; Robins, 1971). Rey (1995) focused on continuities between psychiatric
90 E. de Vito el al.
disorders in adolescents and PD in young adults. In a sample of 145 subjects, adoles-

cents with Disruptive Disorders (Attention Deficit Disorder with Hyperactivity, Oppo-
sitional Disorder, Conduct Disorder, Adjustment Disorder with Disturbance of
Conduct) showed high rates (40%) of PD diagnoses as adults; while subjects with Emo-
tional Disorders (Separation Anxiety Disorder, Other Anxiety Disorder, Dysthymic
Disorder,Adjustment Disorder with mixed emotional features) had lower rates (12%)
of PD diagnoses as adults.
Nevertheless, besides all differences in methodology, sample, etc., there is quite
an agreement on some points: high prevalence of PD during adolescence in non clini-
cal population, low stability of most PD during adolescence, trend to decrease of PD
as adolescence reach adult age, and quite a stability for some PD.
The following study is a first step of a five year follow up research project. The
aim of such research is the evaluation of stability of PD and object relations during
adolescence.
3.2. Subjects and Methods

64 adolescents (F = 25, 39.1 %; M = 39, 60.9%; mean age = 14.25 0.59 years)
were administered the Personality Diagnostic Questionnaire-4 + (PDQ-4+) and the
Bell Object Relations and Reality Testing Inventory (BORRTI). The sample was
recruited in two high schools of Milan. PDQ-4+ is a 99 true/false items, self adminis-
tered questionnaire designed to gather information on the 12 PD listed in the DSM IV.
PDQ-4+ provides both categorial and dimensional PD diagnoses. PDQ-4+ total
score should be considered as a global personality dysfunction measure. The Bell
Object Relations and Reality Testing Inventory (BORRTI) is a 90 true/false item
self administered questionnaire designed to measure 4 dimensions of object relations
and 3 dimensions of reality testing (Alienation: basic lack of trust in relationships which
are superficial with no real sense of connection or belonging. Suspiciousness: anger
and hostile withdrawal, poor empathy. Insecure Attachment: high sensitivity to rejec-
tion, desperate longings for closeness, poor toleration to separation and loneliness.
Egocentricity: mistrust of others' motivation, others exist only in relation to oneself and
to be manipulated. Social Incompetence: shyness, nervousness, uncertainty about how
to interact with members of the opposite sex; avoidance and escape from relationship.
Reality Distortion: severe distortion of external and internal reality (delusions of
influence and paranoid believes) Uncertainty of perception: keen sense of doubt about
their own perception of internal and external reality. Hallucinations and delusions: hal-
lucinatory experiences and paranoid delusions of various types). Questionnaires
have been administered at the beginning of the year of high school (first PDQ and after
one week BORRTI). Re-test has been done after one year. Cohen's statistic has
been used for the evaluation of categorial diagnoses stability. In order to better under-
stand the results, we remind that temporal stability was considered as percentage of
subjects diagnosed at time 1 who met the criteria for the same diagnosis at time 2
corrected for chance agreement. This statistic (K) takes into account stable positive
cases, stable negative cases, the number of cases that remit, and the number of new
cases (e.g. a high number of either remitting or new cases would reduce the K coeffi-
cient). Pearson's r has been used to test dimensional assessment stability. McNemar's
test has been used for the evaluation of significant trends between time 1 and time 2
for categorial diagnoses, while paired sample test t was used for dimensional data.
Nominal alpha level has been Bonferroni corrected.
3.3. Results
Descriptive statistics of PDQ IV and BORRTI are shown in table 1. At time 1,
at least one diagnosis of PD was received by 85.9% of the subjects (N.55), while at time
2 at least one diagnosis of PD was received by 59.4% (N.39). At time 1 mean number
of PDs was 2.441.96, while at time 2 was 1.301.65.
Temporal stability of PD was evaluated for both categorial and dimensional
evaluations (table 2). Considering categorial diagnoses, Paranoid, Schizotypal, Antiso-
cial, Borderline, and Avoidant PD showed moderate stability. Considering dimensional
assessment of PD, Antisocial, Schizotypal, and Borderline PD showed high test-
retest stability, while all the remaining PD showed moderate temporal stability. The
only exception was Schizoid PD which did not show a significant temporal stability,
probably due to the low prevalence of this PD in our sample. Considering categorial
diagnoses, a significant trend towards a decrease was found in Paranoid, Obsessive-
Compulsive, and Dependent PDs. According to a dimensional evaluation, a significant
trend towards a decrease was observed for Paranoid, Schizotypal, Hystrionic, Narcis-
sistic, and Obsessive-Compulsive PDs, except for a trend towards an increase found for
Dependent Po. Dimensional evaluations demonstrated a better temporal stability and
a higher trend towards a decrease than categorial ones. A high level of temporal
stability was found in two BORRTI scales: Egocentricity and Insecure Attachment; a
moderate level was found in all the remaining BORRTI scales. A significant
Table 1. PDQ IV and BORRTI scales: descriptive statistics

Categorial
Dimensional Evaluation Evaluation
Time 1 Time 2 Time 1 Time 2
Mean ± SD Mean ± DS % %
PDQ IV Scales
Paranoid 2.94 ± 1.67 2.20 ± 1.62 42 23
Schizoid 0.95 ± 1.01 1.02 ± 0.98 1.5 0
Schizo typal 2.36 ± 1.70 1.62 ± 1.39 12.5 4.6
Antisocial 1.91 ± 1.77 1.53 ± 1.54 31.2 20.3
Borderline 2.80 ± 1.99 2.27 ± 1.94 23 17.2
Histrionic 2.47 ± 1.33 2.02 ± 1.40 4.6 4.6
Narcissistic 2.41 ± 1.55 1.78 ± 1.54 6.2 6.2
Avoidant 2.17 ± 1.53 1.75 ± 1.69 15.6 17.2
Dependent 1.25 ± 1.51 2.12 ± 1.54 56.2 10.9
Obsessive-Compo 3.25 ± 1.45 2.70 ± 1.53 50 25
Total PDQ IV 26.86 ± 12.37 21.52 ± 11.41 *** ***
Mean N° PD 2.44 ± 1.96 1.30 ± 1.65 *** ***
At least one PD *** *** 85.9 59.4
BORRTI Scales
Alienation -0.28 ± 0.54 -0.31 ± 0.45 *** ***
Egocent. -0.04 ± 0.57 -0.26 ± 0.55 *** ***
Ins. Attachment -0.06 ± 0.57 -0.31 ± 0.62 *** ***
Social Incomp. -0.34 ± 0.55 -0.46 ± 0.48 *** ***
Reality Distortion -0.01 ± 0.58 -0.26 ± 0.40 *** ***
Hall. & Delusions -0.10 ± 0.82 -0.30 ± 0.52 *** ***
Un. of Perception -0.01 ± 0.49 -0.10 ± 0.51 *** ***
.** = Non computable statistic.
92 E. de Vito et aL
Table 2. PDQ IV and BORRTI scales: temporal stability statistics

Categorial Evaluation Dimensional Evaluation
K McNemar r
P·Bonferroni = 0.005 P'Bonferroni = 0.005 p'Bonferroni:: 0.0028 p'Bonferroni == 0.0028
PDQ IV Scales
Paranoid 0.52 0.0028 0.61 4.05
«0.001) «0.001) «0.001)
Schizoid *** *** 0.30 -0.42
(NS) (NS)
Schizotypal 0.52 0.0625 0.69 4.70
(<0.001) (NS) «0.001) (<0.001)
Antisocial 0.56 0.0654 0.76 2.55
«0.001) (NS) «0.001) (NS)
Borderline 0.42 0.3877 0.68 2.72
(<0.001) (NS) «0.001) (NS)
Histrionic -0.05 1.0000 0.59 2.94
(NS) (NS) «0.001) «0.0056)
Narcissistic 0.20 1.0000 0.54 3.39
(NS) (NS) «0.001) «0.0028)
Avoidant 0.49 1.0000 0.56 2.24
«0.001) (NS) «0.001) (NS)
Dependent 0.17 0.0000 0.48 -4.42
(NS) «0.001) «0.001)
Obsessive-Compo 0.25 0.0025 0.49 2.92
(NS) «0.001) (<0.0056)
At least one 1 DP 0.39 0.0000 *** ***
«0.001)
Total PDQ IV *** *** 0.64 5.84
«0.001) «0.001)
BORRTI Scales
Alienation *** *** 0.43 0.52
«0.001) (NS)
Egocentr. *** *** 0.73 4.11
«0.001) «0.001)
Ins. Attachment *** *** 0.67 4.08
(<0.001) «0.001)
Social Incomp. *** *** 0.51 1.81
«0.001) (NS)
Reality Distortion *** *** 0.61 4.17
«0.001) «0.001)
Hall. & Delusions *** *** 0.38 2.08
«0.0028) (NS)
Un. of Perception *** *** 0.44 1.30
«0.001) (NS)
*** = Noncomputable statistic; NS = Nonsignificant.
trend towards a decrease was found for Egocentricity, Insecure Attachment and Reality
Distortion.
3.4. Conclusions
In agreement with previous studies, PDQ-4+ seemed to overdiagnose PD. In fact,
the prevalence of PD in this sample was higher than expected for the general popula-
tion. According to the literature, at one year follow-up we found a significant decrease
in the prevalence of subjects with at least one PD diagnosed and also a decrease of
mean number of po. Levels of stability of PD in our sample are in agreement with
those observed in previous studies. We can distinguish:
1) PD with moderate temporal stability and significant trend toward decrease:

Paranoid and Schizotypal. Moderate temporal stability together with trend
toward decrease of both categorial and dimensional scores could be consid-
ered as an index of PDQ-4+ better diagnostic accuracy in the proceeding of
adolescence. This means, for example, that in the following times of a follow-
up evaluation, there is a decrease in the number of subjects who maintain a
given diagnosis because of false positive decrease, with more accurate iden-
tification of subjects who are likely to have a po. Therefore the diagnoses of
Paranoid PD and Schizotypal PD gathered with PDQ 4 require particular
attention.
2) PD which show moderate stability with no significant decrease: Antisocial,
Borderline, Avoidant. In this group moderate temporal stability is associated
to absence of trend, because of equally frequent of false positive and false
negatives. The stability of Antisocial and Borderline PD confirms literature
and are of relevant clinical interest.
3) PD with no stability and consequently with a low replicability and low clini-
cal usefulness: Schizoid, Histrionic, Narcissistic, Dependant, Obsessive Com-
pUlsive. k value for Dependent PD differs from literature (0.75).
4) As demonstrated in a previous study (Orlandini, 1995), there is a strong asso-
ciation and good predictive efficiency of BORRTI scales on all PD (excepted
Obsessive Compulsive). Therefore BORRTI scales can be considered exter-
nal validators of PD assessment in adolescence. It should be noted the sig-
nificant association between descriptive features and inner distress in
adolescence. In particular a strong association (r adjusted = 0.52; F = 39.2; DF
= 2.69; p. < 0.001) was observed between PDQ-4+ composite score and Inse-
cure Attachment and Reality Distortion. Limits of the present study are non
random sample recruitment, the use of self administered questionnaires, and
the overinclusiveness of PDQ-4+. It is important to underline the perspective
of such diagnoses as the DSM IV one: therefore the above mentioned results
cannot be extended at all to structural diagnoses or psychodinamic concepts,
nor to other standardised interviews. In conclusion, we can affirm that some
of the Axis II diagnoses (see above) can quite accurately describe patholog-
ical pictures during adolescence which can be differentiated in remitting syn-
dromes and long lasting PD only through follow-up studies. In other words,
while a PD diagnosis in adults describes a pathological, stable, and pervasive
syndrome, in adolescence a PD diagnosis requires much more attention and
re-test, due to the psychological distress of the developmental tasks of
adolescence.
Future research should identify stability markers of PD during adolescence such

as psychiatric inheritance, comorbidity between axis I and axis II disorders, psychoso-
cial stressors, and social support. The identification of such markers could help in more
accurate diagnoses of po.
REFERENCES
Aarkrog, T., 1981, The borderline concept in childhood adolescence, and adulthood: borderline adolescent in
psychiatric treatment and 5 years later, Acta Psychiatrica Scandinavica Suppl, 293:1-300.
Adler, G., 1985, Borderline psychopathology and its treatment. New York: Jason Aronson.
Ainsworth, M., Blehar, M.e., Waters, E., and Wall, S., 1978, Patterns of attachment: A psychological study of
the Strange Situation. Hillsdale, NJ: Erlbaum.
Aliprandi, M.T., Pelanda, E., and Senise, T.L., 1991, La psicoterapia breve d'individuazione. Milano: Feltrinelli.
American Psychiatric Association, 1994, Diagnostic and Statistical Manual of Mental Disorders, Fourth
Edition. Washington DC, American Psychiatric Association.
Bell, M., 1991, An introduction to the Bell Object Relations and Reality Testing Inventory (BORRTl) Manual.
Western Psychological Services, Los Angeles California.
Bergeret, J., 1974, La personnalite normale et pathologique. Paris, Dunod.
Bernstein, D.P., Cohen, P., Velez, e.N., et aI., 1993, Prevalence and stability of the DSM- III-R personality
disorders in a community-based survey of adolescents, Am.] Psychiatry, 150:1237-1243.
Bernstein, D.P., Cohen, P., Skodol, A., Bezirganian, S., and Brook, J., 1996, Childhood Antecedents of Adols-
cents Personality Disorders, American Journal of Psychiatry, 153:7, 907-913.
BIos, P., 1967, The second individuation process of adolescence, Psychoanalytic Study of the Child, 22:162-186.
Bowlby, J., 1969, Attachment, Vol. I, Attachment & Loss, London: Hogarth Press.
Bowlby, J., 1973, Separation, Vol. II, Attachment & Loss, New York: Basic Books.
Bowlby, J., 1980, Loss, Vol. III, Attachment & Loss, New York: Basic Books.
de Vito, E. and Muscetta, S., 1992, Attachment and disattachment in adolescence: their relationships with
gender differences. Third International Congress of the International Society for Adolescent Psychia-
try, Chicago, (July 1992).
Fonagy, P. and Higgitt, A.C., 1990, A developmental perspective on borderline personality disorder, Rev. Int.
Psychopathol.,1:125-159.
Fonagy, P., Steele, M., Steele, H., Moran, G.S., and Higgitt, A.e., 1991, The capacity for understanding mental
states: The reflective self in parent and child and its significance for security of attachment, Infant
Mental Health Journal, 12, No 3,201-218.
Fonagy, P., 1995, Psychoanalytic and empirical approaches to developmental psychopatology: an object-
relations perspective, In: T. Shapiro, R.N. Emde (Eds). Research in Psychoanalysis (pp. 245-260).
Madison, International Universities Press.
Freud, S., 1924, Neurosis and psychosis, S.E., 19: 149.148, 1371-1377.
Garnet, K.E., Levy, K.N., and Mattanah, J.J.F., 1994, Borderline personality disorder in adolescents: ubiqui-
tous or specific? Am. ] Psychiatry, 151: 1380--1382.
Giovacchini, P., 1972, Technical difficulties in treating some characterological disorder: countertransference
problems, International Journal of Psychoanalytic Psychotherapy, 1:112-128.
Halperin, D.S., Bouvier, P., Jaffe, P.D., et aI., 1996, Prevalence of child sexual abuse among adolescents in
Geneva. Result of a cross sectional survey, Br. Med. J., 312:1326-1329.
Hyler, S.E., 1994, Personality Diagnostic Questionnaire 4+, New York State Psychiatric Institute, New York.
Johnson, B.A., Brent, D.A., Connolly, J., et aI., 1995, Familial aggregation of adolescent personality disorders,
] Am. Acad. Child Adolesc. Psychiatry, 34:798-804.
Kernberg, 0., 1977, The structural diagnosis of borderline personality organization. In: Borderline personal-
ity disorders, P. Hartocollis ed., New York, International Universities Press, pp. 87-12l.
Kernberg, O.F., 1978, The diagnosis of borderline conditions in adolescence, Adolescent Psychiatry, 5:298-320.
Korenblum, M., Marton, P., Golombek, H., and Stein, B., 1987, Disturbed Personality functioning: patterns
of change from early to middle adolescence, Adolescent Psychiatry, The Chicago University Press, vol
14,26:407-415.
Ladame, F., 1991, Adolescence and the repetition compulsion, Int. ] Psychoanal., 72:253-273.
Ladame, F. and Ottino, J., 1993, Les paradoxes du suicide, Adolescence (Paris), 11:125-136.
Laufer, M. and Laufer, E., 1984, Adolescence and developmental breakdown. A psychoanalytic view. New
Haven and London, Yale University Press.
Mahler, M.S., 1971, A study of the separation-individuation process and it possible application to borderline
phenomena in the psychoanalytic situation, Psychoanalitic Study of the Child. 26:403-424.
Mahler. M.S.• Pine, F., and Bergman, A .• 1975, The Psychological Birth of the Human Infant: Symbiosis and
Individuation, New York: Basic.
Main. M. and Hesse, E .• 1990, Parents' unresolved traumatic experiences are related to infant disorganized
attachment status: is frightened and/or frightening parental behavior the linking mechanism? In M.T.
Greenberg, D. Cicchetti, and E.M. Cummings (Eds.), Attachment in the preschool years: Theory,
research and intervention (pp. 161-182). Chicago: University of Chicago Press.
Main, M. and Goldwyn, R., 1993,Adult attachment scoring and classification system, Unpublished manuscript.
Department of Psychology, University of California at Berkeley.
Masterson, J.F., 1972, Treatment of the borderline Adolescent: A developmental approach. New York: Wiley.
Masterson, J.F. and Rinsley, D.B., 1975, The borderline syndrome: the role of the mother in the genesis and
psychic structure of the borderline personality, International Journal of Psychoanalysis, 56:163-177.
Masterson, J.F. and Costello, J.L., 1980, From borderline adolescent to functioning adult: the test of time,
Brunner & Mazel, New York.
Masterson, J.F., 1986, Sur les traces de la triade borderline. In: La psychiatrie de l'adolescence aujourd'hui.
Quels adolescents soigner et comment? F. Ladame, P. Jeammet eds., Paris, Presses Universitaires de
France, pp. 37-55.
Mattanah, lJ.F., Becker, D.F., Kenneth, N.L., Edell, W.S., and McGlashan, T.H., 1995, Diagnostic stability in
adolescents followed up 2 years after hospitalization, American Journal of Psychiatry, 152:6, 889-894.
Nurnberg, H.G., Raskin, M., Levine, P.E., et aI., 1991, The comorbidity of borderline personality disorder and
other DSM-III-R axis II personality disorders, Am. 1 Psychiatry, 148:1371-1377.
Orlandini, A, Clerici, S., Fossati, A, Vanzulli, L., and Maffei, c., 1995, Assessment of DSM IV personality
traits in adolescence: are they related to Ego Deficits? 4th International Congress on the Disorders of
Personality. Dublin, 21st-24th June.
Pahud, AL., 1996, Relation entre tentatives de suicide multiples a l'adolescence, abus sexuel pendant l'enfance
et trouble de la personnalite borderline, Universite de Geneve, These N° 9716.
Reich, J.H., 1989, Update on instruments to measure DSM III and DSM III-R Personality Disorders,lournal
of Nervous and Mental Diseases, 177:366--370.
Rey, J.M., Morris-Yates, A, Singh, M., Andrews, G., and Stewart, G.w., 1995, Continuities Between Psychi-
atric Disorders in Adolescents and Personality Disorders in Young Adult, American Journal of Psy-
chiatry, 152:6, 895-900.
Rinsley, D.B., 1988, A review of the pathogenesis of borderline and narcisistic personality disorders, Adoles-
cent Psychiatry, 15:387-406.
Robins, L.N., 1996, Deviant children grown up, Baltimore, Williams & Wilkins.
Robins, L.N., Murphy, G.E., Woodrruff, R.A Jr, and King, L.J., 1971, The adult psychiatric status of
blackschool boys, Archives of General Psychiatry, 24:338-345.
Senise, T., 1981, Per l'adolescenza: psicoanalisi 0 analisi del Se. In G. Lanzi: L'adolescenza. Roma, Pensiero
Scientifico Editore.
Tyrer, P., 1994, What are the borders of borderline personality disorder? Acta Psychiatr. Scand., 89(suppl.
379), 38-44.
Winnicott, D.W., 1971, Transitional objects and transitional phenomena, In Playing and Reality, pp. 1-25. 1953.
New York: Basic Books.
Wolff, S. and Chick, J., 1980, Schizoid personality in childhood: a controlled follow-up study, Psychol Med,
10:85-100.
8
INTEGRATIVE PERSPECTIVES ON THE

Theodore Millon
Hegel used a concept many years ago when he spoke of dialectical processes of
analysis (thesis, antithesis), and synthesis; what I want to do in this chapter is a variant
of that dialectic process. It appears to me that we, in our subject areas of personality
and psychopathology, have invested a great deal of our time in the process of analysis.
But not in that of synthesis. We have segmented our subject into many diverse parts,
and each of these parts has been investigated in depth, and with great sensitivity and
critical intelligence. But it appears to me that we are long overdue in addressing syn-
thesis, of putting back together that which we have partitioned these past decades.
There are several general themes that reflect efforts toward integration or synthesis,
and I will address each in turn. Let me summarize a few briefly.
The first deals with the fact that we are ready to bring together what might be
called a clinical science. A clinical science has several components. For the greater part
of our history we have investigated each of the components separately. We have inves-
tigated many diverse psychotherapies, but they have little relationship to the kinds of
diagnostic tools we use to identify the characteristics of our patients. As I will elabo-
rate, there is a need to build a clinical science that coordinates theories, taxonomies,
diagnostic instruments, and psychotherapy.
The second topic I want to address has to do with how we conceptualize the
nature of disorders; here I would like to demonstrative our field's historical develop-
ment and the utility of the multiaxial format employed currently in the DSMs of the
United States. History has shown a progression of knowledge from symptoms to causes
to vulnerabilities. The multiaxial schema leads us first to recognize the presenting symp-
toms (Axis I), subsequently, we look for causes that underlie those symptoms (Axis IV
and Axis III). Only recently have we begun to focus on the patient's vulnerabilities
(Axis II). This is where personality disorders assume a major role.
Third-and a somewhat grandiose theme that I wish to address-has to do with
reconnecting the study of personology to other spheres of nature's sciences. In Hegel's
analysis or thesis phase, we have learned to divide nature into subjects such as physics,
chemistry, biology, psychology, sociology, and so on. It appears to me that there are
certain latent principles and universal laws in common to all of these segmented
sciences. As I see it, all are facets of nature, expressed in one form or another.
Treatment of Personality Disorders, edited by Derksen et aI.
98 T. Millon
Identifying these central principles will demonstratie that a common thread may be
found among physics, chemistry, biology, as well as in our field of personology. It is my
belief that we need to extract these principles that are common to all sciences, I will
contend, further, that these different expressions of nature are all grounded in the basic
laws of evolutionary theory.
Fourth, I should like to close my chapter by arguing in favor of what has been
termed integrative therapy, an approach that is especially suited to the personality dis-
orders. Personality is best defined as representing a complex of characteristics, not just
affective symptoms, or behavioral ones, or cognitive, or interpersonal, and so on. It is
composed of all of them, acting in concert, that is, dynamically interrelating. So too, will
I contend that our psychotherapies should be combined for optimal effect, thereby mir-
roring in its complexity that configurational pattern we call personality.
These are the topics that I wish to cover, one by one. What I hope to do is outline
the rationale for an approach that seeks to reintegrate our field, shaping it into a science
mature, and integrated, and then reconnecting it to other subject domains of nature.
1. TOWARD AN INTEGRATED CLINICAL SCIENCE
All sciences, once they mature sufficiently, must exhibit a strong relationship
between its theories, its taxonomies, and its instruments. A clinical science has an
additional responsibility, the responsibility to intervene, that is, to engage in therapeu-
tic and rehabilitative actions. The history of our clinical field has separated these four
components. Let me elaborate them briefly.
In addition to nature as a whole, integrative cohesiveness is a worthy goal within
each sphere or realm of nature. Particularly relevant in this regard are efforts that seek
to coordinate the often separate realms that comprise a clinical science, namely: its the-
ories. The classification system it follows, the diagnostic tools it employs, and the ther-
apeutic techniques it implements. Rather than developing independently and being left
to stand as autonomous and largely unconnected functions, a truly mature clinical
science will embody: (1) explicit theories, that is, explanatory and heuristic conceptual
schemes that are consistent with established knowledge in both its own and related sci-
ences, and from which reasonably accurate propositions concerning pathological con-
ditions can be both deduced and understood, enabling thereby the development of (2)
a formal nosology, that is, a taxonomie classification of disorders that has been derived
logically from the theory, and is arranged to provide a cohesive organization within
which its major categories can readily be grouped and differentiated, permitting
thereby the development of (3) coordinated instruments. That is, tools that are empir-
ically grounded and sufficiently sensitive quantitatively to enable the theory's propo-
sitions and hypotheses to be adequately investigated and evaluated, and the categories
comprising its nosology to be readily identified (diagnosed) and measured (dimen-
sionalized). Specifying therefrom target areas for (4) interventions, that is, strategies
and techniques of therapy. Designed in accord with the theory and oriented to modify
problematic clinical characteristics consonant with professional standards and social
responsibilities (Millon and Davis, 1996).
The intervention therapies of our day exist largely unto themselves and have little
relationship to the diagnostic tools by which we assess our patients. Moreover, our tax-
onomies-the DSM and lCD-are intentionally atheoretical, that is, they are comprised
of groupings and differentiations that are unrelated to any theoretical grounding. Theory
Integrative Perspectives on the Personality Disorders 99
does not inform our classifications nor the diagnostic instruments that we employ-the
Rorschach, the MMPI, and the like. At best, they are related to theory only incidentally,
haphazardly, and post-facto. The btilliance of theoretical physics is a function of its ability
to apply the equations of theory to a taxonomy of elementary particles. But this could not
be done unless physicists had instruments to test whether or not the theoretically gener-
ated properties were in fact verified. With powerful enough cyclotrons it is possible to
observe the behavior of diverse elementary particles and to check whether what has been
theoretically posited is, in fact, correct.
There is no reason why we in personality pathology cannot systematize our extant
knowledge in a manner akin to advanced sciences. That is, to articulate a series of
theoretical propositions that help explain the latent structure of our subject, group
the constructs of personology into a taxonomic system consistent with the theoretical
model, and then develop instruments which can identify the properties of the taxon-
omy. Ultimately, the goal is to develop intervention techniques consonant with
diagnostic instruments that identify characteristics that will be the target of these tech-
niques. To complete the circle, the theory will become a guide to how the intervention
techniques can best be employed.
We have many tools, we have many therapies, we have had a variety of classifi-
cation systems, we have alternative theoretical models, it is now time to bridge the con-
nections between these components, and to do so in a coordinated way. If we do not
begin to do this task in earnest, we will be in the same position that we have been for
the past three or four thousand years, when brilliant ideas are proposed theoretically,
and serious scholars generate insightful classification systems, highly fruitful therapies
are developed, and diverse diagnostic tools are created, all leading to a babel of con-
fusion, in which little is synthesized logically. Integration among these four components
that define a mature clinical science is a task of the first order toward which we should
direct our efforts.
2. INTEGRATING SYMPTOM, CAUSE, AND VULNERABILITY
The second major theme I want to address relates to our understanding of the
nature of psychopathology and personology.There are three elements that comprise
this conception of our field. Interestingly, they reflect the last century's progress
of medical science; they also mirror the interactive nature of the DSM's multiaxial
system.
To clarify, it may be useful to record two major shifts that have taken place in
medical thinking this past century. They highlight the fact that modern-day health
providers no longer focus on symptoms-as they did a century ago-nor do they focus on
intruding infectious agents-as they did until the last decade or two-but have turned
their attention to the composition and mechanisms of the immune system.
The parallel to Axis I in physical disorders characterizes where medicine was 100
and more years ago; in the early and mid-nineteenth century, physicians labeled, even
defined their patients' ailments in terms of manifest symptomatology-their sneezes
and coughs and boils and fevers, labeling "diseases" with terms such as consumption
and smalipox. Paralleling Axis III and IV of the DSM, the related medical focus, uncov-
ered approximately 100 years ago, was that illness no longer be conceived only in terms
of overt symptomatology, but with reference to minute microbes which intruded upon
and disrupted the body's normal functions; in time, medicine began to employ and
100 T.MiIloD
assign diagnostic names that reflected ostensive etiologies-such as when they

relabeled dementia paralytic to that of neurosyphilis.
Psychopathology in general has progressed in making this shift from symptom to
cause all too slowly. We still focus on what we should do about "dysthymia" or "anxiety,"
giving our prime attention to the surface symptoms that comprise the syndromes of
Axis LAmong those who consider themselves to be "sophisticated" about such matters,
there is a recognition that dysthymia and anxiety are merely a psychic response to life's
early or current stressors, such as those which comprise the DSM -IV's Axis IV-marital
problems, child abuse, and the like-psychic intruders, if you will, that parallel the
infectious microbes of a century ago.
But medicine has progressed in the past decade or two beyond its turn-of-the
century "intrusion disease" model, an advance most striking these last 15 years owing
to the tragedy of the AIDS epidemic. This progression reflects a growing awareness of
the key role of the immune system, the body's intrinsic capacity to contend with the
omnipresent multitude of potentially destructive infectious and carcinogenic agents
that pervade our physical environment. What medicine has learned is that it is not the
symptoms-the sneezes and coughs-not the intruding infections-the viruses and
bacteria-that are the key to health or illness. Rather, the ultimate determinant is the
competence of the immune system. So too, in psychopathology, it is not anxiety or
dysthymia, nor the stressors of early childhood or contemporary life that are the key
to psychic well-being. Rather, it is the mind's equivalent of the body's immune system-
that structure and style of psychic processes that represents our overall capacity to per-
ceive and to cope with our psychosocial world-in other words, the psychological
construct we describe as "personality". Fortunately we have begun to catch up with
medicine this past decade, to turn our attentions from symptoms and stressors
toward "persons," and the psychic structures and styles that signify their disordered
character.
I think we can better understand the DSM multiaxial system by noting that it rec-
ognizes, if only implicitly, that clinical syndromes (Axis I) are largely a product of the
interaction of psychosocial stressors (Axis IV) and a vulnerable personality (Axis II).
Vulnerabilities signify that a person is oversensitized to certain kinds of life events, for
example, interpreting the world cognitively in a trouble-making way, or relating inter-
personally to people in a problem-perpetuating fashion. These vulnerabilities and mal-
adaptive cognitions and behaviors are the patterns the individual has learned as a way
of coping with his or her environment.
As briefly noted above, if we go back about a century or so, most medical ail-
ments were labeled and defined in terms of their clinical symptom picture, for example,
skin pustules such as small pox were a description of the disorder's overt symptoma-
tology. The major revolution in medical thinking around 1870, owing largely to the work
of Pasteur and Semelweiss, reflected the discovery that microbes, termed bacteria, and
viruses today, were the unseen causal agents of these overt symptoms. This first real
advance led to a shift from "understanding" based on overt symptomatology to "infer-
ring" infectious agents within the body, in effect turning attention from symptoms
to causes.
A parallel shift took place in psychopathology. Freud was a student in medical
school when the "germ" theory was first posited. What he learned was that there
were hidden or unseen forces within the body which ultimately produced manifest
symptoms. Through analogy he extended this notion by concluding that there were
psychic events equivalent to infectious agents; that is, unseen emotional forces which
uitimately manifested themselves in psychic symptomatology. Thus, when treating a

conversion hysteria; he did not assume it to be attributable to nerve damage, but rather
the presence of hidden emotional phenomena that could neither be seen or readily
recoverable. In effect, Freud joined the medical revolution with his own revolution in
the psychic sphere. Not only did he identify the role of unconscious' mental forces, but
went on to conclude that powerful influences early in life set the seeds for later
disorders. Beyond disorders such as hysteria and depression, they would manifest
themselves as "character disorders", the parallel to what now refers to as personality
disorders. Thus, the first revolution in medicine had its parallel in personality pathol-
ogy, a shift from looking at the symptom to looking at its underlying and often unseen
causes.
Another second revolution is taking place now, both in medicine and in person-
ality pathology. In the last twenty to thirty years, medicine has begun to recognize
that it is not the symptoms that matter; they are merely the final state in which
problems are displayed. And it is no longer the so-called causes that are central. In
medicine, it is the capability of the immune system to deal effectively with the car-
cinogens and infectious bacteria and viruses to which all of us are SUbjected. Vaccina-
tions of various sorts are designed to strengthen our immune systems. Thus, when
faced with bacterial or viral intruders, the microphages of our immune system set up a
defensive wall, its T-helper cells prepare for action, and the NK killer cells join in
the fray as needed. It is this capacity of the various biophysical components of the
immune system to do their job that determines whether a person will remain healthy
or become ill.
The immune system is the crucial defending complex by which our body per-
ceives, analyzes, and reacts to potentially deleterious intruders. AIDS is a unique. It
affects not only the skin or the upper respiratory system, but signifies that there
has been an attack on the immune system itself. Because of the AIDS epidemic, health
professionals have become even more sensitized than previously to the crucial role
of the immune system in combating intruding agents. This new awareness has resulted
in the second major revolution in medical understanding. It reflects a growing aware-
ness that a vulnerability in the organism's ability to deal with its environment is at
least as important as exposure to potentially dangerous or toxic environments. For sure,
it is not the symptoms-that is a most primitive and unsophisticated way to look at
the problem; nor is it stressful agents such as bacteria or psychic turmoil. In the phys-
ical sphere it is the relative vulnerability of the immune system. And what matters in
the mental sphere is the psychic equivalent of the immune system, the pattern of sen-
sitivities and vulnerabilities, in effect, what we mean when we speak of personality, the
way we perceive, interpret and react to our psychosocial world. What we call person-
ality "disorders" simply signifies the presence of a maladaptive style or structure to
personality, a problematic way of relating to one's psychosoeial world, a manner of per-
ceiving or thinking that will determine whether or not the person is unusually vulner-
able, or will function competently in his/her environment.
The revolution that is now taking place recognizes, therefore, that psychic health
is not the clinical syndrome-the depression or the anxiety-just as it is not the cough
or the rash in the physical sphere. Moreover, it is not merely the stressors of life-the
marital conflicts or economic woes. Rather, it is the vulnerability to these stressors, the
qualities of the psychic "immune system". And since people exhibit different forms of
vulnerability, the same event can evoke diverse, even opposite kinds of responses;
hence, the many variants there are of personality disorder.
102 T.MiIlon
3. ON THE INTEGRATIVE ROLE OF EVOLUTIONARY THEORY

The third area of integration, as touched on earlier, concerns the fact that we have
been studying personality for thousands of years. Brilliant analyses of personality types
have been proposed; numerous theoretiCal hypotheses regarding how we should dif-
ferentiate personalities have been proposed, as have many different methods of analy-
sis been devised, all in the hope of identifying the core characteristics of personality
types or traits. But I think we have reached the end of this period. Brilliant as most of
this work has been, I think it is time to approach the subject from a fresher, and perhaps
significantly different perspective. What I am suggesting is that we turn things around,
to look toward more universal laws that may be found in other sciences, from other
spheres of nature's expression, not only our own.
In 1990, I reconceptualized my model of personality and its disorders to reflect a
reevaluation of what I judged to be the deeper or latent features that undergird human
functioning. For the past twenty to twenty-five centuries, man has attempted to decode
the underlying characteristics of personality functioning by reviewing the diverse forms
in which behaviors, thoughts, and feelings are expressed. Drawing inspiration from
Godel's incompleteness' theorem (1931) that no self-contained system can prove its
own propositions, I made the decision to turn the spotlight away from psychology
proper to expressions of nature that fail outside the field of psychology itself. Just as in
Godel's theorem, I concluded that the deeper laws of human functioning may be best
explicated by examining universal principals derived from other, non-psychological
manifestations of nature, e.g., those expressed in subjects such as, physics, chemistry,
and especially biology. Within these other spheres, I felt that I might uncover more than
just the biophysical underpinnings of psychological functioning, or the unconscious
forms in which experience takes shape, or the phenomenological world of cognitive
experience, or the behavioral consequences of the preceding.
What I deduced from these reevaluations was that the principles and processes
of evolution are essentially universal, expressed in a variety of different forms, as seen
in diverse subjects such as physics, chemistry, biology, and psychology. In my 1990 book,
and the revision of my Disorders text, co-authored with Davis (Millon and Davis, 1996),
I attempted to illustrative the universals that I judged were fundamental to all spheres
of evolution's progression, including those of human experience. What was most
gratifying in this exploratory search was the close correspondence between my 1969
biosocial-Iearning theory (Millon, 1969) and the key elements of this new evolutionary
model. Of additional note, the ontogenetic theory of neuropsychological stages pre-
sented in my 1969 book also paralleled closely the new theoretical conception of
evolutionary phylogenesis. (see Chapter 3 in Millon and Davis, 1995a).
As I have argued this past decade, I believe it was necessary to go beyond tradi-
tional conceptual boundaries in psychology to find a fresh and fundamental grounding
for organizing and understanding both personality and psychopathology. More
specifically, I chose to explore carefully reasoned, as well as "intuitive" hypotheses that
drew their principles, if not their substance, from more established "adjacent" sciences,
such as _evolutionary biology. Not only did such steps bear new conceptual fruits,
but they provided a foundation that both undergirded and guided my theoretical
explorations.
Much of personology, no less psychology as a whole, appears to have been adrift,
divorced from broader spheres of scientific knowledge, isolated from firmly grounded,
if not universal principles, leading psychologists and psychiatrists to continue building
the patchwork qUilt of concepts and data domains that has characterized the field
this past century. Preoccupied with but a small part of that larger puzzle of scientific
endeavors, or fearing accusations of reductionism, most failed to draw on the rich
possibilities to be found in adjacent realms of scholarly pursuit. With few exceptions,
cohering concepts that would connect the subject of personality to those of its sister
sciences have not been developed. The Hebrews taught us a long time ago, that there
is only one God, not many Gods who fought among themselves. In a parallel fashion,
I think, as do most scientists, that there is only one nature; it does not possess a series
of components that lie in opposition or contradiction, nor are they reducible one to
another. Humans are, at one and the same time, each of us, physical, chemical, biolog-
ical, psychic, social, cultural phenomenon. We divide the intrinsic unity of nature
because that is how our pedagogy has been constructed and how our scientific disci-
plines achieve a great measure of analytic precision and depth. But these "schismatic"
efforts have failed to address where the substantive commonalities are among nature's
diverse realms of expression, nor do they enable us to search for the latent principles
that each shares.
In my view, a careful reading of these commonalities suggests they may be found
in what we term "evolution", the manner in which nature has manifested itself over
time in ways that is balanced and progressive. These core themes of evolutionary
thought are found in physics, in cosmogony, of course, in biology, where its features
were first decoded by Darwin and which, I believe, now also can be applied to the
explication of human functioning and behavior. And it is from those common princi-
ples that I believe we can grasp the essentiel processes and laws that underlie human
functioning.
Whereas our psychological and psychiatric theoretical colleagues, past and
present, have sought to understand human functioning by focusing within our own
subject domain, I have decided to turn outward to identify and connect commonalties
that exist in all of nature; as just noted, I judge these to be in the principles of evolu-
tion. If that belief can be sustained, we should be able to derive all variations, adaptive,
and maladaptive, from the same theoretical principles and laws. Thus, we should be able
to explicate what it "means" when one acts in a manner we characterize as a schizoid
personality, or an obsessive-compulsive personality, or an avoidant, a narcissist, and so
on, because each represents a tendency to employ excessively one or another of the
basic processes that evolution has shown to be useful for survival.
As I've stated before, what better sphere is there within the psychological sciences
to undertake such "evolutionary" explorations than with the subject matter of personol-
ogy. Persons are the only organically integrated system in the psychological domain,
evolved through the millennia, and inherently created from birth as natural entities,
rather than culture-bound and experience-derived gestalts. The intrinsic cohesion of
persons is not merely a rhetorical construction, but an authentic substantive unity. Per-
sonologic features may often be dissonant, and may be partitioned conceptually for prag-
matic or scientific purposes, but they are segments of an inseparable biopsychosocial
entity, as well as a natural outgrowth of evolution's progression.
What makes evolutionary principles as relevant as we propose? Owing to the
deductive insights of our colleagues in physics, we have a deeper and clearer sense of
the early evolution and structural relations among matter and energy. So too has knowl-
edge progressed in our studies of physical chemistry, microbiology, population biology,
ecology, and ethology. It is odd that we have only now again begun to investigate-as
we did at the turn of the last century-the interface between the basic building blocks
104 T. Millon
of physical nature and the nature of life as we experience and live it personally? How
much more is known today, yet how hesitant are people to undertake a serious
rapprochement.
It is clear that each evolved species displays commonalities in its adaptive or sur-
vival style. Within each species, however, there are differences in style and differences
in the success with which its various members adapt to the diverse and changing envi-
ronments they face. In these simplest of terms, personality would be conceived as rep-
resenting the more-or-Iess distinctive style of adaptive functioning that an organism of
a particular species exhibits as it relates to its typical range of environments. "Disor-
ders" of personality, so formulated, would represent particular styles of maladaptive
functioning that can be traced to deficiencies, imbalances, or conflicts in a species'
capacity to relate to the environments it faces.
To provide a conceptual background from these sciences, and to furnish a rough
model concerning the styles of personality, normal, and abnormal, four spheres in which
evolutionary and ecological principles can be applied were labeled as Existence, Adap-
tation, Replication, and Abstraction. The first relates to the serendipitous transforma-
tion of random or less organized states into those possessing distinct structures of
greater organization; the second refers to homeostatic processes employed to sustain
survival in open ecosystems; the third pertains to reproductive styles that maximize the
diversification and selection of ecologically effective attributes; and the fourth concerns
the emergence of competencies that foster anticipatory planning and reasoned deci-
sion making. We will restrict this brief discussion of integrating personality and evolu-
tionary theory at this point and suggest that the reader turn to more detailed
presentations in other texts (Milion, 1990; Millon and Davis, 1996).
4. TOWARD THE INTEGRATION OF PSYCHOTHERAPY
Do I think that there is some truth to the integrative faith, that is, that there is a
class of disorders for whom the logic of the integrative mindset is the optimal, if not
the most efficacious therapeutic choice?
Although the approach that has come to be called integrative therapy has its
applications to a variety of diverse clinical-a view I wholeheartedly endorse-I will
seek in this section to outline some reasons why personality disorders may be that
segment of psychopathology for which integrative psychotherapy is ideally and dis-
tinctively suited-in the same sense that behavioral techniques appear most efficacious
in the modification of problematic actions, cognitive methods optimal for reframing
phenomenological distortions, and intrapsychic techniques especially apt in resolving
unconscious processes.
The cohesion (or lack thereof) complex interwoven psychic structures and func-
tions is what distinguishes the disorders of personality from other clinical syndromes
likewise, the orchestration of diverse, yet synthesized techniques of intervention is what
differentiates integrative from other variants of psychotherapy. These two, parallel
constructs, emerging from different traditions and conceived in different venues, reflect
shared philosophical perspectives, one oriented toward the understanding of psy-
chopathology, the other toward effecting its remediation.
It is not that integrative psychotherapies are inapplicable to more focal pathologies,
but rather that these therapies are required, I believe, for the personality disorders
(whereas depression may successfully be treated either cognitively or pharmacologi-
cally); it is the very interwoven nature of the components that comprise personality
disorders that makes a multifaceted and synthesized approach a necessity.
Let me present a few ideas briefly. First, integrative therapies require a founda-
tion in a coordinated theory, that is, they must be more than a schema of eclectic tech-
niques, a hodgepodge of diverse alternatives assembled de novo with each case. Second,
although the diagnostic criteria that comprise leD and DSM syndromes are a decent
first step, they must be comprehensive and comparable, that is, be systematically revised
so as to be genuinely useful for treatment planning. Third, a logical rationale should be
formulated as to how one can and should integrate diversely focused therapies when
treating the personality disorders.
I would like also to comment on some philosophical issues. They bear on a ratio-
nale for developing theory-based treatment techniques, that is, methods that transcend
the merely empirical (e.g., electroconvulsive therapy for depressives). It is my convic-
tion that the theoretical foundations of our science must be further advanced if we are
to succeed in constructing an integrative approach to psychotherapy.
Most current therapeutic schools share a common failure to coordinate the four
components of a clinical science. What differentiaties them has more to do with the fact
that they attend to different levels of data in the natural world. It is to the credit of
those of an eclectic persuasion that they have recognized, albeit in a "fuzzy way," the
arbitrary if not illogical character of such contentions, as well as the need to bridge
schisms that have been constructed less by philosophical considerations or pragmatic
goals than by the accidents of history. There are numerous other knotty issues with
which the nature of personality and integrative therapy must contend (e.g., differing
"worldviews" concerning the essential nature of psychological experience). There is
no problem, as we see it, in encouraging active dialectics among these contenders-
although we personally hold to an "integrative" or "synthetic" view of nature's
phenomena.
What exactly do we mean when we say that therapy should be integrated and
should be grounded in a logical and coordinated theory (Milion, 1995b)? Unfortu-
nately, much of what travels under the "eclectic" or "integrative" banner sounds like
the talk of a "goody goody"-a desire to be nice to all sides, and to say that everybody
is right. These labels have become platitudinous "buzzwords," philosophies with which
open-minded people certainly would wish to ally themselves. But, "integrative theory
and psychotherapy" must signify more than.
First, it is not eclecticism. Perhaps it might be considered post-eclecticism, if we
may borrow a notion used to characterize modern art just a century ago. Eclecticism
is not a matter of choice. We all must be eclectics, engaging in differential and muiti-
modal therapeutics, selecting the techniques that are empirically the most efficacious
for the problems at hand. Moreover, integration is more than the coexistence of two
or three previously discordant orientations or techniques. We cannot simply piece
together the odds and ends of several theoretical schemas, each internally consistent
and oriented to different data domains. Such a hodgepodge will lead only to illusory
syntheses that cannot long hold together. Efforts such as these, meritorious as they may
be in some regards, represent the work of peacemakers, not innovators and not inte-
grationists. Integration is eclectic, of course, but more. It is a synthesized and substan-
tive system.
The personality problems that our patients bring to us are an inextricably linked
nexus of behaviors, cognitions, intrapsychic processes, and so on. They flow through
a tangle of feedback loops and serially unfolding concatenations that emerge at
106 T. MiUoR
different times in dynamic and changing configurations. Each component of these con-
figurations has its role and significance altered by virtue of its place in these continu-
ally evolving constellations. In parallel form, so should integrative psychotherapy be
conceived as a configuration of strategies and tactics in which each intervention tech-
nique is selected not only for its efficacy in resolving particular pathological features
but also for its contribution to the overall constellation of treatment procedures of
which it is but one.
Whether we work with "part functions" that focus on behaviors, or cognitions, or
unconscious processes, or biological defects, and the like, or whether we address con-
textual systems which focus on the larger environment, the family, or the group, or the
socioeconomic and political conditions of life, the crossover point, the place that links
parts to contexts is the person. The individual is the intersecting medium that brings
them together.
But persons are more than just crossover mediums. As noted earlier, they are the
only organically integrated system in the psychological domain, inherently created
from birth as natural entities, rather than experience-derived gestalts constructed via
cognitive attribution. Moreover, it is persons who lie at the heart of the psychothera-
peutic experience, the substantive beings that give meaning and coherence to symp-
toms and traits-be they behaviors, affects, or mechanisms-as well as those beings,
those singular entities, that give life and expression to family interactions and social
processes.
It is our contention that integrative therapists should take cognizance of the
person from the start, for the parts and the contexts take on different meanings, and
call for different interventions in terms of the person to whom they are anchored. To
focus on one social structure or one psychic form of expression, without understand-
ing its undergirding or reference base is, as I see it, to engage in potentially misguided,
if not random, therapeutic techniques.
What I hope perhaps we can achieve is akin to the European Union, or the
Common Market, that is, to come together, to bring our different perspectives within
our science together and begin to utilize the deeper principles that will enable us under-
stand the essential nature of personality and its "disorders".
REFERENCES
Godel, K. (1931). On formally undecidable propositions of principia mathematica and elated systems.
Unpublished doctoral dissertation, University of Vienna.
Millon, Th. (1969). Modern psychopathology: A biosocial approach to maladaptive learning and functioning.
Philadelphia: Saunders.
Millon, Th. (1990). Toward a new personology: an evolutionary model. New York: Wiley.
Millon, Th. and Davis, R. (1996). Disorders of Personality, DSM-/V and Beyond. New York: Wiley.
9
A MULTIDIMENSIONAL APPROACH TO
PERSONALITY DISORDERS AND
THEIR TREATMENT
Joel Paris
Professor of Psychiatry
McGill University
Institute of Community and Family Psychiatry
Sir Mortimer B. Davis-Jewish General Hospital
4333 Chemin de la Cote Ste. Catherine
Montreal QC, H3T1E4, Canada
1. INTRODUCTION
Personality disorders are, by definition, chronic. It is not surprising therefore, that

they are often resistant to change in psychotherapy. Treatment planning requires an ade-
quate etiological model. The therapy of complex pathology can not be based on simple
models of causality. Instead, management must be rooted in a multidimensional model
that takes into account the biological, psychological, and social factors in these disorders.
In this chapter, an approach to the treatment of these disorders is suggested in which
therapists work within the stability of personality, modifying the behavioral expression of
traits sufficiently to help patients use their personality in more adaptive ways.
2. EMPIRICAL FINDINGS ON THE PSYCHOTHERAPY OF

PATIENTS WITH PERSONALITY DISORDERS
By their very definition (American Psychiatric Association, 1994), personality dis-

orders are characterized by long-term chronicity. This observation has been confirmed
by long-term outcome studies, which show that patients with personality disorders
change very slowly over time (Perry, 1993). Those changes that have been observed
consist of gradual reductions in impulsivity (Paris, 1988, 1993).
108 J. Paris
The management of personality disorders has usually fallen under the domain of
psychotherapy, and the primary interests of many psychotherapists have focused on the
treatment of these patients (Gunderson, 1985). Yet there is no consistent evidence that
psychotherapy is predictably effective for most personality disorders (Tyrer, 1988). If
anything, these patients have a reputation for presenting insoluble clinical problems
(Lewis and Appleby, 1988).
The presence of any personality disorder diagnosis presents a challenge to exist-
ing treatment methods. Empirical studies have shown that personality disordered
patients respond poorly to therapies that might otherwise be effective. Thus, when a
patient presents for treatment with an Axis I disorder, the presence of any comorbid
personality disorder makes the relief of symptoms more difficult (Andreoli et aI., 1993).
For example, mood disordered patients who have a personality disorder in addition
to their depression respond more poorly to both psychopharmacology and psy-
chotherapy (Shea et aI., 1990, 1992).
There are few effective and specific treatments for the symptoms of the person-
ality disorders themselves. Thus far pharmacological interventions have been of mar-
ginal value (Soloff, 1993; Coccaro, 1993). It remains possible that there could be drugs
in the future that will be able to modify personality traits (Masters and McGuire, 1994),
but at present, psychosocial interventions remain the backbone of treatment for the
personality disorders.
qinical trials have demonstrated a value, however marginal, for the standard
forms of psychotherapy, in at least some of these patients. In patients with less severe
personality disorders selected for brief or intermediate therapies, patients report symp-
tomatic relief (H0glend et aI., 1992; Winston et aI., 1994; Propst et aI., 1994). In more
severely dysfunctional patients with borderline personality disorder, impulsivity can be
reduced by one year of cognitive-behavioral therapy (Linehan et aI., 1991; Linehan,
1993). However, as shown by a follow-up of this same cohort (Linehan et aI., 1993),
even when symptoms remit, most patients remain with chronic dysfunction.
In view of the long-term nature of personality disorders, some authorities have
recommended that these patients must receive long-term treatment (Kernberg, 1987).
However, there has been no empirical evidence that lengthening therapy produces
better results. In the Menninger study, one of the few research studies on the effects
of longterm treatment, the results were humblingly modest (Kernberg et aI., 1972;
Wallerstein, 1986; Horwitz, 1994).
It is possible that further refinements of psychotherapeutic techniques could
improve this discouraging picture. However, although, cognitive behavioral therapies for
personality disordered patients have been outlined in theoretical and practical detail by
Beck and Freeman (1990), their methods have not been subjected to clinical trials.
3. HOW PERSONALITY PATHOLOGY INTERFERES WITH

THE PROCESS OF PSYCHOTHERAPY
The problems in the psychotherapy of personality disorders are intrinsic to the

nature of these conditions. By definition, personality disordered patients have signifi-
cant difficulties in work and in relationships. Psychotherapy, which is both work and a
relationship, is bound to be difficult.
Every patient in therapy is expected to work with the therapist towards mutually
agreed goals, a process described by the construct of a "therapeutic alliance". Measures
A Multidimensional Approach to Personality Disorders and Their Treatment 109
of the strength of the alliance are robust predictors of the outcome of psychotherapy
(Lubarsky et aI., 1988). However, maladaptive interpersonal behavior patterns outside
therapy interfere with the capacity to develop an alliance inside therapy. For example,
the alliances of patients with borderline personality disorder have been shown to be
unusually fragile (Frank, 1992). Over half of these patients become early dropouts from
psychotherapy (Skodol et aI., 1983; Gunderson et aI., 1989). Although the strength of
a therapeutic alliance is the best predictor of therapy outcome, it can only be reliably
measured after patients have already had a few sessions of treatment, a rather im-
practical requirement if one wishes to carry out differential therapeutics (Frances
et aI., 1984).
"Defense styles" is a useful construct to describe how individuals cope with life
problems. Defenses can be assessed empirically, either through self-report inventories
(Bond et aI., 1983), or through the scoring of vignettes from interviews (Perry and
Cooper, 1989). The different categories of defense styles can also be ranked, from the
most to the least adaptive. Levels of defense maturity have been shown to be corre-
lated with functional levels in personality disordered patients (Bond et aI., 1983). The
most severe personality disorders, such as BPD, have the least adaptive defenses (Bond
et aI., 1994). If defense styles could be shown to be predictive of treatment response,
then we could separate personality disordered patients into those with more mature
defense styles, who would farm stronger treatment alliances, and be more likely to
respond to treatment; and those with more maladaptive defense styles, who would form
weaker treatment alliances, and would be less likely to respond to treatment.
Research has also demonstrated a significant correlation between pre-treatment
functional levels and therapy outcome (Lubarsky et aI., 1988). Since there are wide
individual differences between patients with the same diagnosis, functional level is
scored separately by the DSM system on Axis V. These scores might be used to help
predict treatability in personality disordered patients. In contrast to the alliance, they
can be measured prior to therapy. Although many categories of personality disorder
are consistently associated with lower levels of functioning (Nakao et aI., 1992), it is
possible that by pre-selecting for active treatment those with higher functioning, we
could improve the overall efficacy of therapy.
4. THE NATURE OF PERSONALITY TRAITS
The limited results fore therapy of personality disorders may be accounted for in
part by the stability of traits. Personality tends to endure, even when it is maladaptive.
The same personality traits that are present in late adolescence can be found in old age
(McCrae and Costa, 1990). This stability of personality is probably due to a combina-
tion of genetic influences (Plomin et aI., 1990), the effects of social learning early in life
(Bandura, 1977), and the cyclic and self-reinforcing nature of interpersonal behavior
(Wachtel, 1994). If personality does not change over time, it is not surprising that
disorders show a parallel stability.
Some clinicians have claimed that personality structure can be modified by psy-
chopharmacology (Kramer, 1993), while others have made similar claims for psy-
chotherapy (Kern berg, 1987). At present, there is little solid empirical evidence to
support these ideas. In order to understand the problems of treating personality disor-
dered patients, we may need a different approach, based on the relationship of per-
sonality traits to disorders.
110 J. Paris
Personality traits refer to consistent patterns of behavior, emotion, and cognition

that are characteristic of, and unique to any individual. These traits show a great
deal of variability from one person to another. Both genetic factors, such as tempera-
ment, and environmental factors, such as social learning, playa role in shaping them
(Rutter, 1987).
Temperament describes those behavioral dispositions that are present at birth.
Within normal ranges, infantile temperament is not notably continuous with later per-
sonality (Chess and Thomas, 1990). In general, extreme temperaments are more stable
predictors of personality development (Maziade et aI., 1990; Kagan, 1994). These tem-
peramental effects are not limited to characteristics present at birth, since there may
be a number of genetic effects on personality that only "switch on" at later periods
(Rutter and Rutter, 1994).
Evidence from longitudinal studies shows that temperamental characteristics in
childhood shape the development of adult personality (Rothbart and Ahadi, 1994).
These findings indicate, for example, that children with increased fearfulness and irri-
tability become more neurotic, that children with increased activity level and positive
affectivity become more extraverted, that children with attentional persistence become
more conscientious, and that children who are easily prone to distress have more
difficulties with attachment. Temperamental abnormalities in children may also become
amplified by the difficulties they create for both parents and peers (Rutter and
Quinton, 1984).
Behavioral genetics has produced a large body of research that demonstrates the
strength of biological factors in personality (Plomin et aI., 1990). The most common
method for determining the heritability of traits is by comparing their concordance in
monozygotic and dizygotic twins. These studies show that for almost any personality trait,
monozygotic twins are much more similar. A statistical measure called "heritability" can
be derived from this data (Carey and DiLalla, 1994). Most traits have a heritability of
around 40% to 50%, and about half the variance in most personality traits can be attrib-
uted to genetic influence (Plomin et aI., 1990). Social closeness or intimacy might be an
exception for, with a more moderate (30%) heritability suggesting that these traits are
more open to environmental influence (Livesley et aI., 1993).
Adoption studies yields similar estimates of heritability (Plomin et aI., 1990).
Studies of personality traits in twins raised together and apart has provided a design
that combines the advantages of both twin and adoption studies. This research has pro-
vided striking confirmation for the strength of the genetic influences on personality
(Tellegen et aI., 1988; Bergeman et aI., 1993).
The mechanisms by which genetic factors influence personality are complex. First,
personality traits are influenced not by single genes, but by the interaction of several
genes. Many genetic effects only appear when several genes are all present at the same
time (Lykken et aI., 1992). Second, each genetic effect is itself "pleiotropic", i.e., genes
influence more than one type of behavior. Third, the true proportion of personality
variance accounted for by genetic influence can be underestimated if some of the resid-
ual variance involves interactive effects between genes and environment. Children can
influence the quality of their own environment by shaping the responses of others to
conform to their traits (Scarr and McCartney, 1983).
Two other lines of evidence support the biological nature of personality. The
broadest dimensions of personality are valid in entirely different cultures (Eysenck,
1991). Moreover, some personality traits are associated with biological markers
(Cloninger, 1987; Coccaro et aI., 1989).
Environmental factors still account for 50% of the variance in personality. But
one of the most surprising findings of twin research concerns the source of these
environmental influences. The environmental contribution to personality is largely
"unshared", i.e., not related to living in the same family (Plomin et aI., 1990). Unshared
environmental effects could have a number of possible explanations. One is that sib-
lings may receive differential treatment from their parents; a second is that differences
in traits lead individuals to perceive their environment differently; a third is that impor-
tant influences on personality may be derived from experiences outside the family.
These mechanisms are not easily disentangled. Applying social learning theory
(Bandura, 1977), behavioral patterns in children are shaped by positive and negative
reinforcers, as well as by the modeling of behaviors observed in significant adults in the
child's environment. These influences could come from inside the family, or from peer
groups and the social community (Paris, 1996).
5. THE RELATIONSHIP BETWEEN PERSONALITY TRAITS

AND PERSONALITY DISORDERS
A wide variability in personality traits is fully compatible with normality. It is only

when traits significantly interfere with functioning, and when behaviors are used rigidly
and maladaptively, that we can diagnose a personality disorder. Any set of social behav-
iors could be adaptive in some circumstances, yet become maladaptive when applied
rigidly to situations in which they are not appropriate. Personality disorders can there-
fore best be understood as pathological exaggerations of personality traits. The conti-
nuity between traits and disorders is well supported by empirical research (Livesley
et aI., 1994).
Multiple etiological factors could be required to amplify traits to disorders (Paris,
1993). Some of these could be necessary, but none by themselves would be sufficient.
These etiological influences could be divided into risk factors, which would increase the
likelihood of trait amplification, and protective factors, which would buffer risks and
make the amplification of traits less likely.
The biological influences on the personality disorders remain largely unknown
(Nigg and Goldsmith, 1994). However, only a biological model can account for two
troubling facts: 1) individuals with all the psychosocial risk factors for personality dis-
orders do not necessarily develop them; 2) the same psychosocial risk factors produce
entirely different disorders.
The most comprehensive theoretical model of the personality disorders based on
trait variations is that of Siever and Davis (1991), who suggest that individuals with an
unusually high intensity for one or several temperamental characteristics are most
prone to develop personality pathology. These traits, such as emotional instability,
impulsivity, social anxiety, or cognitive instability, could reflect biological risk factors
for personality disorders. Interactions between these traits would either increase or
decrease the risk: if several maladaptive traits were present, the risk would be higher,
but more adaptive traits could buffer the effects of less adaptive ones.
The evidence for the Siever and Davis model, which depends on finding biologi-
cal markers in patients with personality disorders, is thus far limited. Other lines of
research, in which children with abnormal temperaments are followed into adulthood,
may shed more light on these questions (e.g., Kagan, 1994). .
Existing empirical data suggests that psychosocial risk factors could be the most
112 J.Paris
important determinants of whether traits become amplified to maladaptive levels

(Paris, 1993). The largest amount of research on the nature of these risks concerns the
impulsive cluster of personality disorders. Thus, antisocial personality disorder has been
consistently shown to be associated with criminal behavior in a parent, as well as with
severe family dysfunction (Robins, 1966). Borderline personality disorder has been
shown to be associated with trauma, neglect, parental psychopathology, and family
dysfunction (Paris, 1994). We know less about the psychosocial risks for the other
personality disorders, although some evidence suggests that avoidant and dependent
personality disorders can be associated with insensitive and overprotective parenting
(Arbel and Stravynski, 1991; Bornstein, 1992).
Finally, there are social risk factors for personality disorders, consisting of con-
flicts between traits and social expectations (Paris, 1996). Rapid social change enhances
these conflicts, and some personality types cope poorly with these conditions (Millon,
1987,1993; Paris, 1996).
6. WORKING WITH TRAITS IN THERAPY
In spite of all the difficulties discussed above, psychotherapy could still have a
great deal to offer in personality disorders. We may not be able to cure these patients,
but we can help rehabilitate them and improve their social adaptation. Of course, the
approach proposed here, as with any other method of treatment, would have to be sup-
ported by clinical trials. However, the model will be consistent with the research liter-
ature in that it aims to combines experience with supportive psychodynamic therapies
in personality disorders (see McGlashan, 1993), with empirically validated approaches
using cognitive behavioral methods (see Linehan, 1993).
I will briefly sketch a model that is described in much more detail elsewhere
(Paris, 1996, 1998). It starts with the assumption that personality traits are stable. The
aim of therapy is not to modify personality, but, to reverse the process by which traits
are amplified to disorders. Patients who improve would be expected to retain premor-
bid trait profiles, but at less dysfunctional levels. Treatment would largely consist of
teaching patients how to make better use of their personality traits, so as to make them
more adaptive (Paris, 1998).
This approach is consistent with a biopsychosocial model of the personality dis-
orders (Paris, 1996). The biological factors in personality disorders are temperamental
variations shaping traits. Psychosocial factors affect the threshold at which disorders
appear, but are not specific to any category.
The model leads us to see patients with personality pathology in a different light.
As clinicians know, understanding the etiological influences of one's problems is rarely
sufficient by itself for change (Wachtel, 1977). Psychoanalysts describe the change
process as the "working though" of conflicts. However in practice, this process strongly
resembles behavior therapy, since it consists of examining maladaptive behaviors, extin-
guishing them, and replacing them with more adaptive behaviors.
Treatment methods in the personality disorders might therefore be framed in
terms of rehabilitation (van Reekum et aI., 1993), helping those who have deficits in
adaptive skills conform better to social demands. If the environmental factors in the
personality disorders, including those that precipitated the disorder, as well as those
that prolong and maintain maladaptive patterns, are subject to therapeutic influence,
then it may be possible to reverse the process of trait amplification.
The practice and technique of this trait-oriented approach resembles the princi-
ples of cognitive behavioral therapy, and the goal being proposed here, making use of
one's personality traits in more adaptive ways, has also been advocated by practition-
ers of CBT (Beck and Freeman, 1990). Many of the interventions recommended to
change dysfunctional traits are designed to modulate emotions to optimal intensities,
to limit rigid and inappropriate behaviors, and to expand behavioral repertoires
(Linehan, 1993). Other specific techniques would be focused on the improvement of
current psychosocial functioning, through developing more satisfying social roles, and
through establishing stable social networks.
7. APPLICATION TO THE PERSONALITY

DISORDER CLUSTERS
The treatment of the personality disorders differs a great deal with respect to
diagnosis. We can use the DSM system to describe approaches specific to each cluster
of disorders on Axis II.
"Cluster A" patients may not necessarily be capable of sustained relationships.
Schizoid and schizotypal patients need steady employment, preferably in work settings
that are interpersonally undemanding. Some may have to be encouraged to avoid those
situations which they find most difficult, such as intimate relationships, and to focus on
work, which they can manage with less difficulty.
In "Cluster B", the primary problems are the control of impulsivity and affective
instability. Antisocial patients present with "pure culture" impulsivity (Siever and Davis,
1991), and are therefore relatively untreatable in psychotherapy, and often best managed
in the criminal justice and forensic systems (Yochelson and Samenow, 1976). This conclu-
sion could change dramatically in the future if we were to develop pharmacological
agents to control severe impulsivity. In that scenario, the legal system might eventually
require a major input from psychiatry (Masters and McGuire, 1994).
For the other diagnoses in the impulsive cluster, in which affective dysphoria is a
primary feature, psychotherapy remains the primary means of management. Patients
with histrionic, borderline, and narcissistic personality disorders need to learn to
examine their emotions more critically, to act less on impulse, and to take greater
account of the needs of others. Impulsive cluster patients have trouble with intimacy,
but benefit from developing task orientations through employment, and less demand-
ing interpersonal contact through social networks. It has been shown that borderline
patients have a better outcome when they find these external structures, which can
buffer inner chaos (Bardenstein and McGlashan, 1987; McGlashan, 1993).
Narcissistic personality disorder raises problems of particular interest for working
with traits. Social structures which encourage persistence, competence, and achieve-
ment in work, as well as secure attachments in intimate relationships, can help
these patients be less dependent on the more ephemeral reinforcers on which they
tend to rely, such as sexual attractiveness or power. External structures such as
commitments to work and relationships, can become buffers for their excessive
individualism.
These strategies can help therapists to combat the excessive selfconcern in nar-
cissistic patients. The maintenance of family structures is an important element of the
buffering process, since attachment needs are best met within well-functioning fami-
lies. If, however, psychotherapy encourages patients to be even more individualistic,
114 J. Paris
these networks can break down (Glantz and Pearce, 1989). Psychodynamic therapy,
with its focus on internal emotional states and needs, runs the danger of reinforcing
narcissism. In such cases, individual psychotherapy becomes part of the problem, not
part of the solution. Psychotherapists as a group are often just as influenced as anyone
else, if not more so, by the individualistic values of modern society. The entire thera-
peutic enterprise tends to be oriented to encouraging assertiveness in patients, and to
validating perceived needs. This creates an implicit bias favoring the needs of the indi-
vidual over the family or the larger group. It may be no accident that psychotherapy
as a method appeared at the very moment in history when traditional social structures
were most rapidly breaking down. The individualistic bias intrinsic to psychotherapy
seems to be particularly strong in North America, where it reflects the value system of
the culture as a whole (Frank and Frank, 1991).
Patients with diagnoses in Cluster C tend to deal with anxiety and dysphoria by
avoidance or procrastination. These patients therefore need to be encouraged to take
more risks, so as to become more, not less "impulsive". This is, in fact, the method which
Kagan (1994) described as being used by parents in successfully overcoming behavioral
inhibition in their children. It is also the inverse of the family pattern of overprotec-
tion seen in avoidant and dependent patients (Bornstein, 1992).
Encouraging individualism and emotional expressiveness is most appropriate for
patients in the anxious cluster. These individuals are often over-compliant with other
people, and benefit from getting in touch with their inner selves. But therapists need
to prevent these patients from using their defenses to avoid involvement in relation-
ships. Since psychotherapy can become one more place to hide from life, anxious cluster
pathology calls for an active and confrontational approach.
8. THE GOALS OF PSYCHOTHERAPY IN

The approach presented here has focused on adaptation in work setting and on
the establishment of social networks. Many patients with personality disorders have
difficulties in establishing intimate relationships, but the goals of therapy must often be
limited. In the contemporary world, fewer and fewer individuals live in nuclear fami-
lies (Westen, 1985). Modern social structures, unlike those of traditional societies, do
not insure that all of its members can marry and have children. As a result, more indi-
viduals remain alone in life. Some of them will be content, and may even value and
make creative use of their solitude (Storr, 1988), whereas others will feel emotionally
deprived.
A second limitation concerns the effect of personality traits on the capacity
of individuals to attain intimacy. Those who are not capable of establishing stable inti-
macy need to have other satisfactions in life. They require stable and reasonably
satisfying employment, and a social network consisting of less intimate attachments.
For example, in one long-term outcome study of borderline personality disorder
(Bardenstein and McGlashan, 1988), the ability to work was most strongly associated
with stable recovery, whereas a greater investment in intimate relationships actually
led to relative instability, since patients became dysfunctional if those relationships
broke down.
If treatment goals are not kept modest, therapy can become "interminable". In
part, this phenomenon reflects unrealistic expectations of therapists about what kind
of change is possible in their patients. Thus, studies of borderline patients in psy~

chotherapy (Waldinger and Gunderson, 1984) have found that a large percentage leave
treatment against advice. However, it is possible what this observation really indicates
that borderline patients, who can find continuous therapy debilitating, are wiser than
their therapists!
Reflecting on the findings of long~term outcome research, McGlashan (1993) sug-
gests we should not oppose the tendency for borderline patients to drop out of treat-
ment, but capitalize on it. This would be done primarily by making therapy intermittent
rather than continuous, an idea suggested fifty years ago by Alexander and French
(1946). This principle might apply to a variety of personality disorders. Patients could
be encouraged to leave psychotherapy when they master some of their difficulties, and
then be able to return each time they find that their maladaptive behaviors give them
serious difficulty.
The approach to the personality disorders described in this chapters aims to take
into account biological, psychological, and social dimensions. This model helps explain
therapeutic difficulties, even if it does not claim to resolve them. Therapists need to
accept the limitations set by trait profiles, and concentrate on improvements in social
adaptation. Doing so might help reduce some of the frustration in therapists who treat
personality disordered patients.
REFERENCES
Alexander, F. and French, T. (1946): Psychoanalytic Therapy. New York, Ronald.

American Psychiatric Association (1994): Diagnostic and Statistical Manual of Mental Disorders, 4th Edition,
Washington DC, American Psychiatric Press.
Andreoli, A., Frances, A., Gex-Fabry, M., Aapro, N., Gerin, P., and Dazord, M. (1993): Crisis intervention in
depressed patients with and without DSM-III-R personality disorders. J Nerv Ment Dis 181:732-736.
Arbel, N. and Stravynski, A (1991): A retrospective study of separation in the development of adult avoidant
personality disorder. Acta Psychiatr Scand 83:174-178.
Bandura, A, (1977): Social Learning Theory, Englewood Cliffs, Prentice Hall.
Bardenstein, K.K. and McGlashan, T.H. (1989): The natural history of a residentially treated borderline
sample: gender differences. J Personality Disorders 3:69--83.
Beck, AT. and Freeman, A (1990): Cognitive Therapy of Personality Disorder, New York, Guilford.
Bergeman, C.S., Chipuer, H.M., Plomin, R, Pedersen, N.L., McClearn, G.E., Nesselroade, 1.R, Costa, P.T.,
and McCrae, RR (1993): Genetic and environmental effects on openness to experience, agreeable-
ness, and conscientiousness: an adoption/twin study. J Personality 61:158--179.
Bond, M., Paris, 1., and Zweig-Frank, H. (1994): Defense styles and borderline personality disorder. J Per-
sonality Disorders 8:28--31.
Bond, M.P., Gardner, S., Christian, 1., and Sigal, 1.1. (1983): Empirical study of self-rated defense styles.
Archives Gen Psychiatry 40:33-338.
Bornstein, RF. (1992): The dependent personality: developmental, social, and clinical perspectives. Psychol
Bull 112:3-23.
Carey, G. and DiLalla, D.L. (1994): Personality and psychopathology: genetic perspectives. J Abnormal Psy-
chology 103:32-43.
Chess, S. and Thomas, A (1990): The New York Longitudinal Study: the young adult periods. Can J Psychi-
atry 35:557-561.
Cloninger, C.R (1987): A systematic method for clinical description and classification of personality variants.
Arch Gen Psychiatry 44:579-588.
Coccaro, E.F. (1993): Psychopharmacological studies in patients with personality disorders: review and per-
spective. J Personality Disorders 7:Supp 181-192.
Coccaro, E.F., Siever, L.1., Klar, H.M., Maurer, G., Cochrane, K., Cooper, T.B., Mohs, RC., and Davis K.L.
(1989): Serotonergic studies in patients with affective and personality disorders, Arch Gen Psychiatry
46:587-599. Disorders, New York, Guilford.
116 J.Paris
Eysenck, H.I (1991): Genetic and environmental contributions to individual differences: the three major
dimensions of personality. J Personality 58:245-26l.
Frances, A, Clarkin, I, and Perry, S. (1984): Differential Therapeutics in Psychiatry. New York,
Brunner/Maze!'
Frank, AF. (1992): The therapeutic alliances of borderline patients, in IF. Clarkin, E. Marziali, and H.
Munroe-Blum, eds, Borderline Personality Disorder: Clinical and Empirical Perspectives. New York,
Guilford pp. 220-247.
Frank, ID. and Frank, lB. (1991): Persuasion and Healing, 3rd edition, Baltimore, Johns Hopkins.
Glantz, K and Pearce, IK (1989): Exiles from Eden: Psychotherapy from an Evolutionary Perspective. New
York, Norton.
Gunderson, IG. (1985): Conceptual risks of the Axis I-II division, in Klar H. and Siever L.I, eds, Biological
Response Styles: Clinical Implications. Washington DC, American Psychiatric Press, pp. 81-95.
Gunderson, IG., Frank, AF., Ronningstam, E.F., Wahter, S., Lynch, V.l, and Wolf, P.I (1989): Early discon-
tinuance of borderline patients from psychotherapy. J Nerv Ment Dis 177:38-42.
H\1glend, D., S\1rbye, 0., Sorlie, T., Sorlie, W., Fossum, A, and Englestad, V. (1992): Some criteria for brief
dynamic psychotherapy. Psychotherapy and Psychosomatic 57:67-74.
Horwitz, L. (1974): Clinical Prediction in Psychotherapy, New York, Aronson.
Kagan, I (1994): Galen's Prophecy. New York, Basic.
Kernberg, o.F., Coyne, L., Appelbaum, A, Horwitz, L., and Voth, H. (1972): Final report of the Menninger
Psychotherapy Research Project. Bull Menninger Clin 36:1-275.
Kernberg, O.F. (1987): Severe Personality Disorders. New York, Basic.
Kramer, P. (1993): Listening to Prozac, New York, Viking.
Lewis, L. and Appleby, L. (1988): Personality disorder: the patients psychiatrists dislike. Brit J Psychiatry
153:44-49.
Linehan, M.M., Armstrong, H.E., Suarez, A., Allmon, D., and Heard, H.L. (1991): Cognitive behavioral treat-
ment of chronically parasuicidal borderline patients. Arch Gen Psychiatry 48:1060-1064.
Linehan, M.M. (1993): Cognitive Behavioral Treatment of Borderline Personality Disorder. New York,
Guilford.
Linehan, M.M., Heard, H.L., and Armstrong, H.E. (1993): Naturalistic follow-up of a behavioral treatment
for chronically parasuicidal borderline patients. Arch Gen Psychiatry 50:971-974.
Livesley, W.I, Jang, K, Schroeder, M.L., and Jackson, D.N. (1993): Genetic and environmental factors in per-
sonality dimensions. Am J Psychiatry.
Liveseley, W.l, Schroeder, M.L., Jackson, D.N., and Jang, K (1994): Categorical distinctions in the study of
personality disorder: implications for classification. J Abnormal Psycholl03:6-17.
Luborsky, L., Crits-Christoph, P., Mintz, I, and Auerbach, A (1988): Who will benefit from psychotherapy?
Predicting therapeutic outcomes. New York. Basic.
Lykken, D.T., McGue, M., Tellegen, A, and Bouchard, T.I (1992): Emergenesis: genetic traits which may not
run in families. Amer Psychologist 47:1565-1577.
Masters, RD. and McGuire, M.T. (1994): The Neurotransmitter Revolution: Serotonin, Social Behavior, and
the Law. Carbondale IL, South Illinois Press.
Maziade, M., Caron, c., Cote, R, Boutin, P., and Thivierge, I (1990): Extreme temperament and diagnosis: a
study in a psychiatric sample of consecutive children. Arch Gen Psychiatry 47:477-484.
McCrae, RR and Costa, P.T. (1990): Personality in Adulthood. New York, Guilford.
McGlashan, T.H. (1993): Implications of outcome research for the treatment of borderline personality dis-
order, in Paris I, ed, Borderline Personality Disorder: Etiology and Treatment. Washington DC,
American Psychiatric Press, pp. 235-260.
Millon, T. (1987): On the genesis and prevalence of borderline personality disorder: a social learning thesis.
Journal of Personality Disorders 1:354-372.
Millon, T. (1993): Borderline personality disorder: a psychosocial epidemic, in I Paris, ed, Borderline
Personality Disorder: Etiology and Treatment, Washington DC. American Psychiatric Press, pp.
197-210.
Nakao, K, Gunderson, IG., Phillips, KA, Tanaka, N., Yorifuji, K, Takaishi, I, and Nishimura, T. (1992):
Functional impairment in personality disorders. J Personality Disorders 6:24-33.
Nigg, IT. and Goldsmith, H.H. (1994): Genetics of personality disorders: perspectives from personality and
psychopathology research. Psychol Bull 115:346-380.
Paris, I (1988): Long-term outcome of borderline personality disorder: A critical review. J Personality
Disorders 2:189-197.
Paris, I (1990): Personality disorders and treatment outcome in the NIMH Treatment of Depression Col-
laborative Research Program. Am J Psychiatry 147:711-718.
Paris, J. (1993): Treatment of borderline personality disorder in the light of the research on its long-term
outcome. Can J Psychiat 38:S28-S34.
Paris, J. (1996): Social Factors in the Personality Disorders, New York, Cambrige University Press.
Paris, J. (1998): Working with Traits. Northvale, NJ, Aronson.
Paris, J. (1993): Personality disorders: a biopsychosocial model. J Personality Disorders 7:255-264.
Paris, J. (1994): Borderline Personality Disorder: A Multidimensional Approach. Washington DC, American
Psychiatric Press.
Paris, J., Frank, H., Buonvino, M., and Bond, M. (1991): Recollections of parental behavior and Axis II cluster
diagnosis. J Personality Disorders 5:102-106.
Perry, J.e. (1993): Longitudinal studies of personality disorders. J Personality Disorders, Supp 63-85.
Perry, J.e. and Cooper, S.H. (1989): An empirical study of defense mechanisms. Arch Gen Psychiatry
46:44-452.
Plomin, R, DeFries J.e., and McClearn G.E. (1990): Behavioral Genetics: A Primer. New York, W.H.
Freeman.
Propst, A., Paris, J., and Rosberger, Z. (1994): Do therapist experience, diagnosis and functional level predict
outcome in short-term psychotherapy? Can J Psychiatry 39:178-183.
Robins, L.N. (1966): Deviant Children Grown Up. Baltimore, Williams and Wilkins.
Rothbart, M.K. and Ahadi, S.A (1994): Temperament and the development of personality. J Abnormal Psy-
chology 103:55-66.
Rutter, M. (1987) Temperament, Personality, and Personality Development, Brit J Psychiatry 150:443-448.
Rutter, M. and Quinton, D. (1984): Long-term follow-up of women institutionalized in childhood. Brit J
Developmental PsychoI18:225-234.
Rutter, M. and Rutter, M. (1993): Developing Minds: Challenge and Continuity Across the Life Span. New
York, Basic.
Scarr, S. and McCartney, K. (1983): How people make their own environments: a theory of genotype-
environment effects, Child Development 54:424-435.
Shea, M.T., Pilkonis, P.A., Beckham, E., Collins, J.F., Elikin, E., Sotsky S.M., and Docherty, J.P. (1990): Per-
sonality disorders and treatment outcome in the NIMH Treatment of Depression Collaborative
Research Program. Am J Psychiatry 147:711-718.
Shea, M.T., Widiger, T.A, and Klein, M.H. (1992): Comorbidity of personality disorders and depression: impli-
cations for treatment. J Consulting Clinical Psychol 60:857-868.
Siever, L.J. and Davis, L. (1991): A psychobiological perspective on the personality disorders. Am J Psychi-
atry 148:1647-1658.
Skodol, AE., Buckley, P., and Charles, E. (1983) Is there a characteristic pattern in the treatment history of
clinic outpatients with borderline personality? J Nerv Ment Disease 171:405-410.
Soloff, P.H. (1993): Psychopharmacological intervention in borderline personality disorder, in Paris J., ed,
Borderline Personality Disorder: Etiology and Treatment. Washington DC, American Psychiatric
Press, pp. 319-348.
Storr, A. (1988): Solitude. New York, Free Press.
Tellegen, A, Lykken, D.T., Bouchard, T.J., Wilcox, K.J., Segal, N.L., and Rich, S. (1988): Personality similar-
ity in twins reared apart and together. J Pers Soc PsychoI54:1031-1039.
Tyrer, P. (1988): Personality Disorders, London, Wright.
van Reekum, R, Links, P.S., and Boiago, 1. (1993): Constitutional factors in borderline personality disorder,
in Paris J., ed, Borderline Personality Disorder: Etiology and Treatment. Washington DC, American
Psychiatric Press, pp. 13-38.
Wachtel, P.L. (1994): Cyclical processes in personality and psychopathology. J Abnormal Psychology
103:51-54.
Wachtel, P.L. (1977): Psychoanalysis and Behavior Therapy. New York, Basic.
Waldinger, RJ. and Gunderson, J.G. (1984): Completed psychotherapies with borderline patients. Am J Psy-
chother 38:190-201.
Wallerstein, R (1986): Forty-two Lives in Treatment. New York, Guilford.
Westen, D. (1985): Self and Society: Narcissism, Collectivism and the Development of Morals. New York,
Cambridge University Press.
Winston, A, Laikin, M., Pollack, J., Wallner Samstag, L., McCullough, L., and Muran, J.e. (1994): Short-term
psychotherapy of personality disorders. Am J Psychiatry 151:190-194.
Yochelson, S. and Samenow, S. (1976): The Criminal Personality. New York, Jason Aronson.
10
THE STRUCTURE OF DSM-IV BORDERLINE

PERSONALITY DISORDER AND ITS
IMPLICATIONS FOR TREATMENT
Cesare MaffeP and Andrea Fossati1,2
1Medical Psychology and Psychotherapy Unit

Istituto Scientifico Ospedale San Raffaele
via Stamira d' Ancona 20, 1-20127 Milano, Italy
21nstitue of Psychology
University of Urbino
via Saffi 15, Urbino, Italy
1. INTRODUCTION
Since Stern's observations (1938), the term borderline gained widespread atten-
tion in psychiatric p~actice. For a long time, this term was used in a confusing way, iden-
tifying states, syndromes, personality types, and schizophrenia subgroups. In the '60s,
Kernberg (1967) described Borderline personality as a distinct personality organiza-
tion (BPO), different from Neurotic and Psychotic personality organizations, as well as
from classic psychiatric syndromes (e.g., schizophrenia, mood disorders, etc.). With the
introduction of the Diagnostic and Statistical Manual of Mental Disorder, 3rd Edition
(DSM-III) (APA, 1980) and, later, 3rd Edition, Revised (DSM-III-R) (APA, 1987) the
Borderline Personality Disorder (BPD) was described as one of the Axis II Personal-
ity Disorders (PDs). While Kernberg's BPO heavily relied on a psychoanalytically-
oriented diagnostic framework (i.e., based on the identification of inferred psychic
functioning), DSM-IIIJ-R tried to convert the previous clinical pictures in an atheo-
retical, operationalized set of diagnostic criteria for BPD. In agreement with the
neo-kraepelinian point of view, BPD was considered as a unidimensional, categorial
diagnosis. However, in recent years clinical psychologists built many models normal
and abnormal personality functioning based on dimensional constructs. This raised a
controversy between models considering BPD as a distinct psychiatric disorder and
120 C. Maffei and A. Fossati
those looking at BPD as an extreme variant of a normally distributed personality

dimension. For the former BPD is a complex intermingling of several stable charac-
teristicsof personality (behavioral, cognitive, affective, interpersonal, and tempera-
mental), while for the latter BPD is the extreme expression of a temperamental
disposition, such as Novelty Seeking or Neuroticism. This controversy refers to a deeply
rooted debate, raised in clinical practice, stressing the pros and cons of each approach.
Categorial models are economic, easily transmissible, widely used, and provided with
external validity. However, they are stereotyped in non-prototypical and boundary
cases, and do not explain comorbidity. Dimensional models provide more vivid pictures,
are flexible, and could be converted to categories whenever needed, while the reverse
could not be done. However, they are theoretically-based and there is no agreement
between theorists on the number of dimensions. While in medicine research showed
that some pathologies are better described by categories (e.g., diabetes mellitus), while
others by dimensions (e.g., blood hypertension), up to now no definitive data are avail-
able in psichiatric and psychological research on PDs.
Several authors questioned the validity of the categorial model of PDs, sug-
gesting a shift towards dimensional conceptualization and assessment (Eysenk, 1987;
Livesley, 1991; Widiger, 1991; Oldham, Skodol, Kellman, Hyler, Rosnick, and Davies,
1992; Widiger and Frances, 1994).
As to the BPD, few studies were carried out to specifically test the hypothesis
of a categorial structure (Nurnberg, Hurt, Feldman, and Shu, 1988; Trull, Widiger,
and Guthrie, 1990). Moreover, their results were contrasting (Widiger and Frances,
1994) because they did not support the categorial hypothesis. At the same time, they
did not find clear evidence of a dimensional model. Many taxometric techniques were
used, ranging from cluster analysis to admixture and maximum covariation analysis
(Widiger, 1991; Widiger and Frances, 1994). Latent class analysis, a statistical method
particularly useful in detecting latent taxons when categorial manifest variables (as
DSM-III/-R PD criteria) are used (Everitt, 1993), was never applied to BPD criteria
(Widiger and Frances, 1994). Their results suggested the need for further studies, based
on different statistical techniques, and played a major role in the aims and design of
this study.
A second criticism concerned the unidimensionality of BPD. BPD sub-syndromes
were proposed, implying a multidimensional model of BPD.
Studies based on exploratory factor analysis or cluster analysis questioned the
assumption of unidimensionality of BPD, suggesting the presence of three to four
subsets of DSM-III/-R BPD criteria (Hurt, Clarkin, Widiger, Fyer, Sullivan, Stone, and
Frances, 1990; Livesley and Schroeder, 1991; Clarkin, Hull, and Hurt, 1993). These data
suggested the presence of clinical heterogeneity within the BPD criteria set, as well as
within BPD patients (Trull and McRae, 1994). However, the structure of BPD criteria
subsets could not be completely replicated between different studies due to different
BPD assessment, sample size and characteristics, and statistical analyses. Moreover,
none of these studies statistically assessed the goodness of fit of a specific BPD multi-
dimensional model (i.e., distinct sub-syndromes of BPD), as well as its superiority to
the DSM-IIII-R unidimensional model.
With the introduction of the Fourth Edition of DSM (DSM-IV) (American Psy-
chiatric Association, 1994), the debate on these topics raised to a deeper level. Despite
the suggestion of adopting a dimensional model (Widiger, 1991; Widiger and Frances,
1994), DSM-IV maintained a categorial model for BPD, as well as for other PDs.
According to criticisms to DSM-III-R polythetic format (Widiger, 1991) and psycho-
The Structure of DSM-IV Borderline Personality Disorder 121
metric studies of the diagnostic efficiency of the individual PD criteria (Gunderson,

Zanarini, and Kiesel, 1991), polythetic PD diagnoses were slightly modified in DSM-
IV by entering a hierarchy in PD criteria. Research findings played a major role in
entering anew BPD diagnostic feature (stress-related paranoid ideation or severe dis-
sociative symptoms). Up to now, no study was performed to evaluate the teneability of
DSM -IV hierarchy of BPD revised criteria, based on their diagnostic efficiency.
Starting from these considerations, the present study was carried out in order to:
1. perform an analysis of the diagnostic efficiency of the individual DSM-IV BPD cri-
teria; 2. test the fit of DSM-IV unidimensional model of BPD, and its superiority to
other proposed models; 3. evaluate the presence and number of latent taxons under-
lying the DSM-IV BPD criteria.
2. METHOD
The study group consisted of 564 subjects consecutively admitted from January
1995 to May 1996 to the Medical Psychology and Psychotherapy Unit of the Scientific
Institute H San Raffaele of Milan. None of these subjects met any of the following
exclusion criteria: 1. DSM-IV Axis I diagnosis of Schizophrenia, Schizo affective Dis-
order, Delusional Disorder, or Delirium, Dementia, Amnestic, and Cognitive Disorder
NOS; 2. IQ75; 3. Education level lower than elementary school. 239 (42.4%) subjects
were male and 325 (57.6%) female; mean age was 29.92 (SD = 8.50) years. 368 (65.2 %)
were inpatients and 194 (34.8%) outpatients. 418 (74.2 %) subjects received at least one
DSM-IV Axis I diagnosis; most frequently diagnosed DSM-IV Axis I disorders were
Anxiety Disorders (N = 178,31.6%), Eating Disorders (N = 93, 16.5%), Mood Disor-
ders (N = 63, 11.2%), Substance AbuselDependence Disorders (N = 59, 10.5%), and
Brief/NOS Psychotic Disorder (N = 35, 6.2%). 26 subjects (4.6%) received other Axis
I diagnoses (e.g, Paraphilias,.Sleep Disorders, etc.). The cumulative frequency and per-
centage of subjects with specific DSM-IV Axis I diagnoses exceeded the frequency and
percentage of subjects with at least one DSM-IV Axis I diagnosis because of multiple
Axis I diagnoses. No significant difference was observed between in- and outpatients
with respect to demographic variables; as expected, inpatients showed a significantly
higher frequency of axis I diagnoses (N = 331, 89.9%) than outpatients (N = 87,44.4 %):
Yates corrected chi-square = 135.986, df = 1, P < 0.001. With respect to BPD diagnosis,
no significant difference was observed between inpatients (N = 63, 17.1 %) and out-
patients (N = 37, 18.9%): Yates corrected chi-square = 0.164, df = 1, P > 0.60.
After complete description of the study to the subjects, written informed consent
was obtained.
DSM-IV BPD criteria and diagnosis were assessed using Structured Clinical
Interview for DSM-IV Axis II Personality Disorders, Version 2.0 (SCID-II) (First,
Spitzer, Gibbon, Janet, and Benjamin, 1994), a semistructured interview designed to
diagnose the DSM-IV PDs. SCID-II was preceeded by the administration of its self-
report screening questionnaire (PQ). Subjects with axis I diagnoses were admnistered
SCID-II at acute symptom remission. DSM-IV BPD diagnosis proved to have excel-
lent interrater reliability (dimensional diagnosis: ICC = 0.952-categorial diagnosis: K
= 0.909). Also the individual DSM-IV BPD diagnostic criteria showed adequate inter-
rater reliability (Median K = 0.868-see Table la for detailed listing). All other DSM-
IV PD diagnoses presented satisfactory interrater reliability coefficients (dimensional
diagnoses: median ICC = 0.937-categorial diagnoses: median K = 0.912).
The presence of significant association between DSM-IV BPD and other PDs was
assessed using phi coefficient. Nominal alpha level was controlled by using the Bon-
ferroni procedure (0.05/11 = 0.0045).
DSM-IV BPD criteria diagnostic accuracy was assessed by computing item-total
(n. of criteria) point-biserial correlation coefficient (crpbi) and item-diagnosis phi
coefficient (<Pc), both corrected for overlap (Henrysson, 1963; Guilford and Fruchter,
1978). Within each set of item-total and item-diagnosis comparisons, nominal alpha
level was stabilized using Bonferroni correction (0.05/9 = 0.0056). Correlations
(coefficients) between DSM-IV BPD criteria and other DSM-IV PD diagnoses were
computed to evaluate BPD criteria discriminatory power. Sensitivity, specificity,
efficiency (i.e., the total probability of making a correct statement about the presence
or absence of a particular disease. Youden J was used as efficiency measure),
positive (PPP) and negative (NPP) predictive power of the individual BPD criteria
could not be computed from standard formulas, based on contingency tables, because
of several violations of the assumptions underlying these statistics (e.g., presence of
item-diagnosis overlap, lack of independency among BPD criteria, lack of a "gold stan-
dard" for BPD diagnosis, etc.) (Rindskopf and Rindskopf, 1986). As suggested by
several authors (Rindskopf and Rindskopf, 1986; Young, 1983; Uebersax and Grove,
1990), Latent Class Analysis was used to evaluate BPD criteria diagnostic accuracy.
The formulas for multiple latent classes were used to derive BPD criteria sens-
itivity, specificity, efficiency, PPP and NPP from latent class conditional probabilities
(Uebersax and Grove, 1990).
The hypothesis that one factor was sufficient to explain BPD criteria covariance
was tested using confirmatory factor analysis (CFA). Given the dichotomous assess-
ment of DSM-IV BPD criteria, a weighted least square (WLS) CFA was performed by
using the tetrachoric correlation matrix as input matrix (Bollen, 1989). Covariance
matrix of estimated tetrachoric correlations was used as weight matrix. According to
the hypothesis of this study and the DSM-IV model of BPD, a unidimensional model
with congeneric (i.e., linearly related, with no additional constrain) items was built and
tested against the following alternative models: 1. unidimensional with parallel items
(i.e., items with equal true score and error variance) (compatible with DSM-III-R
model); 2. unidimensional with tau-equivalent items (i.e., items with equal true score
variance but with different error variance) (compatible with DSM-III-R model); 3.
three dimensional (uncertainty about the self and interpersonal difficulties, affect and
mood regulation, and impulsivity) with orthogonal factors, derived from Clarkin et al.
(1993) exploratory study; 4. three dimensional (identity, affect, and impulse clusters)
with orthogonal factors, derived from Hurt et al. (1990) study based on cluster analy-
sis of PDs; 5. four dimensional (uncertainty about self and interpersonal difficulties,
affect and mood regulation, anger, and impulsivity) with orthogonal factors, derived
from Clarkin et al. (1993) study. Published factor loadings were used as starting points
for models derived from Clarkin et al. (1993) study. The identification of the BPD model
based on DSM-IV was assessed using t- and three factor rules. Model goodness-of-fit
was evaluated using WLS asymptotic chi-square (X2) statistic.
The categorial model of BPD hypothesizes the existence of natural clusters of
subjects, i.e., non-artifactual groups where subjects belonging to one group are maxi-
mally similar each other and dissimilar to subjects belonging to the other group(s).
Exploratory LCA was used to identify the number of latent classes underlying DSM-
IV BPD criteria. Best fitting model was identified by using improvement in likelihood
chi-square statistic (U) and bayesian information criterion (BIC).
LCA subject classification was generated using individual latent class member-
ship probabilities. LCA and DSM-IV classifications of BPD were then compared. The
potential confounding role of patient severity on clustering was assessed testing the
association between latent classes and, respectively, in-/outpatient status and presence
of any axis I disorder.
3. RESULTS
According to SCID-II, 370 subjects (65.6%) received at least one DSM-IV PD

diagnosis; mean number of PD diagnoses was 1.14 (SD = 1.11). Mean number of DSM-
IV BPD criteria was 1.71 (SD = 2.37); 280 subjects (49.6%) received no BPD criteria.
DSM-IV BPD was diagnosed in 100 subjects (17.7%). Among these, 39 subjects (39%)
received a "pure" BPD diagnosis, while 61 (61 %) received one or more additional PD
diagnoses (range: 1-4). Significant, even if weak, positive associations were observed
between BPD and Passive-Aggressive (<p = 0.22), Antisocial (<p = 0.19), Histrionic (<p =
0.19), and Narcissistic (<p = 0.18) PDs (all p < 0.0045). DSM-IV BPD diagnosis showed
negative significant correlations with Avoidant (<p = -0.19) and Obsessive-Compulsive
(<p = -0.13) PDs (all p < 0.0045). Mean number of PD co-diagnoses in subjects with
DSM-IV BPD was 1.14 (SD = 1.18). Mean number of BPD criteria among subjects with
BPD diagnosis was 6.21 (SD = 1.32); only 7 subjects could be considered as "proto-
typical" cases (Le., meeting all nine diagnostic criteria) of DSM-IV BPD.
Item analysis results are listed in Table la.
All item-total and item-diagnosis corrected correlation coefficients were large
and significant (all p < 0.0056), as well as larger than the correlation coefficients
Table 1a. Item analysis

DSM-IV BPD Criteria IRR BR cr phi , Item-PD Item-Cluster B PD 
K (N) Median Median
(min.lmax.) (min.lmax.)
1. Frantic Efforts 0.915 0.165 0.54 0.41 0.04 0.11
(93) (-0.08/0.14) (0.04 / 0.14)
2. Unstable Relationships 0.951 0.170 0.72 0.68 0.01 0.12
(96) (-0.13 /0.19) (0.10 / 0.19)
3. Identity Disturbance 0.868 0.200 0.70 0.67 -0.04 0.22
(113) (-0.16/0.23) (0.16/ 0.23)
4. Impulsivity 0.883 0.252 0.64 0.51 -0.03 0.18
(142) (-0.13 / 0.30) (0.18/0.30)
5. Suicidal Behavior 0.873 0.165 0.53 0.43 0.05 0.13
(93) (-0.03 / 0.16) (0.05 /0.16)
6. Affective Instability 0.739 0.216 0.63 0.54 -0.01 0.17
(122) (-0.10/ 0.19) (0.12 / 0.19)
7. Feelings of Emptiness 0.866 0.163 0.61 0.56 -0.03 0.13
(92) (-0.06 / 0.16) (0.08/0.14)
8. Inappropriate Anger 0.861 0.271 0.61 0.44 0.04 0.17
(153) (-0.09 / 0.25) (0)2/0.25)
9. Paranoid Ideation 0.711 0.122 0.54 0.45 0.03 0.12
(69) (-0.09 / 0.13) (0.09 / 0.13)
BPD criteria are listed in DSM-IV order-IRR = Interrater Reliability (Based on 231 observations)-BR = Base Rate-
.,r pm = item-total corrected point-biserial r-, = item-diagnosis corrected coefficient.
124 c. Maffei and A. Fossati
between BPD criteria and other DSM-IV PD diagnoses. However, as expected, differ-
ences in diagnostic efficiency were observed. In particular, considering coefficients cor-
rected for criterion-diagnosis overlap, DSM-IV BPD criteria should be ranked as
follows: 1. Unstable relationships; 2. Identity disturbance; 3. Feelings of emptiness; 4.
Affective instability; 5. Impulsivity; 6. Paranoid ideation; 7. Inappropriate anger; 8.
Suicidal behavior; 9. Frantic efforts. It should be noted that the rank order of BPD
criteria coefficients was independent from the rank order of the interrater reliability
o coefficients: Spearman r = 0.067 P > 0.80.
Sensitivity, specificity, efficiency (Youden J), PPP and NPP of BPD criteria are
listed in Table lb.
As shown, unstable relationships and identity disturbance appeared as the most
relevant diagnostic criteria. Moreover, the data listed in Table la-b suggested a small
heterogeneity of BPD subjects with respect to these two criteria; rather, they appeared
as the two main BPD characteristics. For instance, when the conjoint presence of unsta-
ble relationships and identity disturbance was considered, the sensitivity, specificity,
efficiency, PPP and NPP were as follows: 0.736, 0.998, 0.758, 0.985, 0.952. On the other
hand, frantic efforts to avoid abandonment was one of the criteria provided with the
worst diagnostic accuracy.
All tetrachoric correlation coefficients among BPD criteria were large, positive,
and significant, even after the nominal alpha level Bonferoni correction (0.05/36 =
0.0014). Median tetrachoric correlation was 0.62 (min = 0.48, max. = 0.82). CFA results
are shown in Table 2.
The BPD model derived from DSM-IV (unidimensional/congeneric items)
showed adequate fit and was clearly more valid than other alternative models.
Considering LCA, the unidimensional model with one latent class did not fit ade-
quately (L2=1484.122, df =502, P < 0.001). Adding a second class significantly improved
the model (V difference = 1025.993, df = 10, P < 0.001) and adequately fitted the data
(L2 = 458.129, df = 492, P > 0.80). The model was improved further (V difference =
96.594, df = 10, P < 0.001) when a third latent class was added (goodness of fit: V =
361.535, df = 482, P > 0.90). No significant improvement was observed when a fourth
latent class was added (L2 difference = 14.475, df = 10, P > 0.10). BIC reached its
minimum value (3962.849) in correspondence of the three class model, indicating that
this model is the best fitting one. According to these results, DSM-IV BPD criteria mul-
tivariate distribution could be considered as a mixture of two, or more likely, three mul-
Table lb. Item analysis: diagnostic efficiency statistics

BR-BPD (N) Sensitivity Specificity Youden J PPP NPP
1. Frantic Efforts 0.560 (56) 0.543 0.914 0.457 0.551 0.833
2. Unstable Relationships 0.750 (75) 0.809 0.925 0.766 0.785 0.925
3. Identity Disturbance 0.820 (82) 0.878 0.934 0.812 0.722 0.949
4. Impulsivity 0.810 (81) 0.789 0.863 0.652 0.527 0.912
5. Suicidal Behavior 0.570 (57) 0.534 0.913 0.447 0.544 0.830
6. Affective Instability 0.760 (76) 0.784 0.899 0.684 0.602 0.912
7. Feelings of Emptiness 0.650 (65) 0.663 0.938 0.601 0.672 0.873
8. Inappropriate Anger 0.800 (80) 0.777 0.840 0.617 0.484 0.907
9. Paranoid Ideation 0.480 (48) 0.461 0.947 0.408 0.626 0.812
Derived from LeA best fitting model (see Table 3)-BR-BPD = Base rate of the individual criteria in subjects (N = 100)
with a DSM-IV BPD diagnosis (frequencies are between brackets)-PPP = Positive Predicitve Power-NPP = Negative
Predicitve Power-Youden J = (Sensitivity + Specificity)-l.
Table 2. DSM-IV BPD Criteria: Weighted least square confirmatory factor analysis
Models X' DF P
Unidimensional! Congenericity 18.89 27 0.874
Unidimensional! Tau-Equivalence 80.30 35 <0.001
Unidimensional! Parallelism 80.30 43 <0.001
Three Dimensions (Uncertainty about the self, Affect 1858.31 24 <0.001
regulation, Impulsivity) ! Orthogonal Factors
Three Dimensions (Identity Disorder, Affective Instability, 2889.53 25 <0.001
Impulsivity) ! Orthogonal Factors
Four Dimensions (Uncertainty about the self, Affect regulation, Anger, 2713.85 24 <0.001
Impulsivity) ! Orthogonal Factors
Best Fitting Model

BPD Items Factor Loadings (Completely standardized solution)
1. Frantic Efforts 0.718
2. Unstable Relationships 0.903
3. Identity Disturbance 0.889
4. Impulsivity 0.818
5. Suicidal Behavior 0.703
6. Affective Instability 0.807
7. Feelings of Emptiness 0.788
8. Inappropriate Anger 0.784
9. Paranoid Ideation 0.726
tivariate bernoulli distributions. Conditional probabilities for LCA best fitting model
are shown in Table 3.
Class 1 was mainly composed of subjects (% = 16.1, n = 91) scoring above DSM-
IV diagnostic threshold for BPD; it was characterized by the presence of all DSM-IV
BPD criteria (mean = 6.32, SD = 1.34), with a less clear contribution of criterion 9.
Classes 2 and 3 were composed, respectively of subjects with no (Class 2: % = 56.0, n
Table 3. Latent class analysis: three class model

Latent Class 1 Latent Class 2 Latent Class 3
DSM-IV BPD Items Conditional Probabilities Conditional Probabilities Conditional Probabilities
1. Frantic Efforts 0.543 0.017 0.224
2. Unstable Relationships 0.809 0.000 0.129
3. Identity Disturbance 0.878 0.023 0.151
4. Impulsivity 0.789 0.019 0.375
5. Suicidal Behavior 0.534 0.012 0.237
6. Affective Instability 0.784 0.030 0.242
7. Emptiness 0.663 0.015 0.158
8. Inappropriate Anger 0.777 0.016 0.450
9. Paranoid Ideation 0.461 0.011 0.138
N. of subjects ('Yo) 91 (16.2) 316 (56.0) 157 (27.8)

Mean N. (SD) of BPD items 6.32 (1.34) 0.12 (0.33) 2.25 (1.11)
Subjects were ascribed to latent classes according to their highest individual membership probability. Means and SDs of
BPD criteria are raw means (and SDs) of the nine items.
U6 C. Maffei and A. Fossati
= 316; mean number of BPD criteria = 0.12, SD = 0.33) or few BPD criteria (Class 3:
% = 27.8, n = 157; mean number of BPD criteria = 2.25, SD = 1.11). According to LCA
conditional probabilities, inappropriate anger (0.450) and impulsivity (0.375) were the
two BPD-like personality traits more frequentily diagnosed in non-BPD subjects. As
expected, a highly significant difference was observed between latent classes 1 and 3
when BPD criteria dispersion matrices were compared: Box M = 80.118, chi-square =
76.740, df = 45, P < 0.005 (Latent class 1 dispersion matrix could not be included in the
analysis because singular). Ten subjects diagnosed as having BPD according to DSM-
IV threshold were misclassified by LCA as belonging to class 3, while only one subject
diagnosed as non-BPD according to DSM-IV was classified in class 1 by LCA. When
LCA classes 2 and 3 were grouped to form a "non-BPD" class, the agreement between
DSM-IV and LCA classifications was almost perfect (Cohen K = 0.931 P < 0.001). It
should be noted that the agreement between LCA and DSM-IV classifications of BPD
was substantial also for the less fitting two latent class model (K = 0.794 P < 0.001). No
significant association was observed between latent classes and, respectively, in-/out-
patient status (chi-square = 0.143, df = 2, P > 0.90) and any axis I disorder (chi-square
= 2.091, df = 2, P > 0.35).
4. DISCUSSION
Our data supported the hypothesis of BPD as a unidimensional construct. Item

analysis results confirmed a differential diagnostic efficiency of BPD criteria. Different
from DSM-IV, unstable relationships and identity problems were the main diagnostic
features of BPD. Since these traits were relevant diagnostic markers of BPD, the prob-
ability of finding a BPD subject without any of these two characteristics (or a non-BPD
subject with any of these two characteristics) should be quite low. This result has diag-
nostic and therapeutic implications. Considering sensitivity, specificity, PPP and NPP of
these two features subjects meeting both characteristcs, but scoring below the diag-
nostic threshold should be considered as suspect false negatives, rather than dismissed
simply as "non-BPD". In such situation, longitudinal methods as LEAD or PLASTIC
standards should be used to minimize the probability of a false negative. On the other
hand attention to, and management of interpersonal relationship and identity distur-
bances should be a precocious and major focus of any treatment of BPD. These con-
siderations agree with the clinical literature on the psychotherapeutic treatment of
BPD (Kernberg, 1984). In fact, instability of object relationships and identity, mani-
fested by oscillations from idealization to devaluation (and viceversa) and consequent
rapid shifts in long-term goals, can seriously interfere with therapeutic alliance and
insight (for instance, minor, transient progresses could be considered by the patient as
full remissions). Thus, stable definition of the therapeutic setting and use of confronta-
tive technique seem to be useful in the treatment of BPD sUbjects.
LCA results gave evidence of a categorial structure of BPD. As expected, a
marked heterogeneity was observed among non-BPD subjects, who could be described
as completely different from BPD (latent class 2), or as possessing some BPD-like mal-
adaptive personality features, namely inappropriate anger and impulsivity, which seem
to be non-specific dysfunctional temperamental features (latent class 3). As shown by
dispersion matrix analysis, latent class 1 was not only quantitatively (i.e., n. of BPD cri-
teria), but also qualitatively different from the other latent classes, considering the pres-
ence of a significant difference in the covariation of BPD criteria.
The Structure of DSM·IV Borderline Personality Disorder U7
Thus, despite better reliability of BPD dimensional assessment (Pilkonis, Heape,

Proietti, Clark, McDavid, and Pitts, 1995), the categorial BPD diagnosis seemed more
correspondent to the actual BPD structure. The strong agreement between LCA and
DSM-IV models of BPD seemed to show the adequacy, or at least the empirical repro-
ducibility, of the DSM-IV diagnostic threshold for BPD. It was also interesting to
observe roughly the same average number of BPD diagnostic features (n = 6) in sub-
jects diagnosed as BPD according to DSM-IV cut-off threshold score and in subjects
belonging to latent class 1.
In summary, BPD appeared as a distinct and articulated personality disorder,
rather than the extreme variant of a single personality dimension (or a plus/minus
balance of personality dimensions).
According to diagnostic features, DSM-IV BPD should be described as a unidi-
mensional, categorial PD. This does not mean that subjects with BPD could not be sub-
typed using variables external to diagnostic criteria. Diagnostic homogeneity does not
imply the absence of natural subgroups of BPD subjects when additional characteris-
tics, relevant to personality description and development are considered. In fact, BPD
heterogeneity was shown when temperamental factors (Akiskal, 1981) or variables
related to developmental history were considered (Paris and Zweig-Frank, 1993). For
instance, different subtypes of BPD could need different therapeutic approaches, and
show different clinical course and response to treatment. However, more research on
subtyping BPD according to external variables is needed before drawing any conclu-
sion on this topic.
If these results could be accepted, then two main ways of treating BPD subjects
should be suggested: 1. a symptom-oriented treatment, aiming at reducing specific
symptoms. This goal-oriented approach should be considered when specific BPD char-
acteristics are particularly threatening (e.g., severe impulsivity or sucidal/parasuicidal
behavior) or when patient characteristics do not allow for an insight-oriented treat-
ment. We mean that treating impulsivity or affect dysregulation does not mean treat-
ing the pathological core of BPD. 2. A "radical" treatment of BPD, aiming at changing
the whole maladaptive personality pattern. Unfortunately, these approaches rely on dif-
ferent, and often contrasting views of the etiopathogenesis of BPD.
In our opinion and in more recent clinical developments (Paris, 1996) real treat-
ment conditions need a combination of these two approaches.
REFERENCES
Akiskal, H.S. (1981). Subaffective disorders: dysthymic, cyclothymic, and bipolar II disorders in the "bor-
derline" realm. Psychiatric Clinics of North America, 4, 25-46.
American Psychiatric Association (1980). Diagnostic and statistical manual of mental disorders, 3rd Edition.
Washington, DC: American Psychiatric Press.
American Psychiatric Association (1987). Diagnostic and statistical manual of mental disorders, 3rd Edition,
Revised. Washington, DC: American Psychiatric Press.
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders, 4th edition.
Washington, DC: American Psychiatric Press.
Bollen, K.A. (1989). Structural equations with latent variables. New York, NY: John Wiley and Sons.
Clarkin, IF., Hull, IW., and Hurt, S.W. (1993). Factor structure of borderline personality disorder criteria.
Journal of Personality Disorders, 7,137-143.
Everitt, B.S. (1993). Cluster analysis. Third edition. London, UK: Edward Arnold.
Eysenck, H. (1987) The definition of personality disorders and the criteria appropriate for their description.
First, M.B., Spitzer, RL., Gibbon, M., Janet, B.Y., and Benjamin, L. (1994). Structured Clinical Interview for
DSM-IV Axis II Personality Disorders (SCID-II), Version 2.0. New York, NY: Biometric Research
Department, New York State Psychiatric Institute.
Guilford, IP. and Fruchter, B. (1978). Fundamental statistics in psychology and education, 6th edition. New
York, NY: McGraw-Hill.
Gunderson, IG., Zanarini, M.e., and Kisiel, e.L. (1991). Borderline personality disorder: a review of data on
DSM-III-R descriptions. Journal of Personality Disorders, 5, 340-352.
Henrysson, S. (1963). Correction of item-total correlations in item analysis. Psychometrika, 28, 211-218.
Hurt, S., Clarkin, I, Widiger, T., Fyer, M., Sullivan, T., Stone, M., and Frances, A (1990). Evaluation of DSM-
III decision rules for case detection using joint conditional probability structures. Journal of Person-
ality Disorders, 4,121-130.
Kernberg, O.F. (1967). Borderline personality organization. Journal of the American Psychoanalytic Associ-
ation, 15, 641-{i85.
Kernberg, O.F. (1984). Severe Personality Disorders. New Haven and London: Yale University Press.
Livesley, w.J. (1991). Classifying personality disorders: ideal types, prototypes, or dimensions? Journal of Per-
sonality Disorders, 5, 52-59.
Livesley, W.I and Schroeder, M.L. (1991). Dimensions of personality disorder: the DSM-III-R cluster b diag-
noses. Journal of Nervous and Mental Disease, 179,320-328.
Nurnberg, H.G., Hurt, S.W., Feldman, A, and Shu, R (1988). Evaluation of diagnostic criteria for borderline
personality disorder. American Journal of Psychiatry, 145, 1280-1284.
Oldham, 1M., Skodol, A.E., Kellman, H.D., Hyler, S.E., Rosnick, L., and Davis, M. (1992). Disagnosis of DSM-
III personality disorders by two structured interviews: patterns of comorbidity. American Journal of
Psychiatry, 149, 213-220.
Paris, J. (1996). Social Factors in the Personality Disorders. Cambridge University Press.
Paris, I, Zweig-Frank, H. (1993). Parental bonding in borderline personality disorder. In I Paris (Ed.), Bor-
derline personality disorder. Etiology and treatment. Washington, DC: American Psychiatric Press.
Pilkonis, P.A, Heape, e.L., Proietti, J.M., Clark, S.w., McDavid, J.D., and Pitts, T.E. (1995). The reliability and
validity of two structured diagnostic interviews for personality disorders. Archives of Geneneral Psy-
chiatry, 52, 1025-1033.
Rindskopf, D. and Rindskopf, W. (1986). The value of latent class analysis in medical diagnosis. Statistics in
Medicine, 5, 21-27.
Stern, A (1938). Psychoanalytic investigation of and therapy in the borderline group of neuroses. Psycho-
analytic Quarterly, 7, 467-489.
Trull, T.J. and McCrae, RR (1994). A five-factor perspective on personality disorder research. In P.T. Costa
and T.A Widiger (Eds.), Personality Disorders and the Five-Factor Model of Personality, 2nd Edition.
Washington, DC: American Psychological Association.
Trull, TJ., Widiger, T.A., and Guthrie, P. (1990). The categorical versus dimensional status of borderline per-
sonality disorder. Journal of Abnormal Psychology, 99, 40-48.
Uebersax, IS. and Grove, W.M. (1990). Latent class analysis of diagnostic agreement. Statistics in Medicine,
9,559-572.
Widiger, T.A (1991). Personality disorder dimensional models proposed for DSM-IY. Journal of Personality
Disorders, 5, 386-398.
Widiger, T.A. and Frances, AJ. (1994). Toward a dimensional model for the personality disorders. In P.T.
Costa and T.A. Widiger (Eds.), Personality Disorders and the Five-Factor Model of Personality, 2nd
Edition. Washington, DC: American Psychological Association.
Young, M.A (1983). Evaluating diagnostic criteria: a latent class paradigm. J Psychiatry Research, 17,285-196.
11
THE RELATIONSHIP BETWEEN ANXIETY

DISORDERS AND PERSONALITY
DISORDERS
Prevalence Rates and Comorbidity Models
Carol 1. M. Van Velzen1 and Paul M. G. Emmelkamp2
1Department of Psychiatry
Academical Hospital
Postbox 30.001, 9700 RB Groningen
The Netherlands
2University of Groningen
Department of Clinical Psychology
Academical Hospital
Postbox 30.001, 9700 RB Groningen
The Netherlands
1. INTRODUCTION
The relationship between anxiety disorders and personality disorders (PDs) has
received a great deal of attention by researchers. This chapter describes the comor-
bidity of PDs and anxiety disorders. The data, reviewed in this chapter, were provided
by prevalence and treatment outcome studies and the review will be limited to a
descriptive level of personality pathology (based on self-report and semi-structured
interview). This review includes the following anxiety disorder!!: panic disorder, social
phobia, generalized anxiety disorder, and obsessive-compulsive disorder. Further, addi-
tional Axis I disorders and symptoms will be summarized in order to provide an overall
view on a descriptive level of the co-occurrence of Axis II disorders, anxiety disorders
and related symptomatology. Finally, hypotheses concerning associations between
anxiety disorders and PDs will be discussed and areas for future research will be out-
lined. In this chapter, the term "features" will be used when we refer to the criteria
of the PDs. The term "traits" will be used when dimensional personality traits are
considered.
130 C. J. M. Van Velzen and P. M. G. Emmelkamp
1.1. Anxiety Disorders: The Concept

Since the concept PD is defined elsewhere in this book, we will only describe the
anxiety disorder concept, emphasizing the differences with the PD concept. Whereas
the PDs have been described on the second Axis of the DSM-IV (APA, 1994), the
symptom disorders have been placed on the first Axis. The Axis I disorders refer to a
constellation of symptoms characteristic of the disorder. As for the anxiety disorders,
the most prominent characteristic is the anxiety which is described in behavioural, cog-
nitive, affective, and physiological symptoms. The symptoms are considered to be
merely temporary and to disappear when the patient recovers from the episode.
Whereas symptoms are considered to be abnormal disruptions of behaviour within a
population, pathological personality features are viewed only to be quantitatively dif-
ferent between individuals. Another related issue is the ego-syntonic character of the
PD, versus the ego-dystonic character of the Axis I disorder. In addition, anxiety dis-
orders are assumed to be less stable in time and across situations compared with PDs.
Although the differentiation of Axis I and Axis II makes theoretically sense, in prac-
tice, chronic Axis I disorders and changes in Axis II pathology have been reported in
the literature (e.g. Widiger and Shea, 1991). Furthermore, overlap between criteria of
Axis I and II are found e.g. social phobia and avoidant PD. This chapter will not address
the validity issue which is inherent to any diagnostic classification system, however, the
reader should be aware of this issue.
2. PREVALENCE RATES OF PERSONALITY DISORDERS
In this section, research on the prevalence of PDs in panic disorder with and
without agoraphobia, obsessive-compulsive disorder (OCD), social phobia and gener-
alized anxiety disorder (GAD) will be reviewed. Further, studies that compared preva-
lence rates of PDs in anxiety patients with normal subjects and depressive patients as
well as studies that compared the prevalence rates of PDs across anxiety patients will
be briefly discussed. Table 1 summarizes the results of studies on the prevalence of PD
with panic disorder, social phobia, GAD, and OCD. This review is restricted to those
studies that used standardized assessment of PD, either self-report or structured clini-
cal interviews. Hence, studies in which a clinical diagnosis of PD was provided appar-
ently not based on a structured clinical interview are not included (e.g. Koenigsberg,
Kaplan, Gilmore, and Cooper, 1985; Rasmussen and Tsuang, 1986).
2.1. Prevalence Rates of Personality Disorders in Panic Disorder, Social

Phobia, GAD, and OCD
The median prevalence rate of PDs found in panic disorder with or without ago-
raphobia was 52 (range 24-98 percent). The types of the specific PDs found to be most
prevalent was largely consistent across the studies. Either the avoidant PD, the depen-
dent PD, or both PDs occurred most frequently in 15 of the 23 studies, regardless of
the method of assessment that was used. The obsessive-compulsive and histrionic PDs
were the next most diagnosed PDs. One study (Reich and Braginsky, 1994) focused on
the prevalence of the paranoid PD in panic disorder and found that this PD had the
highest prevalence (in 54 percent of the sample) measured with the PDQ-R
ool
~
=-
Table 1. Prevalence rates of personality disorders with panic disorder, social phobia, generalized anxiety disorder, obsessive-compulsive disorder i:II:I
~
S"
Instrument N/%' AVO DEP COM PAS SZT PAR SZD HIS BOR NAR ANT NOS MPD ::.
=
~
=
Self-report Panic disorder with agoraphobia .g.
PDQ Mavissakalian and Hamann 60/27 17b 15 0 0 3 2 0 12 0 0 12 =-
(1986) i
~
PDQ Reich, Noyes, and Troughton 88/47 13 27 12 34 5 17 16 ~
(1987) =
>
PDQ Mavissakalian, and Hamann 187/40 21 19 0 9 3 0 13 2 0 20 =
~,
(1988) ~
PDQ-R Pollack, Otto, Rosenbaum and 100/42 22 7 21 8 8 19 11 22 23 12 4 19 0
1;]'
Sachs (1992)
PDQ-R Reich and Braginsky (1994) 28/- 46 36 25 54 25
=
a.
~
MCMI-I Wetzler, Kahn, Cahn, Praag, and 20173 13 27 7 33 0 0 13 20 7 7 0 .
Asnis (1990) ;'"
MCMI-I Chambless, Renneberg, 48/98 38 67 4 63 10 4 25 35 48 6 0 =-
Goldstein, and Gracely (1992) ~
.'"
MCMI-I Reich, Noyes, and Troughton 88/61 16 31 11 24 11 12 11 6 5 =
(1987) =
MCMI-II Chambless, Renneberg, 117/91 39 44 25 33 9 7 24 31 18 21 15
$
Goldstein, and Gracely (1992)
0
1;]'
Structured a.=
~
interview .
'"
SCID-II Friedman, Shear, and Frances 26/58 19 4
(1987)
SCID-II Green and Curtis (1988) 25/52 20 8 12 0 8 4 0 4 0 0 4 8 16
SCID-II-R Jansen, Arntz, Merckelbach, and 85/39 23 12 8 2 6 0 3 3 0 14
Mersch (1994)
SCID-I1 Brooks, Baltazar, McDowell, 30/53 27 10 27 10 0 20 0 7 10 10 0 27
Munjack, and Bruns (1991)
SCID-I1 Renneberg, Chambless, and 133/56 32 13 12 7 13 1 13 10 7 4 29
Gracely (1992)
(continued)
..........
....
tH
N
Table 1. (Continued)
Instrument N/% a AVO DEP COM PAS SZT PAR SZD HIS BOR NAR ANT NOS MPD
SCID-II-R Sanderson, Wetzler, Beck, and 46/30 7 4 9 2 0 0 0 2 2 2 0 7 4
Betz (1994) panic disorder
panic disorder with 99124 9 8 8 3 0 0 0 3 1 0 2 11
agoraphobia
SCID-II Hoffart, Thornes, Hedley, and 21152 48 19 10 5 0 0 0 0 10 5 0
Strand (1994) agoraphobia
without panic panic with 57/54 32 18 16 7 0 12 0 14 9 2 4
agoraphobia
SCID-II-R Dreessen, Arntz, Luttels, and
Sallaerts (1994) 31145 19 19 6 3 0 13 0 0 6 0 0
SIDP Reich, Noyes, and Troughton 88143 20 18 8 2 0 6 0 10 7 0
(1987)
SIDP Alnaes and Torgersen (1988)
panic disorder 39m 56 43 8 8 13 0 0 10 13 5 0
agoraphobia without panic 16/87 81 50 37 19 12 12 0 12 12 0 0
SIDP-R Scuito, Diaferia, Battaglia, 48/60 10 12 6 10 2 12 6 27 4 4 0 12 27
Gabriella, and Bellodi (1991)
~
SIDP-R Diaferia, Scuito, Perna, 101m 18 19 14 12 3 15 2 35 6 16 16 !-'
Bernardeschi, Rusmini, and ~
Bellodi (1993)
PDE Mauri, Sarno, Rossi, Armani, 40/50 18 18 5 15 5 5 0 2 25 7 0 30 ~
Zambotto, Cassano, and ~
Akiskal (1992) if1:1
PDE Crino and Andrews (1996) 10913 1 0 0 0 0 0 0 1 1 0 0 0
Q.
Social Phobia =
:-=
Self-report ~
PDQ Reich, Noyes, and Yates (1989) 141? 21 21 42 0 57 14 0 21 21 0 0 ~
MCMI Tran and Chambless (1995) 461? 37 24 15 11 11 0 22 2 6 2 0 t!ll
Structured ~
interview
t
-=
..,;2
SIDP Alnaes and Torgersen (1988) 101100 90 100 10 20 20 20 0 0 20 10 0
SCID-II Turner, Beidel, Borden, Stanley, 68/37 22 13 0 0 0 4 1 3
..,~
and Jacob (1991) ~
SCID-II-R Jansen, Arntz, Merckelbach, and 32/56 31 31 16 0 3 12 0 6 6 6 0 =
'"
Mersch (1994) l-
SCID-II-R Sanderson, Wetzler et aL (1994) 51161 37 18 4 2 0 2 0 2 0 4 0 8 16 =-~
SCID-II-R Mersch, Jansen, and Arntz (1995) 34/23 18 3 0 0 0 6 0 0 0 0 0
PDE Crino and Andrews (1996) 69/33 29 4 0 1 0 3 3 0 9
..,~
Generalized Anxiety Disorder
=
>
Self-report
=
..,l:!.
~
PDQ Mavissakalian, Hamann, Haidan, 39/36 26 5 3 3 10 10 0 8 0 0 0 0
~.
and De Groot (1993) =
Structured
.
..,=-
interview
.
'"
=
SIDP Alnaes and Torgersen (1988) 11172 45 18 9 0 18 9 9 18 27 0 0 =-=
SIDP Gasperini, Battaglia, Diaferia, 46/58 4 6 15 2 6 2 2 17 4 4 0 39 ~
:;!
and Bellodi (1990) =
PDE Mauri, et aL (1992) 37/41 19 22 8 3 0 8 3 0 11 3 0 25 =
e:.
SCID-II Mancuso, Townsend, and 44175 14 7 20 0 0 11 7 11 0 4 0 q
Mercante (1993) 0
~.
SCID-II-R Sanderson, Wetzler et aL (1994) 78/49 13 5 23 5 0 4 0 4 4 4 0 5 18 =

.
SCID-II-R Sanderson, Beck, and McGinn 32/50 16 13 3 0 0 3 0 3 0 0 0 13 ..,=-
:;!
(1994)
Obsessive-compulsive disorder
self-report
PDQ Black, Yates, Noyes, Pfohl, and 21133 0 24 0 0 14 0 0 24 24 0 0
Kelley (1989)
PDQ Mavissakalian, Hamann, and 43/53 30 19 2 0 16 7 0 26 5 0 30
Jones (1990a)
PDQ Mavissakalian, Hamann, and 51149 26 20 2 0 14 20 0 24 4 0 2 28
Jones (1990b)
(continued)
....
....
....
~
Table 1. (Continued)
Instrument N/%a AVO DEP COM PAS SZT PAR SZD HIS BOR NAR ANT NOS MPD
PDQ Steketee (1990) 26/50 27 38 4 8 35 11 0 31 11 0 0 42
MCMI Joffe, Swinson, and Regan (1988) 23/83 56 56 4 61 17 17 26 17 39 9 13
Structured
Interview
SCID-II-R Stanley, Turner, and Borden 25/48 12 4 28 0 8 4 4 12 0 0 0 30
(1990)
SCID-II-R Sanderson, Wetzler et al. (1994) 21/24 5 5 5 0 0 0 0 5 0 0 0 5 5
SIDP Alnaes and Torgersen (1988) 9/67 44 11 44 22 0 0 0 11 11 22 0
SIDP Baer, Jenike, Ricciardi, Holland, 96/52 5 12 6 0 5 5 1 9 4 0 0 15 6
Seymour, Minichiello, and
Buttolphi (1990)
SIDP-R Sciuto, et al. (1991) 30/69 27 13 3 13 0 20 0 23 3 3 0 10 33
SIDP-R Baer, Jenike, Black, Treece, 55/60 25 24 16 16 9 7 2 7 9 0 0 7 27
Rosenfeld, and Griest (1992) il
SIPD Black, Noyes, Pfohl, Glodstein, 32/87 22 50 28 47 19 16 0 9 19 6 0 9 !-
and Blum (1993) ~
PDE Crino and Andrews, 1996 80/9 0 7 0 0 0 0 0 0 0
~
Note N-sample size; AVO-avoidant; DEP-dependent; COM-obsessive-compulsive; PAS-passive-aggressive; SZT-schizotypal; PAR-paranoid; SZD-schizoid; HIS-histrionic; BOR- ~
borderline; NAR-narcissistic; ANT-antisocial; NOS-not otherwise specified; MPD-more than one personality disorder; [-]-not reported; PDQ (R)-Personality Diagnostic Question- if
naire (Revised) (Hyler, Skodol, Kellman et aI., 1990); MCMI-Millon Clinical Multiaxial Inventory (Millon, 1983); SCID-II (R)-Structured Clinical Interview for DSM-III (R) Personality
Disorders (Spitzer, Williams, Gibbon, and First, 1990); SIDP (R)-Structured Interview for the DSM-III (R) Personality Disorders (Stangl, Pfohl, Zimmerman et aI., 1985); PDE-Personality
=
=
~
=
Disorder Examination (Loranger, Susman, Oldham, and Russakoff, 1987).
'percentage of the sample with at least one personality disorder.
:-=
bpercentage of sample with the specific personality disorder; This figure may not add up to the combined total of the specific PD prevalence rates due to the fact that more than one diagnosis ~
can be given to one patient. ~
t'!J
aa
e~
"C
The Relationship between Anxiety Disorders and Personality Disorders 135
(Personality Diagnostic Questionnaire-Revised; Hyler, Skodol, Kellman, Oldham, and

Rosnick, 1990) followed by the avoidant and dependent PD (respectively 46 and 36
percent). This is a rather intriguing finding, because in none of the other studies the
paranoid PD was the most frequently diagnosed PD in their panic sample. Reich and
Braginsky explained the high prevalence of paranoid PD in terms of patients charac-
teristics: the patients were drawn from a tertiary outpatient care system "for patients
who had nowhere else to go" (Reich and Braginsky, 1994, p. 262).
Eight studies have reported prevalence rates of PDs in social phobia. Of these
studies, six reported the percentages of the presence of one or more PDs, ranging
from 23 percent (Mersch, Jansen, and Arntz, 1995) to 90 percent (Alnaes and
Torgersen, 1988), with a median of 46. Most PD-diagnosed patients had an avoidant
PD; the dependent PD was also frequently present, followed by the obsessive-com-
pUlsive PD and the paranoid PD. The two Dutch studies (Mersch et aI., 1995; Jansen,
Arntz, Merckelbach, and Mersch, 1994) showed considerable differences in prevalence
rates while using the same instrument i.e. the SCID-II-R (Structured Clinical Interview
for DSM-III-R Personality Disorders; Spitzer, Williams, Gibbon, and First, 1990). This
might be due to differences in referral source: advertisement in local newspaper
(Mersch et aI., 1995) versus an outpatient sample at a mental health service (Jansen
et aI., 1994), the latter group having more personality pathology. Reich, Noyes, and
Yates (1989) assessed PDs with the PDQ-R in a small group of social phobics (four-
teen patients): 57 percent of this group received a schizotypal PD diagnosis, 42 percent
a obsessive-compulsive PD diagnosis, and only 21 percent an avoidant PD diagnosis.
This unique finding might be related to the use of this self-report questionnaire; Reich,
Noyes, and Troughton (1987) also found these PDs to be most prevalent in a panic
sample. Further, two studies (Schneier, Spitzer, Gibbon, Fyer, and Liebowitz, 1991;
Herbert, Hope, and Bellack, 1992) assessed the prevalence rate of the avoidant PD in
samples of social phobic patients using the SCID-II, the rates found were respectively
70 and 61 percent.
Seven studies have been published that examined the prevalence rates of PD with
GAD. The studies reported rates varying from 36 percent (Mavissakalian, Hamann,
Haidar, and Groot, 1993) to 75 percent (Mancuso, Townsend, and Mercante, 1993), with
a median of 50 percent. In the Alnaes and Torgersen (1988), the Mauri et ai. (Mauri,
Sarno, Rossi,Armani et aI., 1992) and the Sanderson, Beck, and McGinn (1994) studies,
the most common PDs were the avoidant and the dependent PD. In the study of Mavis-
sakalian et ai. (1993) the avoidant PD was most prevalent, followed by the schizotypal
and paranoid PD. Gasperini, Battaglia, Diaferia, and Bellodi (1990) found a majority
of PD NOS (Not Otherwise Specified; DSM-IV, APA, 1994) in their sample, the PD
NOS diagnosis was the second most diagnosed PD (besides the dependent PD) in the
sample of Sanderson, Beck, and McGinn (1994). Finally, Sanderson, Wetzler, Beck, and
Betz (1994) and Mancuso et ai. (1993) found the highest rate of obsessive-compulsive
PD, followed by the avoidant PD.
The PD rates found in samples of patients with OCD varied between nine and
87 percent in the samples studied (with a median of 52 percent). The avoidant, depen-
dent, and histrionic PD were most frequently found. Further, the obsessive-compulsive
PO was also frequently found when POs were assessed with a semi-structured inter-
view, whereas borderline and schizotypal PO occurred more often when the PO diag-
nosis was made on the basis of a self-report questionnaire.
To conclude, about half of the anxiety patients receive one or more PD diagnoses.
The POs that have predominantly been assessed in these samples were the avoidant
and dependent PD. As shown in Table 1, there are large discrepancies among the preva-
lence rates of specific PDs within each anxiety disorder category. Although the findings
are fairly consistent across the studies in finding the same specific PDs to be most preva-
lent (avoidant and dependent), the number of PDs diagnosed, varies enormously. These
discrepancies may be partly due to differences in assessment method. Other factors
which may influence the prevalence rates of PD diagnoses are related to sample char-
acteristics. In most studies, the assessment of PDs was part of a comparative treatment
outcome study. Inclusion criteria may be more or less stringent or the source of refer-
ral may be biased which may lead to differences in sample characteristics. Although
not reported in every study, most studies reviewed in this chapter studied outpatients
or patients who sought treatment. Exclusion criteria were rarely mentioned in these
reports.
2.2. Comparisons with Normal SUbjects and Depressive Patients

Comparisons of prevalence rates of PDs between anxiety patients and normal
subjects or depressive patients may shed a light on the questions whether and which
specific PDs are more prevalent in anxiety patients compared to other samples. Studies
that compared prevalence rates of PDs between the anxiety disorders and normal con-
trols showed a greater risk of personality pathology in panic disorder (Reich and
Troughton, 1988; Wetzler, Kahn, Cahn, Praag, and Asnis, 1990; Noyes, Reich, Suelzer,
and Christiansen, 1991; Diaferia, Sciuto, Perna et aI., 1993), social phobia (Reich, Noyes,
and Yates, 1989), GAD (Gasperini et aI., 1990) and OCD (Black, Yates, Noyes, Pfohl,
and Kelley, 1989; Black, Noyes, Pfohl, Goldstein, and Blum, 1993). The findings tenta-
tively suggest that differences in personality pathology were not related to one or more
specific PDs, but to the whole range of PDs. In addition, PD diagnoses and features
from cluster C (avoidant, dependent, obsessive-compulsive, and passive-aggressive PD)
were most often reported to be more prevalent in anxious patients compared with
normal controls. The study of Wetzler et al. was unique in finding more narcissistic PD
in the normal control group (44 percent versus seven percent) which they related to
the method of assessment (the MCMI, Millon, 1983).
As for comparisons between anxiety disorders and depressive disorders, the
prevalence rates of PDs in social phobia has not yet been compared with depressive
disorder. Two studies have been done on respectively OCD (Joffe, Swinson, and Regan,
1988) and GAD (Mauri et aI., 1992). Both studies did not find any difference in preva-
lence rates of PD diagnoses in the anxiety sample compared to the depressive sample.
A number of studies examined the prevalence rates between panic disorder and depres-
sive disorder (Reich and Noyes, 1987; Alnaes and Torgerson, 1990; Wetzler et aI., 1990;
Hoffart and Martinsen, 1992; Mauri et aI., 1992; Mellman, Leverich, Hauser et aI., 1992;
Flick, Roy-Byrne, Cowley, Shores, and Dunner, 1993) and a number of studies have
compared mixed anxiety samples with depressive samples (Tyrer, Casey, and Gall, 1983;
Flick et aI., 1993). The differences in findings across these studies preclude any con-
clusions on associations between specific PDs and panic disorder or anxiety disorders
in general. A critical remark should be made here. A substantial number of anxiety
patients are themselves depressed or have had past episodes of depression. This
issue has been neglected in most of the studies discussed above, however, it could
explain the differences in findings. Two studies did take into account lifetime additional
diagnoses (Alnaes and Torgersen, 1990; Garyfallos, Adamopoulou, Voikli et aI., 1994)
and both studies compared the prevalence rates of PDs in anxious and depressive
patients without a lifetime additional (respectively depressive and anxious) diagnosis.

Garyfallos et ai. found no differences in percentages of PDs between the samples.
Alnaes and Torgerson found more obsessive-compulsive PDs in the major depressive
group compared with the panic group but not compared with a mixed anxiety group.
Mauri et ai. (1992) excluded patients with an additional Axis I disorder from their
depressive, panic, and GAD samples (Mauri et ai., 1992) and no differences were found
in prevalence rates of PDs between the depressive and both anxiety groups. In sum,
the findings suggest that PDs co-occur as frequently with depressive disorders as with
anxiety disorders, however, more studies are needed before any firm conclusions can
be drawn.
2.3. Comparison of Personality Disorders across Anxiety Disorders

Although a preponderance of avoidant and dependent PD was found in all
anxiety samples, comparing the prevalence rates of PDs across the anxiety disorders
might demonstrate a specific relationship between an anxiety disorder and one of these
particular PDs. Although the number of studies that compared the PDs across the
anxiety disorders are limited, some preliminary conclusions can be drawn. Comparisons
among the anxiety disorders have resulted in a few differences which were mostly related
to PD features and not to a specific PD diagnosis, indicating that the personality patterns
of patients with specific anxiety disorders are largely similar (cluster C) (Sciuto, Diaferia,
Battaglia et ai., 1991; Mauri et ai., 1992; Mavissakalian et ai., 1993; Flick et ai., 1993;
Blashfield, Noyes, Reich, Woodman, Cook, and Garvey, 1994; Sanderson, Wetzler et ai.,
1994). In addition to social phobia, as far as it has been compared with other anxiety
disorders, this anxiety disorder was found to have more personality pathology compared
to the other anxiety disorders (Alnaes and Torgerson, 1988; Reich et ai., 1989; Jansen et
ai., 1994; Sanderson, Wetzler et ai., 1994; Noyes, Woodman, Holt, Reich, and Zimmerman,
1995; Crino and Andrews, 1996). However, no specific associations can be pointed out,
since the PDs found to be more prevalent differed per study.
The comorbidity among the anxiety disorders has largely been neglected in the
above mentioned studies. Several studies indicated that most patients fulfil diagnostic
criteria for more than one anxiety disorder (e.g. Ameringen, Mancini, Styan, and
Donison, 1991; Rapee, Sanderson, and Barlow, 1988). In particular, social phobia and
simple phobia were often found to co-occur with other anxiety disorders (Brown and
Barlow, 1992). This means that comparisons of PD rates among the various anxiety dis-
orders are difficult to interpret unless the comorbidity of other anxiety disorders is
controlled for. Now that we have some idea of the comorbidity of PD with anxiety dis-
orders, the next question is: What does it mean when half of the anxiety patients receive
at least one PD diagnosis? Are there any differences between anxious patients with
and without a PD? And, if we find any differences, how should we explain them? In
the next section, an attempt will be made to answer these questions.
3. DIFFERENCES BETWEEN ANXIOUS PATIENTS WITH AND

WITHOUT A PERSONALITY DISORDER
After the separation of the PDs (Axis II) from symptom disorders (Axis I), an
increasing number of studies have examined the co-occurrence of Axis I and Axis II
138 C. J. M. Van Velzen and P. M. G. Enunelkamp
disorders. In addition, the clinical impression that anxiety patients who have a comor-
bid PD are difficult to treat could be systematically studied. For example, it might be
that the co-occurrence of a (specific) PD lead to differences in clinical features in
anxiety patients compared with anxious patients without that PD.
There is some evidence that the presence of a PD influence treatment outcome
of Axis I disorders negatively. Reich and Green (1991) reviewed the influence of PDs
on treatment outcome for various Axis I samples and found that the presence of a PD
was negatively related to treatment outcome. However, they did not take into account
co-occurring symptoms, such as depressive features, which may have also had a nega-
tive influence on treatment of the anxiety disorder as has been reported elsewhere (e.g.
Wittchen, Essau, and Krieg, 1991). Most of the studies did not properly take into
account the severity of the Axis I disorders or co-occurrence of additional Axis I dis-
orders which might also have implications for treatment outcome (Brooks, Baltazar,
and Munjack, 1989). In the next section, we will first examine whether patients with a
PD differ from patients without a PD with respect to reported symptomatology, related
or unrelated to the primary anxiety disorder. Next, we summarize the co-occurrence of
additional anxiety and depressive disorders in both groups.
3.1. Differences in Symptomatology

Table 2 provides an overview of differences between patients with and without a
PD on reported symptomatology. It should be noted that this overview is restricted to
those symptoms that were assessed in the studies. Consequently, the resulting symptom
patterns reported in most of these studies probably reflect a lower rate of concurrent
symptomatology than would have been observed using a broader range of psy-
chopathology measures.
Most studies were conducted with patients with panic disorder with or without
agoraphobia. Concerning the relationship between PDs and severity of the primary
Axis I disorder, the reported data are inconclusive. Some studies did find that
panic patients with a PD report more severe panic and agoraphobic symptoms
(Mavissakalian and Hamann, 1988; Noyes, Reich, Christiansen et aI., 1990; Noyes et aI.,
1995) but other studies found no differences in severity of panic and agoraphobic symp-
toms (Reich and Troughton, 1988; Chambless, Renneberg, Goldstein, and Gracely, 1992;
Renneberg, Chambless, and Gracely, 1992; Mellman et aI., 1992; Dreessen, Arntz,
Luttels, and Sallaerts, 1994). Hoffart, Thomes, Hedley, and Strand (1994) found that
avoidant and dependent PDs were related to symptom severity. Further, in six studies,
social phobic and/or depressive symptoms were found to be more severe in panic
patients with a PD, in one study a non-significant difference in depressive symptoms
was found (Dreessen et aI., 1994). Mellman et ai. (1992) found no differences in number
of depressive episodes. In the remaining studies, social phobic symptoms or depressive
symptoms were not measured. As far as it has been measured, the relationship between
PDs and social phobic, and depressive symptoms has consistently been shown in panic
samples. This might be explained by the fact that the avoidant PD is highly prevalent
among these patients. As will be outlined in more detail in paragraph 3.1., social phobia
and avoidant PD overlap to a great extent. The findings of Chambless et ai. (1992)
support this view: In their study, social phobic symptoms and depressive symptoms were
related to the avoidant PD. Reich and Braginsky (1994) compared a sample of panic
patients (82 percent had also agoraphobic complaints) with or without paranoid PD.
The comorbid group had an earlier onset and longer duration of illness, they received
Table 2. Comparison of symptomatology between patients with and without a personality
disorder (PD)
Study N Fmdings
Panic Disorder with or without Agoraphobia
Mavissakalian et aI. (1988) high
PD features vs low PD 188 high features group reported more overall symptomatology;
features group predictor for PD features: depressive symptoms and social anxiety
Noyes et al. (1990) 89 more severe anxiety symptoms and phobic avoidance in PD group
Reich and Throughton (1988) 204 predictor for the presence of any PD: state depression; PD was not
related to state anxiety
Chambless et al. (1992) 165 social anxiety and depressive symptoms were related to the presence
of avoidance PD; depressive symptoms were also related to
antisocial and paranoid PD. No differences in number of panic
attacks or severity of agoraphobia
Renneberg et aI. (1992) 133 more depressive and social anxiety symptoms in PD group; no
differences in severity or duration of agoraphobia or frequency of
panic attacks
Mellman et aI. (1992) 23 no differences in chronicity or duration nor in severity of
agoraphobia, frequency of panic attacks or depressive episodes
Dreessen et aI. (1994) 31 no differences, although a trend towards more trait anxiety and
depressive symptoms in the PD group
Noyes et al. (1995) 72 phobic anxiety was associated with most PDs
Hoffart et al. (1994) 78 avoidant and dependent traits were related to symptom severity;
more agoraphobic cognitions concerning sociailbehavioral control,
more trait anxiety, more depressive symptoms in PD group
Social Phobia
Throer et al. (1991) 71 more severe depressive symptomatology in all comorbid groups;
1) social phobia and additional more severe anxiety symptoms in the additional Axis I group
axis I (regardless of the presence of a PD)
2) social phobia and axis II
3) social phobia and axis I and
II
4) social phobia
Mersch et aI. (1995) 34 more severe target symptoms in PD group; no difference in overall
symptomatology, although a trend toward more overall
symptomatology in PD group
Noyes et aI. (1995) 46 phobic anxiety was associated with most PDs from cluster A and C
Obsessive Compulsive Disorder (OCD)
Mavissakalian et al. (1990a) 43 depressive symptoms: best overall predictor for PD features; PD
features were related to more severe OCD symptomatology,
dysphoric mood and more overall psychopathology
Bear et al. (1990) 96 presence of PD was not related to severity of OCD
Steketee (1990) 26 no differences on target symptoms; associations between target
symptoms and passive-aggressive PD, depressive symptoms and
paranoid PD, trait anxiety and antisocial, borderline and avoidant
PD
Bear et aI. (1992) 55 number of PDs and cluster a and b related to OCD symptomatology
number of PDs was not related to depressive symptoms
Mixed Anxiety Disorders
Dreessen et aI. (1994) 57 more overall psychopathology and more depressive symptoms in PD
group
Sanderson, Wetzler et al. (1994) 347 more depressive symptoms in comorbid group, however,looking at
the specific anxiety disorder, this only holds true for panic disorder
with agoraphobia; no differences in reported anxiety between both
groups
Mixed Anxiety and Depressive Patients
Flick et al. (1993) 352 comorbid PD was related to more depressive and anxious symptoms
more PD diagnoses, reported more overall pathology, interpersonal sensitivity, hostil-

ity, and depressive symptoms. No differences were found concerning general anxiety
or phobic anxiety.
Three studies compared the differences between social phobics with any PD and
social phobics without any PD. Some evidence was found for more severe symptoma-
tology in the comorbid group (Mersch et aI., 1995; Noyes et aI., 1995). Turner, Beidel,
Borden, Stanley, and Jacob (1991) compared social phobics with or without additional
Axis I and/or comorbid PD on severity of symptomatology (see Table 2). They found
that social phobics with comorbid Axis I and/or Axis II diagnoses reported more severe
depressive symptomatology compared with the social phobics without a comorbid diag-
nosis. Social phobics with a PD, regardless of the presence of an additional Axis I dis-
order, did not differ on anxious symptomatology. In contrast, social phobics with an
additional Axis I disorder, regardless of the presence of a PD, reported more severe
anxious symptomatology.
A number of the studies have compared social phobia and avoidant PD. This
is not surprising given the overlap in criteria between generalized social phobia
and avoidant PD: The main characteristic of both disorders is a fear of being negatively
evaluated, resulting in avoidance of social situations or feeling extremely uncomfort-
able in social situations. Further, both disorders are chronic conditions, possibly
starting at the same developmental phase in life (DSM-IV, APA, 1994). Differences in
social skills have been thought to be the distinguishing factor between social phobia
and avoidant PD (Brooks et aI., 1989). Some studies found indeed that avoidant
PD patients were less socially skilled than the social phobic patients (Turner, Beidel,
Dancu, and Keys, 1986; Marks, 1985; Greenberg and Stravynski, 1983). More recently,
however, these findings have not been replicated (Turner, Beidel, and Townsley, 1992;
Herbert et aI., 1992). This could be due to the fact that the more recent studies used
the DSM-III-R criteria in which the boundaries between social phobia (generalized
type) and avoidant PD are even less clear (Turner et aI., 1992) than under DSM-III
criteria.
Following the severity continuum hypothesis (Liebowitz, Gorman, Feyer, and
Klein, 1985; Reich and Yates, 1988), a series of studies have been conducted to examine
the hypothesis that avoidant PD and social phobia only differ in severity rather than
being conceptual different. Early results revealed that individuals with avoidant PD
reported higher levels of emotional distress and interpersonal sensitivity (Turner et aI.,
1986) and were more socially avoidant (Turner et aI., 1986; Greenberg and Stravynski,
1983) than social phobics without an avoidant PD. Heimberg, Hope, Dodge, and Becker
(1990) found that generalized social phobics were more anxious, more depressed and
performed more poorly on behavioral tests than a specific subtype, the public speak-
ing phobics. More recently, the problem has been addressed in four studies which com-
pared discrete social phobia with generalized social phobia with or without the avoidant
PD (Holt, Heimberg, and Hope, 1992; Herbert et aI., 1992; Turner et aI., 1992; Tran and
Chambless, 1995) on physiological, behavioral, cognitive, and sUbjective dimensions.
The results of these studies all failed to show a qualitative distinction between the gen-
eralized social phobia and the avoidant PD on these dimensions. The differences found
between the generalized social phobia and the avoidant PDs were related to severity
of the disorder (e.g. general distress, severity of the social anxiety or social function-
ing), and depressive symptomatology but these differences were also found between
the generalized social phobia and the discrete social phobia, the generalized social
phobics (without avoidant PD) being more generally distressed (Turner et aI., 1992).
These findings are further supported by treatment outcome studies that examined the
differential treatment outcome in (generalized) social phobia with or without an
avoidant PD (e.g. Brown, Heimberg, and Juster, 1995; Feske, Perry, Chambless,
Renneberg, and Goldstein, 1996). These studies reported that social phobic patients
with an avoidant PD had a more severe symptom pattern (before and after treatment)
than social phobics without an avoidant PD. In addition, in these studies and in the
study of Schneier et al. (1991) only a minority of discrete social phobics were diagnosed
as avoidant PD compared with generalized social phobics.
Interpreting these findings in the light of the severity continuum hypothesis, the
avoidant PD appears to be the most severe diagnosis, followed by the generalized social
phobia, the discrete social phobia being the least impaired social anxious disorder. An
argument in favour of two separate categories of avoidant PD and generalized social
phobia is the high prevalence rates of avoidant PD in the other anxiety disorders.
However, taking in account the frequent occurrence of social phobia as an additional
diagnosis, the question raises, what has been assessed? Social phobia or the avoidant
PD, thus, the issue of two distinct categories remains questionable.
In two OCD studies, no differences were found between OCD patients with or
without a PD on severity of the OCD symptoms (Baer, Jenike, Ricciardi et aI., 1990;
Baer, Jenike, Black, Treece, Rosenfeld, and Greist, 1992). However, in the second study,
clusters A and B were associated with more severe OCD symptomatology. Also, the
number of PDs per individual was positively related to severity of OCD symptoms but
not to depressive symptoms (Baer et aI., 1992). Steketee (1990) found no differences
in severity on target situations between patients with or without a PD, although
specific PD features were related to specific symptoms. In contrast, Mavissakalian,
Hamann, and Jones (1990a) found that both severity of the OCD symptoms, overall
level of psychopathology and depressive symptoms were related to the presence of a
PD. Stanley, Turner, and Borden (1990) compared OCD patients with threshold (n =
2) and subtreshold (n = 7) schizotypal PD with OCD patients with other or no PD (n
= 18). They found that the schizotypal group received more often a diagnosis of social
phobia, reported more psychotic-like experiences, and obsessions regarding self-
evaluative concerns. The level of reported general anxiety, depression, and obsessive-
compulsive symptoms did not differ between both groups which could be related to
the fact that in the non-schizotypal group, other PDs were included.
3.2. Differences in Additional Axis I Disorders

In Table 3, anxious patients with and without a PD have been compared on addi-
tional Axis I disorders. As can be seen, anxious patients with a PD tend to receive more
often additional Axis I disorders, especially depressive disorders. Most studies have
been done with panic disorder or among social phobics with or without an avoidant
PD. The results of the limited number of studies on OCD and mixed anxiety disorders
suggest a similar pattern of comorbidity.
Three studies have compared patients with a pure anxiety disorder, with patients
with a pure depressive disorder and patients with both anxiety and depressive disor-
der on the prevalence rates of PDs. In the comorbid groups more avoidant, dependent,
paranoid, and borderline PD diagnoses were found (Alnaes and Torgersen, 1990;
Hoffart and Martinsen, 1992; Voikli et aI., 1994). This finding emphasizes the impor-
tance of controlling for additional depressive disorders when comparing PD prevalence
rates among anxious samples. It is important to learn how these symptoms and features
142 C. I. M. Van Velzen and P. M. G. Emmelkamp
Table 3. Secondary Axis I diagnoses associated with personality disorder (PD)

Study N Findings
Panic Disorder with or without Agoraphobia
Reich et al. (1987) 88 major depression
Noyes et al. (1990) 89 current and past major depression
Pollack et al. (1992) 100 major depression
social phobia
generalized anxiety disorder
simple phobia
Renneberg et al. (1992) 133 social phobia
simple phobia
dysthymia
Friedman et al. (1987) 26 no differences in additional Axis I disorders
Social Phobia
Turner et al. (1991) 66 35% only social phobia
15% comorbid PD
23% additional Axis I
27% Axis I and PD; no associations reported concerning additional Axis I and
PDs
Holt et al. (1992) 33 more depressive disorders in social phobics with avoidant PD; no differences
in percentages of additional anxiety disorders;
Herbert et al. (1992) 23 more additional Axis I and PDs in social phobics with avoidant PD
Brown et al. (1995) 102 more mood disorders in social phobics with avoidant PD; more additional
anxiety disorders in generalized social phobics without an avoidant PD
(compared to generalized social phobics with an avoidant PD and discrete
social phobics)
Hofmann et al. (1995) 30 87 percent of the social phobics with an avoidant PD versus 50 percent of the
social phobics without an avoidant PD received an additional diagnosis (not
significant)
Mixed Anxiety and Depressive Patients
Flick et al. (1993) 352 lifetime Axis I diagnoses
are related in order to better understand and more effectively treat the disorders.
However, the relationship between anxiety and depression is still an issue of consider-
able debate and warrants further study. It is still unclear whether both constructs can
be meaningfully separated (Stavrakaki and Vargo, 1986) or whether a third intermedi-
ary diagnostic category may be included with a mixed anxiety-depression syndrome
(Dobson, 1985).
To conclude, the available data suggest that anxious patients with a PD report
more depressive symptoms and disorders than anxious patients without a PD. In studies
on panic disorder, additional social phobic symptoms are found to be related to the
presence of a PD. Further, the studies are inconclusive with respect to the relationship
between severity of the primary Axis I disorder symptoms and the presence of a PD,
except for social phobia and a comorbid avoidant PD. The latter group was consistently
found to have more severe social phobic and depressive symptoms. Comparisons
between specific PDs and anxiety disorders await further systematic study. A more
fruitful approach to examine this issue is to compare within a specific anxiety disorder,
patients with a specific PD with patients without any PD, in stead of lumping the
remaining PDs together with patients without a PDs in one group. Further, additional
Axis I disorders need to be taken into account.
4. COMORBIDITY MODELS
Several models have been proposed for the comorbidity of PDs and Axis I dis-
orders (Docherty, Fiester, and Shea, 1986), PDs and depressive disorders (Farmer and
Nelson-Gray, 1990), PDs and panic disorder and/or agoraphobia (Starcevic, 1992), PDs
and anxiety disorders (Stein, Hollander, and Skodol, 1993) and personality traits and
Axis I disorders (Clark, Watson, and Mineka, 1994). Both descriptive as well as causal
models have been suggested (Farmer and Nelson-Gray, 1990).
- Predisposition or vulnerability model: Certain PDs may predispose to the
development of an anxiety disorder;
- Continuity model: Certain PDs are viewed as the subclinical manifestation of
a slowly developing anxiety disorder;
- Complication model: PDs develop as a result of an enduring anxiety
disorder;
- Co-effect model: Certain co-occurring PDs and anxiety disorders are two sep-
arate psychobiological structures, but co-occur as a result of a third common
factor or causal process.
- Attenuation model: Both disorders are alternative expressions of the same
genetic or constitutional liability.
The above mentioned models are of a causal nature. The following models are of
a descriptive nature and especially relevant to keep in mind when cross-sectional
studies will be done on comorbidity of anxiety disorders and PDs.
- The pathoplasty or modification model: An interaction is assumed between
certain PDs and Axis I disorder, both occurring at the same time and this
interaction is manifested in specific symptomatology, specific course of illness
and prognosis;
- Orthogonality hypothesis: Comorbidity of PDs and anxiety disorders is the
result of chance co-occurrence;
- Overlapping symptomatology hypothesis: Comorbidity of PDs and anxiety
disorders are supposed to be the artifact of overlapping criteria of both dis-
orders;
- Heterogeneity hypothesis: This suggest in fact that in a given heterogenous
(sub)population, support for any of the models or hypotheses mentioned
above might be found, depending on the sample that is studied.
To date, the latter hypothesis seems to be the best supported (e.g. Farmer and
Nelson-Gray, 1990; Clark et al., 1994). This hypothesis stresses the importance of taking
into account the sample characteristics. Sample characteristics might be influenced by
the design of the study and the site were the samples were drawn from. Most studies
examined the PD prevalence rates in patients who sought treatment at an (specialized)
institute and these patients had to fulfil the inclusion criteria for a treatment study.
These inclusion criteria might differ among studies, e.g. exclusion of depressive disor-
ders according to the particular objectives of the study. Also, the referral source may
lead to differences in sample characteristics. Differences in sample characteristics might
also be reflected in the prevalence rates of PDs, the severity of the Axis I symptoma-
tology or additional symptomatology. Unfortunately, this hypothesis does not widen
our understanding of the relationship between anxiety disorders and PDs, however, it
does encourage us to focus on specific PDs and specific anxiety disorders.
4.1. Avoidant and Dependent Personality Disorders and

the Anxiety Disorders
Given the fact that avoidant PD is not specifically related to social phobia but
also to panic disorder, OeD, and GAD, a nonspecific link between avoidant PD and
panic disorder with or without agoraphobia, OeD, and GAD might exist. Mavis-
sakalian et al. (1993) found five features to be most common in both OeD, GAD, and
panic patients, three of these features were from the avoidant PD category: Desire for
affection, low self-esteem, and hypersensitivity to rejection. This finding indicates that
the relationship between some avoidant features and any anxiety disorder is of a more
global nature, rather than specific. It could be that these features are predisposing con-
ditions to the development of any anxiety disorder, while other factors may contribute
to the specificity of the anxiety disorder. Alternatively, these features may be the result
of a chronic anxiety disorder, that is to say that changes in personality can occur because
of the enduring anxiety disorder or the personality change can be seen as an epiphe-
nomenon of the present anxiety disorder. Some support for the latter is found in that
some avoidant personality features were no longer present after treatment (e.g. Noyes
et aI., 1991; Loranger, Lenzenweger, Gartner et aI., 1991).
Another, more interesting, hypothesis is the attenuation hypothesis: The anxiety
disorders and avoidant PD are related to a common underlying third factor, namely
social anxiety (Stein et aI., 1993). One of the most apparent features of avoidant PD is
the presence of social anxiety, which has also been associated with panic and agora-
phobia (Arrindell and Emmelkamp, 1987; Green and Curtis, 1988) and with OCD
(Steketee, Grayson, and Foa, 1987). This is consistent with the findings of Renneberg
et al. (1992), Chambless et al. (1992), and Mavissakalian and Hamann (1988) discussed
earlier. If this underlying factor accounts for the relationship between avoidant PD
and the anxiety disorders, it may also explain the frequent occurring of social phobia
as a secondary condition (additional diagnosis or at subclinical level) in the anxiety
disorders.
The severity continuum hypothesis was suggested to explain the relationship
between social phobia and avoidant Po. In terms of the above mentioned models, the
severity continuum hypothesis can be viewed as a specification of the overlap in symp-
tomatology model. This is in particular apparent for the relationship between general-
ized social phobia and avoidant PD. For the discrete social phobia, being a more
circumscribed fear (in terms of a limited number of situations that are feared or
avoided), it is less clear. When the discrete social phobia is considered to be a distinct
Axis I disorder, the attenuation hypothesis may also be applied to the relationship
between discrete social phobia and avoidant Po.
Dependent PDs were also found to co-occur frequently in anxiety disorders.
Reich (1990) found that dependent and avoidant PD overlap considerably and he sug-
gested that these PDs should be pooled in one category with three subtypes: depen-
dent, avoidant, and a mixed variant of both. This co-occurrence of avoidant and
dependent PD might also offer a plausible explanation for the comorbidity of the
dependent PD and anxiety disorders. Dependent features may be the result of coping
with (chronic) anxiety and avoidance, the main characteristics of both the avoidant PD
and the anxiety disorders.
The hypotheses suggested here should be studied extensively in large sarriples not
only with anxiety disorders but also in other Axis I disorders in order to examine spe-
cific associations between Axis I and Axis II disorders. Prospective longitudinal studies
would be especially useful in this regard.
4.2. Obsessive-Compulsive Personality Disorder and

the Anxiety Disorders
The obsessive-compulsive PD has also been frequently found in the different
anxiety samples after the avoidant and the dependent PD. The key feature of the obses-
sive-compulsive PD is a preoccupation with orderliness, perfectionism, and mental and
interpersonal control, at the expense of flexibility, openness, and efficiency (APA, DSM-
IV, 1994). It has been suggested by Brooks et ai. (Brooks, Baltazar, Mcdowell, Munjack,
and Bruns, 1991) that a perfectionistic personality style is related to panic disorder.
Panic patients may overcontrol and overreact to their physiological symptoms in
certain situations, thereby fuelling obsessive and catastrophic thoughts associated with
fear of dying or fainting (Brooks et aI., 1991). A related mechanism has been described
by Brown and Barlow (1992) as the process of anxious apprehension which denotes "a
state of persistent overarousal associated with a preparatory and hypervigilant style
concerning upcoming negative events that one may not be able to cope with or control"
which is considered to be inherent to all anxiety disorders (Brown and Barlow, p 837,
1992). This mechanism may lead to a perfectionistic personality style, which, in turn,
maintains the process of anxious apprehension. This process is the central characteris-
tic of GAD (DSM-IV, APA, 1994). In contrast to the other anxiety disorders, avoid-
ance of situations does not have to be present for the diagnosis of GAD. Table 1 showed
that in two of the seven studies, the obsessive-compulsive PD was the most frequently
assessed specific PD in GAD (Mancuso et aI., 1993; Sanderson, Wetzler et aI., 1994)
and in one study, it was the second most frequently diagnosed PD, after the histrionic
PD (Gasperini et aI., 1990). The obsessive-compulsive features may protect against
developing avoidance behaviour, but on the other hand, may predispose the individual
to worrying. Beck and Freeman (1990) and Brooks et ai. (1991) suggested that obses-
sive-compulsive personality features create a vulnerability to anxiety states, especially
obsessive thinking related to a specific fear, which subsequently causes bodily symp-
toms and more obsessional thoughts about these symptoms. Alternatively, the perfec-
tionistic personality style does not have to be the vulnerability factor itself but may
develop as a way of coping with the process of anxious apprehension. Another way of
coping with this anxiety is avoidance of the (anticipated) situations, leading to avoidant
PD features. Both coping styles, avoidance and perfectionism are not mutually exclu-
sive but may be related to the patients' estimation of controllability in a given life
period or situation. In social phobia, reactions to physiological symptoms such as blush-
ing, trembling or sweating, and catastrophical thoughts about these symptoms lead to
the fear of being embarrassed in front of others (Scholing and Emmelkamp, 1993).
These patients are preoccupied with avoiding or trying to control these symptoms. This
preoccupation leads to more anxiety. Some support for this perfectionistic style in
public speaking social phobics has been found (Heimberg et aI., 1990).
No differences in obsessive-compulsive features were found in normal controls
or depressive patients compared with GAD, panic disorder, and OeD. It could be that
some obsessive-compulsive features co-occur with all anxiety disorders, depressive
disorders, and also in non-pathological samples. In the latter samples, obsessive-
compulsive features do appear frequently in a more adaptive way (Pollak, 1987), and
it could be that during anxiety states, these "normal" obsessive-compulsive features
become more apparent as a way of keeping control over oneself and the anxiety. There-
fore, obsessive-compulsive features as defined by DSM-IV (maladaptive, causing social,
and occupational impairment) may be the result of the anxiety disorder, or, these fea-
tures may become more pronounced when an anxiety disorder is present and/or certain
stressful life-events occur. A second hypothesis was already mentioned earlier; namely
that obsessive-compulsive features might be in particularly prominent in anxious
patients who do not avoid many situations. Friedman, Shear, and Frances (1987) sug-
gested that obsessive-compulsive features may protect against the development of
agoraphobia whereas avoidant PD predisposes an individual to it. In line with this argu-
ment, more obsessive-compulsive features should be found in the discrete social
phobics compared with the generalized social phobics. Obsessive-compulsive features
in anxiety patients may force the patient to attend to his responsibilities concerning
work or housekeeping. Consequently, avoidance of situations remains limited although
the patient still may fear a panic attack or may worry about possible dangers that might
occur.
To conclude, not a single model can account for the relationship between the
anxiety disorders and PDs. Several relationships have been suggested, depending on
the specific anxiety disorder and comorbid PD. Although other hypotheses may be
relevant at the same time (e.g. the pathoplasty hypothesis), the most interesting and
compelling hypothesis is the attenuation hypothesis. This hypothesis suggests that
anxiety disorders and avoidant and obsessive-compulsive PDs are associated through
underlying genetic or constitutional constructs. In case of anxiety disorders and
avoidant PD, social anxiety and perfectionism were pointed out as the underlying per-
sonality constructs. In the following section, we shall elaborate on the attenuation
hypothesis in an attempt to integrate PDs and anxiety disorders in a hierarchical model
of personality.
4.3. Relationship between Anxiety Disorders, Personality Disorders, and

Personality Traits
A promising dimensional model recently applied in the research on PDs, is the
Five-Factor Model of personality (Costa and McCrae, 1992). The five dimensions are:
Extraversion, Agreeableness, Conscientiousness, Neuroticism, and Openness to Expe-
rience. These five dimensions have been extensively studied in normal populations and
empirical evidence supports the usefulness of this model as a hierarchical organization
of normal personality traits (McCrae and John, 1992). A number of studies (Costa and
McCrae, 1990; Soldz, Budman, Demby, and Merry, 1993; Trull, 1992; Wiggins and Pincus,
1989; Yeung, Lyons, Waternaux, Faraone, and Tsuang, 1993) showed a clear relationship
between the five dimensions and the PDs, although these dimensions did not fully
explain the variance in the PDs. Neuroticism, Extraversion, and Conscientiousness
explained a large portion of the variance in the studies, Agreeableness and Openness
appeared to be less clearly related to the PDs. These five dimensions are further defined
by specific traits, the so-called facets (Costa and Widiger, 1994). The extent in which
(the facets of) the five dimensions were able to differentiate between the PDs is encour-
aging and warrants further research.
Factor analytic studies on data from both clinical and nonclinical samples have
found evidence for a general distress dimension which has been put forward as a poten-
tial significant vulnerability factor for anxious and depressive disorders. This factor has
been identified by Clark and Watson (1991) as the temperamental core of Negative
Affect or Neuroticism. They suggested a tripartite model of anxiety and depression:
General distress is shared by both disorders and positive affectivity is the tempera-
mental core of Extraversion and is (negatively) related to depressive disorders. Auto-
nomic hyperarousal (d. process of anxious apprehension, Brown and Barlow, 1992) is
related to the anxiety disorders although the relationship with a broad personality
dimension is less clear: It is related to Neuroticism but does not fully represent this per-
sonality trait (Clark et aI., 1994).
The attenuation hypothesis with social anxiety as the common underlying factor
in anxiety disorders and the avoidant PD is consistent with a hierarchical model of per-
sonality. Social anxiety can be viewed in terms of the Five-Factor Model as a combi-
nation of low Extraversion (i.e. high Introversion) and high Neuroticism (i.e. low
Emotional Stability). Widiger, Trull, Clarkin, Sanderson, and Costa (1994) stated that
extreme scores on the five factors place individuals at risk for certain PDs. In addition,
they provided a Five-Factor translation of the PDs based upon the DSM-III-R descrip-
tions and the available literature on the respective PDs. Most of the facets of Extra-
version and Emotional Stability dimensions were negatively related to the avoidant PD
(Widiger et aI., 1994). The obsessive-compulsive features or perfectionistic personality
style may also be related to the underlying personality factors found in nonclinical pop-
ulations, namely high Conscientiousness in combination with low Emotional Stability.
Widiger et ai. (1994) also placed obsessive-compulsive PD low on some facets of agree-
ableness and openness. .
Another argument for the use of a dimensional model is provided by Reich and
Vasile (1993) and Shea, Klein and Widiger (1992), who reviewed the impact of PDs on
the treatment outcome of Axis I disorders. These authors emphasized the role of per-
sonality traits in stead of the categorical PDs in predicting outcome. These personality
traits, in interaction with specific environmental factors, may play an important role in
the etiology of both Axis I and Axis II disorders. Both the personality traits as well as
the PDs may each account for some of the variation in treatment-outcome or relapse
of Axis I disorders. It may be hypothesized that the more extreme the underlying per-
sonality trait, and in particular in a context of certain negative parental rearing styles
or life-events in childhood, the greater the probability of developing a PD. The PDs
may be viewed as inflexible ways of coping with a given constellation of extreme per-
sonality traits in the context of certain environmental factors. In addition, these PDs
can be conceptualized as maladaptive, deeply ingrained patterns of thinking, feeling,
and behaving characteristic of the individuals functioning since early adulthood. In con-
trast, the symptom disorders can be considered as episodic changes in thinking, feeling,
and behaving when certain life events or circumstances can not be effectively coped
with at any stage in life. Moreover, because the coping strategies of PDs are, per defi-
nition, maladaptive, more negative life experiences will be encountered and, conse-
quently, leading to more symptomatology and (additional) Axis I disorders. In this
context, patients without PDs may present less symptomatology because of an overall
more adaptive and flexible way of coping with ongoing life-events. In addition, it can
be argued that patients without a PD will also have a better prognosis after treatment.
Alternatively, the complication hypothesis may also explain the comorbidity of PDs in
some anxious patients: As a consequence of a chronic (untreated) Axis I disorder
and/or negative long-lasting life-events after adolescence, the personality will change
during adulthood, in order to cope with the Axis I disorder or chronic circumstances
(panic disorder-dependent PD). The Five-Factor model seems to be a promising

model to explain the associations between some PDs from the anxious cluster and the
anxiety disorders on a phenomenological level.
5. CONCLUDING REMARKS
Research on the comorbidity of PD and anxiety disorders suggest that a common

personality pattern, with predominantly cluster C PDs, is characteristic for all anxiety
disorders. Further, anxious patients with a PD appear to have more psychopathologi-
cal symptomatology and additional Axis I disorders compared to patients without a
PD. Several models have been suggested for the relationships between avoidant PD,
dependent PD and obsessive-compulsive PD and the anxiety disorders. The attenua-
tion hypothesis, in particular, might explain the comorbidity between avoidant PD and
the anxiety disorders and obsessive-compulsive PD and the anxiety disorders: Both the
PD and the anxiety disorder are related to higher order personality traits. These traits
were found in The Five-Factor Model of Costa and McCrae (1992) in terms of inter-
actions of Neuroticism and Extraversion or Conscientiousness. The Five-Factor Model
does not include any biogenetic factors that might be related to the development of
PDs, anxiety disorders and personality traits.
Biological mechanisms have been associated with the tripartite model of person-
ality, mood, and anxiety disorders (Clark et al. 1994). Further, the psychobiological per-
sonality models of Cloninger (1987) and Siever and Davis (1991) have related PDs and
Axis I disorders to their dimensional personality model. These models incorporate bio-
logical correlates, social learning, and environmental factors. This line of research can
contribute to the understanding of the relationship between personality traits, PDs, and
anxiety disorders on a more fundamental level.
Several areas for future research have been indicated throughout this chapter,
some of these issues will repeated here. Two major research lines can be pointed out.
First, the causal models outlined earlier in this and the previous section should be
studied, especially important are prospective longitudinal studies were the causal
hypotheses can be examined properly. The hypotheses mentioned in this chapter
covered the co morbidity of avoidant, dependent and obsessive-compulsive PDs and
the anxiety disorders. However, the relationship of anxiety disorders with the other
PDs, although less frequently co-occurring in anxiety samples, should also be further
explored. Moreover, where should the depressive disorders be placed in the model? As
it appeared that depressive disorders frequently co-occur in patients with both an
anxiety disorder and a PD, this also deserves further attention. Clark and Watson's tri-
partite model (1991) suggest that the dimension of Neuroticism (Negative Affectivity)
is related to both anxiety and depressive disorders. Further, low on Positive Affectiv-
ity (Extraversion) is especially related to depressive disorders. The second line of
research concerns the comorbidity of Axis I and Axis II in relation to course of illness,
specific symptomatology, treatment outcome, and prognosis. Cross-sectional prevalence
studies are needed comparing different samples (inpatients and outpatients, normal
controls and specific Axis I disorders, different anxiety disorders, and between anxiety
disorders and other Axis I disorders) to study the pathoplasty hypothesis. Treatment
outcome studies should include long-term evaluation in order to examine the influence
of PDs, not only during treatment, but also on the longer term. Patients with specific
PDs should not be lumped together and then compared with patients without a PD;
specific PDs should be compared with other specific PDs and with patients without a
PD. Within such a design, the specific course and (additional) symptomatology of the
anxiety disorder related to the specific PD can be examined more clearly. For example,
the co-occurrence of certain borderline features may have a different impact on course
or outcome of the anxiety disorder then certain avoidant features. Besides the PD diag-
noses, where the cut-off point determines whether the patient has the PD or not, also
personality features from the different categories should be included. Then this arbi-
trary decision is not in question and more information can be used. Moreover, per-
sonality traits need to be included in these research programs.
Three confounding factors need to be pointed out. These are the influence of
mood on diagnosing a PD, additional Axis I disorders and multiple Axis II disorders.
The assessment of PD may be influenced by the Axis I disorder (Hirschfeld, Klerman,
Clayton et aI., 1983; Loranger et aI., 1991; Reich, Noyes, Coryell, and O'Gorman, 1986).
Most studies assessed the PD before treatment, when the patient was in an anxious
and/or depressed mood. Further studies are needed to determine the effect of the mood
of the patient on the diagnosis of PD. Studies investigating the PDs when the same
patient is in an anxious or depressed mood compared with a non-anxious or neutral
mood are required. Further, especially important will be to control for additional Axis
I disorders, since these disorders may also influence treatment outcome or course of
illness. Third, separate analyses of specific PDs should be conducted whenever possi-
ble. Consequently, some PDs are more likely to be studied separately (e.g. avoidant,
dependent, and compulsive) then others (schizoid or anti-social) because of the number
of patients with these PDs in anxious populations. However, samples may also consist
solely of specific PDs, like anti-social PDs in forensic psychiatric centres, and in these
populations, the prevalence of Axis I disorders can be studied.
Finally, some basic methodological problems arise from these designs which
are inherent to the study of categorical entities: What kind of criteria do we employ for
the differentiation of the primary from the secondary or additional Axis I disorder?
How to handle patients who fulfil the criteria for more than one PD diagnosis? Theory-
driven guidelines should be provided for how to deal with these issues. In addition,
researchers should be aware of the impact the chosen guidelines will have on the
findings.
REFERENCES
Alnaes, R and Torgersen, S. (1988). The relationship between DSM-III symptom disorders (axis 1) and
personality disorders (axis II) in an outpatient popUlation. Acta Psychiatrica Scandinavia, 78, 485-492.
Alnaes, R and Torgersen, S. (1990). DSM-III personality disorders among patients with major depression,
anxiety disorders, and mixed conditions. The Journal of Nervous and Mental Disease, 178, 693-698.
Ame,rican Psychological Association. (1994). Diagnostic and Statistical Manual of Mental Disorders (fourth
ed.). Washington, DC: Author.
Ameringen, M., Van, Mancini, c., Styan, G., and Donison, D. (1991). Relationship of social phobia with other
psychiatric illness. Journal of Affective Disorders, 21, 93-99.
Arrindell, WA and Emmelkamp, P.M.G. (1987). Psychological states and traits in female agoraphobics: a
controlled study. Journal of Psychopathology and Behavioral Assessment, 9, 237-253.
Baer, L., Jenike, M.A, Ricciardi II, 1.N., Holland, AD., Seymour, R.1., Minichiello, WE., and Buttolphi, M.L.
(1990). Standardized assessment of personality disorders in obsessive-compulsive disorder. Archives
of General Psychiatry, 47, 826--830.
Baer, L., Jenike, M.A, Black, D.W., Treece, c., Rosenfeld, R, and Greist, 1. (1992). Effect of Axis II
diagnoses on treatment outcome with clomipramine in 55 patients with obsessive-compulsive
disorder. Archives of General Psychiatry, 49, 862-866.
Beck, A.T. and Freeman, A. (1990). Cognitive Therapy of Personality Disorders. New York: Guilford Press.
Black, D.W, Yates, WR, Noyes, R, Pfohl, B., and Kelley, M. (1989). DSM-III personality disorders in
obsessive-compulsive study volunteers: a controlled study. Journal of Personality Disorders, 3, 58-62.
Black, D.W, Noyes, R, Pfohl, B., Goldstein, RB., and Blum, N. (1993). Personality Disorder in Obsessive-
compulsive volunteers, well comparison subjects, and their first-degree relatives. American Journal of
Psychiatry, 150, 1226-1232.
Blashfield, R, Noyes, R, Reich, 1, Woodman, c., Cook, B.L., and Garvey, MJ. (1994). Personality disorder
traits in generalized anxiety and panic disorder patients. Comprehensive Psychiatry, 35, 329-334.
Brooks, RB., Baltazar, P.L., and Munjack, D.I (1989). Co-occurrence of personality disorders with panic
disorder, social phobia, and generalized anxiety disorder: a review of the literature. Journal ofAnxiety
Brooks, RB., Baltazar, P.L., Mcdowell, D.E., Munjack, D.I, and Bruns, IR (1991). Personality disorders
co-occurring with panic disorder with agoraphobia. Journal of Personality Disorders, 5, 328-336.
Brown, T.A. and Barlow, D.H. (1992). Comorbidity among anxiety disorders: implications for treatment and
DSM-IY. Journal of Consulting and Clinical Psychology, 60, 835-844.
Brown, E.I, Heimberg, RG., and Juster, H.R (1995). Social phobia subtype and avoidant personality
disorder: Effect on severity of social phobia, impairment, and outcome of cognitive behavioral
treatment. Behaviour Therapy, 26, 467-486.
Chambless, D.L., Renneberg, B., Goldstein, A., and Gracely, EJ. (1992). MCMI-diagnosed personality
disorders among agoraphobic outpatients: Prevalence and relationship to severity and treatment
outcome. Journal of Anxiety Disorders, 6, 193-211.
Clark, L.A. and Watson, D. (1991). Tripartite model of anxiety and depression: psychometric evidence and
taxonomic implications. Journal of Abnormal Psychology, 100,316--336.
Clark, L.A., Watson, D., and Mineka, S. (1994). Temperament, personality, and the mood and anxiety
disorders. Journal of Abnormal Psychology, 103, 103-116.
Cloninger, C.R (1987). A systematic method for clinical description and classification of personality variants.
Costa, P.T. and McCrae, RR (1990). Personality disorders and the Five-Factor Model of personality. Journal
of Personality Disorders, 4, 362-371.
Costa, P.T. and McCrae, RR (1992). The Five-Factor Model of personality and its relevance to personality
disorders. Journal of Personality Disorders, 6, 343-359.
Costa, P.T. and Widiger, T.A. (1994) Summaries and unresolved issues. In P.T. Costa and T.A. Widiger (Eds.),
Personality disorders and The Five-Factor Model of Personality. Washington: American Psychological
Association.
Crino, RD. and Andrews, G. (1996). Personality disorder in obsessive-compulsive disorder: a controlled study.
Psychiatric Research, 30, 29-38.
Diaferia, G., Sciuto, G., Perna, G., Bernardeschi, L., Battaglia, M., Rusmini, S., and Bellodi, L. (1993)
DSM-III-R personality disorders in panic disorder. Journal of Anxiety Disorders, 7,153-161.
Dobson, K.S. (1985). The relationship between anxiety and depression. Clinical Psychology Review, 5,
307-324.
Docherty, J.P., Fiester, S.l, and Shea, T. (1986). Syndrome diagnosis and personality disorder. In A.I Frances
and RE. Hales (Eds.), American Psychiatric Association Annual Review, Vol. 5, 315-355. Washington,
DC: American Psychiatric Press.
Dreessen, L., Arntz, A., Luttels, c., and Sallaerts, S. (1994). Personality disorders do not influence the results
of cognitive behaviour therapies for anxiety disorders. Comprehensive Psychiatry, 35, 265-274.
Farmer, R and Nelson-Gray, R.O. (1990). Personality and depression: hypothetical relations, empirical
findings, and methodological considerations. Clinical Psychology Review, 10,453-476.
Feske, u., Perry, K.I, Chambless, D.L., Renneberg, B., and Goldstein, A.I (1996). Avoidant personality
disorder as a predictor for treatment outcome among generalized social phobics. Journal of
Personality Disorders, 10, 174-184.
Flick, S.N., Roy-Byrne, P.P., Cowley, D.S., Shores, M.M., and Dunner, D.L. (1993). DSM-III-R personality
disorders in a mood and anxiety disorders clinic: prevalence, comorbidity, and clinical correlates.
Journal of Affective Disorders, 27, 71-79.
Friedman, C.l, Shear, M.K., and Frances, A. (1987). DSM-III personality disorders in panic patients. Journal
Garyfallos, G., Adamopoulou, A., Voikli, M., Saitis, M., Kirtsos, P., and Moutzoukis, C. (1994). DSM-III-R
personality disorders among patients with depressive and/or anxiety disorders. Journal of Personality
Gasperini, M., Battaglia, M., Diaferia, G., and Bellodi, L. (1990). Personality features related to generalized
anxiety disorder. Comprehensive Psychiatry, 31, 363-368.
Green, M.A and Curtis, G.c. (1988). Personality disorders in panic patients: Response to termination of
antipanic medication. Journal of Personality Disorders, 2, 303-314.
Greenberg, D. and Stravynski, A (1983). Social phobia. [Letter to the editor]. British Journal of Psychiatry,
143,526.
Heimberg, RG., Hope, D.A, Dodge, C.S., and Becker, RE. (1990). DSM-III-R subtypes of social phobia.
Comparison of generalized social phobics and public speaking phobics. The Journal of Mental and
Nervous Disease, 178, 172-179.
Herbert, J.D., Hope, D.A, and Bellack, AS. (1992). Validity of the distinction between generalized social
phobia and avoidant personality disorder. Journal of Abnormal Psychology, 101,332-339.
Hirschfeld, RM.A., Klerman, G.L., Clayton, P.J., Keller, M.B., McDonald-Scott, P., and Larkin, B.H. (1983).
Assessing personality: effects of the depressive state on trait measurement. American Journal of
Hoffart, A and Martinsen, E.W. (1992). Personality disorders in panic with agoraphobia and major depres-
sion. British Journal of Clinical Psychology, 31, 213-214.
Hoffart, A, Thomes, K., Hedley, L.M., and Strand, J. (1994). DSM-III-R Axis I and II disorders
in agoraphobic patients with and without panic disorder. Acta Psychiatrica Scandinavia, 89,
186-19l.
Hofmann, S.o., Newman, M.G., Becker, E., Taylor, C.B., and Roth, w.T. (1995). Social phobia with and without
avoidant personality disorder: preliminary behaviour therapy outcome findings. Journal of Anxiety
Holt, C.S., Heimberg, Ro., and Hope, D.A (1992). Avoidant personality disorder and the generalized subtype
of social phobia. Journal of Abnormal Psychology, 1, 318-325.
Hyler, S.E., Skodol, AE., Kellman, H.D., Oldham, J.M., and Rosnick, L. (1990). Validity of the Personality
Diagnostic Questionnaire-revised: Comparison with two structured interviews. American Journal of
Psychiatry, 147, 1043-1048.
Jansen,M.A,Arntz,A, Merckelbach, H., and Mersch, P.A (1994). Personality disorders and features in social
phobia and panic disorder. Journal of Abnormal Psychology, 103, 391-395.
Joffe, RT., Swinson, RP., and Regan, J.J. (1988). Personality features of obsessive-compulsive disorder.
American Journal of Psychiatry, 145, 1127-1129.
Koenigsberg, H.W., Kaplan, RD., Gilmore, M.M., and Cooper, AM. (1985). The relationship between
syndrome and personality disorders in DSM-III: Experience with 2462 patients. American Journal of
Psychiatry, 142,207-212.
Liebowitz, M.R, Gorman, J.M., Feyer, AJ., and Klein, D.E (1985). Social phobia: Review of a neglected
anxiety disorder. Archives of General Psychiatry, 42, 729-736.
Loranger, A.W., Lenzenweger, M.E, Gartner, A.E, Susman, Y.L., Herzig, J., Zammit, O.K., Gartner, J.D.,
Abrams, RC., and Young, RC. (1991). Trait-state artifacts and the diagnosis of personality disorders.
Loranger, AW., Susman, UL., Oldham, J.M., and Russakoff, L.M. (1987). The personality disorders
Examination: A preliminary report. Journal of Personality Disorders, 1, 1-13, Mancuso, D.M.,
Townsend, M.H., and Mercante, D.E. (1993). Long-term follow-up of generalized anxiety disorder.
Comprehensive Psychiatry, 34, 441-446.
Marks, I.M. (1985). Behavioral treatment of social phobia. Psychopharmacology Bulletin, 21, 615-618.
Mauri, M., Sarno, N., Rossi, Y.M., Armani, A, Zambotto, S., Cassano, G.B., and Akiskal, H.S. (1992).
Personality disorders associated with generalized anxiety, panic, and occurrent depressive disorders.
Mavissakalian, M. and Hamann, M.S. (1986). DSM-III Personality disorder in agoraphobia. Comprensive
Mavissakalian, M., Hamann, M.S., and Jones, B. (1990a). Correlates of DSM-III personality disorder in
obsessive-compulsive disorder. Comprehensive Psychiatry, 31, 481-489.
Mavissakalian, M., Hamann, M.S., and Jones, B. (1990b). A comparison of DSM-III personality disorders in
panic/agoraphobia and obsessive-compulsive disorder. Comprehensive Psychiatry, 31, 238-244.
Mavissakalian, M. and Hamann, M.S. (1988). Correlates of DSM-III personality disorder in panic disorder
and agoraphobia. Comprehensive Psychiatry, 29, 535-544.
Mavissakalian, M.R, Hamann, M.S., Haidar, S.A., and De Groot, C. (1993). DSM-III personality disorders
in generalized anxiety, panic/agoraphobia, and obsessive-compulsive disorders. Comprehensive
McCrae, RR and John, O.P. (1992). An introduction to the Five-Factor Model and its applications. Journal
of Personality, 60, 175-215.
Mellman, T.A., Leverich, G.S., Hauser, P., Kramlinger, K., Post, RM., and Uhde, T.W (1992). Axis II
pathology in panic and affective disorders: relationship to diagnosis, course of illness, and treatment
response. Journal of Personality Disorders, 6, 53-63.
Mersch, P.A., Jansen, M.A., and Arntz, A. (1995). Social phobia and personality disorder: Severity of
complaint and treatment effectiveness. Journal of Personality Disorders, 9; 143-159.
Millon, T. (1983). Millon Clinical Multiaxial Inventory Manual (3rd ed.). Minneapolis, MN: National
Computer Systems.
Noyes, R, Reich, J., Christiansen, J., Suelzer, M., Pfohl, B., and Coryell, W.A. (1990). Outcome of panic
disorder. Relationship to diagnostic subtypes and comorbidity. Archives of General Psychiatry, 47,
809-818.
Noyes, R, Reich, IH., Suelzer, M., and Christiansen, 1 (1991). Personality traits associated with panic
disorder: Change associated with treatment. Comprehensive Psychiatry, 32, 283-294.
Noyes, R, Woodman, e.L., Holt, e.S., Reich, IH., and Zimmerman, M.B. (1995). Avoidant personality traits
distinguish social phobic and panic disorder subjects. The Journal of Nervous and Mental Disease, 183,
145-153.
Pollack, M.H., Otto, M.W., Rosenbaum, J.F, and Sachs, G.S. (1992). Personality disorders in patients
with panic disorder: association with childhood anxiety disorders, early trauma, comorbidity, and
chronicity. Comprehensive Psychiatry, 33, 78-83.
Pollak,I (1987). Relationship of obsessive-compulsive personality to obsessive-compulsive disorder: a review
of literature. Journal of Psychology, 121, 137-148.
Rapee, RM., Sanderson, We., and Barlow, D.H. (1988). Social phobia features across the DSM-III-R anxiety
disorders. Journal of Psychopathology and Behavioral Assessment, 10,287-299.
Rasmussen, S.A. and Tsuang, H.T. (1986). Clinical characteristics and family history in DSM-III obsessive-
compulsive disorder. American Journal of Psychiatry, 143,317-322.
Reich, 1H. (1989). Update on instruments to measure DSM-III and DSM-III-R personality disorders. The
Journal of Nervous and Mental Disease, 177,366-370.
Reich,1 (1990). Relationship between DSM-III avoidant and dependent personality disorders. Psychiatry
Research, 34, 281-292.
Reich, 1H. and Braginsky, Y. (1994). Paranoid personality traits in a panic disorder population: a pilot study.
Comprehensive Psychiatry, 35, 260--264.
Reich, 1H. and Green, A.I. (1991). Effect of personality disorders on outcome of treatment. The Journal of
Nervous and Mental Disease, 179,74-82.
Reich, IH. and Noyes, R (1987). A comparison of DSM-III personality disorders in acutely ill patients and
depressed patients. Journal of Anxiety Disorders, 1, 123-131.
Reich, 1 and Troughton, E. (1988). Frequency of DSM-III personality disorders in patients with panic
disorder: comparison with psychiatric and normal control subjects. Psychiatry Research, 26, 89-100.
Reich, IH. and Vasile, RG. (1993). Effect of personality disorders on the treatment outcome of Axis I
conditions: an update. The Journal of Nervous and Mental Disease, 181,475-484.
Reich, 1, Noyes, R, Coryell, W., and O'Gorman, T.W (1986). The effect of state anxiety on personality
measurement. American Journal of Psychiatry, 143,760-763.
Reich, 1, Noyes, R., and Troughton, E. (1987). Dependent personality disorder associated with phobic
avoidance in patients with panic disorder. American Journal of Psychiatry, 144,323-326.
Reich, 1, Noyes, R, and Yates, W. (1989). Alprazolam treatment of avoidant personality traits in social phobia
patients. Journal of Clinical Psychiatry, 50, 91-95.
Reich,1 and Yates, W. (1988). A pilot study of treatment of social phobia with alprazolam. American Journal
of Psychiatry, 145,590--594.
Renneberg, B., Chambless, D.L., and Gracely, E.I (1992). Prevalence of SCID-diagnosed personality
disorders in agoraphobic outpatients. Journal of Anxiety Disorders, 6,111-118.
Sanderson, We., Beck, A.T., and McGinn, L.K. (1994). Cognitive therapy for generalized anxiety disorder:
Significance of comorbid personality disorders. Journal of Cognitive Psychotherapy, 8, 13-18.
Sanderson, W.e., Wetzler, S., Beck, A.T., and Betz, F (1994). Prevalence of personality disorders among
patients with anxiety disorders. Psychiatry Research, 51,167-174.
Schneier, FR, Spitzer, RL., Gibbon, M., Fyer, A.I, and Liebowitz, M.R (1991). The relationship of social
phobia subtypes and avoidant personality disorder. Comprehensive Psychiatry, 32, 496-502.
Scholing, A. and Emmelkamp, P.M.G. (1993). Cognitive and behavioral treatments of fear of blushing,
sweating or trembling. Behaviour Research and Therapy, 31, 155-170.
Sciuto, G., Diaferia, G., Battaglia, M., Perna, G., Gabriele, A, and Bellodi, L. (1991). DSM-III-R personality
disorders in panic and obsessive-compulsive disorder: A comparison study. Comprehensive Psychia-
try, 32, 450-457.
Shea, M.T., Klein, M.H., and Widiger, T.A (1992). Comorbidity of personality disorders and depression:
Implications for treatment. Journal of Consulting and Clinical Psychology, 60, 857-868.
Siever, L.I and Davis, K.L. (1991). A psychobiological perspective on the personality disorders. American
Journal of Psychiatry, 148, 1647-1658.
Soldz, S., Budman, S., Demby, A, and Merry, I (1993). Representation of personality disorders in circumplex
and Five-Factor space: Explorations with a clinical sample. Psychological Assessment, 5, 41-52.
Spitzer, RL., Williams, IB.W, Gibbon, M., and First, M.B. (1990). Structured Clinical Interview for DSM-III-
R-personality disorders (SCID-II, Version 1.0), Washington, DC: American Psychiatric Press.
Stangl, D., Pfohl, B., Zimmerman, M., Bowers, W, and Corenthal, C (1985). A Structured Interview for the
DSM-III Personality Disorders. Archives of General Psychiatry, 42, 591-596.
Stanley, M.A, Turner, S.M., and Borden, IW (1990). Schizotypal features in obsessive-compulsive disorder.
Comprehensive Psychiatry, 31, 511-518.
Starcevic, V (1992). Comorbidity models of panic disorder/agoraphobia and personality disturbance. Journal
Stavrakaki, C and Vargo, B. (1986). The relationship of anxiety and depression: a review of the literature.
British Journal of Psychiatry, 149,7-16.
Stein, D.l., Hollander, E., and Skodol, AE. (1993). Anxiety disorders and personality disorders: a review.
Journal of Personality Disorders, 7, 87-104.
Steketee, G. (1990). Personality traits and disorders in obsessive-compulsives. Journal of Anxiety Disorders,
4,351-364.
Steketee, G., Grayson, IB., and Foa, E.B. (1987). A comparison of characteristics of obsessive-compulsive
disorder and other anxiety disorders. Journal of Anxiety Disorders, 1, 325-335.
Tran, G.Q. and Chambless, D.L. (1995). Psychopathol~gy of social phobia: effects of subtype and of avoidant
personality disorder. Journal of Anxiety Disorders, 9, 489-501.
Trull, T.I (1992). DSM-III-R personality disorders and the Five-Factor Model of personality: an empirical
comparison. Journal of Abnormal Psychology, 101,553-560.
Turner, S.M., Beidel, D.C, Borden, IW, Stanley, M.A., and Jacob, RG. (1991). Social Phobia: Axis I and II
correlates. Journal of Abnormal Psychology, 100, 102-106.
Turner, S.M., Beidel, D.C, Daneu, C.v., and Keys, D.l. (1986). Psychopathology of social phobia and
comparison to avoidant personality disorder. Journal of Abnormal Psychology, 95, 389-394.
Turner, S.M., Beidel, D.C., and Townsley, RM. (1992). Social Phobia: a comparison of specific and
generalized subtypes and avoidant personality disorder. Journal of Abnormal Psychology, 101,
326-331.
Tyrer, P., Casey, P., and Gall, l. (1983). Relationship between neurosis and personality disorder. British Journal
of Psychiatry, 142,404-408.
Wetzler, S.W, Kahn, RS., Cahn, W, Praag, H.M., and Asnis, G.M. (1990). Psychological test characteristics
of depressed and panic patients. Psychiatry Research, 31, 179-192.
Widiger, T.A. and Shea, T. (1991). Differentation of Axis I and Axis II disorders. Journal of Abnormal
Psychology, 100, 399-406.
Widiger, T.A, Trull, T.I, Clarkin, IF., Sanderson, C, and Costa, P.T. (1994). A description of the DSM-III-R
and DSM-IV personality disorders with the Five-Factor Model of personality. In P.T. Costa and
T.A Widiger (Eds.), Personality Disorders and The Five-Factor Model of Personality. Washington:
American Psychological Association.
Wiggins, IS. and Pincus, A.L. (1989). Conceptions of personality disorders and dimensions of personality.
Psychological Assessment, 1,305-316.
Wittchen, H.-D., Essau, CA, and Krieg, I-C (1991). Anxiety disorders: similarities and differences of
comorbidity in treated and untreated groups. British Journal of Psychiatry, 159,23-33.
Yeung, AS., Lyons, M.I, Waternaux, CH., Faraone, S.v., and Tsuang, M.T. (1993). The relationship between
DSM-III personality disorders and the Five-Factor Model of personality. Comprehensive Psychiatry,
34, 227-234.
12
PSYCHODIAGNOSTICS AND INDICATIONS

FOR TREATMENT IN CASES OF
PERSONALITY DISORDER
Some Pitfalls
J. Derksen and H. Sloore
1. INTRODUCTION
In both clinical practice and scientific research, personality disorders have

acquired a stable position. Clinicians diagnose personality disorders quite frequently,
and researchers have been enthralled with the topic since the appearance of a sepa-
rate axis for personality disorders in DSM-III (APA, 1980). In bygone years, numerous
discussions have occurred on, among other things, the definition of what constitutes a
personality disorder and whether the notion of a personality disorder belongs to a cat-
egorical, dimensional or prototypical classification. A number of key characteristics
stand out in all of the discussions of personality disorder. Personality disorders are char-
acterised by:
a. their early onset;
b. their stable and persistent character;
c. the fact that they influence several different domains of behaviour (e.g., work,
relationships, free time);
d. their interpersonal locus, which means primary expression in an interpersonal
context;
e. a significant degree of disturbance of the personality, which means that the
pathology is present to a considerable degree.
The distinction between Axis I and Axis II disorders is not always accepted,
however. The distinction is also sometimes quite difficult to make in actual practice.
Consider, for example, the sometimes difficult distinction between the antisocial per-
sonality disorder and disorders which can accompany the use of particular substances
or the distinction between the schizo typal personality disorder and schizophrenia
(Widiger and Shea, 1991). In the present chapter, the problems which present them-
selves during the diagnosis and assessment of the symptoms of personality disorders
156 J. Derksen and H. Sloore
will be discussed. Just how the clinician should proceed, which instruments may
be particularly helpful, and the problems which can arise on this front will also be
considered.
2. THE DISTINCTION BETWEEN AXIS I AND AXIS II
In a fascinating article by Hirschfield (1993) (see Derksen, 1995, for extended dis-
cussion), the question of whether or not personality disorders can be adequately dis-
tinguished from other mental disorders is considered. Put differently, Hirschfield asks
whether the distinction between Axis I and Axis II is justified or not. The differences
observed by Hirschfield have now been accepted on a fairly wide basis and allow us to
draw a number of practical conclusions with regard to the diagnosis and measurement
of personality disorders. The most important differences are as follows.
A. Personality disorders are chronic and protracted conditions while Axis I dis-
orders are rather episodic in nature. The chronic character can be detected
in a diagnostic procedure and perhaps most easily with the use of an inter-
view as opposed to a questionnaire.
B. Personality disorders reflect more basic emotional dysfunctions and are thus
deeply rooted, which constitutes one of the most difficult problems when it
comes to diagnosis in our opinion. We no longer find ourselves at the level of
discernible facts; the clinician can only make deductions from a set of data,
which makes the use of questionnaires to diagnose personality disorders
much more difficult.
C. Personality disorders are almost always complex and thus more difficult to
treat than many of the disorders associated with Axis I of DSM -IV (e.g.,
depression, certain phobias). From a diagnostic perspective, it is therefore
critical that the presence (or absence) of a personality disorder be identified.
In such a manner, the treatment for the problem(s) at hand can be specially
adjusted to the individual in question and, in cases of an integrated treatment
approach (Derksen, 1996), such diagnosis should definitely be undertaken
prior to treatment.
D. Personality disorders are ego-syntonic , which means that the individual is
not particularly aware of his or her problem, that the person feels relatively
good and that others are more likely to experience the problems than the
person in question. Other mental disorders tend to be ego-dystonic, which
means that pain and suffering on the part of the patient accompany them.
The present distinction nevertheless depends on the nature of the personal-
ity disorder. The obsessive-compulsive person will usually not "suffer" much
from the disorder while the borderline patient may experience, among other
things, suicidal inclinations and a generally disturbing instability.
In light of these observations, it is clear that an immense problem confronts us at
the level of diagnosis and certainly when we operate with a questionnaire or self-
description method. The problems associated with the accurate diagnosis of a person-
ality disorder as opposed to some other mental disorder may also help explain the
frequent discrepancy between results attained using questionnaires and results attained
using clinical opinion (Oldham, 1991).
The following arguments are often used to deny the necessity of diagnosing per-
sonality disorders separately from other disorders.
Psychodiagnostics and Indications for Treatment of Personality Disorder 157
• Personality disorders are not as well-known as Axis I disorders (e.g., schizo-

phrenia, depression).
• Personality disorders are not always viewed as "real" mental disorders. Some
authors pose the transition from "normal" personality to what can be con-
sidered a personality disorder as an explicit problem and thereby question
the need to distinguish between the two DSM-IV axes (Widiger and Costa,
1994).
• The description of personality disorders as long-term problems may itself
make the diagnosis unpopular among therapy-oriented clinicians. The prob-
lems associated with such a diagnosis are clearly considered difficult to treat
(Reich and Vasile, 1993).
• The diagnosis of personality disorders is usually not easy and is frequently
influenced by aspects of Axis I mental disorders (e.g., a borderline personal-
ity disorder may be camouflaged by depression and fear). This is the problem
of comorbidity between not only the Axis I and Axis II disorders but also
between the Axis II disorders themselves (Klein, 1993).
The diagnosis of personality disorders is essential for the planning of accurate and
effective treatment. Specific identification of the relevant disorder can be of ultimate
importance for the choice of therapy and evaluation of the therapy. A few brief exam-
ples may illustrate this point. The presence of an Axis II (personality) disorder can
strongly curb the effectiveness of short-term behaviour-modification techniques. Simi-
larly, the diagnosis of a personality disorder can make us aware of the fact that treat~
ment of the symptom may produce little result in light of the more general underlying
syndrome. Finally, the conclusion that the patient has little empathic capacity will have
much broader significance when the diagnosis is an antisocial personality disorder as
opposed to no underlying personality disorder.
In addition to clinical experience, the-as yet quite limited-scientific research
along these lines has also demonstrated the importance of early and accurate diag-
nosis of personality disorders (Garyfallos, Adamopoulou, Voikli, Saitis, Kirtsos, and
Moutzoukis, 1994; Farmer and Nelson-Gray, 1990; Hoffart, 1994; Johnson, Hyler,
Skodol, Bornstein, and Sherman, 1995; Klein, Wonderlich, and Shea, 1993; Reich and
Green, 1991; Reich and Vasile, 1993; Smith, Deutch, Schwartz, and Terkelsen, 1993;
Stein, Hollander, and Skodol, 1993; van Velzen and Emmelkamp, 1997). It can gener-
ally be concluded on the basis of the research to date that the presence of a comorbid
Axis II personality disorder will greatly complicate treatment in many cases, require
more time for treatment and increase the probability of (quick) relapse. Extra
investment in diagnosis may thus be not only helpful but also necessary under such
circumstances.
With the aforementioned considerations in mind, we will now describe the diag-
nostic procedures available to both the clinician and the researcher for facilitating indi-
cation for treatment.
3. ESTABLISHING A DIAGNOSIS
In diagnosing personality disorders, one could-in principle-make things rela-

tively easy and simply leave the diagnosis to the "expert" judgement of the clinician.
In doing this, however, we may lose track of an old but nevertheless still current debate,
namely: actuarial versus clinical prediction. This discussion gained impetus in the field
158 J. Derksen and H. Sioore
of psychology with the work of P. Meehl in 1954. Dawes, Faust, and Meehl (1989) later
observed in an article in Science that more than 100 studies in the social sciences have
shown the actuarial method to be consistently equal or superior to the clinical method.
This clearly suggests that we should not settle for clinical judgement alone. We must
also develop techniques (e.g., structured interviews and questionnaires) to support
and/or supplemental clinical judgement.
Research has shown the following problems to occur (among others) with
reliance on clinical judgement alone.
A. Clinicians have trouble distinguishing valid variables from invalid variables
and can thereby develop false associations between variables.
B. Clinicians tend to make a diagnosis much too quickly ("premature closure")
on the basis of only those few elements which stand out (e.g., a suicide attempt
and generally unstable relationships can quickly lead to a borderline diag-
nosis).
C. Every clinician has a limited amount of experience which is quite specific,
complicates the assessment of certain relations and makes the evaluation of
certain criteria for their "correct" values most difficult.
D. The clinician tends to encounter those people with the most severe or
"heaviest" pathology, which can create a distorted picture of the relevant
disorders.
Most clinicians have an exaggerated degree of trust in their clinical judgement.
It should be evident that we are speaking of the results of general research and
that individual differences may clearly exist with regard to clinical effectiveness and
accuracy.
What are the methods available to us for the measurement of personality disor-
ders? Vertommen (1994) recently presented an extensive overview of the most fre-
quently used questionnaires. In the present discussion, we will limit ourselves to a brief
summary of the most frequently used methods. The creation of an unambiguous clas-
sification of the methods is not a simple matter, however. Among the most frequently
used methods are the following.
Semi-structured or structured interviews such as the Structured Interview for
DSM-W Personality, SIDP-W (Pfohl, Blum, and Zimmerman, 1995); the Personality
Disorder Examination, PDE-R (Loranger et aI., 1987); or the Diagnostic Interview for
Personality Disorders, DIPD (Zanarini et aI., 1989). In all of these interviews, either the
DSM criteria or the ICD criteria are simply translated into concrete questions or the
items are re-grouped into a number of domains relevant to personality disorders (e.g.,
interpersonal relations, work, emotions, etc.) and questions are then formulated.
Self-report questionnaires such as the Personality Diagnosis Questionnaire
(PDQ-R) from Hyler and Rieder (1987) and the adaptation of this for the Dutch lan-
guage by Vertommen et ai. (1993); the Millon Clinical Multiaxial Inventory (MCMI-
III) from Millon (1994: Sloore and Derksen, 1997); the personality disorder scales
developed by Morey, Waugh, and Blashfield (1985) on the basis of the MMPI item pool;
more recent and adapted to the DSM-IV, the PD-scales for the MMPI-2 from Somwaru
and Ben-Porath (1995); and the Questionnaire Regarding Characteristics of the Per-
sonality adapted for Dutch by Duijsens, Eurelings-Bontekoe, Diekstra, and Ouwersloot
(1995).
There are also combinations of the two aforementioned methods: The Structured
Clinical Interview for DSM-W (Spitzer, 1983; Spitzer at aI., 1995), which is better known
as SCID-II. In this case, the patient completes the questionnaire and the clinician can
undertake a completely structured interview or semi-structured interview with regard
to those questions which the patient responded "yes" to.
There are also theory-based techniques such as the MCMI-III (1994) or Kern-
berg's Structured Interview (Kernberg, 1981; Derksen, 1986) and also techniques which
are only bound to the DSM criteria or ICD criteria (e.g., the MMPI-PD scales from
the Questionnaire Regarding Characteristics of Personality).
Some techniques only measure certain personality disorders, such as the Diag-
nostic Interview for Borderline Patients, DIB (Zanarini et aI., 1989); other techniques
measure the full range of personality disorders (PDQ-R) or also measure Axis I dis-
orders in addition to personality disorders (MCMI-III) and other pathological charac-
teristics (MMPI-2).
Other self-evaluation instruments search for the underlying dimensions of the
personality disorder; one example is the Temperament and Character Inventory (TCI)
based on the biosocial theory of Cloninger (1991) and, of course, all of the question-
naires based on the Big Five Personality Factors (Costa and Widiger, 1994).
For researchers, the semi-structured interview has proved to be the most reliable
and valid instrument. The questionnaire method produces specific problems. A few
examples are the following. When information on the possible early onset of a per-
sonality disorder is sought using a questionnaire, questions must be posed with regard
to not only the current condition, behaviour, problems, etc. but also with regard to
earlier periods in the patient's life. This is certainly possible but nevertheless not always
easy to realise. The stable and persistent nature of the disorder makes the distinction
between the more temporary symptoms and the long-term syndrome very difficult to
maintain at the level of the questionnaire. Many of the questions must contain very
specific temporal qualifications, and the question becomes just which temporal dis-
tinctions should be used? A consequence is that the majority of the questionnaires in
the area of personality disorders over-report the presence of a personality disorder.
For the clinician, research instruments (and the semi-structured interview in par-
ticular) have little value beyond the DSM classification which they produce. Given this
classification, the clinician clearly needs additional diagnostic information pertaining
to the particularly weak and strong aspects of the individual's personality along with
an inventory of the possibilities for treatment and the difficulties which can be expected
during intervention.
In our opinion, the clinician's manner of operating should be characterised by a
series of exploratory interviews and the collection of a heterogeneous case history fol-
lowing intake and prior to intervention (e.g., Derksen, 1983, 1993). As necessary and
commonly the case with personality disorders, the interviews are supplemented with
psychological testis selected from the battery of available clinical instruments. During
this research phase, hypotheses are formulated for evaluation during the treatment
phase. Within this approach, diagnosis and treatment cannot be separated although they
can be distinguished. Assessment is much more extensive at the beginning of the clin-
ical contact than further up in the process but never disappears completely. According
to this line of thought, continuous diagnosis is the guide to treatment. Following intake,
the patients are informed that there will be a series of exploratory conversations in
order to get to know them better and determine the suitable course of treatment or
treatment recommendation. During these conversations, the client's life course is exam-
ined along with his or her relationships and work. The course of the psychiatric and
possibly somatic complaints is considered in detail. With the permission of the patient,
the partner is also invited on one occasion. This conversation allows the clinician to
observe how the patient reacts and can thereby clarify many future (im)possibilities.
The psychodynamic techniques which Kernberg describes in his structural interview
(1981) can also be integrated into such conversations. In addition to providing the nec-
essary factual information, this research phase also enables examination of the various
emotional aspects of the patient's life course. A patient can dryly talk about his late
father, for whom he had great respect, but quite different information may emerge
when this same patient is asked to close his eyes, envision his father before him and
relate the feelings which he then experiences in the here and now. Such diagnostic
material is quite dynamic. Indeed the information in the conversations remains unclear,
raises questions or points to some other gaps, supplemental cognitive testing tech-
niques, questionnaire methods, and projective techniques can still be applied.
The underlying line of reasoning here is that a symptom only has significance
within the general psychological framework of the person in question. People give
expression to what they otherwise cannot express at that moment via their complaints.
The personality is the psychological basis for the behaviour of an individual. A phobia
can have a completely different function at both the conscious and unconscious levels
for one person when compared to another person. The phobia for stores encountered
in a person with a borderline personality disorder may serve to maintain the psycho-
logical balance of the personality much more than in a person without such striking
personality characteristics. An adequate diagnosis calls up hypotheses in this light and
allows these to playa role in treatment specification. Clinical experience with patients
with compulsive symptoms shows the symptom to often avert psychotic decompensa-
tion and the ego structure of the patient to often be structurally damaged. A compul-
sive symptom in an otherwise well-integrated person cannot be compared to the
preceding, as discerning psychotherapeutic experience with these patients often shows
the symptom to replace, for example, the expression of feelings of rage. It is thus
obvious that this way of diagnosing will substantially influence the choice of treatment
under such circumstances.
A protracted examination phase is certainly attractive but in actual practice a
clear treatment policy is often expected more or less directly following the initial intake
interview. When further psychological examination (read: testing) is requested follow-
ing the intake interview, this is usually accompanied by insufficient case material, an
insufficient breadth of material and an unclear formulation of the question(s) to be
answered. The psychological testing is also often requested by a non-psychologist who
cannot, thus, evaluate the possible implications of the test results in light of the exist-
ing or ongoing diagnosis. It is striking in this context that those in training (e.g., psy-
chiatrists, clinical psychologists, social workers, and social-psychiatric nurses) tend to
be placed at the gates of an institution rather than the most experienced forces.
The preceding approach is, in our opinion, called for when the intake with an
experienced clinician suggests a severe personality disorder. In the case of psycholog-
ical testing, moreover, the clinician is usually helped most by the use of clinical as
opposed to research instruments (Beutler and Berren, 1995). Clinical instruments such
as the Wais, MMPI-2, Rorschach, and the MCMI offer a large amount of extra infor-
mation because they generally involve a broad range of techniques and are not, in con-
trast to most research instruments, geared to the reliable and valid measurement of a
descriptive concept such as a certain type of personality disorder in terms of the DSM.
When applied creatively, moreover, the interviews and testing supply numerous details
on a patient's perceptions, experiences, weaknesses, crooked feelings, unfelt conflicts,
and compensation possibilities. Such details clearly come in handy for the formulation
of theory and treatment of the person in question, and they can also reduce the risk of
drop-out and optimise the relation between therapist and patient (Derksen, 1995).
When one discerns that a patient is not particularly grateful for the help being offered
him or her, for example, the contribution of the therapist may need to be consciously
minimised and the accent placed on the progress being made by the patient him/herself
in order to prevent drop-out. The narcissistic vulnerability of the patient can then be
addressed in a later phase when the patient can gradually bear it.
The following elements have been found to be a critical part of the attitude of a
diagnostician.
• Make the patient as opposed to the reference framework, psychotherapeutic

school or clinical protocol the centre of attention.
• Feel free to undertake a semi-structured interview in order to determine
whether the diagnosis of a personality disorder on the basis of the clinical
contact is justified or not.
• Use that reference framework which fits the needs of the patient in both
the diagnostic and treatment phases; the relevant framework may be psy-
chodynamic, behavioural therapeutic, cognitive or humanistic. One frame-
work may supplement or follow the other. Consideration of a complaint
without accurate and complete functional and behavioural analysis is unac-
ceptable. Similarly, initiation of treatment for an anxiety disorder without
considering the possibility of a mood disorder is also inadequate even when
the patient does not mention or possibly feel the latter. Anxiety for sharp
knives should only be treated after the moral consicence of the patient in
question has been assessed. This can prevent us from teaching, for example,
someone with an antisocial personality disorder to handle knives in a more
supple manner.
• Prepare yourself for continuous assessment. Use a questionnaire such as the
MMPI-2 on a regular basis to map the course of the treatment and compare
one's impressions with this information.
• Be direct, open and honest in reporting the examination results at the close
of the diagnostic phase and initiation of treatment.
4. TREATMENT DETERMINATION: INDIVIDUAL OR

STANDARDIZED?
The approach described in the preceding assumes a model geared to the individ-
ual. In the ongoing scientific discussion, this stands in marked contrast to the standard
procedure. Let us therefore consider the arguments of the empiricists who prefer a stan-
dard treatment procedure over a more individualised treatment procedure. From a
standardised treatment perspective, extensive consideration of the different treatment
possibilities is simply considered lost time and the academic, heavily experimental,
behavioural-therapeutic study by Schulte, Kunzel, Pepping, and Schulte-Bahrenberg
(1992) with regard to "tailor-made versus standardised" treatment is often brought
forth. On the basis of this study, it is argued that extensive consideration of the various
treatment alternatives and individualisation of the treatment process produces poorer
results than following standard treatment procedures. The dominant line of reasoning
162 J. Derksen and H. Sloore
behind this position is that the diagnosis of a personality disorder is simply not rele-
vant to treatment. Following brief consideration of the symptom(s) and diagnosis, a
treatment protocol should be applied with an eye to gaining time and efficiency. In our
opinion, this is not the correct conclusion for a number of reasons.
To start with, the study by Schulte et al. (1992) involved only phobic patients. The
patients could only have one or more phobias and thus no other complaints.
In the second place, the standard treatment consisted of exposure in vivo, which
is known to be an outstanding treatment strategy for phobic complaints in particular.
Even the most orthodox psychoanalyst would agree to this today. Live exposure was
compared to treatment involving less exposure and greater time for panic management
and cognitive techniques. Comparison of the group averages showed a relatively small
difference with standard treatment generating about 77% improvement and the
individualised treatment (read: a splash of cognitive techniques) generating 60% im-
provement. A third group received an individualised program but not applied to the
appropriate patient and showed 57% improvement. Under treatment indication in clin-
ical practice, however, we understand something different than in the study by Schulte
et al. (1992). Treatment indication is generally not the choice of techniques from a par-
ticular therapeutic approach but the choice of a particular therapeutic approach or
combination of approaches. A team does not opt for panic management or live expo-
sure but, rather, for behaviour therapy or psychoanalysis.
In the third place, the study by Schulte et al. (1992) provides no information
on the development of the individualised treatment procedure. Given a behavioural-
therapeutic reference framework, for example, one can expect an extended functional
and topographic analysis. In the study by Schulte et aI., however, nothing along these
lines is reported.
In the fourth place, it is striking that the therapists in the study by Schulte et al.
were young (just how young?) psychologists with some of them having no treatment
experience. The median number of patients previously treated by the psychologists was
9, which is comparable to the figure for psychology practicum students in The Nether-
lands. A large degree of drop-out was also observed among both the patients and ther-
apists. Research raising arguments against indication for treatment was not involved
here, however.
An associated problem, which also proved difficult to solve in the Schulte et al.
study, was the "manual fidelity" of the therapists or degree to which the therapist
adhered to the protocol. For standardised treatment, only 46% of the therapists
adhered to the protocol. All kinds of adjustments were made for the patient in ques-
tion. Similarly, in interviews we conducted with colleagues participating in protocol
treatment within the framework of a research project, comparable results were found.
The interviews showed a considerable difference between the interventions as they
were expected to occur and how they actually occurred. Put concretely, this means that
the therapists did much more and also other things such as giving advice, discussing the
background to the complaints and extensive consideration of the patient's life history.
Studies with protocols often involve techniques which otherwise elicit little or no
discussion with regard to their utility. Live exposure and panic management in the case
of a panic disorder with agoraphobia is quite often useful, for example. As Korrelboom
(1995) recommends in his study, however, comorbidity must also be taken into con-
sideration and this does not occur sufficiently in a treatment approach based on a pro-
tocol. Korrelboom speaks primarily of comorbid symptomatology such as relational
problems and depressions, although this also holds for personality disorders.
The preceding is one side of the story. The other side is that the protocol makers
may be right in some cases. Why can't treatment be shorter? There are, after all, long
waiting lists and many people actually seek very brief, complaint-oriented help. There
are also complaints for which brief complaint-oriented intervention provides clear
relief. Such protocols meet not only the demands of the insurers but also play an impor-
tant role, which should not be forgotten, in clinical practice.
A familiar example of a treatment protocol is that described by Craske and
Barlow (1993) for panic disorders with agoraphobia. During assessment, the primary
concern is clear identification of the anxiety disorder. Other related DSM Axis I dis-
orders such as depression and somatoform disorders are considered; a functional analy-
sis is undertaken; a medical examination is performed; an inventory is made of various
physiological variables; and the patient is also allowed to contribute via self-reporting
and monitoring. Additional DSM Axis II disorders are simply not screened for,
however. Little or no attention is paid to the life course of the patient or to the patient's
capacity to form relationships. The course of the protocol sessions is followed in utmost
detail but also filled with numerous individual adaptations. The treatment package is
extensive; both cognitive therapy and various forms of behaviour therapy are put to
use. In fact, the package resembles the individualised treatment program in the study
by Schulte et al. more than the standardised treatment program. In any practice, more-
over, people are likely to be jealous of the precision and structure with which such
standardised treatment occurs. In their conclusions, Craske and Barlow, nevertheless
suggest that the protocol should be further individualised to increase its effectiveness.
Some 50% of the patients still report anxiety complaints following treatment for a panic
disorder with agoraphobia. After generalisation of a protocol, in other words, it must
again be individualised. For complex complaints other than the familiar anxiety disor-
ders and certainly in cases with comorbidity, moreover, the situation may be even
worse.
On the basis of our own practice, we can illustrate the importance of a diagnos-
tic process which also takes the possibility of personality disorders into consideration.
In an exploratory study, we asked patients who had received short-term treatment con-
sisting of an average of 16 sessions in our practice between four and six years prior if
they were willing to talk to a researcher about how things were going and to take a
test. A total of 46 patients were willing to help us. After completion of the interview
with the researcher, their case files were examined for the diagnosis established
following assessment. With respect to the reliability of the DSM classifications, the
following can be observed. The diagnostic examination was extensive; intake was per-
formed by an experienced clinician; a number of diagnostic interviews took place; and
a hetero-anamesis was collected. The reporting on the patient together with the DSM
classification was done by at least two co-operating psychologists, with at least one of
these having full clinical registration. Of the 20 females and 26 males, 30 appeared to
have only an Axis I classification and 16 both an Axis I and Axis II personality disor-
der diagnosis. The most frequent Axis II classification was a mixed personality disor-
der. The second most frequent Axis II classification was the avoidant personality
disorder. A variety of Axis I complaints occurred: in particular, anxiety disorders (11),
somatoform disorders (7), adjustment disorders (7), and mood disorders (5). During
the interviews with the 46 patients, we determined their current GAF score and their
general functioning which meant assessment of how things were going with them, how
they were functioning, use of medications and possible treatment since four to six years
prior. They also completed the MMPI-2.
The patients were divided into two groups, those diagnosed either with or without
a personality disorder at the time. The average GAF scores showed those patients pre-
viously diagnosed as having a personality disorder to function worse than those without
such a diagnosis. Females with prior diagnosis of a personality disorder received the
lowest GAF score. It should be noted that the GAF score is not generated using a reli-
able and valid measurement instrument; it predominantly reflects clinical impression.
The group with a previous diagnosis of personality disorder was also found to use more
medication than the group without such a prior diagnosis. The group with a previous
diagnosis of a personality disorder had also undergone more treatment than the other
group. Some 80% of the females previously diagnosed with a personality disorder
had sought additional treatment. The average MMPI-2 profiles for the patients at the
present time confirm this picture. Patient without a personality disorder diagnosis pro-
duced MMPI-2 profiles (clinical and content scales) within the normal range (all scales
below T = 60) and the patients who were diagnosed with a personality disorder pro-
duced higher profiles. This was specially true for women. Two of the clinical scales (para-
noia and schizophrenia) were above 65.
5. CONCLUSION
We have hopefully made the importance of diagnosing personality disorders

in addition to the other symptoms and syndromes needed to determine appropriate
psychological treatment clear in the preceding. If an Axis II disorder is not diagnosed,
then one can always-in consultation with the individual seeking help-choose the
most efficient/economic option: short-term, protocol-based treatment. The examination
interviews conducted in the diagnostic phase can also be minimised. If Axis II prob-
lems are indeed detected, then these can be confirmed or refuted using a research
instrument such as the SIDP-IV, PDE-R or SCID-II. An extended diagnostic phase
including the use of further psychological tests may then be not only clinically useful
but necessary. The treatment will be more complex than in the absence of Axis II
disorders, which also means that extensive diagnosis and a thorough search for the most
appropriate intervention will be needed. Future research along these lines is also
recommended.
REFERENCES
American Psychiatric Association. (1952, 1968, 1980, 1987, 1994). Diagnostic and statistical manual of mental
disorders. Washington, DC: Author.
Beutler, L.E. and BeTTen, M.R (1995). Integrative Assessment of Adult Personality. New York: The guilford
Press.
Cloninger, C.R (1987). Tridimensional Personality Questionnaire (TPQ). St-Louis, MO: Department of
Psychiatry, Washington University School of Medecine.
Cloninger, C.R, Przybeck, T.R, and Svrakic, D.M. (1991). The tridimensional personality questionnaire: U.S.
normative data. Psychological Reports, 69, 1047-1057.
Costa, P.T. and Widiger, T.A. (Eds.) (1994). Personality disorders and the jive-factor model of personality.
Washington, DC: American Psychological Association.
Craske, M. and Barlow, D. (1993). Panic disorder and agorafobia. In D. Barlow (Ed.), Clincial Handbook of
Psychological Disorders. New York: The Guilford Press, p. 1-47.
Dawes, RM., Faust, D., and Meehl, P.E. (1989). Clinical versus actuarial judgement. Science, 243, 1668-1674.
Derksen, 1. (1983). Psychotherapieen in het geding. Baam: Nelissen.
Derksen, 1. (1986). Structurele diagnostiek van psychische stoornissen: neurose, borderline, psychose. Baarn:
Nelissen.
Derksen, 1. (1995). Personality disorders, Clinical & Social Perspectives. Chichester-New York: John Wiley &
Sons. Handboek Persoonlijkheidsstoornissen. Utrecht: De Tijdstroom.
Derksen, 1. (1996). The complex relationship between assessment and treatment of personality disorders: a
practical guideline for the clinicians. 2nd European Congress on Personality Disorders. Milan.
Derksen, 1., Mey, H., de, Sloore, H., and Hellenbosch, G. (1995). MMPI-2: Handleiding bij afname, scoring en
interpretatie. Nijmegen: Pen Tests Publisher, 2e geheel herziene druk.
Duijsens, 1.1., Eurelings-Bontekoe, E., Diekstra, R, and Ouwersloot, G. (1995). Vragenlijst voor kenmerken
van de persoonlijkheid. Lisse: Swets Test Services.
Farmer, R and Nelson-Gray, R (1990). Personality disorders and depression: Hypothetical relations,
empirical findings, and methodological considerations. Clinical Psychological Review, 10, 453-476.
Garyfallos, G., Adamopoulou, A, Voikli, M., Saitis, M., Kirtsos, P., and Moutzoukis, Ch. (1994). DSM-III-R
personality disorders among patients with depressive andor anxienty disorders. Journal of Personal-
ity Disorders, 8, 320--332.
Hirschfield, RM.A (1993). Personality disorders: definition and diagnosis. Journal of Personality Disorders,
7,4--8.
Hoffart, A (1994). State and personality in agoraphobic outpatients. Journal of Personality Disorders, 8,
333-341.
Hyler, S.E. and Rieder, Ro. (1987). PDQ-R: Personality Diagnostic Questionnaire-Revised. New York: New
York State Psychiatric Institute.
Johnson, 1., Hyler, S., Skodol, A., Bornstein, R, and Sherman, M. (1995). Personality disorders symptoma-
tology associated with adolescent depresiion and substance abuse. Journal of Personality Disorders,
9,318--329.
Kernberg, 0. (1981). Structured interviewing. Psychiatric Clinic of North America, 4,169-195.
Klein, M.H. (1993). Issues in the assessment of personality disorders. Journal of Personality Disorders.
Supplement to Spring, 18-33.
Klein, M., Wonderlich, S., and Shea, M. (1993). Models of relationships between personality and depression:
Toward a framework for theory and research. In: M. Klein, D. Kupfer, and M. Shea (Eds.), Personal-
ity and depression: A current view. New York: The Guilford Press, 1-54.
Korrelboom, C. (1995). Paniekmanagement en exposure in vivo bij paniekstoornis met agorafobie. Proefschrift
Katholieke Universiteit Nijmegen.
Loranger, Susman, Oldham, and Russakoff. (1987). The personality disorder examination: a preliminary
report. Journal of Personality Disorders, 1, 1-13.
Meehl, P.E. (1954). Clinical versus statistical prediction. Minneapolis, MN: University of Minnesota Press.
Millon, T., Millon, c., and Davis, RP. (1994). Millon Clinical Multiaxial Inventory-III. Minneapolis: National
Computer Systems.
Morey, L.c., Waugh, M.H., and Blashfield, RB. (1985). MMPI scales for DSM-III personality disorders: their
derivation and correlates. Journal of Personality Assessment, 49, 245-251.
Oldham, 1.M. (Ed.) (1991). Personality disorders: new perspectives on diagnostic validity. Washington, DC:
American Psychiatric Press.
Pfohl, B., Blum, N., and Zimmerman, M. (1995). Structured interview for DSM-IV personality, SIDP-IV. Iowa
City, IA: Department of Psychiatry, University of Iowa.
Reich, 1. and Green, A (1991). Effect of personality disorders on outcome of treatment. The Journal of
Nervous and Mental Disease, 179,74--82.
Reich, 1. and Vasile, RG. (1993). Effect of personality disorders on the treatment outcome of Axis I
conditions, an update. Journal of Nervous and Mental Disease, 181, 475-484.
Schulte, D., Kunzel, R., Pepping, G., and Schulte-Bahrenberg, Th. (1992). Tailer-made versus standardized
therapy of phobic patients. Advances in Behavior research and Therapy, 14, 67-92.
Sloore, H. and Derksen, 1. (In press). Issues and procedures in MCMI translations. In:Th. Millon, The Millon
inventories. New York: The Guilford Press.
Smith, T., Deutch, A, Schwartz, F., and Terkelsen, K. (1993). The role of personality in the treatment of
schizofrenicand schizoaffective disorder inpatients: A pilot study. Bulletin of the Menniger Clinic, 57,
88--99.
Somwaru, D. and Ben-Porath, Y. (1995). Assessment of Personality Disorders with the MMPI-2. 30th
Annual Symposium on Recent Developments in the Use of the MMPI-2 and MMPI-A. St.-Petersburg,
Fl.
Spitzer, R.L. (1983). Psychiatric diagnosis: are clinicians still necessary? Comprehensive Psychiatry, 24,
399-411.
Spitzer, R.L., First, M., Gibbon, M., and Williams, IB. (1995). Structured Clinical Interview for the DSM-JV
Personality Disorders. New York, New York State Psychiatric Institute.
Stein, D., Hollander, E., and Skodol, A. (1993). Anxiety disorders and personality disorders: A review. Journal
Van Velzen, C. and Emmelkamp, P. (1999). The relationship between anxiety disorders and personality
disorders: Prevalence rates and comorbidity models. In: I Derksen, C. Maffei, and H. Groen (Eds.),
Personality Disorders: Diagnosis and Treatment. New York: Kluwer Academic Plenum Publishers,
129-154.
Vertommen, H. (1994). Nieuwe perspectieven in de vragenlijstdiagnostiek van persoonlijkheidsstoornissen.
Tijdschrift klinische Psychologie, 24, 2, 101-119.
Vertommen, H., Van den Brink, w., De Doncker, D., Schotte, K., Ouwersloot, G., and Loontjes, P. (1993). De
PDQ-R: een standaard nederlandstalige versie van de Personality Disorder Questionnaire-revised.
Intern rapport. Leuven: Centrum Klinische Psychodiagnostiek.
Widiger, T.A. and Costa, P.T. (1994). Personality and personality disorders. Journal ofAbnormal Psychology,
103, 1, 78-91.
Widiger, T.A. and Shea, T. (1991). Differentiation of Axis I and Axis II disorders. Journal of Abnormal Psy-
chology, 100, 3, 399-406.
Zanarini, M.C., Gunderson, IG., Frankenburg, F.R., and Chauncey, D.L. (1989).'The revised diagnostic inter-
view for borderlines: discriminating BPD from other Axis II disorders. Journal of Personality Disor-
ders, 3, 10--16.
13
THE PSYCHOTHERAPEUTIC TREATMENT

OF BORDERLINE PATIENTS
Otto F. Kernberg
Cornell University Medical College

New York, New York
1. INTRODUCTION
Under the impact of new clinical experiences and empirical research in the
last twenty years, the psychodynamic psychotherapy for borderline patients has
evolved into a more exploratory, expressive direction. The supportive psychotherapy
approaches that were formerly recommended as treatment of choice have been losing
their appeal. Differences persist, however, regarding the extent to which the psy-
chotherapy should be purely analytic, exploratory, or expressive, or should combine
expressive and supportive features, at least in the initial stages of treatment.
Robert Waldinger's (1987) comprehensive review provides what is at once a syn-
thesis and summary of the differences and common features among the leading con-
tributors in this field, including Buie and Adler (1982), Chessick (1977), Giovacchini
(1979), Gunderson (1984), Kernberg (1982), and Masterson (1976). All of these adopt
an essentially interpretive or expressive approach, although they vary in the degree (1)
to which they consider supportive techniques helpful or even central in the early stages
of treatment; (2) to which providing a holding environment early in the treatment
is crucial; and (3) to which the therapist needs to structure the boundaries and the
framework of the treatment.
Waldinger and Gunderson (1987), in their conclusions regarding the study of five
borderline patients treated with long-term psychoanalytic psychotherapy, propose the
following general strategies: the treatment must include a stable framework that defines
the boundaries of its setting; the therapist must be more active than he would be with neu-
rotic patients because of the borderline patient's problems in reality testing, projective
mechanisms, and distortions; the negative transference has to be tolerated; the patient
must be helped via interpretation to establish bridges between his actions and feelings;
selfdestructive behavior needs to be discouraged by clarification and confrontation;
168 O. F. Kernberg
acting out has to be blocked by setting limits on actions that endanger the patient, others,
or the treatment; in the early phases of treatment clarifying and interpreting the
transference in the here-and-now is preferable to genetic interpretations; and counter-
transference analysis is a crucial aspect of the interpretive work.
Adler (1989), in comparing the models of Kernberg (1984), Masterson (1981) and
Rinsley (1982),Adler (1985) and Buie and Adler (1982), Gunderson (1984), and Searles
(1986), also stresses the importance of analyzing of countertransference, the processes
and problems in building the therapeutic alliance, the problems posed by getting the
patient to comply with treatment schedule and boundaries, and the complications pre-
sented by self-destructive actions and inadequate social and family support. He con-
cludes that, in contrast to Kernberg's "relatively clear distinctions between supportive
and expressive psychotherapy, most of the literature defines a mixture of supportive
and expressive exploratory techniques p. 62)."
In what follows, and in order to highlight some potentially alternative strategies
within the common frame of psychodynamic psychotherapy for borderline patients, I
am selecting a few representatives from this field. Because of the basic similarity
between my own approach to that of Gunderson and Waldinger and of Masterson and
Rinsley, I shall utilize my own approach as representing that general strategy, while
underlining the fact that there are significant differences between what I do and what
others do in the early stages of the treatment: These differences reside in my empha-
sis on early interpretations of the negative transference; while the others stress the
importance of the therapist's holding function and of building up the therapeutic
alliance at that time.
Because Vamik Volkan (1987) has integrated my ideas with ideas stemming from
Winnicott (1953, 1956, 1960), I shall outline his approach. The significant difference
between Adler and Buie's approach to mine has led me to choose Adler's (1985) for-
mulations for a more detailed analysis; and finally, because of Harold Searles (1986)
significant utilization of concepts from interpersonal psychoanalysis I have selected him
as another major representative of contemporary psychodynamic psychotherapy with
borderline patients. In the final section of this chapter, I summarize some contributions
to the psychodynamic psychotherapy of borderline patients derived from my recent
work.
2. SUMMARY OF THE APPROACH OF THE WESTMINSTER

DIVISION-NEW YORK HOSPITAL BORDERLINE
PSYCHOTHERAPY RESEARCH PROJECT
We have constructed a theory of psychodynamic treatment of borderline patients

that derives from the theory of psychoanalytic technique, and modifies this technique
by creating a general strategy aimed at resolving the specific disturbances of border-
line patients (Kernberg, 1984; Kernberg et al., 1989). Our basic objective is the diag-
nosis and psychotherapeutic resolution of the syndrome of identity diffusion, and, in
the process, resolution of primitive defensive operations characteristic of these patients,
and their primitive internalized part-object relationships into "total" object relation-
ships characteristic of more advanced, neurotic, and normal functioning. Primitive
internalized object relations are constituted by part-self representations relating to
part-object representations in the context of a primitive, all-good or all-bad, affect state,
and they are part-object relations precisely because the representation of self and of
The Psychotherapeutic Treatment of Borderline Patieuts 169
object have been split into an idealized and persecutory component-in contrast to the
normal integration of good or loving, and bad or hateful representations of self and
significant others. These primitive or part-object relations emerge in the treatment
situation in the form of primitive transferences characterized by the activation of such
self and object representations and their corresponding affect as a transference "unit"
enacted defensively against an opposite primitive transference unit under completely
opposite affect valence or dominance.
In essence, the psychotherapeutic strategy in the psychodynamic treatment of
borderline patients consists of a three-step procedure: Step 1 is the diagnosis of an
emerging primitive part-object relationship in the transference, and the interpreta-
tive analysis of the dominant unconscious fantasy structure that corresponds to this
particular transference activation. For example, the therapist may point out to the
patient that their momentary relationship resembles that of a sadistic prison guard
and a paralyzed, frightened victim.
Step 2 of this strategy is to identify the self and the object representation of this
particular primitive transference, and the typically oscillating or alternating attribution
of self and object representation by the patient to himself and to the therapist. For
example, the therapist may point out, in expanding the previous intervention, that it is
as if the patient experienced himself as a frightened, paralyzed victim, while attribut-
ing to the therapist the behavior of a sadistic prison guard. Later on in the same session
the therapist may point out to the patient that now the situation has become reversed
in that the patient behaves like a sadistic prison guard while the therapist has been
placed in the role of the patient as a frightened victim.
Step 3 of this interpretative intervention would be to link this particular object
relationship activated in the transference and an entirely opposite one activated at
other times but constituting the split-off idealized counterpart to this particular, per-
secutory object relationship. For example, if at other times the patient has experienced
the therapist as a perfect, all-giving mother, while experiencing himself as a satisfied,
happy, loved baby who is the exclusive objective of mother's attention, the therapist
might point out that the persecutory prison guard is really a bad, frustrating, teasing
and rejecting mother, the victim an enraged baby who wants to take revenge but is
afraid of being destroyed because of the projection of his own rage onto mother. The
therapist might add that this terrible mother-infant relationship is kept completely
separate from the idealized relationship out of the fear of contaminating the idealized
one with the persecutory one, and of the destruction of all hope that, in spite of the
rageful, revengeful attacks on the bad mother the relationship with the ideal mother
might be recovered.
The successful integration of mutually dissociated or splitoff, all-good and all-bad
primitive object relations in the transference includes the integration not only of the
corresponding self and object representations, but also of primitive affects, leading to
affect modulation, to an increase in the capacity for affect control, to a heightened
capacity for empathy with both self and others, and a corresponding deepening and
maturing of all object relations.
This psychotherapeutic strategy also includes a particular modification of three
basic tools derived from standard psychoanalytic technique. First is, interpretation, that
is, establishing hypotheses about unconscious determinants of the patient's behavior.
In contrast to standard psychoanalysis, however, interpretation here involves mostly
the preliminary phases of interpretative interventions, that is, a systematic clarification
of the patient's sUbjective experience, the tactful confrontation of the meanings of those
170 O. F. Kemberg
aspects of his subjective experience, verbal communication, nonverbal behavior, and

total interaction with the therapist that express further aspects of the transference, and
a restriction of the unconscious aspects of interpretation to the unconscious meanings
in the here-and-now only. In contrast to standard psychoanalysis, where interpretation
centers on unconscious meanings both in the here-and-now and the there-and-then of
the unconscious past, in the psychodynamic psychotherapy of borderline patients
psychodynamic interpretations of the unconscious past are reserved for relatively
advanced stages of the treatment, when the integration of primitive transferences has
transformed primitive into advance transferences (more characteristic of neurotic
functioning, and more directly reflective of actual experiences from the past).
Another difference from standard psychoanalysis is the modification of transfer-
ence analysis, in each session, derived from the therapist's ongoing attention to the
long-range treatment goals with any particular patient and the dominant, current con-
flicts in the patient's life outside the sessions. In order that the treatment not gratify
excessively the patient's transference-thus undermining the patient's initial motiva-
tion and treatment objectives, the therapist has to keep in touch with longrange treat-
ment goals. Also, in order to prevent splitting-off of external reality from the treatment
situation-and severe acting out expressed by such dissociation between external
reality and the treatment hours, transference interpretation has to be linked closely to
the present realities in the patient's life. In short, then, in contrast to psychoanalysis
(where a systematic focus on the transference is a major treatment strategy), in the psy-
chodynamic psychotherapy of borderline patients transference analysis is modified by
attention to initial treatment goals and present external reality.
Thirdly, insofar as interpretations require a position of technical neutrality (the
therapist's equidistance from the forces in mutual conflict in the patient's mind),
technical neutrality is an important aspect of the psychodynamic psychotherapy of
borderline patients-as well as of standard psychoanalysis. However, given the severe
acting out of borderline patients inside and outside the treatment hours, technical neu-
trality may have to be limited by indispensable structuring (limit setting) of the treat-
ment situation, which (at least temporarily) reduces technical neutrality and requires
its reinstatement by means of interpretations of the reasons for which the therapist
moved away from a position of technical neutrality.
Our strategy also requires a set of tactical considerations regarding the inter-
ventions in each treatment hour, which give a particular coloring to this psycho-
therapy that differentiates it both from standard psychoanalysis and from supportive
psychotherapy.
In contrast to supportive psychotherapy, the therapist refrains, as much as pos-
sible, from technical interventions such as affective and cognitive support, guidance and
giving advice, direct environmental intervention, and any other technical maneuver that
would reduce technical neutrality-with the exception of necessary structuring or limit
setting in or outside the treatment hours. As mentioned before, whenever such limit
setting becomes necessary, it also requires a systematic effort to analyze the reasons
for setting these limits, the interpretation of the underlying transference conflicts, and
the gradual resolution of the need to set such limits by means of interpretative reso-
lution of these underlying conflicts. In contrast to supportive psychotherapy, transfer-
ences are not utilized for enhancing the therapeutic alliance, patient compliance,
or symptom resolution, but both positive and negative transferences are interpreted
(with the exception of milder aspects of the positive transference, which may be left
untouched, particularly in early stages of the treatment, in order to foster the thera-
The Psychotherapeutic Treatment of Borderline Patients 171
peutic alliance). Our treatment approach includes a systematic effort to interpret

the patient's primitive idealizations of the therapist because of their counterpart to
dissociated primitive negative transferences.
In each session, it is important to assess the patient's capacity to differentiate
fantasy from reality, and to carry out interpretation of unconscious meanings only after
the confirmation of commonly shared views of reality on the part of the patient and
therapist. This may require consistent and tactful confrontation of the patient with
immediate reality before interpreting its unconscious meaning. The patient's attribu-
tion of fantastic meanings to the therapist's interpretive interventions also needs to be
clarified and interpreted. It is important to assess secondary gain of severe symptoms
and behaviors, to interpret such secondary gain, and, if necessary, reduce or eliminate
it by limit setting, with the corresponding need to reassess and interpret any slippage
of technical neutrality. The analysis of unconscious sexual conflicts must include the
analysis of contamination of sexuality with aggression in order to help the patient to
free his sexual behavior from the control by aggressive impulses.
The interpretation of unconscious meanings must often start out being formu-
lated in a metaphorical, "as if', or atemporal way before linking it with a particular
aspect of the patient's past. For example, to say "It is now as if an enraged child were
defiantly provoking his father to show that he is not afraid of him" may be a better way
of initiating an interpretation that links present and past than a direct implication that
the patient is concretely repeating an experience of the past. Premature and forceful
interpretation of the unconscious past may reinforce the confusion between present
and past and induce transference psychosis, not to mention the danger of artificially
indoctrinating the patient. In contrast, the formulation of interpretations in a metaphor-
ical, atemporal, or "as if' mode facilitates the emergence of new, previously unsuspected
material.
Another crucial aspect of tactical interventions is the need to interpret primitive
defenses systematically as they emerge in each hour: the interpretation of primitive
defenses tends to strengthen reality testing and overall ego functioning. In contrast to
previous conceptions of primitive defenses, these do not "strengthen" the frail ego of
the borderline patient but are the very cause of chronic ego weakness. Interpretation
of primitive defenses is a major tool for increasing ego strength and reality testing, and
facilitates the interpretation of primitive transferences. In this connection, primitive
transferences and more advanced or realistic transferences that characterize later
stages of the treatment may alternate, so that patients sometimes present primitive and
advanced transferences within the same session. The general principle is to interpret,
as much as possible, primitive transferences before the advanced ones.
Given the severe tendencies toward acting out on the part of borderline patients,
dangerous complications in their treatment may derive from their characterologically
based, "nondepressive" suicide attempts, drug abuse, self-mutilating and other self-
destructive behaviors, and aggressive behaviors that may be life-threatening to them-
selves and others. An important aspect of each session, and not just a part of the overall
structuring of the treatment, is the assessment of whether there are emergency situa-
tions that require immediate interventions. On the basis of our general treatment strat-
egy and the specific experiences in the treatment of severely ill borderline patients, we
have constructed the following set of priorities of intervention that reflect the need to
assess, diagnose and treat these and other complications.
A threat of imminent suicidal or homicidal behavior has the highest priority in
each session. If there seem to be immediate threats to the continuity of the treatment,
172 o. F. Kemberg
these constitute the second highest priority to be taken by the therapist. If the patient
appears to be communicating in deceptive or dishonest ways, this constitutes the
third highest priority: psychodynamic psychotherapy demands honest communication
between patient and therapist, and, by the same token, the interpretation of the trans-
ference meanings that underlie the patient's dishonesty or deceptiveness. Acting out in
the sessions as well as outside the sessions constitutes the next highest priority, already
signaled in our earlier statement that transference analysis must include the con-
sideration of dominant conflicts outside the hours.
With these priorities considered, the therapist may then concentrate fully on the
analysis of the transference along the lines already outlined. There are times when the
dominant affects in the hours are linked with developments outside the hours: that is,
"affective dominance" may not always center in the transference, and if that is so, affec-
tive dominance determines the focus of the therapist's attention-with an awareness
that affectladen conflicts outside the treatment situation also may have transference
implications that may become clearer and dominant later on. Finally, trivialization of
the patient's communication requires attention: the therapist's sense that "nothing is
going on" in the hours implies either trivialization, deceptiveness, a dominance of nar-
cissistic resistances, or perhaps a period of respite and working through-or simply the
therapist's temporary disorientation. The therapist has to decide which of these causes
of real or apparent trivialization in the hours he should take up.
This psychodynamic psychotherapy for borderline personality organization
expects the therapist to see the patient in face-to-face interviews, to maintain his natural
behavior but to refrain from communicating anything about his private life to the
patient as part of the general psychodynamic principle of not gratifying transference
needs but analyzing them. To be technically neutral and to maintain a natural, objec-
tive stance while remaining "in role" does not mean "studied indifference": to the con-
trary, the therapist must examine his own emotional reactions to the patient on an
ongoing basis in order to detect undue involvement or undue distancing on his part
and to be able to diagnose any countertransference implications. The analysis of
what Racker (1968) described as concordant and complementary identifications in the
countertransference is an important aspect of the psychodynamic psychotherapy of
borderline patients. We assume that, particularly under conditions of complementary
countertransference, (that is, when the therapist's countertransference reflects his
unconscious identification with a repressed, dissociated, or projected self representa-
tion or object representation of the patient,) the diagnosis of this identification in the
countertransference is an extremely helpful tool for diagnosing the dominant primitive
object relation activated in the transference. The therapist's utilization of the counter-
transference consists in his silent analysis of the corresponding meanings, whether it
stems from the patient or from himself, and the utilization of this understanding in his
interpretative interventions. Rather than communicating the countertransference
directly to the patient, the therapist utilizes it in the formulation of his transference
interpretations.
Patients with borderline personality organization, as is true for all patients with
severe character pathology, tend to communicate significantly in nonverbal ways-in
addition to the content of their verbal communications. The therapist must be alert to the
content of the patient's communications, to his nonverbal communication, and to the
total atmosphere expressed in the moment-to-moment relationship between patient and
therapist. We also assume that an underlying, nonverbalized, constant aspect of the
relationship between patient and therapist is an important potential "channel" for trans-
ference communication, together with the moment-to-moment shifting verbal and non-
verbal communications. The therapist has to be alert to all these sources of information:
they constitute the raw material for his constructions of interpretative interventions. This
therapeutic alertness toward the total communication of the patient is particularly
helpful when the sessions have prolonged silences, a resistance that may be interpreta-
tively resolved by the analysis, at such points, of predominately nonverbal and general
relational "channels" of communication.
Our approach also considers the analysis of severe regressions in the transfer-
ence, particularly strong paranoid regressions, microparanoid psychotic episodes, and
transference psychosis in general. In so far as under such conditions reality testing may
be temporarily lost in the therapeutic communication, a tactical intervention mentioned
before, that is, to clarify reality fully before interpreting the unconscious elements in
the transference now becomes operational. Therefore, when the patient is deeply
regressed, most of the psychotherapeutic interventions may be focused on analyzing
the distortions of immediate reality in the patient-therapist interaction and their poten-
tial meanings. Transforming a breakdown of common reality boundaries into the
hypothesized activation of an unconscious object relationship involving patient and
therapist at that point is one more application of step 1 of the strategic analysis of
primitive transferences.
Advanced stages of the psychodynamic psychotherapy of borderline patients
resemble the psychodynamic psychotherapy and even the psychoanalytic treatment of
neurotic patients. At such advanced stages of the treatment more subtle and specific
character pathology and corresponding transference resistances may be explored and
resolved, particularly the more subtle aspect of narcissistic pathology. Severe negative
therapeutic reactions, particularly in patients with significant narcissistic pathology may
complicate the treatment, and require focusing once more on the more primitive trans-
ference developments (for example, unconscious envy of the therapist) implied in these
reactions. In short, the analysis of primitive transferences, while gradually receding
throughout the duration of the treatment, still remains as a therapeutic task to the very
end of the treatment.
3. THE APPROACH OF GERALD ADLER
Adler (1985) divides treatment of borderline psychopathology into three succes-

sive phases. In phase I, the therapist tries to establish and maintain a dyadic relation-
ship in which he can be steadily used over time by the patient as a holding selfobject.
The aim is to allow the patient to acquire a solid evocative memory of the therapist as
sustaining holder, which serves as a substrate out of which the patient can then form
adequate holding introjects.
The inevitable development of rage, however, interferes with this process. This rage
has three sources: first, holding is never enough to meet the patient's felt need to assuage
aloneness, and the patient is inclined to vengefully destroy the offending therapist.
Hence, the patient fears that he has lost or killed the therapist and that he is threatened
by the therapist's responding to his rage by turning from "good" to "bad." Second, the
frustrating holding selfobject is distorted by projection of "hostile introjects," so that the
patient "carries out what he experiences as an exchange of destructiveness in a mutually
hostile relationship" (p. 50). Third, the patient's envy of the object that is so endowed with
the potential for holding sustenance produces hateful destructive impulses toward the
174 o. F. Kemberg
therapist. Finally, patients may also feel a primitive gUilt because they feel undeserving
of the therapist's help owing to their own evilness.
Treatment in phase I aims to reduce these impediments by means of clarification,
confrontation, and interpretation. The therapist, as a holding selfobject, provides "tran-
sitional objects" (a term Adler uses to refer to the therapist's providing the patients,
for example, with vacation addresses and postcards), telephone calls, and extra appoint-
ments to reaffirm that the therapist continues to exist. The patient thus learns that
the therapist is an enduring and reliable holding selfobject, or that the therapist is
indestructible as a "good object" (here Adler refers to Winnicott [1958]).
In phase II, the major task is to help the patient, by means of what Kohut (1971) has
called "optimal disillusionment," to gradually become aware of the unrealistic aspects of
the holding introjects established in phase one. While the patient is still heavily depen-
dent on a continuing relationship with holding selfobjects, he needs to be weaned for
a viable setup for adult life "in which selfobjects cannot realistically be consistently
available and must over the years be lost in considerable number" (p. 59).
In phase III, to help the patient become optimally autonomous in regard to secure
holding and a sense of worth, the focus is on helping him develop a realistic superego
that is not inappropriately harsh. Also, the patient's ego must develop a capacity for
pleasurable confidence in the self and for directing love toward the self that is similar
to the affectionate nature of object love rather than narcissistic love.
4. THE APPROACH OF VAMIK VOLKAN
Volkan (1987) agrees, essentially, with Kernberg's (1975, 1984) formulations

regarding the predominance of primitive defensive operations in these patients, particu-
larly what Volkan calls "defensive splitting"-in contrast to "developmental splitting"-
as a central feature of the most severely ill borderline patients. Volkan points out that
splitting is not limited to the borderline patient, and that it finds expression in adult life,
particularly under socially facilitating circumstances involving ethnicity, nationality, etc.
By the same token, borderline patients also present repressive mechanisms together with
splitting although they utilize splitting more than repression. Regarding the question
of etiology he also agrees with Kernberg regarding the condensation of Oedipal and
preoedipal conflicts in the borderline patient.
Volkan focuses particularly on the psychosis-prone borderline patient, and dis-
cusses nine such patients in some detail. These nine patients were treated over several
years, seven of them successfully, and all nine showed marked improvement in stabi-
lizing their object relationships, vocations, and learning to tolerate being alone. The
seven patients who were successfully treated showed, in addition, significant resolu-
tion of their object conflicts, advanced along the preoedipal-oedipal continuum, and
gradually shifted into a more advanced defensive organization.
Volkan specifically tolerates and welcomes a transference regression in his
patients, acknowledging his debt to Boyer (1983), Giovacchini (1979,1986), and Searles
(1979,1986). His assumption is that such regression brings about a redifferentiation of
self and object representations, similar to what occurs in a transference psychosis. This
Redifferentiation is then followed by a progressive development in which self and
object representations are differentiated, and the patient experiences "developmental
splitting" in the transference instead of the previous defensive splitting.
Volkan outlines six steps in treatment. Step one is establishing with the patient a
reality base for the treatment. Volkan provides explicit information about the nature
of the treatment, the task for the two participants, and sets necessary limits in a tactful,
"non drastic" way.
Step two refers to the appearance of the first split transference. Here Volkan is
alert to early manifestations of splitting in the transference, brings his observation to
the patient's attention without genetic interpretations or an active effort to resolve this
splitting process, and concentrates on preserving the therapeutic relationship, and main-
taining a "holding function" by focusing on the patient's psychic operations in the
here-and-now. The therapist helps the patient see how sharply split percepts of others
are a characteristic phenomenon in the patient's experience, without attempting to
interpret and resolve them at this stage. He also agrees with Rosenfeld (1987) in
avoiding interpretation of manifest Oedipal material at this stage, which might serve a
defensive function as an "upward resistance."
Volkan also pays attention to the patient's experience of the therapist's analyzing
functions in order to clarify significant distortions in the perception of the therapist.
Step three is represented by a focalized psychotic transference leading to reacti-
vated and transference-related transitional phenomena. It is at this point that Volkan
stresses the importance of tolerating and facilitating psychotic regression in the trans-
ference. He uses the concept of psychotic "therapeutic stories," an affectively lived
drama, a here-and-now version of a real or fantasized event in the past that includes
considerable action inside and outside the sessions. Now transference psychosis and
delusional relatedness may emerge, processes that may last for a few weeks or months.
Optimally, the patients now develop new transitional objects or phenomena, which
have the potential to become a new bridge to reality. Volkan here uses concepts of
Winnicott (1953, 1956, 1960) and Greenacre (1970).
Step four refers to the second split transference, which now becomes the focus of
the work, and brings the possibility of mending the patient's opposing object-relations
units. At this point, the interpretation of the meaning of splitting of all-good from all-
bad images includes genetic material as it appears in the transference, in the patient's
daily activities, and in dreams. The emphasis now is also on identification with the
analyst'S integrative functions.
Step five consists in the development of the transference neurosis. The patient
may develop profound depression as part of the integration of mutually split-off object
representations in the transference, and the vicissitudes of the Oedipus complex begin
to dominate the material. Volkan has found that with the development and resolution
of these issues patients show an increased capacity for repression, and that some of the
elements of split-transference manifestations that are not mended are repressed.
Finally, step six includes the third split transference and termination. It is a stage
of repetition and working through of previously explored issues. This step approximates
that of classical analysis and working through of the termination phase of treatment.
Volkan quotes Modell's (1976) writing on the termination stages with patients with nar-
cissistic character disorder; he agrees that, in a similar way, borderline patients may
present atypical terminations characterized by the return of primitive splitting.
5. THE APPROACH OF HAROLD SEARLES

Perhaps the most central concept in Searles' (1986) formulations is the idea that
the borderline patient, by means of projective identification, projects his pathogenic
176 O. F. Kernberg
introjects onto the therapist, who, experiencing the consequences of these projective
mechanisms in his countertransference, tends in turn to activate, as part of his coun-
tertransference regression, his own primitive dissociated introjects. The condensation
of projected transference elements (the patient's introjects) and the therapist's reacti-
vated transference dispositions (the therapist's primitive layer of pathogenic introjects)
permits the therapist to identify with experiences that the patient cannot tolerate in
himself, and to share with the patient his understanding of the nature of these primi-
tive introjects. In interpreting the transference, the therapist shares with the patient
aspects of his countertransference that reflect his understanding and acceptance of this
joint world of projected and reactivated introjects.
On the negative side, Searles goes on, the danger exists that the therapist may not
be able to tolerate the reactivation of his own past introjects within his counter-
transference. This may lead to premature interpretations of the transference and a
consequent perpetuation of projection and reprojection by interpretation, leading to
stalemates in the treatment. Another negative possibility is that the therapist may
become fixated in his own countertransference regression and fail to extricate himself
interpretively from that situation.
Searles warns against the danger of "brainwashing" the patient with premature
interpretations, but also refers to the patient's temptation to project his own brain-
washing tendencies onto the therapist. He stresses the need for the therapist to accept
the countertransference consequences of the patient transferences as an indispensable
precondition for the patient's eventual acceptance of the therapist's interpretations.
Searles offers abundant clinical evidence throughout to illustrate his proposal that
every bit of psychopathology in the patient has a counterpart somewhere in the
therapist's functioning, and that there are always nuclei of realintuitions in the trans-
ference developments.
Another, and related, central concept is that of the therapeutic symbiosis as an
indispensable precondition for psychotherapeutic work to proceed. Searles's concept
of symbiosis is a relatively broad one, including both actual merger phenomena (such
as are characteristic of transference regression in schizophrenic patients) and those
confusions or interchanges of aspects of their personality that are products of projec-
tive identification operating mostly in the patient but also in the therapist.
The therapeutic symbiosis, in Searles's view, includes the patient's multiple
identifications with self and objects, projected and introjected, and is characterized
by profound ambivalence because of the mutually contradictory nature of these iden-
tifications. This ambivalent symbiosis serves both as a defense against threatening prim-
itive aggression associated with the relationship to pathogenic introjects, and as a
potential for emotional growth as tolerance and working through of these mutually
split-off internalized object relations are achieved in the transference experience
and analysis. Searles illustrates the many and complex forms these transference-
countertransference developments may take: for example, the patient's defensive
detachment with the therapist's approaching vacations may reflect, in part, his identi-
fication with the analyst'S defensive detachment because the latter cannot fully
acknowledge his own dependency on his patient. The patient's narcissistic withdrawal
therefore interacts with the analyst'S defenses against his own dependency needs
toward the patient as a transference object. Or perhaps, oedipal-triangular relation-
ships may be played out by the patient's jealousy of a thriving plant in the therapist's
office. At times the therapist who feels "irrelevant" to his patient may represent the
patient's projected own "irrelevant" self. The therapist, in short, may become the
patient's self, his multiple introjects, live and dead, human and nonhuman. Searles
reminds us of his earlier work regarding the differentiation of the animate and inani-
mate world on the basis of the normal introjection of mother during the symbiotic
period of development; he points to the deepest levels of loss of orientation toward
external reality when the most primitive maternal introjects are split off and projected.
Searles questions some standard assumptions of psychoanalytic psychotherapy,
such as the autonomous development of the "real" relationship to the therapist in the
early years of treatment. He points out that this "real" relationship more often than
not may be a split-off, symbolically meaningful transference relationship that differs
from other, dissociated, transference relationships: the "real" relationship can emerge
only in advanced stages of the treatment, as the consequence of working through
split-off, primitive object relations in the transference.
In an attempt to provide some general characteristics of Searles's technique, I
would stress his emphasis on the communication of countertransference; his very
careful and gradual development of transference interpretations against the back-
ground of the tolerance and silent analysis within the therapist's mind of the symbiotic
relationship; a general tendency to tolerate lengthy silences; a long-range time frame
for giving the patient the "space" to experience himself as different from the therapist
in the relationship, to gradually hear more realistically what is being said, to listen
to it, and to associate to it; and the therapist's parallel process of growth in his
capacity to hear, listen, and understand. Searles's emphasis, that one can interpret only
within the symbiosis-to do otherwise is to risk sterile intellectualizations-is the
counterpart to his stress on the therapist's need to tolerate split-off internalized, pro-
jected object relations from the patient in his countertransference before interpre-
tation can proceed.
Searles explores the very limits of the psychotherapeutic relationship in propos-
ing that the patient can abandon his illness only if the therapist has come to cherish it
too, so that the patient's improvement becomes an experience of loss for both patient
and therapist: this idea touches on some of the most complex and subtle aspects of the
developments in long-term treatment of severely regressed patients. He also suggests
that, just as the therapist must become the patient's mothering shield against the
outside world-so that the patient can submerge himself protectively within the
therapist-the therapist must be able to submerge himself in the patient, the patient
thus assuming the function of the maternal shield for the therapist as a precondition
for the patient's abandonment of his own autistic stance. In other words, within the
therapeutic symbiosis, the patient, by identification with the therapist, must accept his
roles as both protected infant and protective mother who helps him to differentiate
and individuate.
6. RECENT DEVELOPMENTS IN MY APPROACH:

THE TRANSFERENCE IMPLICATIONS OF SEVERE
SUPEREGO PATHOLOGY
Some degree of superego pathology is frequent in the personality structure of

patients with borderline personality organization. The severe distortions in ego devel-
opment, the predominance of splitting mechanisms, and the lack of integration of
internalized object relations influence the setting up of various layers of the superego.
The normal integration and mutual toning down of idealized pre oedipal superego
178 o. F. Kemberg
precursors and of aggressive, persecutory pre oedipal superego precursors fails to some
extent, with a consequent weakening of the internalization of the later, more realistic
superego internalizations of the Oedipal period. The effects of these distortions in
superego developments include some degree of failure to develop stable, integrated
value systems, the ordinarily solid fundament of built in autonomous morality.
These superego failures cause a weakness of normal superego regulation by
means of differentiated self-criticism, a tendency toward regulation of self-esteem by
violent mood swings, and a dissociation between states of intense exaggerated guilt feel-
ings and despair, on the one hand, and, on the other, rageful, inconsiderate, self-serving
flaunting of ordinary considerations for other peoples' rights and for objective consid-
erations of fairness and justice. In addition, borderline patients may present some
degree of passive or parasitic and aggressive antisocial behavior, which is of particular
importance regarding the prognosis for their psychotherapeutic treatment. In fact, as I
have stressed throughout my contributions to this field, the quality of object relations
and extent to which antisocial features are present are the two overriding prognostic
factors in the treatment of borderline personality organization.
Superego pathology becomes an important issue in the treatment strategy, par-
ticularly in dealing with transferences that evolve as a consequence of its expression.
As mentioned before, in our research project on the psychodynamic psychotherapy of
borderline patients, we have established a list of priorities of subject matters that
require urgent attention. The highest priority is dealing with imminent threats to the
life of the patient, of other individuals, or the therapist, so that violent and potentially
destructive behavior needs to be focused upon and dealt with first.
A second priority is represented by indications that, probably, the treatment is
about to be disrupted, and it is essential that the therapist focus on the transference
implications of such potentially imminent breakdown of the therapeutic relationship.
In my view, two major factors involved in the high rate of early dropout in the psy-
chotherapy of borderline patients are lack of attention to setting up the initial treat-
ment contract, and lack of attention to the transference implications of threats of
disruptions of the treatment.
The third highest priority, finally, is the evidence of deceptiveness in the patients'
communication, that is, conscious suppression or alteration of essential information to
the therapist, so that the therapist must necessarily be in error in his assessment of the
patient's present emotional state and reality. This deceptiveness may take the form of
suppression of information, of feeding the therapist false data, that is, outright lying,
and/or manipulative behavior intended to disorient the therapist or to exploit him in
some way. All of these behaviors are carried out in clear consciousness by the patient,
and are not a consequence of unconscious denial or confusion.
It is striking how difficult it is for therapists to acknowledge to themselves and to
their patients that their patients are lying to them or treating them in a dishonest, decep-
tive way. Typically, patients who develop such behaviors also project such tendencies
onto the therapist. Indeed, the more dishonest the patient, the more dishonest he
believes his therapist to be, and the less he can trust what his therapist says to him. In
some cases, verbal communication is vitiated to such an extent that it become a
mockery of ordinary psychotherapeutic communication.
I have coined the term psychopathic transference to refer to periods in the treat-
ment when such conditions of deceptiveness prevail in the transference. In my view, it
is essential to explore such transferences in great detail, and to resolve them interpre-
tively before proceeding with other material (except the two higher priorities men-
tioned earlier). The psychopathic transferences tend to infiltrate and corrupt the entire
psychotherapeutic process, and is a major reason for psychotherapeutic stalemates and
failure. To treat a patient psychotherapeutically requires total communication between
patient and therapist, and it is for this technical reason-and not for any "moralistic"
one-that the therapist has to work on opening the field of communication by
resolving psychopathic transferences.
The therapist should share with the patient his concern at the patient's lying or
consciously distorting information. This confrontation may bring about an immediate,
angry attack on the part of the patient, who may in turn accuse the therapist of aggres-
sion or dishonesty. By means of projective identification and omnipotent control, the
patient may unconsciously try to provoke the therapist to behavior that the patient may
then interpret as dishonest.
These cases illustrate patients' intrapsychic conflicts between a desire for honesty
and a corruption of this desire, conflicts that usually reflect an unconscious identifica-
tion with a parental image perceived as profoundly inconsistent or dishonest. In cases
with narcissistic personality disorder, the enactment of a sadistically infiltrated, patho-
logically grandiose self that operates against the healthy, dependent part of the patient's
self constitutes a frequent dynamic underlying psychopathic transferences.
In other cases, a stubborn and silent protracted tendency to lying may defy the
therapist's efforts to explore the very reasons for this deceptiveness. Still other patients
may insist, over an extended period of time, that there are issues they will not discuss
with the therapist, which is honest and may permit the therapist to analyze the reasons
for that fearfulness and distrust.
In somewhat different yet related cases, there seems to be open communication,
except that the patient treats all other people with total ruthlessness and lack of
consideration, expects the psychotherapist to treat him in the same way, and acts as if
there were no such thing as an honest mutual commitment between two people. Here,
rather than deceptiveness, the patient's assumption is that any closeness or com-
mitment is deceptive, and that the therapist, by pretending to be interested-beyond
any financial, scientific, or prestige benefits he may gain from the patient-is
really dishonest. This may be an unconscious dynamic as well as a consciously experi-
enced fear.
What I have been describing are patients suffering from a deep corruption of the
capacity for closeness, dependency, emotional commitment, and love. Typically, follow-
ing the exploration of the origins of these psychopathic transferences and their effects
on the therapeutic relationship, they tend to shift, after a time, into a different trans-
ference disposition. The patient gradually begins to understand that complete openness
may be necessary for psychotherapeutic work to proceed, but that this exposes him to
the danger of rejection, criticism, and attack, as he sees it, on the part of the therapist.
The patient who, after maintaining crucial information secret for an extended period
of time, finally "confesses" what he did not dare to discuss before, typically experiences
fears of attack, depreciation, or abandonment on the part of the therapist. The patient
who treats all other people as "objects," typically fears that the abandonment of that
protective distance from the therapist will endanger his security. In short, psychopathic
transferences gradually are transformed into paranoid ones.
There are many patients with borderline personality organization whose pre-
dominately negative transferences contain strong paranoid elements from the begin-
ning of the treatment. Although, on the surface, severely paranoid borderline patients
may appear to be more difficult treatment challenges than more smoothly functioning
180 O. F. Kernberg
patients with psychopathic transferences, it is much easier to explore the correspond-

ing projective identifications of primitive internalized persecutory objects in the para-
noid transferences. What is of particular importance here is that, where such paranoid
transferences are the outcome of previously worked through psychopathic transfer-
ences, the paranoid elements may be particularly powerful and be expressed as serious
distortions in the therapeutic relationship, even to the extent of the development of
transference psychosis.
When this occurs, the technique I have described in earlier work (1984) to deal
with paranoid regression in the transference may be particularly helpful. If the patient
seriously distorts the reality of the therapist's behavior, the therapist may communi-
cate to the patient that, in the therapist's view, the reality of their interaction is com-
pletely different, and that, in some very specific ways, the therapist is convinced of
something diametrically opposite to the patient's conviction. Simultaneously, the
therapist should go on, he accepts the patient's conviction regarding that issue, so that
they have reached a state in which both are convinced of what turn out to be "incom-
patible realities." This reproduces the situation that occurs when a "mad" person and
a "normal" person try to communicate with each other without an outside witness or
arbitrator to clarify the reality of the situation. The only alternative, of course, is that
the therapist might be lying to the patient, and if the patient is convinced of that, that
would need to be explored further. The analysis of the meanings, to the patient, of the
assumed dishonesty of the therapist may lead directly to very primitive psychopathic
transferences and their antecedent object relationships in the patient's early life. In
other cases, the therapist now must examine the patient's paranoid regression in
terms of the activation of a "psychotic nucleus" in the transference, sorting out the
extent to which unresolved psychopathic transferences (the patient's assumption that
the therapist is lying) need to be reexplored later on.
In my experience, when the therapist communicates his tolerance of incom-
patible realities, and fully examines the patient-therapist relationship under such con-
ditions, it may gradually lead to the resolution of the psychotic nucleus and of the
paranoid transference itself. Paranoid transferences typically reveal the presence of
severe primitive aggression in the form of "purified" aggressive internalized part-object
relations, split off from the patient's idealized self and object representations.
The decrease of projective mechanisms and of paranoid transferences by inter-
pretation brings about a gradual recognition by the patient of the intrapsychic sources
of his own aggression, and the development, for the first time in the treatment, of
authentic experiences of guilt, remorse, concern for the therapist, and anxiety over the
possibility of repairing their relationship. The patient becomes aware that his attacks
were directed not at the bad, sadistic, tyrannical, or dishonest therapist, but at the good
therapist who was trying to help him. This development marks the beginning of depres-
sive transferences, characteristic of advanced stages of the psychodynamic psycho-
therapy of borderline patients, indicating that a significant degree of integration is
taking place. At this point in the treatment, patients begin to be able to reflect on the
implications of their own behavior, and to integrate the previously split off images
of the idealized and the persecutory therapist, in the context of also developing an inte-
grated view of their parental images in terms of the idealized and persecutory aspects
of their representations.
The most important problem, at this advanced stage of the treatment may be the
therapist'S unawareness of the beginning of change in the patient. The early manifes-
tations of such a depressive potential may show in the patient's more considerate
behavior toward others, in his sublimatory functioning outside the treatment situation.
The therapist may miss the development of such a new potential, especially in patients
who had evinced severely paranoid transferences over an extended period of time. One
additional reason why such an improvement may go undetected is because the patient
may develop negative therapeutic reactions from an unconscious sense of guilt. This is
a higher level negative therapeutic reaction than that which obtains in the case of
narcissistic personalities, in whom negative therapeutic reactions usually reflect un-
conscious envy of the therapist.
The most dramatic indicators of depressive transferences in the advanced stages
of the treatment of borderline patients will be the growing evidence of the patients'
capacity to empathize with feeling states of the therapist-sometimes they develop an
uncanny capacity to interpret the therapist's behavior-their concern for "maintaining
alive" what is being learned in the psychotherapy, their capacity for independent work
on the issues developed within the treatment outside the treatment hours, and their
expression of dependency on and love for the therapist rather than a superficial "as if"
show of cooperation in the search for additional gratifications.
The outline I have presented is necessarily schematic and oversimplified. Given
the periods of chaotic condensation of transferences from many sources and levels
of development in the treatment of borderline patients, depressive, paranoid, and
psychopathic features may coexist or intermingle. The importance of the outlined
sequence, however, lies in orienting the therapist to the order of priority in which he
should explore such chaotic transferences: I have found it extremely helpful to first take
up and resolve psychopathic transferences before focusing on the paranoid aspects of
the material, and then to resolve persistent paranoid elements before examining the
depressive developments in the transference.
REFERENCES
Adler, 0. (1985). Borderline Psychopathology and Its Treatment. New York: Jason Aronson.
Adler, 0. (1989). Psychodynamic Therapies in Borderline Personality Disorder. In Review of Psychiatry.
Volume 8. Washington, DC: American Psychiatric Press, pp. 49-64.
Boyer, L.B. (1983). The Regressed Patient. New York: Jason Aronson.
Buie, D. and Adler, 0. (1982). The definitive treatment of the borderline patient. International Journal of
Psychoanalytic Psychotherapy 9:51-87.
Chessick, R.D. (1977). Intensive Psychotherapy of the Borderline Patient. New York: Jason Aronson.
Giovacchini, P.L. (1979). Treatment of Primitive Mental States. New York: Jason Aronson.
Giovacchini, R.D. (1986). Developmental Disorders: The Transitional Space in Mental Breakdown and
Creative Integration. New York: Jason Aronson.
Greenacre, P. (1970). The transitional object and the fetish: with special reference to the role of illusion.
International Journal of Psychoanalysis 51:447-456.
Gunderson, 10. (1984). Borderline Personality Disorder. Washington, DC: American Psychiatric Press.
Kernberg, O.F. (1975). Borderline Conditions and Pathological Narcissism. New York: Jason Aronson.
Kernberg, 0.F. (1982). The psychotherapeutic treatment of borderline personalities. In Psychiatry 1982:
The American Psychiatric Association Annual Review. Edited by Grinspoon, L. Washington, DC:
American Psychiatric Press.
Kernberg, O.F. (1984). Severe Personality Disorders: Psychotherapeutic Strategies. New Haven: Yale Univer-
sity Press.
Kernberg, O.F., Seizer, M.A, Koenigsberg, H.W., Carr, AC., and Appelbaum, AH. (1989). Psychodynamic
Psychotherapy of Borderline Patients. New York: Basic Books.
Kohut, H. (1971). The analysis of the self. New York: International Universities Press.
Masterson, IF. (1976). Psychotherapv of the Borderline Adult. New York: Brunner/Maze!.
Masterson, IF. (1981). The Narcissistic and Borderline Disorders. New York: Brunner/Maze!.
182 o. F. Kemberg
Modell, A.H. (1976). The "holding environment" and the therapeutic action of psychoanalysis. Journal of the
American Psychoanalvtic Association 24:285-307.
Racker, H. (1968). Transference and countertransference. New York: International Universities Press.
Rinsley, D. (1982). Borderline and Other Self Disorders. New York: Jason Aronson.
Rosenfeld H. (1987). Impasse and Interpretation. London: Tavistock.
Searles, H.F. (1979). Countertransference and Related Subjects. New York: International Universities Press.
Searles, H.F. (1986). My Work with Borderline Patients. New York: Jason Aronson.
Volkan, V. (1987). Six Steps in the Treatment of Borderline Personality Organization. Northvale, NJ: Jason
Aronson.
Waldinger, R.I (1987). Intensive Psychodynamic Therapy With Borderline Patients: An Overview. In: The
American Journal of Psychiatry 144:3,267-274.
Waldinger, R.I and Gunderson, IG. (1987). Effective Psychotherapy with Borderline Patients. New
York:Macmillan.
Winnicott, D.W. (1953). Transitional objects and· transitional phenomena. International Journal of
Psycho-Analysis 34:89-97.
Winnicott, D.W. (1956). On transference. International Journal of PsychoAnalysis 37:386-388.
Winnicott, D.W. (1958). Collected papers: through paediatrics to psychoanalysis. New York: Basic Books.
Winnicott, D.W. (1960). The theory of the parent-infant relationship. In The Maturational Processes and the
Facilitating Environment, pp. 37-55. New York: International Universities Press, 1965.
14
FUNCTIONAL ANALYSIS OF BORDERLINE

PERSONALITY DISORDER BEHAVIORAL
CRITERION PATTERNS
Links to Treatment
Jennifer Waltz1* and Marsha M. Linehan2
1 University of Montana
2 University of Washington
INTRODUCTION
Borderline Personality Disorder (BPD), as an entity, has a history of being

associated with a wide-ranging set of diagnostic criteria and symptoms. A variety of
theories, primarily psychodynamic, have attempted to explain the interpersonal and
self-related difficulties clients with BPD experience (Gunderson, 1984; Kernberg, 1975).
Linehan's (1993) Dialectical Behavior Therapy (DBT) and her formulation of BPD are
centrally informed by a behavioral model. DBT seeks to understand the behavior of
people meeting criteria for BPD using a functional analytic approach. Although fre-
quently confusing and self-destructive, the behavior of individuals with BPD is viewed
as following the same behavioral principles that govern all behavior, and particular
emphasis is placed on understanding the behavior within the context.
Linehan's (1993) biosocial theory also posits a central role for emotion dysregu-
lation in the etiology and maintenance of BPD. This transactional model suggests that
biologically based emotion dysregulation interacts over time with an invalidating
environment to produce the symptoms associated with BPD. Linehan (1993) further
elaborates on several behavioral patterns associated with BPD that stem directly
from emotion dysregulation and the experience of pervasive invalidation. These behav-
ioral patterns integrate difficulties in the areas of emotion regulation, interpersonal
* Address correspondence to: Jennifer Waltz, Ph.D., Department of Psychology, University of Montana,
Missoula MT 59812, Phone: (406) 243-4523, Fax: (406) 243-6366, e-mail: [email protected]

184 J. Waltz and M. M. Linehan
behavior, sense of self, and other behavioral deficits to provide a conceptualization of

many of the problematic behaviors people with BPD experience. This chapter provides
an overview of the biosocial theory informing DBT, a description of the behavioral pat-
terns identified by Linehan (1993), and a behavioral formulation of those patterns. An
overview of DBT, the role of the behavioral patterns and how they are addressed in
treatment is also discussed.
The development and testing of DBT represent perhaps the first large-scale, sys-
tematic attempts to empirically validate a treatment for chronically parasuicidal clients
with BPD using randomized clinical trials (Koons, Robins, Bishop, Morse, Tweed,
Lynch, Gonzalez, Butterfield, and Bastian, 1998; Linehan, Armstrong, Suarez, Allmon,
and Heard, 1991; Linehan, Dimeff, Comtois, and Kantor, 1998). This approach contin-
ues to be tested, and to show promise in its application to a range of difficult-to-treat
populations (Linehan and Dimeff, 1995).
BIOSOCIAL THEORY
Linehan (1993) has developed a biosocial theory that incorporates both biologi-
cal and environmental components to explain a variety of behaviors commonly asso-
ciated with BPD. This theory posits that the various maladaptive behaviors that make
up BPD represent different types of dysregulation. These maladaptive behaviors reflect
ways that the individual is emotionally, cognitively, interpersonally, and/or behaviorally
dysregulated. Because emotional responses are understood to be full system responses,
including not only the phenomenological experience and physiological response, but
also the cognitive, behavioral, and/or action urge responses, emotion dysregulation in
particular plays a central role in the biosocial theory underlying DBT.
EMOTION DYSREGULATION
People experiencing emotion dysregulation have overwhelming, unpredictable,

and uncontrollable emotions. They frequently experience very painful, "negative"
emotions, without being able to reliably predict when these will occur, or to effectively
modulate or change them. The person's emotions are intense and out of control, often
precipitating extreme behaviors that are the only effective means of getting relief. This
conceptualization is supported by empirical evidence that individuals meeting criteria
for BPD experience elevated levels of affective instability (Stein, 1996).
Emotion dysregulation involves the person a) being extremely vulnerable to
emotion-generating stimuli, and b) having poor emotion regulation capacity. Sensitiv-
ity to emotion-generating stimuli is characteristic of most individuals with BPD. They
have emotional responses to events, and interactions that the average person would
have very little or no reaction to. When emotional reactions are triggered, they tend to
be intense and long-lasting. The level of physiological arousal, experiential distress,
cognitive involvement, and intensity of action urges all are likely to be strong. Unlike
the average person, for whom many emotional reactions are transient, the borderline
individual is likely to experience the emotion and all its concomitants for a relatively
extended period of time. Rather than having a sense that "this will pass," there is a
sense that the emotion may last "forever", or become unbearable. The biosocial theory
contends that these emotional reactions, though in response to more subtle stimuli than
Functional Analysis of Borderline Personality Disorder Behavioral Criterion Patterns 185
the average person would respond to, are nonetheless "real" in the same sense as the
non-dysregulat..:d individual's emotional responses are "real."
In addition to being vulnerable to emotions, emotionally dysregulated individ-
uals also have difficulty regulating or modulating their emotions. In other words, they
have difficulty changing their emotional state from an intensely positive or negative
one, to a more neutral one, or from a negative state to a positive state. There are a
number of ways to modulate emotional responses; for example, through changing
physiology directly, changing focus of attention, allowing the emotion to occur without
attempting to block it, or responding behaviorally in a way that is likely to induce some
alternate emotional state. People engage in these types of behaviors regularly as they
cope with emotional responses that occur throughout their daily lives. For example,
if I receive news on a voice mail message that makes me angry, I may decide to focus
my attention on other calls before dealing with that particular one, I may take a deep
breath and relax my shoulders and neck, I may allow myself to experience the anger
until it subsides, or I may decide to continue with a task that needs to be done.
Alternatively, I may take a more cognitive approach and tell myself that I can handle
whatever the problem is, that it is solvable, that it is not a crisis. Individuals with BPD
generally have a great deal of difficulty employing any of these approaches to modu-
lating their own affect. This is probably the case for a number of reasons. First, the
intensity of the emotional responses borderlines experience makes those responses dif-
ficult to modulate. Clearly it is easier to modulate or change a low-level or subtle emo-
tional reaction than an intense or strong one. Second, most individuals with BPD have
not learned effective means of modulation, because of deficits in their family environ-
ments growing up. Learning how to regulate emotions is an important childhood
developmental task, and cannot be accomplished without the necessary input from the
environment. Third, many BPD individuals have learned maladaptive means of mod-
ulating negative affect; these means tend to be quite powerful, but inconsistent with
more adaptive responses. For example, drug use, suicidal behavior, risky sexual or crim-
inal behavior, gambling and the like can be powerful means to modulate negative affect.
These methods tend to be more "effective" in the short run than many more adaptive
alternatives, and consequently are difficult to change. Finally, the dysregulated behav-
ior of the borderline is likely to increase the likelihood that more painful emotional
experiences will occur. For example, the BPD individual's dysregulated behavior is
likely to produce problematic interpersonal situations that lead to further negative
affect. This cyclic pattern can be very difficult to break out of.
The biosocial theory suggests that emotion dysregulation is caused and main-
tained, at least in part, by some biological factor. No one particular etiological factor
or pathway is proposed to cause emotion dysregulation, rather, it is likely that there
are a wide variety of ways that emotion dysregulation may occur. These may include
genetic factors, compromised prenatal environment, or severe, early trauma that may
have a physiological impact. Further research is needed to determine whether emotion
dysregulation in BPD individuals actually has a biological basis, and what the causal
factors are.
INVALIDATING ENVIRONMENT
Linehan's biosocial theory proposes that a certain type of context is required, in

addition to the presence of emotion dysregulation, in order for a person to develop the
behaviors that make up BPD. This context has been labeled the "invalidating environ-
ment". The invalidating environment, as described by Linehan (1993), is one in which
the person does not receive basic communications of the reasonableness or appropri-
ateness of his/her responses. Instead, the environment communicates that the person,
his/her behavior, thoughts, feelings, or responses are flawed or pathological. Receiving
validation is probably one of the most basic, essential parts of healthy development.
Children must learn that their private experiences (thoughts, feelings, etc.) are reliably
and meaningfully providing them information about what is happening in the envi-
ronment. Through validation, the child learns to see his/her reactions to the world as
valid indicators of what is real or true.
In contrast, an invalidating environment results in the opposite experience. The
child is told that his/her reactions are not like other people's, and not realistic or appro-
priate. The family communicates that the child's private experience is faulty and not
to be trusted, and that valid indicators of how to respond can only be found in the
external world, through watching others. The child is likely to feel ashamed of
his/her responses, confused about how to respond, and a sense that he/she is somehow
"different" from important others.
In addition, invalidating families tend to minimize problems, and suggest that they
should be easy to resolve. They emphasize cognitive control of emotion, for example,
taking a "just don't think about it," or "look on the bright side" approach to painful or dif-
ficult life experiences. These approaches may be effective in modulating emotions for
some people, but tend not to work for the emotionally dysregulated child. Consequently,
he/she does not learn more effective means of changing his/her emotional state.
The invalidating environments that many BPD clients grow up in are particularly
likely to invalidate self-generated behavior. This type of invalidation has some unique
and important consequences. Self-generated behavior can be understood as behavior
that is not under the influence of aversive forms of control, and that is inherently rein-
forcing for the individual. Although it may be shaped by interpersonal consequences,
it is behavior that, even in the absence of immediate interpersonal reinforcement or
punishment, would be naturally reinforcing to the individual. For example, if a person
enjoys playing piano, writing, giving gifts, or hiking, these are probably self-generated
behaviors for that person. It may be that the person has received praise, criticism, or
other social consequences for playing these behaviors, but they can be understood to
be self-generated if, had no other contingencies ever come into play, the person would
have continued to engage in them because of their inherently reinforcing properties.
The punishment of self-generated behavior seems to have a profound effect on the
development of sense of self. Much of what we might think of as sense of self has to
do with our understanding of our own self-generated behaviors and responses. If these
responses are punished, the person does not have access to knowledge of their own
likes, wants, and other dispositional responses.
Although much of the invalidation reported by clients occurs within the family
context, invalidation also occurs on a broader social scale. Being part of a marginalized
or devalued group results in chronic, pervasive invalidation on a societal level. For
example, research suggests that being part of a marginalized group, namely being a fat
woman, puts one at increased risk for borderline symptomatology (Sansone, Sansone,
and Fine, 1995). Although these data are correlational, it is interesting to note that
borderline symptoms are associated with one of the more accepted forms of bias in our
culture, bias against fat people. This may be due to the invalidation that accompanies
this prejudice.
BIOSOCIAL THEORY AS A TRANSACTIONAL MODEL
The biosocial theory underlying DBT is a transactional one: biological and envi-
ronmental factors are proposed to affect each other in a continuous, on-going process.
Neither factor is sufficient to produce a pattern of BPD behaviors, both are viewed as
crucial to its development. In addition, emotion dysregulation and the invalidating envi-
ronment are understood to interact in an on-going way, each affecting the other over
time. For example, a child who is born with some biological vulnerability to emotion
dysregulation is likely to be much more difficult for the family to cope with than chil-
dren who are not similarly dysregulated. The family may single the child out as the
"problem child," the one who is somehow defective or bad because of the intensity of
his/her emotional responses and concomitant behavior. This labeling by the family is
typical invalidating behavior, and may be experienced by the emotionally dysregulated
child as particularly painful and hard to cope with, resulting in an exacerbation of dys-
regulated behavior. The escalation in dysregulated behavior serves to verify the family's
view of the child. They become increasingly frustrated with the child and escalate their
own invalidation of him/her in attempts to further control his/her behavior. Conse-
quently both the environment and the biological vulnerability of the child interact
in a continuous flow of influence in both directions, often serving to exacerbate the
problems on either side.
BPD AS A DISORDER OF DYSREGULATION
From a DBT perspective, the organizing feature of the various maladaptive

behaviors associated with BPD is that they all represent forms of dysregulation.
Linehan (1993) identifies each BPD symptom with a particular type of dysregulation.
Affective lability and difficulties with anger represent forms of affective dysregula-
tion. Linehan (1993) notes that these two manifestations of emotion dysregulation are
probably not the only ones, and it is somewhat unclear why the diagnostic criteria focus
in particular on anger. Many BPD clients seem to have as much or more difficulty with
other intense emotions, such as shame or sadness. From a DBT perspective, the central
feature here is that the person has difficulty with affective regulation, and the role and
importance of specific emotions is assessed on an individual basis.
ImpUlsive and suicidal behaviors represent forms of behavioral dysregulation.
"Frantic efforts to avoid abandonment" and unstable, intense relationships are under-
stood as forms of interpersonal dysregulation. Dissociation and paranoid ideation are
types of cognitive dysregulation. And, finally, emptiness and identity disturbance can
be understood as forms of self dysregulation. In each of these cases, the person has dif-
ficulty changing, modulating, or bringing the behavior, thought, feeling, etc., under voli-
tional control. He/she is likely to become extreme in some way, and have trouble
returning to a more moderate state or behavior.
From a DBT perspective, behavioral, cognitive, interpersonal, and self dysregu-
lation are seen as flowing from emotional dysregulation. Although individual assess-
ment is done on a case-by-case basis to determine whether this is the direction of
influence, what is often found is that affective dysregulation is a precipitant of other
forms of dysregulation. It is quite readily apparent how emotional dysregulation can
lead to dysregulation in the other areas. Stability in relationships, for example, requires
some ability to modulate affective responding, since relationships are such rich sources
of triggers for emotion. To participate in a stable relationship one must be able to mod-
ulate emotions so that one can problem-solve, focus on the other person at times, meet
one's own emotional needs at times, and so on.
As a more extended example, if the BPD client becomes affectively dysregulated,
and experiences some intense, negative affect, he/she is likely to seek help or assistance
from someone he/she is in a relationship with, in order to modulate the negative affect.
It is likely that the client will be feeling desperate for this help, and may fear that the
other person will not come through. Because of the emotional dysregulation, the person
may seek the help in a way that decreases the likelihood of actually getting it. He/she
may stake all hope on getting help from this particular person, and feel intensely con-
nected and cared about if the help is forthcoming. Alternatively, he/she is at risk for
intense feelings of abandonment or rejection if the help is not forthcoming, given the
level of vulnerability he/she has. It is easy to see how unstable relationships, and
attempts to avoid abandonment could result, and be a direct outgrowth of emotion
dysregulation.
Similar processes occur in the relationship between emotional dysregulation and
the other forms of dysregulation associated with BPD (cognitive, behavioral, and self).
BPD clients are likely to become cognitively or behaviorally dysregulated in the
context of extreme emotions that they cannot modulate. For example, parasuicide or
other maladaptive and/or impulsive behaviors frequently occur in the context of
emotion dysregulation, either intense negative affect or a numbed state of absence
of affect. Similarly, cognitive dysregulation, such as being unable to think clearly, or
thinking extreme or unrealistic thoughts, is likely to happen in the context of extreme
emotion. Over time, emotion dysregulation may also lead to self dysregulation. Having
a coherent sense of self is likely dependent on having a consistent sense of one's
own emotional reactions and experience. In contrast, the borderline individual may
have difficulty having a sense of "who she is," in part because of the lack of coher-
ence in her emotional life, given that her emotional responses are unpredictable and
uncontrollable.
BEHAVIORAL PATTERNS
Borderline individuals tend to vacillate between behavioral extremes: between

expression and inhibition of emotion, between blaming themselves and blaming others,
between holding unrealistically high expectations of themselves and assuming they are
incapable. Linehan (1993) has organized several of these polarities of behavior into
behavioral patterns, or "dialectical dilemmas," that reflect opposite ends of poles. These
behavioral patterns may not be universal to persons with BPD, nor are they supposed
to be exhaustive descriptions. Rather, they are clinically useful observations of behav-
ioral patterns that frequently interfere with quality of life and with learning new, more
adaptive behavior. In addition, they provide a conceptualization of the origins of these
types of behaviors. In this section, the behavioral patterns will be described, and some
functional analysis of how these patterns develop and are maintained will be discussed.
It is important to keep in mind that DBT does not assume that all clients with BPD
have all of these behavioral patterns, nor is a specific explanation of the origins of the
pattern provided. Individual assessment is necessary to determine whether these pat-
terns are present, and what seems to be maintaining them. The factors described here
are offered for heuristic purposes.
EMOTIONAL VULNERABILITY
The most important and central dialectical dilemma for the borderline individual
involves the interplay between emotional vulnerability and self-invalidation. Individ-
uals with BPD experience intense, painful, difficult to regulate emotional responses on
a frequent basis. The person experiences his/her emotional responses as overwhelming,
uncontrollable and unbearable, but is unable to regulate or change this state of affairs.
Often the individual panics about losing control or becomes despondent and apathetic
with hopelessness. The DBT model suggests that this emotional vulnerability, includ-
ing high sensitivity, high intensity of emotional response, and slow return to baseline,
are primarily biologically based. From a DBT perspective, this emotional vulnerability
is at the core of many problems associated with DBT.
When functioning at the extreme end of this dialectic, the person is likely to view
their problems as unsolvable, and to feel that others simply cannot understand how bad
things really are. He/she is likely to be "immersed" in the intense emotion, to be behav-
ing consistent with the emotion in an uncontrollable way, and to be unable to organize
his/her behavior to cope more adaptively. He/she may be angry at others for not real-
izing how horrible things really are, and for not helping. The emotionally vulnerable
person may fear new situations or situations in which his/her emotions have gotten out
of control in the past.
Although viewed as primarily biologically based, some behaviors associated with
the emotionally vulnerable pattern may at times serve the function of validating the
person's own emotional experience. If a person experiences an intense emotion, but
has been told that he/she is overreacting or should not be feeling that way, escalating
his/her communication of the negative affect may provide evidence for the validity of
the emotion. For example, if the client has extreme distress in response to a particular
loss, but the environment does not provide validation for the response, the person may
engage in parasuicide which serves the function of communicating to the person
him/herself that things truly are bad, that the emotions are valid.
Shame can also play a role in emotionally dysregulated behavior. When the
person experiences intense negative affect, but is aware that others are not likely to be
responding as strongly, he/she may feel ashamed of his/her response. Escalating emo-
tional communication may actually serve as a means of avoiding the feelings of shame
in the short run. Immersing oneself in the primary emotions associated with some neg-
ative event or experience may be less aversive and more validating than experiencing
secondary shame. Although reinforcing in the immediate sense, this strategy does not
work well in the long run, since the escalated display can then also become a further
source of shame.
SELF-INVALIDATION
In contrast to the emotion dysregulation pattern, when the borderline is operat-

ing at the other end of this pole, in self-invalidation, he/she blames him/herself for
all problems, and thinks that if he/she just were not so reactive, illogical or crazy,
everything would be fine. In the self-invalidating mode, the person views his/her own
emotional responses as completely unreasonable, and decides the whole problem is
him/herself. He/she does not acknowledge the realities of the difficult or painful
situation he/she may be in. The person basically adopts the views of the invalidating
environment. The person in this mode is likely to make sweeping, unrealistic resolu-
tions to change.
For example, the borderline who responds to criticism with feelings of shame or
self-consciousness may tell herself she is stupid for feeling that way, is too sensitive, is
overreacting. Benjamin and Wonderlich (1994), for example, found that borderline
subjects engage in more self-attack than bi-polar subjects. This self-invalidation may be
relatively chronic and low-level, with the person being essentially unaware that it is
happening. Self-invalidation may also reach a very intense level of self-beratement and
feelings of self-hatred.
From a behavioral perspective, how do we understand a behavior such as self-
invalidation, and how it develops and is maintained? First, the person may simply have
never, or rarely, been exposed to and reinforced for self-validation. In other words,
the person may simply never have learned to validate him/herself. Self-validation is a
verbal behavior, often private, that must be learned like any other verbal behavior. In
an ideal family situation, a child is validated by others frequently for his/her responses,
and he/she is also reinforced for self-validation. For example, the child may describe
some way in which he responded, indicating that his response was reasonable. The
parent may reinforce this self-validation by concurring or agreeing with the child. Not
having this type of experience may create a skills deficit in self-validation. Self-
invalidation may be maintained into adulthood because of the continued presence of
contingencies that do not support self-validation. The person may be living in an inter-
personal environment that continues to punish self-validation and to reinforcement
self-invalidation. Alternatively, self-invalidation may become an automatic or habitual
response.
Self-invalidation may also function as a means to avoid anxiety associated with
alternative behavior, in particular, self-generated behaviors. Self-generated behaviors
are likely to be anxiety-provoking if they have been punished in the past. It may be
that the BPD individual escapes or avoids negative affect associated with self-
generated behaviors through shifting to self-invalidation. For example, the person may
express a thought or feeling that reflects their actual experience, then feel guilty,
ashamed or anxious, and begin to self-invalidate. The self-invalidation functions
as avoidance behavior with reference to the painful feelings associated with self-
generated behavior.
Invalidating environments can also lead to self-invalidation more indirectly,
through the impact on the person's view of him/herself. Criticism, blame and general
communication that the person is unworthy lead to a negative view of the self, which
can then generalize. The person responds to life problems and disappointments with
the assumption that all problems represent personal failures caused by character flaws
of the individual. Lacking a sense of self as worthy and valuable, the individual cannot
generate responses to adversity that include the idea that "although I may have failed
in this situation, I am not a bad person" or "although this negative thing happened, it
was not all my responsibility".
ACTIVE PASSIVITY
The term "active passivity" refers to the tendency of BPD individuals to not only
fail to engage in active attempts at solving their own problems, but to attempt to get
others to solve their problems for them. Thus, they are actively involved in taking a
passive approach to their problems. This behavior pattern is familiar to most clinicians
working with BPD clients, since the therapist is often the person the client will be
attempting to get help from. Therapists may feel a pull to be responsive to the active
passivity of the client, attempting to solve the client's problems for him/her. Alterna-
tively, the therapist may feel resentful of the client's failure to do things for him/herself,
ascribing it to lack of motivation, and may feel unwilling to offer the help the client is
requesting.
From a behavioral perspective, there are of course a number of possible ways that
active-passivity behavior can develop. It is fairly easy to imagine that for some people,
early attempts at actively and directly solving problems are punished, or at least not
reinforced. This is likely to be the case for BPD individuals, whose attempts at chang-
ing or coping with their intense, difficult to regulate emotions are likely to be unsuc-
cessful. The invalidating environment tends not reinforce the individual if they are only
partially successful, and does not reinforce effort that is less than perfect. In addition,
the invalidating environments do not teach, model or reinforce effective problem-
solving. Instead these environments tend to minimize the difficulty of solving life prob-
lems. Consequently, the BPD individual is not likely to have learned effective means
of solving his/her own problems.
On the other hand, the client may have a history of experiences where others,
helping professionals in particular, may have actively intervened on their behalf, result-
ing in resolution of a problem that he/she felt unable to resolve him/herself. For
example, if an adolescent client living in an abusive family situation is unable to change
that situation directly, but she eventually becomes suicidal. This client subsequently
began very frequent help-seeking behavior from this physician; she engaged in para-
suicidal behavior, which was reinforced by contact with the physician. A totally over-
whelming and incredibly painful life situation was suddenly ameliorated following her
expression of suicidal intent to the physician. This experience was an extremely salient
instance of active-passivity being reinforced in a powerful way. Her own attempts at
changing her situation had all failed, but her help-seeking was reinforced in an
extremely salient way.
It is important to keep in mind that a more interpersonally oriented style
of problem-solving is not inherently dysfunctional. Relying on others to provide help
with resolving problems probably works well in cultures that value and support
this approach. To a large extent, the difficulty for borderline clients may be a lack of
fit with the cultural context; this interpersonal approach is not widely maintained
in our culture, and to a large extent is devalued in Western cultures. From a
behavioral perspective, understanding the role of context is central, and the broader
cultural context seems to play an important part in this situation. The expectation of
staunch self-sufficiency present in Western cultures is not a good fit for many people
with BPD.
Linehan (1993) also proposes that there may be a temperamental basis for the
passive problem-solving approach. Some people may be predisposed to rely more on
others for managing their emotional responses. It is likely that active-passivity results
from a combination of both temperamental and environmental factors. It is easy to
imagine how being predisposed to a more passive approach, in combination with little
reinforcement from attempts at active problem-solving, would lead to a passive
approach, which is likely to be at least intermittently effective.
APPARENT COMPETENCE
Individuals with BPD at times function in a very effective, skillful way, perform-
ing remarkably well in a job, as a friend or parent. As is the case with everyone, this
behavior is context specific, meaning that under different conditions, the person
may not be able to function in an effective way. The context includes both the envi-
ronmental context, who is present, what the situation is, what is required, etc., and the
private context, the person's mood, thoughts, physiological state, etc. Because the
private context shifts so dramatically for borderline individuals, their overt behavioral
competence also changes drastically. For example, the person is likely to function very
well when not emotionally aroused, but very poorly when distressed. Although this may
be the case for most of us, the difference is that individuals with BPD get more unpre-
dictably emotionally aroused and the outside observer is likely not to recognize it. The
triggers for the emotion may be subtle so that the observer is unaware of what has
produced the emotional arousal.
The word "apparent" is used in labeling this pattern because the skillful behav-
ior has the appearance of being a consistent part of that person's repertoire. The person
is apparently good at solving problems, functioning in the world and so on; however,
because the competent behavior is so context specific, the expectation that the
behavior will always be available is erroneous. One of the most difficult problems
caused by this situation is that people tend to expect others to behave in a relatively
consistent way, and to be able to cope at a level that they have seen in the past.
Unless a change in context is very obvious (e.g. hislher mother just died), the
expectation is for level of functioning to be relatively consistent. This poses a real
dilemma for the person with BPD. Functioning at a high level sets her up for inevitable
failure, since she knows she will not be able to maintain that level of functioning. The
result is likely to be frustration, disappointment or anger on the part of others, and feel-
ings of shame or guilt herself when things change and the behavior can no longer be
maintained.
Apparent competence is therefore defined in DBT as the tendency of others to
over ascribe competence to the individual with BPD, which is caused by two factors.
First, the competence of the individual with BPD is likely not to generalize in the same
way that it does for others. Others expect the competencies of individuals with BPD
to be consistent over time, and fail to take into account the effect of private experi-
ences that have a strong impact on competency, as well as the extreme sensitivity and
lability of the emotional life of most people with BPD.
A second factor that leads others to over-ascribe competence is the desynchrony
between the client's emotional experiencing and emotional expression. People with
BPD at times inhibit expression of intense negative affect, so they may look as though
they are not experiencing painful emotions. This calm exterior presentation is misin-
terpretted by others as reflecting a competence in coping that crosses situations and
emotional states. Alternatively, the individual with BPD may express negative affect
verbally, but without accompanying nonverbals that match his/her level of distress. For
example, he/she may casually or even jokingly mention that he/she is thinking about
suicide, when in fact he/she is feeling extreme distress. Many observers are not likely
to pick up on the true level of pain the person is experiencing, and again to misattribute
a level of consistency in competence that is not accurate.
Apparent competence is often related to emotional dysregulation, in that the
competent behavior is likely to occur when the person is at an emotional calm or base-
line state. Because of emotional vulnerability, the BPD person is easily thrown out of
that emotionally neutral state, and is likely to lose his/her ability to function at the same
level. Environmental factors are also important, in that certain contexts are likely to
be emotionally dysregulating. For example, the person may be able to function well in
the context of therapy, where he/she feels safe, supported and trusting; however, he/she
may not function well in non-therapeutic relationships, where criticism, conflict and
disagreement feel more threatening.
ACTIVE PASSIVITY VS. APPARENT COMPETENCE
These two patterns, active passivity and apparent competence, form a "dialecti-
cal dilemma" for the client. When the person is in "apparently competent" mode, others
are likely to build up expectations of him/her that he/she will later not be able to meet.
Others are all the more exasperated when the individual then shifts to active passivity,
asking others to solve problems that he/she appeared perfectly capable of solving the
day before. The dilemma for the client involves being in a position of both needing help
and not needing help, being able to inhibit emotions at times, but not predict when she
will be able to function well. When feeling good, the person with BPD remains aware
that at some point in the future she will again find it very difficult to regulate her emo-
tions, and consequently feels an extreme fear of helping, supportive people leaving or
not being available.
UNRELENTING CRISIS
Therapists working with BPD clients are familiar with the pattern of on-going,
multiple crises that make up the day-to-day life of individuals with BPD. BPD clients
are likely to react to stressful life events and situations in ways that prolong and esca-
late problems. Their intense emotional responses and tendency to be volatile and impul-
sive make it likely that difficult or painful events will be exacerbated. The life situations
of many of these clients also put them at increased risk for negative events. Many of
our clients are poor, on disability, living in low-income or subsidized housing, often in
unsafe areas, or working in low-paying and unsatisfying jobs. They often lack financial
or social support resources that make difficult events easier to solve. Thus, the combi-
nation of living in conditions of increased risk for negative events, along with the ten-
dency to engage in behaviors that increase negative events, results in a chronically high
frequency of such events.
The unrelenting crises many individuals with BPD experience prevent them from
effectively recovering from anyone event. Because the next event so often follows close
on the heels of the previous one, there is not time to completely resolve the emotions
associated with a particular stressful experience. Life becomes a stream of painful expe-
riences and emotions, that fluctuate between extremely negative to more mildly nega-
tive and back. It is easy to see how the client with BPD becomes frustrated and hopeless
in such a context.
Although the pattern of unrelenting crisis is likely due in part to environmental
stressors, there appears to be a behavioral component as well. This in part seems to be
due to a skills deficit: the individual with BPD often lacks the needed skills to deal with
stressful events and living situations. It is exacerbated by the emotional dysregulation
the person experiences in these situations. In addition, it may be that being engaged in
life crises at times serve the function of avoidance for the individual. Given the inten-
sity of the pain that negative emotions produce for most people with BPD, many of
them avoid painful emotions at all costs. The powerful nature of their emotions neces-
sitates extreme behavior to allow the person to avoid. Many of the maladaptive behav-
iors borderlines engage in reflect emotional avoidance: parasuicide, drug abuse,
dissociation, etc. Participating in life crises situations may at times, paradoxically, play
a similar role. Being involved in a major upset over a friend's betrayal, a boss's insen-
sitivity, quitting a job, or losing housing may at times serve the function of allowing the
person to avoid even more painful emotions resulting from other life events or traumas.
This is not to suggest that all crises are generated by the individual, nor that crises are
not painful in and of themselves, but simply that in certain cases, they may serve the
function of avoidance of even more painful affect.
Finally, unrelenting crises may be triggered by discomfort associated with life
being calm or uneventful. Many clients with BPD have had the experience of things
"blowing up" when life seemed to be going well. The experience of life calm is associ-
ated with out-of-control eruptions or losses. Life calm therefore triggers anxiety and a
feeling of waiting for "the other shoe to drop." This anxiety may then lead to crisis-
generating behaviors as a means of avoidance, or crisis-generating behaviors as a means
of eliciting assistance in escapating.
INHIBITED GRIEVING
The vast majority of individuals with BPD have histories of trauma and/or sig-
nificant loss (Goodwin, Cheeves, and Connell, 1990; Laporte and Guttman, 1996; Ogata,
Silk, Goodrich, Lohr, Westen, and Hill, 1990; Wagner and Linehan, in press). Coming
through a loss or trauma generally requires that the person experience his/her feelings
about the event in a relatively complete way. This is critically difficult for individuals
with BPD, since their emotional responses are so powerful and difficult to modulate.
Allowing him/herself to feel the emotions associated with a trauma or loss may be expe-
rienced as impossibly dangerous and risky. The individual may, in many cases accu-
rately, perceive that she does not have the resources or abilities to experience such
intense emotions and continue to function. Consequently, many people with BPD
inhibit experiencing of emotions associated with trauma and loss. They are frequently
exposed to cues that trigger loss-related emotions, and automatically inhibit or avoid
these feelings. The result is that they become even more sensitized, and do not process
the primary emotions of sadness and fear.
UNRELENTING CRISIS VS. INHIBITED GRIEVING

There is an important interplay between the inhibited grieving and unrelenting
crisis patterns. The individual with BPD experiences an on-going, high rate of painful
experiences and automatically blocks his/her emotional responses to these experiences.
Inhibited grieving makes it unlikely that anyone crisis will be successfully processed
at an emotional level. Likewise, the unrelenting nature of the crises borderlines expe-
rience increases the likelihood of resorting to blocking emotions in order to cope.
Having not processed the painful emotions, the person is likely at even higher risk for
Functional Analysis of Borderline Personality Disorder Behavioral Criterion Patterns 19S
behaving in such a way as to produce yet more crises in his/her life. This complex
interplay between these two patterns keeps the person locked into them.
BEHAVIORAL THEORY IN DBT
Case conceptualization and intervention in DBT are based largely on behavioral

theory. It is extremely important to recognize that a behavioral approach requires indi-
vidual assessment of clients' specific behavioral patterns. Although DBT offers descrip-
tions of some common behavioral patterns seen in individuals diagnosed with BPD,
the clinician doing DBT does not assume that these patterns are present in all indi-
viduals, nor does he/she assume that a given behavior that looks like one described in
the theory is necessarily caused or maintained for the reasons the theory suggests. DBT
theory provides suggestions and frequently seen patterns and hypothesized associated
factors, but these are not to be substituted for careful behavioral analysis of
problematic behaviors being addressed in treatment. The descriptions are to be used
for heuristic purposes.
A thorough understanding of behavioral theory and concepts are crucial to the
conduct of DBT. One of the most frequent interventions is the carrying out of behav-
ioral analyses. A behavioral analysis is an approach to understanding a problematic
behavior, that involves detailed, step-by-step analysis of the problematic behavior,
the context and events leading up to it, the relationships between these precipitant
factors, and finally, the consequences of the behavior. Understanding concepts such as
avoidance, reinforcement, punishment, shaping, and so on are essential to conducting
a thorough behavioral analysis and subsequent solution analysis, in which solutions or
possible strategies for addressing the problematic behavior are developed.
A behavioral approach makes the assumption that all difficulties in life represent
problems to be solved (Linehan, 1993). DBT therapists maintain this problem-solving
stance, and teach it to clients. Behavioral approaches also are typified by a collabora-
tive therapeutic relationship, in which the therapist, although still an expert, shares
information with the client about what he/she sees the problem to be, and what possi-
ble solutions are. The therapist educates the client about behavioral theory, and about
the application of the theory to the client's particular problems. The therapist and client
then work together on implementing solutions.
After conducting a behavioral analysis, the therapist generates a series of
hypotheses about what might be contributing to or maintaining the problematic behav-
ior. There are likely to be multiple hypotheses, often representing different stages of
the "chain" of events and behaviors leading up to the problematic behavior of inter-
est. In DBT, these hypotheses often include a focus on specific behaviors and ways of
conceptualizing associated with DBT theory; for example, the therapist is likely to be
cognizant of the roles of emotion dysregulation, self-invalidation, and so on.
A behavioral formulation typically points to one of four different types of
processes that maintain problematic behavior. First, a problematic behavior may occur
because the client's behavioral repertoire is lacking in alternatives. He/she may simply
not have learned to engage in alternate, more adaptive behaviors. If this is the case, the
therapist is likely to suggest or engage in skills training with the client.
Alternatively, an adaptive behavior may be within the client's repertoire, but
the contingencies present in the environment may not be supporting the client in
engaging in those behaviors. For example, the client may know how to ask for help
when emotionally distressed, but these requests may be ignored or punished. The
person may, however, be reinforced for escalating the intensity of their request and
their expression of pain. Suicidal behavior can at times be reinforced in this way,
with more adaptive, alternative ways of expressing distress not reinforced by the
environment.
Third, a client may be blocked from engaging in adaptive behavior by intense
emotions, such as fear or guilt. For example, a client may know how to interview for a
job, but she may experience such intense anxiety when she begins thinking about or
taking steps toward applying for a job, that she ultimately avoids the situation and is
negatively reinforced by reduction in anxiety. There is now a great deal of evidence
supporting the efficacy of exposure-based treatments for anxiety-related problems. The
DBT therapist is likely to use exposure-based techniques with a client who is blocked
from an adaptive behavior by feelings such as fear, guilt or shame.
Finally, the client may be influenced by maladaptive cognitions or beliefs. For
example, she may have difficulty going for the job interview and interacting skillfully
because she believes she is unworthy of getting the job. If such cognitions appear to be
interfering, the DBT therapist will likely use cognitive therapy techniques to address
them. This could include having the client gather information relevant to his/her beliefs
or directly challenging the beliefs.
APPLICATION TO THERAPY
Many BPD clients present with multiple diagnoses and problems, a variety of
interpersonal difficulties that make the process of therapy difficult, and frequent crises.
These factors make treatment complicated; DBT is correspondingly complex, as it has
been designed for multi-problem, difficult to treat client populations. Any brief
overview of DBT is necessarily incomplete; however, we attempt here to provide a
sense of how DBT is organized, and focus specifically on how DBT addresses the
behavioral patterns.
FUNCTIONS AND MODES
DBT is organized around five primary functions, or things that must be accom-
plished as part of the treatment. These functions are divided among several modes of
treatment (e.g. individual therapy, group skills training, phone consultation to the
patient, etc.). The five functions include 1) enhancing client capabilites, 2) decreasing
factors that interfere with motivation to change, 3) generalization from therapy to the
natural environment, 4) supporting the therapist in a way that allows him/her to carry
out effective therapy, and 5) modifying the client's environment such that it supports
adaptive behaviors. Addressing the five functions is necessary to provide a compre-
hensive DBT program; however, the various functions can be carried out in a variety
of different modes. For example, the function of enhancing client capabilities (e.g.
acquisition of new behavioral skills) can be accomplished in a skills group, or through
individual skills training. The function is what is critical to DBT, however the mode
must be carefully considered to ensure that it is appropriate to that function.
The DBT model assumes that individuals meeting criteria for BPD are lacking
essential behaviors and skills. The primary ones are behaviors that allow them to suc-
cessfully cope with their own emotions. Consequently, one of the functions of DBT is
to increase clients' skills and help them develop new behaviors. In standard DBT, this
is done in a group skills training format, and all clients are required to participate in
one year of skills training. A group format is used because it is efficient, and because
it helps create a structure that supports teaching and learning of skills, rather than
processing life events or other topics that are often a part of therapy. Skills are also
taught and reinforced by the individual therapist, as the client and therapist work on
individual therapy goals.
There are four types of DBT skills taught, presented as four modules: core mind-
fulness, emotion regulation, distress tolerance, and interpersonal effectiveness. Core
mindfulness skills are based on a zen mindfulness tradition, and are designed to increase
client's control over emotional processes, ability to observe their own thoughts and feel-
ings, and ability to be present to the moment without attempting to avoid or escape.
Emotion regulation skills focus on preventing negative affective states, and increasing
ability to change from a negative to a more neutral or positive affective state. In con-
trast, distress tolerance skills focus on tolerating negative affective states without
attempting to change them. The goal is to learn ways to get through times of crisis and
intense pain without making things worse by engaging in maladaptive behaviors.
Finally, interpersonal effectiveness skills focus on learning to balance getting what you
want, maintaining self-respect, and maintaining good relationships. Assertiveness skills
are central to this module.
A second function of DBT is to determine what factors are interfering with the
development of new behaviors and skills, and to intervene to change these factors. For
example, clients may have low motivation or they be inhibited from trying new behav-
iors by fear or other emotions. Addressing these issues is primarily done by the indi-
vidual therapist for several reasons. The individual therapist is likely to know the client
the best, and have the clearest understanding of what is interfering. Also, the thera-
peutic relationship is viewed as a primary reinforcer in DBT, and is used in contingency
management to help insure that adaptive client behaviors are reinforced. The individ-
ual therapist may also use exposure-based techniques or cognitive interventions to
address factors interfering with the development of new behaviors.
Ensuring generalization of new behaviors, a third function of DBT, is also pri-
marily done by the individual therapist. In standard DBT, generalization work often
takes place in the phone consultation mode of treatment. Therapist and client have
contact by phone to talk about how the client can apply skills in day-to-day life, when
difficulties come up. This is very important because many clients can grasp and under-
stand skills in therapy, but have difficulty applying them, particularly when they are
under emotional duress. Having to wait until their next therapy appointment makes it
unlikely that they will get the opportunity to practice in a given situation, whereas being
able to have a brief phone consult allows greater possibility for the development of
new behaviors in the natural environment.
DBT puts a great deal of emphasis on the therapist's need for validation and assis-
tance in working with BPD clients. Provision of this validation and consultation to the
therapist are critical components of the treatment: one in fact cannot do DBT without
participating in a DBT consultation group, as this is not ancillary, but an essential part
of the treatment. The purpose of the consultation group is to provide skills training in
how to do DBT, reinforcement for doing it competently, and validation for the thera-
pist. BPD clients are likely to punish therapists for doing DBT well, and reinforce them
for doing DBT poorly, because doing DBT well involves pushing the client to do
painful, difficult things. If the therapist's immediate therapy-relevant context only

includes the client, he/she is inevitably shaped into drifting from the treatment proto-
col. The task of the consultation group is to provide reinforcement for sticking with the
treatment.
Finally, the DBT treatment milieu must be set up in a way that reinforces adap-
tive behavior. The director of the program or clinic is largely responsible for this com-
ponent of the therapy. If the program or setting is arranged in a way that reinforcement
is more available for maladaptive behavior, the treatment is not likely to be success-
ful. The program or organization must examine its policies and rules in order to arrange
contingencies in the most effective manner.
STAGES AND TARGETS
DBT is a stage model of treatment, with the stage being determined by the client's
current behavior and degree of commitment to change. Clients enter treatment at the
pre-treatment stage, where the focus is on developing and attaining a commitment to
working on problematic behaviors. Clients may return to the pretreatment stage
at other times if commitment lapses. Stage 1 of treatment is focused on reducing 1)
suicidal and life-threatening behaviors such as parasuicide and suicide crises behaviors,
2) behaviors that interfere with the process of therapy, 3) quality of life interfering
behaviors (e.g. substance abuse, homelessness, etc.). In addition, acquiring and strength-
ing of skills is also a target of stage 1. These Stage 1 targets are organized hierarchi-
cally from highest to lowest priority. Stage 2 of DBT focuses on reducing symptoms
associated with trauma, and an exposure-based model is used. At Stage 3, clients work
on more ordinary problems of living, such as relationship or career difficulties, depres-
sion, and so on. Finally, at Stage 4, if a person wishes to pursue it, therapy focuses on
existential issues, developing a sense of meaning in life and the ability to experience
true joy.
The stage model used in DBT delineates that therapy continues to focus on
targets for a particular stage, such as quality of life interfering behaviors at stage 1 until
they have been successfully dealt with before moving to stage 2. The stages are not rigid
in the sense that a client may at one time move on to stage 2, but then begin engaging
in stage 1 target behaviors again, at which point therapy would return to Stage 1. The
DBT models tested to date have been almost exclusively with clients and pretreatment
and Stage 1 levels.
In addition to the targets described above, DBT also focuses on secondary targets,
which are the behavioral patterns discussed here. The behavioral patterns are only tar-
geted as they relate to the primary targets. For example, self-invalidation is targeted if
it occurs in the context of one of the primary targets, such as parasuicide, therapy
interfering behavior, or quality of life interfering behavior.
ACCEPTANCE VS. CHANGE ORIENTATION AND

TREATMENT STRATEGIES
DBT combines includes two primary types of treatment interventions or strate-

gies: those focused on change, and those focused on acceptance. All of the treatment
strategies reflect one or the other of these stances. Change oriented strategies promote
client behavior change in a direct way. They include teaching, encouraging, modeling,
roleplaying, problem-solving, using exposure-based interventions or cognitive inter-
ventions and generally pushing the client to make change. Acceptance oriented strate-
gies, on the other hand, are focused on validation of the client's experience, acceptance
of reality as it is without effort to change it, allowing emotions to exist without attempt-
ing to change them. The dichotomy between acceptance and change is in some sense
false, in that for many people, engaging in acceptance actually constitutes change. Like-
wise, being able to make changes is contingent on first accepting the problem for
what it is.
DBT therapists make extensive use of problem-solving strategies, including
behavioral analysis, to address problematic behaviors that have been targeted for
change. A behavioral analysis involves an exploration of the precipitating factors
contributing to some problematic behavior, as well as of the consequences of the
behavior. A behavioral analysis may be brief and relatively informal, or may be more
extensive and take one or more sessions. Behavioral analysis is a change-oriented inter-
vention because the point is to develop an understanding of the factors contributing to
the problematic behavior, and then to work on changing them. This may take the form
of teaching and practicing of new behavioral skills, doing an exposure-based procedure,
carrying out a contingency-management procedure, or some other approach.
DBT therapists engage in several other important change-oriented intervention
strategies. One of these, irreverent communication, is an aspect of the style in which
the therapist interacts with the client. When a therapist uses irreverent communication,
he/she may be blunt, off-beat, extreme or startling in some way. The communication
style is designed to keep the therapeutic interaction from becoming bogged down or
"stuck," to keep the client's attention and to open up new possibilities. The therapist
may play the "devil's advocate," or talk about topics in a direct, matter-of-fact manner
that are usually avoided.
The DBT therapist also makes use of change-oriented approaches to helping the
client deal with other professionals involved with the client, in that he/she encourages
the client to manage his/her own mental health and medical care, rather than
managing it for the client. In general, the DBT therapist takes the stance that his/her
job is to help the client become effective and skillful at interacting with other profes-
sionals, rather than intervening on the client's behalf with other professionals. Using
the consultation-to-the-client strategy, the therapist avoids stepping in to change
other people in the environment, but focuses on helping the client do what he/she needs
to do to effectively deal with the people in his/her life, including other mental health
professionals.
Finally, two of the skill modules are primarily change-oriented. The interpersonal
skills and emotion regulation skills both emphasize clients changing their own behav-
ior in order to become more effective in their lives. The interpersonal skills focus on
changing interpersonal behavior, to be assertive, sensitive to the needs of the situation
as well as one's own wants, and skillful in interacting with others. The emotion regula-
tion skills are designed to help the client shift from experiencing predominantly nega-
tive affective states, to experiencing more neutral to positive affective states, through
use of a wide variety of skills designed to decrease vulnerability to negative affect,
increasing pleasurable events, acting opposite to negative emotions and increasing
experiencing of pleasurable emotions.
Balancing these change-oriented strategies are a set of acceptance-oriented inter-
vention strategies, with validation being the central one. Validation is communication
from the therapist that he/she both understands the client's experience, and thinks that
it makes sense or is a reasonable response. Validation is crucial to the DBT therapy
process, because BPD clients have generally experienced so much invalidation that they
are very much in need of a relationship in which they are validated. Validation also
facilitates behavioral change, because the client is more likely to be able to focus on
making changes if he/she feels that the therapist truly understands his/her situation. In
DBT, six levels of validation have been identified (Linehan, in press); these include
accurate reflection of what the client is thinking, feeling or doing, "staying awake" by
being actively attentive and engaged, verbalizing unspoken feelings or reactions of
the client, validating in terms of past learning experience or biological dysfunction,
validating in terms of present circumstances, and being "radically genuine". Radical
genuineness involves dropping the "mask" of therapist, and simply being a person in a
deeply genuine way. Validation can be direct and verbal, or can be inherent in how the
therapist responds to the client's problems and requests.
Balancing the irreverent communication style is an alternate style in which the
therapist is warm, responsive and in tune with the client. In DBT this is referred to
as the reciprocal communication strategy. This strategy also includes judicious use of
self-disclosure, primarily to give the client information about how he/she is affecting the
therapist or coming across to others, or for the therapist to provide examples or ways
that he/she has made use of skills in his/her own life. The reciprocal style is designed to
mitigate the inherent power differential in the therapeutic relationship, to emphasize
the "realness" of the relationship and of the therapist's feelings for the client.
DBT therapists also promote acceptance in clients by balancing change-oriented
skills with acceptance-oriented skills. Mindfulness and distress tolerance skills both are
ways that clients can work on accepting that reality is what it is, tolerating negative
affect rather than fighting against or avoiding it, being present to the current moment
without judgment. It is easy for clinicians to get caught up in a sense of urgency around
things always needing to be different, to agree with the client that the client cannot
tolerate things as they are. DBT attempts to balance striving for change with a will-
ingness to also consider the option of tolerating painful affect in the moment.
Finally, acceptance also informs how DBT therapists interact with other members
of the client's personal and professional network. The counterpart to the change-
oriented consultation-to-the-patient strategies are the environmental intervention
strategies. The environmental intervention strategies involve advocating or intervening
on the client's behalf, in essence accepting that he is unable to, in situations where the
client is genuinely incapable of achieving some important outcome himself. For
example, if the client is a minor, is incapacitated or if the person or institution has a
great deal of power over the client and is unresponsive to him. Principles of shaping
must be applied when attempting to balance the consultation-to-the-patient and envi-
ronmental intervention strategies, such that the therapist does not require the client to
do something so far out of his reach that he will necessarily fail, on the other hand not
doing something for the client that he could actually use as a reasonable opportunity
to learn and engage in new behavior.
BEHAVIORAL PATTERNS-LINKS TO TREATMENT

The behavioral patterns are directly targeted for change in individual DBT when
they seem to be playing a role in the maintenance of one of the primary targets. For
example, behavioral analyses of parasuicide for a particular client may reveal that he
frequently engages in parasuicide in the context of active passivity. The client might
show a pattern of attempting to get others to solve problems for him when he runs into
difficulties. Lacking adequate interpersonal skills, his attempts at getting help alienate
others, leading him to feel rejected and alone. Noting this pattern, the therapist
may conceptualize a part of the client's problem as being actively passive in the face
of life problems. The first goal is to stop the parasuicide, but in that process the thera-
pist may target one of the behavioral patterns if that behavior seems linked to suicidal
behavior.
In general in DBT, any problem targeted for change, including the behavioral
patterns, is openly discussed with the client. The therapist describes what the problem
is, how he/she sees it affecting the client's life and why he/she is suggesting that it be
changed. This orienting is viewed as crucial to the change process. DBT therapists also
work with the client on making a commitment to changing the behavior. The thera-
pist's goal is to get the client to make the firmest possible commitment, to increase the
likelihood that the client will actually participate fully in changing the behavior. For
example, the therapist might explain what active-passivity is, explain what he/she sees
the client doing that constitutes active-passivity, engage the client in a discussion about
the behavior and how it is affecting the client's life. The therapist will encourage the
client to think about how the behavior is functioning is the client's life, in what ways it
is helpful or harmful in terms of the client achieving his/her ultimate goals. Commit-
ment strategies such as examining pro's and con's, the therapist playing the devil's advo-
cate, or using shaping to increase commitment, or other DBT commitment strategies
are likely to be used. In the case of the above example, the therapist may ask the client
if he wants to work on active-passivity, why he would want to change it, what the pro's
and con's of changing v.s. staying the same are, and so forth.
If the client commits to working on the problem, the therapist would then conduct
a more complete behavioral analysis of the problem behavior. Such a behavioral analy-
sis would most likely be conducted on several occasions when the behavior has
occurred, and the therapist would begin to conceptualize what factors are influencing
or maintaining the behavior, in order to plan solutions. The therapist would also watch
for the occurrence of active passivity in the session, and use instances of the behavior
occurring to examine factors that seemed to influence it in the moment, looking at what
was going on in the session before the active-passivity appeared. The therapist would
begin to conceptualize what the controlling variables are for the behavioral pattern:
under what circumstances does this client become actively passive, what seems to
trigger it, what is maintaining it, what other skills is the client lacking as alternative,
more adaptive behaviors. This process leads directly to the generation of solutions to
address the problem.
All of the four problem-solving strategies described (skills training, contingency
management, exposure and cognitive modification), are applied to the behavioral
patterns. Although not typically the primary source of skills training, individual DBT
therapists do reinforce and promote generalization of skills. This may involve coaching
the client during times of crisis, modeling, roleplaying, and discussing the application
of skills to specific situations.
In terms of contingency management, DBT therapists try to ensure that behav-
iors that represent adaptive alternatives to the behavioral patterns (i.e. self-validation,
adaptive experiencing rather than avoidance of painful affect, appropriate communi-
cation of distress, etc.) are reinforced by the therapist, and that pattern-consistent
behavior is not reinforced. This means being alert to occurrences of either, and aware
of how the therapist's responses may be reinforcing or punishing the behavior of inter-
est. For example, if a therapist consistently becomes more interested, attentive, and
involved when a client is in a crisis, and more laid-back; distant or passive when
the client is doing well, he/she may be inadvertently reinforcing unrelenting crisis
behavior.
Exposure may be used to address the behavioral patterns, in that some prob-
lematic emotion may trigger the behavioral pattern, and exposure to cues triggering
that emotion will reduce the occurrence of the pattern. For example, if a client engages
in apparent competence because of shame associated with expressing his distress and
asking for help, exposure to this cue may reduce the shame and thereby prevent the
occurrence of apparently competent behavior. This involves presenting or not remov-
ing cues for negative emotions when they occur, encouraging and actively promoting
client experiencing of negative affect without escape, and encouraging and coaching
alternative behavioral responses. In the current example, after orienting and commit-
ment occur, the client would be encouraged to express distress and ask for help repeat-
edly in ways that elicit the sense of shame. The client would be instructed not to engage
in shame-consistent behavior (i.e. hiding, withdrawing, etc.), and coached in being direct
and forthright in his requests for help. Repeated exposure to the cues triggering shame
should eventually reduce the intensity of the shame, thereby reducing the likelihood
of apparently competent behavior.
DBT therapists may also use cognitive modification procedures to address behav-
ioral pattern related behavior, if dysfunctional thoughts seem to be involved in eliciting
or maintaining the behavior. It is very common for all of the patterns to be associated with
some sort of dysfunctional thoughts. For example, cognitive modification is likely to be
used to address dysfunctional, self-invalidating thoughts.
In addition to the problem-solving strategies, DBT therapists make frequent use
of validation strategies, particularly attempting to validate the client's reactions to
his/her current circumstances as reasonable and normal, when they are. Validation is
often essential to the change process, in that the client needs to know that the thera-
pist understands precisely where she is at, in order to engage in the process of search-
ing for a way out. For example, if the client does not feel that a therapist understands
her experience of apparent competence, she is not likely to be actively engaged in trying
to do things differently around this pattern.
Validation is particularly important to addressing the behavioral pattern of self-
invalidation, in that it directly counters the client's experience of the invalidating envi-
ronment. Validation from the therapist provides the client with the often new experience
of being taken seriously, of having her experience of the world and her self acknowl-
edged. These are the first steps toward changing chronic self-invalidation. For a client to
begin to validate her own experience, she must hear from her therapist that her experi-
ence is in fact valid, that her reactions are meaningful and make sense. The therapist can
then work with the client on actively challenging her own self-invalidating thoughts, and
generating of more self-validating ones.
Validation is likely to be involved in addressing the other behavioral patterns in
a variety of ways. The therapist may validate the emotions prompting the pattern or
the difficulty of changing pattern-related behavior. The therapist may validate the
occurrence of the behavior, in light of the client's learning history. For example, saying
"it makes sense that you want to avoid feeling your feelings around painful things like
losing your girlfriend (inhibited grieving), given the number of painful things you've
Functional Analysis of BorderHne Personality Disorder Behavioral Criterion Patterns 203
had happen, and your fear that you'll be overwhelmed by the emotions". The DBT
therapist would be likely to balance this with a statement about how avoidance is also
problematic and why.
Validation generally has a calming effect on a client who is in the midst of emo-
tional dysregulation, which can allow the client to continue to focus and actively work.
DBT therapists use shaping to gradually promote client self-soothing; the therapist is
more likely to actively validate to reduce client emotional dysregulation early on in
treatment, and to teach the client to rely more on self-validation over time.
DBT therapists also attend specifically to in-session behavior. Behavioral theory
and research suggest that immediate contingencies are the most powerful, and thera-
pists only have direct control over contingencies within the session; therefore, DBT
therapists focus on modifying behavioral patterns as they occur within the therapy
session (Kohlenberg and Tsai, 1991). This will often involve pointing out when the client
is engaging in a behavior consistent with one of the behavioral patterns. For example,
the therapist may point out to a client who is trying to get the therapist to call the
client's pharmacotherapist rather than the client doing so him/herself that the client is
being actively passive. Drawing the client's attention to the behavior needs to be done
non-punitively. The therapist may then initiate a discussion of how the behavior is
helpful to the client and how it interferes with achieving hislher long-term goals. If it
appears that the behavior is problematic, the therapist will attempt to engage the client
in committing to change it, and will then suggest alternatives, and/or provide opportu-
nities for the client to do things differently in the moment. The therapist will reinforce
new behaviors, keeping principles of shaping in mind. For example, the actively passive
client who wants the therapist to call the pharmacotherapist rather than doing it
him/herself may suggest and reinforce alternatives, such as having the client call during
the session, making the call together, etc., depending on what the client is capable of
doing. The idea is to elicit new, more maladaptive behavior at all costs.
DBT SKILLS TRAINING APPLIED TO BEHAVIORAL PATTERNS
The skills taught in DBT are directly applicable to changing the behavioral pat-
terns. For example, core mindfulness skills are used to teach clients to observe and
describe their own emotions, sensations, and thoughts. This process of observing and
describing is generally necessary to changing any of the behavioral patterns, in that the
client first must be able to recognize and describe the pattern when it occurs. He/she
may also need to learn to observe and describe the emotions and thoughts accompa-
nying the pattern, in order to figure out what is maintaining the pattern. In addition,
core mindfulness skills can be used as alternative responses to the behaviors associated
with the behavioral pattern. For example, rather than acting as if she is fine and coping
well when she is actually distressed (apparent competence) a client may learn to be
more mindful of her actual emotional state, so that she can act more consistently with
it (i.e. actively seek to solve whatever problem is producing the distress, communicate
her distress to others).
Core mindfulness skills also teach clients about "emotion mind," "rational mind,"
and "wise mind". Emotion mind is in essence the state of mind one is in when being
controlled by one's emotions. Clients are taught ways to reduce vulnerability to
emotion mind, and ways to move out of emotion mind. "Rational mind" is the state of
paying attention only to reasoning and intellect, ignoring emotions and feelings. Wise
mind is the intersection of emotional and rational minds; it is a more holistic state in
which both reasoning and feeling influence behavior and decision-making. It is a more
intuitive, centered place from which to act. These concepts apply to the behavioral
patterns in a variety of ways. An emotionally vulnerable person may be making
maladaptive choices purely from emotional mind. Core mindfulness skills would help
this person recognize that she is in emotion mind, and help her learn to move to wise
mind. Alternatively, the apparently competent person may be in rational mind, ignor-
ing his/her emotions. Again, the skills of moving into wise mind may help this person
do better self-care and make more effective decisions.
Emotion regulation skills focus on helping clients learn to prevent negative emo-
tional states through active self-care, and to change negative emotional states through
opposite action. For example, ample evidence supports the notion that acting opposite
to depression, through being active, engaging in pleasant events, and mastery experi-
ences, is an effective means of changing depressed mood. These skills are clearly very
applicable to the behavioral pattern of emotional vulnerability. Often the emotionally
vulnerable person has very few skills to change his emotional state, and thus experi-
ences him to be a victim of his emotions, at the whim of mood and unable to control
his emotion-driven behavior. DBT emotion regulation skills address these exact prob-
lems through helping the person learn things he can do to avoid getting into an emo-
tionally dysregulated state, and helping him change negative emotions.
Emotion regulation skills are also applicable to the other behavioral patterns in
that many of them occur in the context of intense emotion. For example, the person
involved in a pattern of unrelenting crisis may be behaving in a way that increases crises
in her life because of an inability to regulate or change negative emotions. The actively
passive person may engage in that pattern in the context of feeling depressed, hope-
less or angry. Emotion regulation skills may help such a person change her emotional
state, thereby decreasing the likelihood of becoming actively passive.
Distress tolerance skills focus on teaching clients to get through times of intense,
painful emotions without exacerbating the situation. Rather than trying to change emo-
tional state, these skills help the person experience the moment as it is, getting through
the moment without blunting or escalating the negative affect. Distress tolerance skills
are most applicable to times when the person is in the midst of a distressing situation
that cannot be changed. Distress tolerance skills are applicable to the behavioral pat-
terns in that often times these patterns represent attempts to avoid or escape from emo-
tions or emotional pain. For example, inhibited grieving is a pattern in which the person
blocks experiencing of painful emotions. Unrelenting crisis behavior can reflect
attempts to avoid distress by creating distracting crisis situations. Distress tolerance
skills teach the individual to accept the current moment as it is. This does not mean
that the person passively gives up, but rather that he/she simply acknowledges that what
is, is. The person acknowledges her own pain, without the addition of a "and this has
to end" or "and I can't stand this". Developing the ability to tolerate distress obviates
the need for avoidance behaviors such as those sometimes involved in unrelenting crisis
or inhibited grieving patterns.
The behavioral patterns often are triggered by, or related to, interpersonal events
and difficulties. For example, active passivity often involves unskillful attempts to get
others to solve problems. Apparent competence can result from deficits in the area of
appropriate communication of degree of distress, and avoidance of asking for help
directly, sometimes because of lack of skill. DBT interpersonal effectiveness skills are
designed to teach clients how and when to ask for help in ways that will not destroy
important relationships. More generally, interpersonal effectiveness skills should lead

to fewer relationship difficulties that trigger negative affective states, making the person
less emotionally vulnerable.
SUMMARY
The biosocial theory underlying DBT emphasizes the role of emotion dysregula-
tion and an invalidating environment in the development and maintenance of border-
line personality disorder. This model, in conjunction with behavioral theory, has given
rise to a treatment model that has demonstrated effectiveness in reducing parasuicidal
behavior and increasing general functioning in individuals who meet criteria for BPD.
The DBT approach continues to be investigated, and is currently being tested with
several additional populations: borderlines who also abuse substances, suicidal adoles-
cents, eating disordered clients. The DBT model promotes a compassionate but effec-
tive stance with difficult to treat clients. It combines emphases on in-depth
understanding of the client's pain, with relentless effort to get the client to change, in
order to make his/her life better.
REFERENCES
Benjamin, L.S. and Wonderlich, S.A (1994). Social perceptions and borderline personality disorder: The rela-
tion to mood disorders. Journal of Abnormal Psychology, ]03,610-624.
Goodwin, 1M., Cheeves, K, and Connel, V. (1990). Borderline and other severe symptoms in adult survivors
of incestuous abuse. Psychiatric Annals, 20, 22-32.
Gunderson, 10. (1984). Borderline Personality Disorder. Washington, D.c.: American Psychiatric Press.
Kernberg, O. (1975). Borderline Conditions and Pathological Narcissism. Northvale, NJ: Aronson.
Koerner, K and Linehan, M.M. (1996). Case formulation in Dialectical Behavior Therapy for Borderline
Personality Disorder. In T. Eells (Ed.), Handbook of psychotherapy case formulation. New York:
Guilford Press.
Kohlenburg, R.l and Tsai, M. (1991). Functional Analytic Psychotherapy: Creating intense and curative
therapeutic relationships. New York: Plenum.
Koons, C.R., Robins, C.l, Bishop, G.K, Morse, lQ., Tweed, lL., Lynch, T.R., Gonzalez, AM., Butterfield,
M.I., and Bastian, L.A (1998). Efficacy of Dialectical Behavior Therapy with borderline women
veterans: A randomized controlled trial. Paper presented at the 32nd Annual Association for the
Advancement of Behavior Therapy Convention. Washington, D.C.
Laporte, L. and Guttman, H. (1996). Traumatic childhood experiences as risk factors for borderline and other
personality disorders. Journal of Personality Disorders, ZO,247-259.
Linehan, M.M. (1993). Cognitive-behavioral treatment of borderline personality disorder. New York:
Guilford.
Linehan, M.M. (in press). Validation and psychotherapy. In A Bohart and L.S. Greenberg (Eds.), Empathy
and psychotherapy: New directions to theory, research, and practice. Washington, DC: American
Psychological Association.
Linehan, M.M., Armstrong, H.E., Suarez, A, Allmon, D., and Heard, H.L. (1991). Cognitive-behavioral
treatment of chronically parasuicidal borderline patients. Archives of General Psychiatry, 48,
1060-1064.
Linehan, M.M. and Dimeff, L.A (1995). Extension of Standard Dialectical Behavior Therapy to Treatment
of Substance Abusers with Borderline Personality Disorder. Unpublished manual.
Linehan, M.M., Dimeff, L.A, Comtois, KA., and Kantor, 1 (1998). Dialectical behavior therapy for opiate
abusers with borderline personality disorder: Findings to date from two randomized controlled
trials. Paper presented at the 32nd Annual Association for the Advancement of Behavior Therapy
Convention. Washington, D.C.
Linehan, M.M., Thtek, B.A., Heard, B.A., and Armstrong, H.E. (1994). Interpersonal outcome of cognitive
behavioral treatment for chronically suicidal borderline patients. The American Journal of
Psychiatry, 151,1771-1776.
Stein, K.F. (1996). Affect instability in adults with a borderline personality disorder. Archives of Psychiatric
Nursing, 1, 32-40.
Strelau, 1., Farley, F.H., and Gale, A. (Ed.). (1986). The biological bases of personality and behavior: Vol. 2.
Psychophysiology: Performance and applications. Washington, D.C.: Hemisphere.
Ogata, S.N., Silk, K.R., Goodrich, S., Lohyr, N.E, Weston, D., and Hill, E.M. (1990). Childhood sexual and
physical abuse in adult patients with borderline personality disorder. American Journal of Psychiatry,
147,1008-1013.
Sansone, R.A., Sansone, L.A., and Fine, M.A. (1995). The relationship of obesity to borderline personality
symptomatology, self-harm behaviors, and sexual abuse in female subjects in a primary-care setting.
Wagner, A.W. and Linehan, M.M. (in press). The relationship between childhood sexual abuse and suicidal
behaviors in borderline patients. In M. Zanarini (Ed.), The Role of Sexual Abuse in the Etiology
of Borderline Personality Disorder.
15
PSYCHOPHARMACOLOGICAL TREATMENT
OF PERSONALITY DISORDERS
A Review
Peter Moleman,1* Karin van Dam,2 andVeron Dings2
1Moleman Research BV
Burg. Jhr. H. v.d. Boschstraat 11
3958 CA AMERONGEN
The Netherlands
2Afdeling Klinische Psychologie en Persoonlijkheidsleer
Katholieke Universiteit Nijmegen
Nijmegen, The Netherlands
1. SUMMARY
Studies for each of five (groups) of psychopharmaca are reviewed: antipsychotics,

antidepressants, benzodiazepines, lithium, and carbamazepine. Large placebo effects
are observed in all studies and show a positive effect on general level of funcioning.
Antipsychotics certainly are effective in borderline personality disorder. Classical (tri-
cyclic) antidepressants are found to be detrimental for borderline and/or schizotypal
patients, while MAO-I show beneficial effects in patients with borderline personality
disorder. Treating patients with personality disorders with benzodiazepines (anxiolyt-
ics) has a frail scientific basis. Lithium appears to be usefullness in a diversity of sub-
jects with aggression and/or impulsivity. Carbamazepine (an anti-epilepticum) may be
effective in patients with symptoms of behavioural dyscontrol and impulsiveness.
Issues relevant in our opinion to put the data into perspective are discussed at
the end of each section. At the end conclusions for treatment of personality disorders
with psychopharmacological agents are presented.
* Tel: (343) 461424, Fax: (343) 461080, e-mail: [email protected]

208 P. Moleman et al.
2. INTRODUCTION
Psychopharmaca are increasingly being used in treating patients with personality

disorders. These drugs have been developed for and their names coined after axis I
disorders (according to DSM terminology). The selection of patients to treat in that
case is a disease category. Personality disorders, however, often are dimensionally
described: in terms of complaints (either of the patients themselves or, more often, of
the family or "society") in different domains such as affective, impulsivity, and cogni-
tive domains. This raises the question whether psychopharmaca can be used to treat
such domains of complaints and how this relates to axis-I categories treated with these
drugs.
Another issue is related to the classification of personality disorders. We will
restrict this review to DSM (Diagnostic and Statistical Manual of mental disorders)
categories of personality disorders, where possible. Firstly, because of the advantage of
using a broad, well-known classification, which enhances an uncomplicated communi-
cation with other disciplines like psychology. Secondly, because most researchers have
used DSM-categories to describe their subjects and the outcomes of their studies. Thus,
using the DSM makes generalization of different studies possible.
3. METHODOLOGICAL ISSUES
Clinical studies with personality disorder patients pose methodological problems

in addition to those with other psychiatric patients. More than with axis I syndromes
(according to DSM terminology) the delineation of the syndrome can be problematic.
Connected to this is the problem of outcome measurement, because most personality
disorders comprise many symptoms, and it is not always clear whether the core symp-
toms are those which respond to treatment. What is reported as significant improve-
ment can consist of very different effects in terms of symptoms or symptomclusters,
and is in part dependent on the presence or absence of these symptoms at baseline.
Another problem is the high placebo response, which is connected in part to the atten-
tion and structure provided by the clinical study per se. For this reason, studies with
placebo control and a double"blind design are imperative for conclusions to be drawn.
Non blind studies can only be used as an indication that double-blind studies are war-
ranted. Studies without placebo control only permit conclusions if one of the treat-
ments is significantly superior to the other, because this excludes the possibility of all
treatments being equal in efficacy to placebo.
The different studies will be divided into two sections: early and uncontrolled
studies, and controlled studies. Early studies are reviewed together with uncontrolled
studies, because in studies before circa 1975 the DSM was not used. This may lead to
problems with generalisation. Controlled studies have to be double-blind and with
random treatment allocation and in most cases also placebo controlled.
4. ANTIPSYCHOTICS
The tranquillizing effect chlorpromazine was discovered by cOIncidence. It was

administered for the first time by the Frenchman Delay and Deniker to patients with
mania and later to schizophrenics. The drug was called "neurolepticum" because of
Psychopharmacological Treatment of Personality Disorders 209
the taming effect in the absence of clouding of consciousness. Later, the induction
of extrapyramidal side effects was also subsumed under the term "neuroleptic". In
addition, the drug appeared to suppress hallucinations and delusions as well. This
led to the most important indication of the neuroleptics: the treatment of psychoses.
Nowadays the term "neuroleptics" is replaced by "antipsychotics" to aknowledge their
therapeutic effect in a more restricted sense. Antipsychotics suppress psychotic symp-
toms regardless of the ethiology or diagnose (Molenman, 1992). This implies that
antipsychotics are used for symptomatic treatment and not for treatment of the
syndrome.
In addition, the expressions "typical" and "atypical" are used. In first instance, this
referred to antipsychotics with ("typical") and without ("atypical") extrapyramdal side-
effects. Nowadays, "typical" and "atypical" are sometimes added to indicate similarity
or dissimilarity with standard antipsychotics like haloperidol with regard to supposed
differences in efficacy against e.g. negative symptoms of schizophrenia. It is, however,
not clear what really differentiates atypical from typical antipsychotics (Casey, 1992).
However, it is not clear what properties are neccessary and beneficial in the treatment
of personality disorders. Often it is supposed that these are the same as in axis-I dis-
orders, but this can not be taken for granted, as will be discussed in this review. We
restrict ourselves for this reason to well characterized pharmacological properties of
antipsychotics if differences between them are to be discussed.
The anti psychotics available for treatment differ in a number of aspects. Most of
them are dopamine (D2) antagonists, but many possess also other properties with clin-
ical effects related to them (table 1). For most psychotic patients the antidopaminergic
effect is probably essential in controlling psychotic symptoms. Unfortunately, also the
most cumbersome side-effects-extrapyramidal-originate from this mechanism of
action, due to blockade of dopamine receptors in the extrapyramidal system. This often
causes side-effects like acute dystonia, Parkinsonism, rabbit-syndrome and acathisia,
and in the long run may result in tardive dyskinesia. The more or less pure dopamine
antagonists like haloperidol, fluphenazine, perphenazine, and trifluoperazine are most
likely to cause acute extrapyramidal side-effects.
Antipsychotics also block other receptors in different degrees. Strong antihista-
minerg effects can cause hynosedation and also weightgain, as with chlorpromazine,
chlorprotixene, clozapine, loxapine, olanzapine, and risperidone. Antipsychotics like
chlorpromazine, chlorprotixene, clozapine, olanzapine, periciazine, pipamperone,
respiridone, sertindole, and thioridazine have marked to pronounced aI-receptor
blocking properties (table 1). This can result in orthostatic hypotension, generalized
hypotension, and some hypnosedation. Blockade of muscarinic receptors is most pro-
nounced with clozapine, olanzapine, and thioridazine, which may cause a manifold of
anticholinergic side-effects, but also counteracts acute extrapyramidal side-effects
(table 2).
In the next section, the treatment of the borderline personality disorder with
antipsychotics is described seperately, because most studies are concerned with patients
with borderline personality disorder.
4.1. Antipsychotics: Early and Uncontrolled Studies

Reyntjens described in an open pilot study in 1972 the effect of pimozide in the
treatment of personality disorders. Subjects were 120 outpatients: 43 schizoId, 31 para-
nOId, 14 compulsive, 14 hysterical, 12 borderline, 4 "inadequate", and 2 with sexual devi-
210 P. Moleman et aL
Table 1. Effects of selected antipsychotics on receptors responsible for some desirable and
undesirable effects (see text)
receptor-antagonism
Generic name Alpha-1 His-1 Muscarine 5HT-2
benperidol -1 -2 -2 -1
broomperidol -1 -2 -2 -2
chloorpromazine +1 +1
chloorprotixeen +1 +1 +1
cJozapine +2 +2 +1 +2
Droperidol -2 -2 -1
flufenazine -1 -2 -1
flupentixol -1 -1 -1
haloperidol -1 -2 -2 -2
loxapine +1 -1 +1
olanzapine +1 +2 +1 +1
perfenazine -1 -2 -1
periciazine +1 -1 -2 +1
pimozide -1 -2 -2 -2
pipamperon +1 -2 -2 +2
risperidon +1 +1 -2 +2
sertindole +1 -2 -2 +2
sulpiride -2 -2 -2 -2
thioridazine +1
tiotixeen -1 -1 -2 -1
trifluoperazine -1 -1 -2 -1
zuclopentixol -1 -2 -1
Legend
Alpha-1: alpha-l adrenergic receptors.
His-l: histamine-l receptors.
Muscarine: muscarinic cholinergic receptors.
5HT-2: serotonin-2 receptors.
normal doses: up to 15 mg per day of haloperidol (see tabel 2 for equivalent doses).
-2: of no relevance at normal doses.
-1: weak at normal doses (some effects probable in vulnerable patients).
=: moderate at normal doses (effects probable in some patients).
+1: strong at normal doses (effects probable in many patients).
+2: very strong at normal doses (e.g. with +2 for His-l it can be difficult to achieve normal doses because of overwhelming
sedation).
(adapted from P. Moleman. Praktische psychofarmacologie. Bohn, Stafleu, van Loghem, Houten, 1992.)
ations. Global improvement with a dose of 3 mg per day of pimozide was reported as
excellent or good in 69% of the patients, as moderate in 27% and as poor in 8.5%.
Social and familial integration was found to be markedly improved, as was professional
performance (though to a lesser extend). All effects were significant if compared to the
pre-treatment level. Reyntjens concluded that pimozide might be highly effective in
the treatment of a broad range of personality disorders. Collard (1979) reached the
same conclusion in a non-controlled study on forty patients with some "social malad-
justments" in personality disorders. Improvement was observed especially in anxiety
and paranoidism.
Hymowitz et al. (1991) treated 17 patients meeting DSM-III criteria for schizo-
typal personality disorder (with no axes I diagnoses of schizophrenia or major affec-
tive illness) with low doses of haloperidol in a single-blind trial. The results reflect mild
to moderate improvement in approximately half of the patients. Improvement was
noted on overall schizotypal symptoms as assesed on the SIB (Schedule for interview-
Psycbopbarmacological Treatment of Personality Disorders 211
Table 2. Some anticholinergic (anti-muscarinic) effects

blurred vision
dry mouth
tachycardia
disturbed miction (urinary rention)
erection problems
obstipation (paralytic ileus)
glaucoma
sedation, tiredness, sleepiness
dizziness
nausea
vomiting
amelioration of acute extrapyramidal symptoms (acute dystonia, Parkinsonism, acathisia)
worsening of tardive dyskinesia
worsening of psychotic symptoms
anticholinergic syndrome:
.central: anxiety, agitation, disorientation, hallucinations, delirium, dysarthria, myoclonia, seizures
peripheral: tachycardia, arrhythmias, dilated pupils, warm dry skin, fever, dry mouth
(adapted from P. Moleman. Praktische psychofarmacologie. Bohn, Stafieu, van Loghem, Houten, 1992.)
ing borderlines), especially on the scales measuring ideas of reference, odd communi-
cation, and social isolation. This points to effects on psychotic symptoms proper. The
study also underscored serious compliance problems: 50% of the patients were unwill·
ing or unable to complete the full 6-week medication trial, due to extensive complaints
about side-effects, particualarly sedation. This high drop-out rate hampers interpreta-
tion of the study, because it may have introduced a serious bias.
Jensen and Anderson (1989) observed in an open study with 10 patients that
amoxapine improved schizophrenic-like and depressive symptoms in patients with
schizotypal personality disorders. No effects could be shown in patients with border-
line disorders.
4.1.1. Discussion. These poorly controlled studies do not allow any conclusions
to be drawn as to the usefulness of antipsychotics. They are cited here, because they
provide the only information on antipsychotic treatment in personality disorders other
than borderline disorder, although it is not clear what personality disorders in fact were
tr.eated. It appears prudent to restrict antipsychotic treatment to patients with unequiv-
ocal psychotic symptoms.
4.2. Antipsychotics: Early and Uncontrolled Studies in the Borderline

Personality Disorder
Brinkley et al. (1979) in a study of 5 cases reported possitive effects of antipsy-
chotics and emphasized that titration is a critical aspect of neuroleptic treatment, even
with the low doses used. The margin between doses resulting in improvement of symp-
toms and resulting in onset of side·effects appeared to be small. Also, neuroleptics that
lack sedating properties, such as thiothixene, trifluoperazine, and fluphenazine appear
to be more useful. Out of the five patients described, in the end two patients were
treated with perhenazine (6mg and 8mg per day, respectively), two with thiothixene
(lOmg and 4mg per day, respectively), and one with 2mg perfenazine per day (after a
possitive response on thioridazine, 25 mg nocte).
In another study (Cole et at., 1984) "core borderliners"-those who resembled

neither patients with depression nor with schizophrenia according to the authors-were
reported to fail to respond to drug therapy. On the other hand, borderline patients
meeting criteria for Major Depressive Disorder or with clear schizophrenic-like symp-
toms were likely to respond to medications appropriate to their axes I condition. The
general proposition by Brinkley et al. (1979) that borderline patients respond to low
doses of antipsychotics was supported in that 10 of the 17 patients improved.
4.3. Controlled Studies with Antipsychotics in the Borderline

Personality Disorder
Leone (1982) compared the effects loxapine succinate and chlorpromazine on
acute, disruptive symptoms in two groups of 40 borderline patients in a double-blind
study. Both groups improved significantly, the loxapine patients significantly more than
the chlorpromazine patients in several symptom areas. For example, loxapine was
rapidly effictive in controlling anxiety, hostility, suspiciousness, and depressed mood.
Montgomery and Montgomery (1982) carried out two parallel studies on the pre-
vention of suicidal behaviour. Most subjects had BPD, but a few had histrionic or
dependent personality disorders instead. Flupenthixol in a low dose of 20mg every 4
weeks was compared to placebo. There was a significant reduction in the number of
suicidal acts.
Serban and Siegel (1984) treated 52 borderline and schizotypal patients (and one-
third "mixed", i.e. fulfilling criteria for both diagnoses) with thiothixene (mean =
9.4mg) or haloperidol (mean = 3mg) in a double-blind study. Thiothixene resulted in
significantly more improvement of general, depressive, and paranoid symptoms in bor-
derline patients. Overall, 84 % patients were markedly to moderately improved at 3-
month follow-up. Areas of positive response were those of cognitive disturbance,
derealization, ideas of reference, anxiety, and depression. There was also a significant
improvement in self-image and social functioning, as measured by self-rating on the
Borderline Syndrome Index.
Goldberg et al. (1986) randomized fifty outpatients with borderline or schizo typal
personality disorder to thiothixene or placebo. Significant drug-placebo differences
were observed for psychotic symptoms like "dellusions", "ideas of reference", "psy-
chotism", and other symptoms, i.e. "obsessive-compulsive symptoms" and "phobic
anxiety", but not for "depression". Thiothixene mean dose was 8.7mg per day. Patients
responding to thiothixene were more severely ill at baseline with regard to delusions,
ideas of reference, psychotisism, phobic anxiety, and obsessive-compulsivity, which are
the symptoms that also responded best to thiothixene.
In a double-blind crossover trial with 16 borderline patients by Cowdrey and
Gardner (1988), alprazolam, carbamazepine, trifluoperazine, tranylcypromine, and
placebo were compared. The population was described as a homogeneous group of out-
patient women with BPD (DSM-III and Gunderson criteria), characterized in addition
by "hysteroid dysphoria " (closely related to "atypical depression" according to
Columbia University criteria) and serious behavioural dyscontrol (e.g. multiple over-
doses, selfmutilation, or physical violence). Patients were switched from one treatment
phase to another according to a restricted randomization scheme. Among those who
completed the trifluoperazine trial (50% )-i.e. who tolerated the antipsychotic-
superiority to placebo was shown, although the improvement was less than with
carbamazepine in the same patients. Thus, the effects of trifluoperazine observed were
restricted, although qualitatively parallel to those reported by Soloff et al. (1986). This
may be related to the selection of patients, and may not be generalizable to patients
with less severe BPD or to patients with prominent schizotypal symptoms.
Soloff et ai. (1986,1989) treated in a double-blind study 90 patients, diagnosed by
the Diagnostic Interview for Borderlines (1981) with haloperidol (mean dose 4.8mg by
day 35, range 4-16mg), amitriptyline (mean dose 149mg by day 35, range 100-175mg)
or placebo. As is observed in all studies with borderline patients, response on placebo
was considerable. Haloperidol was superior to both amitriptyline and placebo. There
was no difference between amitryptiline and placebo in general. Haloperidol produced
significant improvement over placebo in global functioning, depression, hostility,
schizotypal symptoms, and impulsive behaviour.
The authors detected three symptom change patterns: the "global depression",
"hostile depression", and "schizotypal" change factors. The identification of these
factors indicates that the effects of the treatments over time can be separated into more
general patterns. The global depression change factor accounted for 75% of the overall
variance in scores, but no differences were detected between the three treatments. This
indicates that in this area of most pronounced improvement placebo effects prevail, i.e.
hospitalization, milieu care, and structure provided, among others, by the study.
Haloperidol was superior on the hostile depression and schizotypal factor. Amitrypti-
line for the whole group was equal to placebo on the hostile depression factor, but this
response was composed of an amelioration in about half of the patients and a some-
times considerable worsening in the other half. On the schizotypal factor amitriptyline
was worse than placebo, if anything.
In another randomized, double-blind, placebo-controlled study Soloff et al. (1993)
treated 108 borderline inpatients (Gunderson's Criteria (1981)) with phenelzine,
haloperidol, or placebo with the aim to dissect apart affective and schizotypal symptom
sUbtypes. No consistent effect of haloperidol was observed. This result was surprising,
because the authors did not reproduce their own previous results. Differences between
studies in sample characteristics and study design may have contributed to the failure
to replicate. To cite the authors: "A retrospective comparison of the current patient
sample with our previous study sample demonstrates greater symptom severity in the
previous study (Solof et ai., 1989), with statistically significant differences in psychoti-
cism, schizotypal symptom severity, disruptive ward behaviour, and overall global
impairment. Severity of schizo typal symptoms was a predictor of favorable response
to haloperidol in the previous study" (Soloff et ai., 1993).
A 16-week follow-up of 54 patients from this 5-week study with 108 patients was
reported seperately (Cornelius et ai., 1993). Patients with at least some improvement
during the 5-week study were elegible for the follow-up study. Doses of haloperidol
and phenelzine were unchanged in general. Results showed few effects of haloperidol
or phenelzine. The drop-out rate in haloperidol was more than 50%, indicating the drug
was not well tolerated. To extend the study beyond the acute treatment phase of 5
weeks is of utmost importance in a chronic disorder like borderline personality disor-
der. However, it is not surprising that phenelzine did not show continued improvement
beyond 5 weeks, and that haloperidol did not show effects, since it was rather ineffec-
tive during the first 5 weeks. Placebo-controlled discontinuation is called for to evalu-
ate continued efficacy. The authors' conclusion that their findings suggest that there is
as yet no clear pharmacological treatment of choice for the continuation therapy of
borderline personality disorder is unsubstantial for this reason.
Table 3. Placebo-controlled studies with antipsychotics in borderline personality disorder

Patients Mean dose
Publication and design Diagnosis Drugs per day Duration Outcome
Goldberg n=50 BPD/SPD with thiothixene 9mg 12 weeks thiothix.
et al. inpatients; psychotic placebo >
(1986) parallel symptoms placebo
groups
Cowdry & n= 16 BPD/SPD with alprazolam 4.7mg 6 weeks tranyl. >
Gardner outpatients; loss of control carbamazepine 820mg carba. >
(1988) cross- and "hysteroid tranylcypromine 40mg trifluo. >
over dysphoria" trifluoperazine 7.8mg placebo
placebo > alpraz.
Soloff et al. n= 89 BPD/SPD haloperidol 4.8mg 5 weeks halop. >
(1989) inpatients (57% mixed) amitriptyline 149mg amitrip. =
placebo placebo
Soloff et al. n= 108 BPD/SPD haloperidol 4.0mg 5 weeks phenel. >
(1993) inpatients (61 % mixed) phenelzine 60mg halop. =
placebo placebo
4.3.1. Discussion. Antipsychotics have been shown to be instrumental in improv-

ing borderline patients in a number of well controlled studies (Gitlin, 1993; New et aI.,
1994). As with all treatments, the use of antipsychotics meets with considerable prob-
lems in borderline patients and the significant results cited above do not give a defi-
nite answer to the question what the proper clinical use of antipsychotics in borderline
patients is. A number of issues will be discussed to put the data into perspective: 1)
doses of antipsychotics, 2) severity of the disorder, 3) the categorical versus the dimen-
sional approach of the disorder, which is related to the question of treating the disor-
der or symptomclusters, and 4) the question which antipsychotic to choose for
borderline patients.
1) It is generally accepted that doses of antipsychotics for borderline patients
have to be low, i.e. in the range of several milligrams per day of haloperidol or 100-200
mg per day of chlorpromazine. These are low doses compared to those used for acute
and chronic treatment of schizophrenic patients until recently. However, it appears that
dosing for schizophrenic patients has been rather too high. Studies of McEvoy (1991)
and Vanputten (Vanputten et aI., 1992; Marder et aI., 1991) in acute treatment and of
Marder in preventive treatment (1991) have shown that minimum effective doses of
haloperidol are around 2.5 mg per day. This is in perfect agreement with recent positron
emission tomographic (PET) studies showing that these kind of doses result in exten-
sive ocupation of dopamine receptors (Kapur et aI., 1996; Rosenberg, 1994). In the light
of these data the doses used in borderline patients do not appear to be low, and it would
be worthwile to try even lower doses for borderline patients, particularly because of
the high susceptibility for side-effects, especially extrapyramidal. Whether this high sus-
ceptibility is of pharmacological or psychological origin can not be clarified at this stage;
it has not been objectivated whether borderline patients experience more side-effects
or whether the burden of-objectively not different-side-effects is heavier. Taken
together borderline patients should be treated-if at all-with doses of haloperidol in
the range of 0.5 to 5mg per day (the equivalent of 25 to 250mg chlorpromazine) or
equivalent doses of other antipsychotics (table 4).
Table 4. Equivalent doses of some selected antipsychotics

Generic name dose-equivalence
broomperidol 1
chloorpromazine 50
chloorprotixeen 50
clozapine 50--1oo?
flufenazine 0,5-2
flupentixol 1
haloperidol 1
loxapine 200
perfenazine 2-8
periciazine 5
pimozide (0,25-1)
risperidon 2?
sulpiride 200
thioridazine 50
tiotixeen 2
trifluoperazine 3-5
zuclopentixol 5
(adapted from P. Moleman. Praktische psychofarmacologie. Bohn, Stafleu, van
Loghem, Houten, 1992.)
2) It is quite clear from the well controlled studies that antipsychotics are indi-
cated only for the more severe forms of borderline disorder. Unequivocal effects were
observed in studies including a large proportion of patients with "mixed" borderline
disorder, i.e. fulfilling full criteria for schizotypal personality disorder. Indication for
antipsychotics may be best applied in a way similar to other disorders: they are in place
if and when the patient experiences psychotic symptoms and not if the symptoms are
only "quasi" psychotic. This may also explain why Soloff et aI. observed significant
effects of haloperidol in one study (1986), but not in another study (1993) with patients
with less psychoticism, schizotypal symptom severity, disruptive ward behaviour, and
overall global impairment.
The conclusion that effects are most pronounced in acute states has to be post-
poned until valuable follow-up studies are available. Anyway, significant treatment
effects were reported in a well controled study of 12 weeks duration (Goldberg et aI.,
1986).
3) What is being treated with antipsychotics in borderline patients? Soloff et aI.
(1993) in an elegant effort did not succeed in pharmacological dissection of the syn-
drome into an affective and a schizotypal subtype. We agree that this challenges the
basic assumption that the symptom domains-affective, cognitive, and impulsive-
represent biologically based trait vulnerabilities. But this not neccessarily "suggests that
low-dose neuroleptic functions as a nonspecific tranquilizer in BPD, reducing overall
acute sympom severity in the most impaired inpatients but contributing little in a less
impaired population"(Solof et aI., 1993) (cursivation by the present authors). When
patients respond to antipsychotics, not only psychotic, but also other symptoms
improve, like suicidality, aggression, and not in the least the affective symptoms. Gitlin
takes this as suporting the model that antipsychotics treat the disorder itself (1993).
However, the data can be incorporated into a dimensional model of treating symptom
clusters. It is plausible to assume that antipsychotics primarily act on psychotic symp-
toms, as discussed above. If or at the moment these symptoms are an important part
of what a patient experiences, suppression not only results in reduction of psychotic

symptoms, but also of other, connected symptoms. This hypothesis leans on the idea of
complex dynamic interactions between the different symptom domains. Affective, cog-
nitive and impulsive symptoms may to a certain degree be independent, because they
can but do not have to cooccur in a patient. The relative independent occurence of the
symptom domains induced the search for categorical sUbtypes. However, in an organ-
ism (patient) these domains interact dynamically, which can result in changes in one
domain if another domain is affected by treatment. Probably this only holds true if 1)
the domain treated is disturbed enough to impair proper functioning of other domains
of the patient, and 2) if treatment has strong enough effects on the domain that is being
treated. Thus, in a borderline patient with evident psychotic symptoms antipsychotics
may suppress the psychotic symptoms, which in its turn results in amelioration of the
affective, impulsive, and perhaps other symptoms. An important practical consequence
would be that in the selection of pharmacological treatment the symptoms are an
important guide. Thus, borderline patients should only be treated with antipsychotics
if and when they show psychotic symptoms. But in the evaluation of the efficacy of the
treatment differentiation of symptoms is not neccessarily very helpful, and effects do
not have to be expected to be restricted to psychotic symptoms.
Though purely hypothetical at this stage, parallels can be found in other disor-
ders. For instance, patients with severe depression can be treated effectively with
antidepressants. If a severely depressed patient has psychotic symptoms-like delu-
sions-addition of an antipsychotic is neccessary, and can be instrumental in the success
of the treatment. In that case not only the psychotic symptoms subside, but also the
affective symptoms. This has been taken by some authors to indicate that antipsycho-
tics can be antidepressant. We think this expresses a wrong concept. The proper
concept-we hypothesize-is that the antipsychotic only ameliorates the psychotic
symptoms and by way of doing so can induce amelioration of the depressive symptoms
too. A drug (e.g. an antipsychotic) should only be taken as having antidepressant prop-
erties if it ameliorates symptoms of depression in only, or primarily depressed patients.
This concept may not only have heuristic value for the study of the complex rela-
tionships between biological and psychosocial personality traits in borderline disorder
and its treatment, but also for the clinical treatment of patients. It fortifies the idea of
only prescribing antipsychotics to borderline patients with psychotic symptoms, as sug-
gested also in point 2). The fact that other symptom domains can also improve on
antipsychotics does not justify prescribing them to patients not experiencing psychotic
symptoms. This restriction may also provide a conceptual framework against
"medicalization of interpersonal problems" (Rosenberg, 1994) in that improvement
on all points of the patients responding to a treatment does not justify prescription of
that drug to all patients with the same diagnosis.
4) If the observation is correct that sedating antipsychotics are less effective in
borderline patients (Brinkley et al., 1979), those with antihistaminic properties should
be avoided. From this inventory of pharmacological properties it is evident that antipsy-
chotics like haloperidol are not very well tolerated by patients sensitive to extra-
pyramidal side-effects. The alternative is an antipsychotic with more anticholinergic
properties (see table 2). Whether borderline patients in general or some specific bor-
derline patients are sensitive to anticholinergic side-effects is unknown to us. Other
properties may also be of relevance in the treatment of borderline patients, not only
with regard to side-effects. Blockade of serotonin-2 receptors has been implicated in
reduced Parkinsonism, but also in anti-agressive effects. This could explain the supe-
rior efficicay of loxapine over chlorpromazine in one study (Montgomery et aI., 1982).
Because of the propensity of extrapyramidal side-effects, alternatives for haloperidol
have been looked for among the atypical antipsychotics (Dulz et aI., 1996). However,
controlled studies are lacking to evaluate their value in treatment. It may be relevant
that most of the atypical antipsychotics also possess strong anticholinergic, antihista-
minergic, or antiserotonergic properties to different degrees (e.g. table 1). It deserves
investigation whether reduced extrapyramidal side-effects or antihistaminergic or anti-
serotonergic properties contribute to possitive effects of antipsychotics in borderline
disorder.
A sensible approach in practice would be to treat a patient unresponsive to one
antipsychotic with one with different properties as summarized in table 1.
5. ANTIDEPRESSANTS
The most important categories of antidepressants were discovered by coInci-

dence. The Swiss Kuhn discovered in 1957 imipramine, the first tricyclic antidepressant
(or classical antidepressant). In the search of a cure for tuberculosis, iproniazide, the
first monoamine-oxidase-inhibitor (MAO-I), was discovered. MAO-I's inhibit the
enzymes monamine-oxidase A and B, which are responsible for the degradation of
the neurotransmitters noradrenaline, dopamine, and serotinine. Recently moclobemide
has been introduced, which is a relatively selective and reversible inhibitor of MAO-
A, which is responsible for the degradation of serotonine and noradrenaline. Also
more recently the selective serotine reuptake-inhibitors like fluoxetine have been
introduced.
The classical antidepressants like imipramine have many negative side-effects,
such as anticholinerge side-effects (dry mounth, blurred vision, obstipation, and
diffilculty urinating), cardiovasculair effects, hypnosedative effects, loss of libido,
tremors.
MAO-I can cause several side-effects like orthostatical hypotension and changes
in sleeppatterns. Most important is the risk of a hypertensive crisis due to inhibited
elimination of tyramine.
5.1. Antidepressants: Early and Uncontrolled Studies

Klein and Fink in 1962 reported an open trial with imipramine in 13 subjects with
"histronic labile affects and manipulative character", subjects that we would nowadays
probably classify as cluster-B personality disorders according to DSM-IV. The results
were dissapointing, the subjects responded poorly.
In a retrospective study using available clinical records of 62 patients, Cole et aI.
(1984) found that patients diagnosed as having borderline personality disorder and
meeting the criteria for major depressive disorder tricyclic antidepressants showed
some improvement.
In a group of 5 patients with borderline personality disorder and 5 patients with
schizotypical personality disorder, an open trial with amoxapine (with oxazepam)
showed some effects in the schizotypical subjects (Jensen et aI., 1989). Amoxapine may,
however, have some antidopaminergic, i.e. antipsychotic properties.
Three open, uncontrolled trials with the selective serotonin reuptake inhibitor
(SSRI) fluoxetine have been reported. Norden (1989) administered fluoxetine to 12
borderline patients not suffering from a major depression. Doses ranged from 5 mg to
40mg daily, depending on individual requirements. Eight patients were rated as much
improved or even as very much improved, although in three patients excessive agita-
tion occurred. This problem was reported to be resolved by reducing the doses.
Cornelius et aI. (1991) described five borde line personality disorder patients
with severe symptoms resistent to phenelzine and neuroleptics who responded well to
ftuoxetine, 20-40mg per day. They emphasized the effect on depressive and impulsive
symptoms. The latter are supposed to be associated with serotonin dysfunction.
Markovitz et aI. (1991) treated 22 patients with borderline or schizo typical per-
sonality disorder or both with ftuoxetine for twelve weeks. Doses were increased every
three weeks from 20mg to 80mg daily. Statistically and clinically significant improve-
ment on the Hopkins Symptom Checklist (SCL-90) were observed.
In a study with the mixed noradrenaline-serotonine reuptake inhibitor venlafax-
ine Markovitz et aI. (1995) treated 45 patients with borderline personality disorder for
12 weeks. Doses ranged from 200 to 400mg per day after the initial titration phase.
Again statistically and clinically significant improvement was reported; some but not
all patients responded to treatment.
5.2. Controlled Studies with Antidepressants in Borderline Disorder

A number of double blind studies have been performed in patients with person-
ality disorder and/or depression. The present concept is to consider these patients as
suffering from depressive disorders, especially dysthymia, and to connect therapy with
antidepressants to this disorder to the exclusion of personality disorders. This point of
view is summarized by the WPA dysthymia working group (Akiskal et aI., 1995). It is
beyond the scope of this review to enter into the discussion, but it implies that almost
nothing can be said about treatment of cluster B (depressive and related) personality
disorders with antidepressants. .
The only remaining personality disorder in which antidepressants have been
tested extensively is borderline personality disorder.
Montgomery and Montgomery (1982) carried out two parallel studies. Most sub-
jects had BPD but some had histrionic or dependent personality disorders. In one study,
mianserin was given in a dose of 30mg nightly and compared to placebo. In the patients
treated with mianserin there were slightly less suicidal acts compared with placebo but
was not significant.
The study of Soloff et aI. (1986, 1989) with amitryptiline and haloperidol in 90
patients with borderline personality disorder (Gunderson et aI., 1981) has been dis-
cussed under the subject heading of the antipsychotics. Amitriptyline, 100--175 mg per
day, showed little benefit. Some limited improvement on measures of depression-
although less than with haloperidol-was observed, but as many patients worsened as
improved. Earlier, Soloff et aI. (1995) elaborated on the paradoxal effects of amitripty-
line in 15 patients, consisting of paranoid ideation, impUlsive behaviour, and threaten-
ing with suicide.
The double-blind cross-over study with 16 borderline patients by Cowdry and
Gardner (1988) has also been described above (subject heading of the antipsychotics).
Tranylcypromine (a MAO-I), in an avarage dose of 40mg per day and carbamazepine
had the highest completion rates, which indicates they were best tolerated. The sub-
jects rated themselves as significantly improved only on tranylcypromine. Significant
improvement was ascertained by physicans on different scales like depression, anger,
rejection sensitivity, euphoria, capacity to experience pleasure, impulsivity, suicidality,

and global functioning. The improvement on anxiety was not significant, while, on the
contrary, the improvement on the depression scale was most significant.
Links et ai. (1990) compared desipramine and lithium with placebo. In this small
study with 17 patients with borderline personality disorder, lithium was superior to
placebo, especially on impulsivity and selfdestructive behavior. Desipramine (mean
dose 163mg per day) was not different from placebo.
Soloff et ai. (1993) compared phenelzine sulfate (a MAO-I) and haloperidol in
borderline personality disorder patients, as described above (subject heading of the
antipsychotics). The dose of phenelzine was stabilised at 60mg per day during the
second week and could be increased up to 90mg depending on the patients need. While
haloperidol surprisingly had no beneficial effects, phenelzine was slightly superior to
placebo on measures of depression, borderline psychopathologic symptoms, anxiety,
anger, and hostility, but phenelzine was not superior on measures of general pathology.
No efficacy of phenelzine was observed against atypical depression or hysteriod
dysphoria.
Recently, a double-blind placebo-controlled study of 12 weeks treatment with the
SSRI fluoxetine has been reported (Salzman et aI., 1995). Only 13 of the 22 subjects
fulfilled criteria for borderline personality disorder, the other subjects only exhibited
traits. Subjects were recruited by way of advertisement. This was, therefore, a study of
volunteers with mild to moderate symptoms. Response on 20-60mg per day of fluox-
etine was slightly better than on placebo. Only global mood and functioning, anger and
depression were assessed pre- and posttreatment. This small study in a peculiar popu-
lation does not allow for conclusions, but makes a controlled, double-blind study on a
larger scale in borderline personality disorder obligatory.
5.2.1. Discussion. Two studies with tricyclic-classical-antidepressants did not

show any beneficial effect, one with desipramine and one study with amitriptyline. In
fact, a considerable number of patients treated with amitriptyline experienced a dete-
rioration. It would appear that tricyclic antidepressants have a small place in the treat-
ment of borderline personality disorder, if any. Tricyclic antidepressants are mixed
noradrenaline- and serotonine reuptake-inhibitors with a number of side-effects, such
as anticholinerge side-effects (see table 4), cardiovasculair effects like orthostatic
hypotension and conduction abnormalities and hypnosedative effects.
Two studies with MAO-I, on the other hand, showed clear beneficial effects com-
pared to placebo, one with tranylcypromine and one study with phenelzine. These
studies with MAO-I merit a more extensive discussion. To put the data into perspec-
tive the following points will be discussed: 1) the effecacy of MAO-I in depressive dis-
orders,2) the patients selected for these positive studies, 3) the categorical versus the
dimensional approach of the disorder, which is related to the question of treating the
disorder or symptomclusters.
1) Recently, it has become clearer that there is some difference between the
depressed patients most benefitting from MAO-I compared to classical antidepressants.
While severely depressed inpatients appear to benefit less from MAO-I's, patients with
atypical features do appear to benefit more than from classical antidepressants (Thase
et aI., 1995). The core feature of atypical depression consists of "mood reactivity", i.e.
patients can still enjoy previously pleasant events while being depressed. In addition,
they exhibit at least two of the following symptoms: oversleeping, overeating, leaden
paralysis, and rejection sensitivity. About 75% of patients with this kind of atypical
Table 5. Placebo-controlled studies with antidepressants in personality disorders

Patients Mean dose
Publication Design Diagnosis Drugs Per day Duration Outcome
Cowdry & n= 16 BPD/SPD with alprazolam 4.7mg 6 weeks tranyl. >
Gardner outpatients loss of control carbamazepine 820mg carba. >
(1988) cross-over and "hysteroid tranylcypromine 40mg trifluo. >
dysphoria" trifluoperazine 7.8mg placebo
placebo > alpraz.
Soloff et al. n= 89 BPD/SPD haloperidol 4.8mg 5 weeks halop. >
(1989) inpatients (57% mixed) amitriptyline 149mg amitrip. =
placebo placebo
Links et al. n=17in-1 BPD desipramine 163mg 6 weeks lithium ~
(1990) outpatients lithium 986mg desipr. =
cross-over placebo placebo
Soloff et al. n = 108 BPD/SPD haloperidol 4.0mg 5 weeks phenel. ~
(1993) inpatients (61 % mixed) phenelzine 60mg halop.=
or BPD traits placebo placebo
depression respond to the MAO-I phenelzine, compared to about 50% response to

imipramine and about 25% response to placebo (Quitkin et aI., 1993).
2) The patients recruited for the Cowdry & Gardner (1988) study were suffering
from borderline personality disorder with loss of control and "hysteroid dysphoria".
The latter means they had symptoms similar to those of atypical depression. The fact
that the MAO-I tranylcypromine produced the most positive response in this study
could be dependent on this selection of patients. Soloff et al. (1993) took this as a
hypothesis for their study comparing haloperidol with the MAO-I phenelzine. In fact
they tried do dissect the borderline syndrome into an affective and a schizotypal sub-
syndrome, apparently by predicting response of a schizotypal subsyndrome to haloperi-
dol and of an affective sub syndrome to phenelzine. However, phenelzine was in general
not superior to placebo, but only on a scale for anger and hostility. The response was
certainly not specific against atypical depression and "hysteroid dysphoria". The
authors argue that this may be due to their focus being on the symptoms of atypical
depression rather than on the syndrome. However, the inclusion of less severe patients
in this study compared to previous ones may be of as much relevance.
3) While classical antidepressants were of little therapeutic value and often even
worsened symptoms, another class of antidepressants-MAO-I's-had beneficial
effects. In the study of Cowdry and Gardner (1988) these were more pronounced than
in the study of Soloff et aI. (1993). The patients in the Cowdry and Gardner study all
had "hysteroid dysphoria". It does not seem productive to invoke categories such as
DSM axis I disorders to explain these results. A dimensional approach seems to be
more in place. Probably, MAO-I's are most effective in borderline patients with atyp-
ical symptoms, but improvement is not restricted to these symptoms in a responding
patient. The studies with fluoxetine, although inconclusive, point in the same direction
as with the MAO-I's. If proof can be obtained that SSRI's are similarly effective as
MAO-I's, this would be of important practical significance in view of their compara-
tively benign side effects. MAO-I's can cause side-effects like orthostatic hypotension,
agitation, and a hypertensive crisis if tyramine-rich food is consumed. Puzzling and as
yet unexplained is the fact that response to phenelzine in the study of Soloff et aI. (1993)
was probably weak, because patients with milder symptoms were selected, while
response to ftuoxetine was observed in patients with borderline disorder or even bor-
derline traits of mild to moderate severity. This neccessitates replication, before con-
clusions for clinical practice can be drawn.
6. ANXIOLYTICS
Benzodiazepines are the most important anxiolytics. These drugs were discovered
in the fifties in the laboratories of Hoffmann-La Roche in Switserland. Benzodiazepines
activate the benzodiazepine-receptor which is connected to the GABA-receptor (0-
amino-butyric-acid receptor). These receptors are found in different areas of the
nervous system, such as the spinal cord, the limbic system, and the brains tern. The
administration of benzodiazepines amplifies the activity of GABA which results in
muscle-relaxation, fear reduction, anti-anxiety, anti-epileptic, and sedative effects.
The most important side-effects of benzodiazepines consist of sedation, muscle-
weakness, anterograde amnesia, and dependence. Important in the treatment of per-
sonality disorders are the paradoxical effects of benzodiazepines such as irritability,
outbursts of anger, suicidality, and panic-attacks.
The beta-blockers form another subgroup of the anxiolitics. Their use is restricted
to patients with social phobia, such as stage fear. In treatment of chronical anxiety their
efficacy has not been proven.
In a double blind, controlled trial chlordiazepoxide, oxazepam or placebo were
randomly assigned to 65 subjects with symptoms of temper outburst, irritability, hos-
tility, assaultive behaviour, and impulsiveness associated with anxiety. Compared to
placebo, significant reduction of anxiety, irritability, and hostility was found for
oxazepam. Chlordiazepoxide was less effective, showing only some improvement in irri-
tability. No paradoxical responses such as an increase in hostility, were observed (Lion,
1979).
In the double-blind, placebo-controlled, randomized trial by Cowdry and
Gardner (1988) in borderline personality disorder alprazolam (average dose 4.6 mg per
day) caused an increase in the severity of the episodes of serious dyscontrol these
patients had at baseline, and it caused increased suicidality. The authors comment that
this probably also holds for other benzodiazepines. For two of the 16 patients, though,
alprazolam caused clear cut benificial effects, superior to all other drugs tested.
6.1.1. Discussion. Data on the effects of benzodiazepines are very scarce. Few
beneficial effects have been reported. Frequently problems such as increased dyscon-
trol and rage reactions were related to the administration of benzodiazepines. Besides,
benzodiazepines can cause physical dependence, which is a problem particularly in
patients with personality disorders (Murphy et aI., 1991). Although frequently used, the
scientific basis for the use of benzodiazepines in personality disorders is frail.
7. LITHIUM
The effect of lithium was more or less discovered by coincidence by the

Australian Cade. Systematic application in psychiatry was prevented as a consequence
of fatal intoxications in the U.S.A. during the use of lithium chloride as a replacement
for common table salt. Later Schou discovered that lithium is a very effective mood
222 P. Moleman el aL
stabilizer, i.e. it can both treat the depressive as well as the manic episodes in bipolar
(manic-depressive) disorder. The intoxications could be prevented by attuning the dose
to the concentration of lithium in the blood.
Lithium can cause side-effects like trembling, tremor, nausea, skin reactions, and
weight increase. These side-effects generally have no serious consequences: they dis-
appear when the dose is being lowered or with the discontinuation of treatment. Also
neuropsychological side-effects have been described, especially concentration prob-
lems and memory disturban.ces. The most serious side-effects are concerned with the
kidneys and the thyroid gland. With approximately 5% of the patients, lithium inhibits
the action of the thyroid gland, which can give utterence to slowness and even to
depression. This can be treated with thyroid hormones.
7.1. Lithium: Early and Uncontrolled Studies

Rifkin et al. (1980) conducted a six-week, double-blind, random assignment
placebo-controlled study of lithium in 21 patients with an "emotionally unstable char-
acter disorder". The distance between the the extreme high and low mood within each
day, the "within-day fluctuation", was significant lower during treatment with lithium
carbonate than with placebo. According to Rifkin, the diagnose of "emotionally unsta-
ble character disorder" applies to patients with chronic maladaptive behavior patterns,
such as poor acceptance of reasonable authority, truancy, poor work history, manipu-
lativeness, with a core psychopathological disturbance of depressive, and hypomanic
mood swings that last hours to days. In the DSM-U, this diagnoses was subsumed by
hysterical or explosive character disorder. It has some resemblences with the border-
line personality disorder as described in the DSM-IV. With the antisocial personality
disorder it has socially unacceptable behaviour (delinquent mischief, irresponsibility:
avoidence of school or work, poor acceptance of reasonable authority) in common.
Also, there are traits of the histrionic personality disorder to be found in emotionally
unstable character disorder, like a tendency to sexually promiscuous behaviour and
overreactivity. Overall, the emotionally unstable character disorder is not fully compa-
rable to one of the present DSM categories, but it most closely resembles cluster
B-personality disorders.
Dale (1980) reported in 1980 on lithium treatment of fifteen aggressive,
mentally subnormal patients. In eleven cases, aggressive behaviour decreased. There
was no change in three cases and a deterioration in one patient. If the patients
responded well to the treatment, they did so remarkably quickly, within one or two
weeks. Unfortunately, this was not a double-blind study and the number of patients was
small.
Tupin et al. (1973) studied 27 male inmates of the psychiatric institution of the
California Department of Corrections who were all referred because of recurrent
violent behaviour. Twelve of them had a disorder of a mainly explosive type ("socio-
pathic"), three had other diagnoses, 8 schizophrenia, and four possible schizophrenia;
they all had a pattern of recurring, easily triggered violence. As result of 1,5 years of
treatment with lithium carbonate, fifteen of 22 subjects showed a marked decrease in
the number of violent prison infractions, 3 showed an increase, and 4 showed no change.
Many subjects reported an increased capacity to reflect, a frame of mind Monroe (1970)
calls "reflective delay". One prisoner for example stated:" Now I can think about
whether to hit (him) or not". Also, they felt an increased capacity to control angry
feelings when provoked and there was a diminished intensity of angry affect and, finally,
an increased capacity to reflect on the consequences of actions. Unfortunately, there

was no placebogroup involved and the study was not blind. However, this study
suggests that lithium may be effective for long-term control of aggressive, violent
behaviour.
7.2. Controlled Double-Blind Studies with Lithium

A double-blind placebo controlled study by Sheard et al. (1976) 66 prisoners
ranging in age from 16 to 24 years with histories of chronic impulsive behaviour (and
continued chronic assaultive behaviour while in prison) received lithium for up to 3
months. The lithium group had significantly fewer major infractions (serious threaten-
ing behaviour or actual assaults) and fewer total infractions (less serious offenses such
as being out of place or possession of contraband) than the placebo group. Minor infrac-
tions showed a tendency to increase during the medication period for both placebo and
lithium groups. This indicates that lithium did more than globally inhibiting behaviour.
On stopping the lithium there was an immediate rebound increase in the number
of major infractions. The authors suggest that lithium can have a clinically useful effect
upon impulsive aggressive behaviour when this behaviour is not associated with
psychosis.
Links et al. (1990) reported preliminary results of a double-blind placebo-
controlled cross-over trial with lithium (mean dose 986mg per day) and desipramine
in 17 patients with borderline personality disorder. The results hinted at superiority of
lithium, especially on anger and suicidal symptoms. Effects were superior to placebo
when rated by therapists but not by the patients.
7.2.1. Discussion. From the studies available and from case reports it is reason-
able to conclude that lithium is effective in reducing aggression and probably also self-
mutilating behaviour (Stein, 1992). It appears that this holds for different individuals,
such as mentally handicapped, psychotic or delinquent SUbjects. Therefore, patients with
serious episodes of aggressive behavior extending over years and unresponsive to other
treatments should be given the benefit of a three-month trial of lithium. Useful point-
ers to lithium responsiveness include mood disturbance, aggressive behaviour in the
context of anger, a family history of classic affective disorder, and a personal or family
history of alcoholism (Stein, 1992).
Because of the risk of intoxication, especially when the patient is prone to chaotic
intake of the drug (Thiel et al., 1993), it would seem to us, that treatment with lithium
should be reserved for patients with severe symptomatology and unresponsive to other
treatments. Lithium should always be given under close supervision.
8. CARBAMAZEPINE
Carbamazepine is an anti-convulsant which was observed to have possitive effects

on behaviour of epileptic patients. This led to trials with carbamazepine in psychiatric
and other patients with disruptive behaviour. Nowadays, carbamazepine is an accepted
alternative for lithium in the treatment of mania; and in the treatment of resistent
depressions as addition to an antidepressant. It is also added to lithium in the preven-
tive treatment of bipolar disorder, if lithium alone is not effective or in the case of rapid
cycling (Dardennes et al., 1995; Albani et al., 1995; Calabrese et al., 1996).
8.1. Carbamazepine: Early and Uncontrolled Studies

Tunks and Dermer (1977) reported the first succesful use of carbamazepine (800
mg per day) in a case of episodic behavioural dyscontrol, which was reportedly related
to limbic dysfunction.
In a study of eighty patients with diverse diagnoses, Mattes (1984) reported a
decrease of temper-outbursts on carbamazepine in those patients that were diagnosed
as intermittent explosive disorder according to DSM-III.
8.2. Controlled Studies with Carbamazepine

In the double-blind, cross-over study of 16 borderline patients by Cowdry and
Gardner (1988), carbamazepine showed some efficacy. Symptoms of aggression and
impulsiveness improved, although the anti-psychotic drug trifiuoperaze led to more
beneficial effects.
8.2.1. Discussion. Several trials have shown carbamazepine to be effective in

treating symptoms of behavioural dyscontrol and impulsiveness in patients with
hetrogeneous psychiatric diagnoses, including borderline and anti-social personality
disorders.
Neppe (1983) and Gardner and Cowdry (1986) reported that carbamazepine may
create some time to weigh decisions, instead of acting impulsively.
Considering symptoms of impulsiveness and behavioural dyscontrol, the use of
carbamazepine should be a standerd consideration in the treatment of patients with
personality disorders, in particular borderline and anti-social personality disorders.
9. GENERAL CONCLUSIONS
Controlled studies of pharmacotherapy in personality disorders is almost entirely

restricted to borderline personality or closely related disorders. Only lithium has been
studied to some degree in other personality disorders. Lithium will, therefore, be dis-
cussed under this general heading, and all other drugs under the heading of borderline
personality disorder.
9.1. Lithium
Lithium was superior to placebo in one small study. In view of its side-effects and
the risk of intoxication in patients with poor compliance it should probably be tried
only in patients with serious episodes of aggression extending over years.
10. CONCLUSIONS ON BORDERLINE PERSONALITY

DISORDERS
10.1. Antipsychotics
The four placebo-controlled trials with antipsychotics, in our opinion, prove effi-
cacy beyond reasonable doubt. The dissection of borderline disorder into subcategories
has not been succesful, but the data point to characteristics of patients responsive to
antipsychotics. It appears that antipsychotics should be reserved for patients with any
or all of the following: 1) severe symptoms, 2) a co-diagnosis of schizotypal personal-
ity disorder, or 3) with unequivocal psychotic symptoms.
Although antipsychotics have been evaluated in clinical trials of short duration
of e.g. 5 weeks, and not in continuation-or preventive trials of longer duration, they
have not been tested for use in situations of emergency or crisis.
If a patient responds to an antipsychotic, not only psychotic symptoms, but also
affective, cognitive and impulse-related symptoms ameliorate. Thus, in clinical practice
a treatment with antipsychotics should only be regarded as fully succesful, if more than
only psychotic symptoms respond.
Doses of antipsychotics to be used are around 0.5-5mg of haloperidol or 50-250
mg of chlorpromazine per day. In the light of present knowledge that this is also the
optimal dose-range for many schizophrenic patients, higher doses should only be used
in exceptional cases.
The choice between antipsychotics can not be guided by data from research, since
few antipsychotics have been studied. Since it is not known whether other properties
besides antidopaminergic effects are of relevance for borderline patients, it is advisable
to try an antipsychotic with different pharmacological properties in a patient not
responding to the present antipsychotic.
10.2. Antidepressants
From the few placebo-controlled studies it appears that classical antidepressants
like amitryptiline and desipramine are ineffective in borderline patients, and can better
be avoided because of the risk of paradoxal effects.
On the other hand, both MAO-I tested-tranylcypromine and phenelzine-have
shown unequivocal beneficial effects. As with the antipsychotics, especialy the more
severe patients seem to benefit and, if a patient responds, not only affective symptoms
ameliorate. It appears that not only-and in fact perhaps not specifically-patients with
depressive symptoms or with symptoms of "hysteroid dysphoria" or atypical depres-
sion respond.
A problem with the MAO-I's is the risk of a hypertensive crisis if the tyramine-
restricted diet is not adhered to, a risk not to be disregarded in borderline patients.
However, an alternative in the selective MAO-I's or in the SSRI's has not yet been
ascertained, due to the lack of controlled studies.
10.3. Benzodiazepines
The only benzodiazepine tested has been alprazolam in rather high doses
around 4.6mg per day. It had more detrimental than beneficial effects in borderline
patients.
10.4. Carbamazepine
Carbamazepine showed some promIsmg results, especially on symptoms of
aggression and impulsiveness. As with lithium, it should probably be restricted to
patients with severe symptoms. Carbamazepine has a less narrow therapeutic window
than lithium, but it has also severe side-effects.
226 P. Moleman et 01.
REFERENCES
Akiskal H.S., Silva IA.C.E., Frances A. et al. (1995) Dysthymia in clinical practice. Br. 1. Psychiatry,
166:174-183.
Albani F., Riva R, and Baruzzi A. (1995) Carbamazepine clinical pharmacology: A review. Pharmacopsy-
chiatry,28:235-244.
Brinkley IR, Beitman B.D., and Friedel RO. (1979) Low-dose neuroleptic regimens in the treatment of
borderline patients. Arch. Gen. Psychiatry, 36:319-326.
Calabrese IR, Fatemi H., Kujawa M. et al. (1996) Predictors of response to mood stabilizers. 1. Clin.
Psychopharmacol.,6:S24--S3l.
Casey D.E. (1992) What makes a neuroleptic atypical? In: Novel antipsychotic drugs. New York: Raven Press,
241-25l.
Cole 10., Salomon M., Gunderson I et al. (1984) Drug therapy in borderline patients. Comprehensive
Psychiatry, 25:249-254.
Collard I (1979) Pimozide in the treatment of some "social maladjustments" in "personality disorders". Acta
Psychiatrica Belgica, 79:686-703.
Cornelius IR, Soloff P.H., Perel 1M. et al. (1991) A Preliminary Trial of Fluoxetine in Refractory Border-
line Patients. 1. Clin. Psychopharmacol., 11:116-120.
Cornelius IR, Soloff P.H., Perel 1M. et al. (1993) Continuation Pharmacotherapy of Borderline Personality
Disorder with Haloperidol and Phenelzine. Am. 1. Psychiatry, 150:1843-1843.
Cowdry RW. and Gardner D.L. (1988) Pharmacotherapy of borderline personality disorder; alprazolam,
carbamazepine, trifluoperazine, and tranylcypromine. Arch. Gen. Psychiatry, 45:111-119.
Dale P.G. (1980) Lithium therapy in aggressive mentally subnormal patients. Br. 1. Psychiatry, 137:469-474.
Dardennes R, Even c., Bange F. et al. (1995) Comparison of carbamazepine and lithium in the prophylaxis
of bipolar disorders-a meta-analysis. Br. 1. Psychiatry, 166:378-38l.
Dulz B., and Schneider A. boek, editor. (1996) Borderline-StOrungen; Theorie und Therapie. 2nd: Stuttgart:
SchattauerVeriagsgesellschaft; 110.
Gitlin M.I (1993) Pharmacotherapy of Personality Disorders-Conceptual Framework and Clinical
Strategies. 1. Clin. Psychopharmacol., 13:343-353.
Goldberg S.c., Schulz S.c., Schulz P.M. et al. (1986) Borderline and schizotypal personality disorders treated
with low-dose thiothixene vs placebo. Arch. Gen. Psychiatry, 43:680-686.
Gunderson IG., Kolb IE., and Austin V. (1981) The diagnostic interview for borderline patients. Am. 1.
Hymowitz P., Frances A., Jacobsberg L.B. et al. (1991) Neuroleptic treatment of schizotypal personality
disorders. Comprehensive Psychiatry, 48:1064--1067.
Jensen H.Y. and Andersen I (1989) An open, noncomparative study of amoxapine in borderline disorders.
Acta Psychiatr. Scand., 79:89-93.
Kapur S., Remington G, Jones C. et al. (1996) High levels of dopamine D-2 receptor occupancy with
low-dose haloperidol treatment: A PET study. Am. 1. Psychiatry, 153:948-529.
Klein D.F. and Fink M. (1962) Psychiatric reaction patterns to imipramine. Am. 1. Psychiatry, 119:432-438.
Leone N.F. (1982) Response of borderline patients to loxapine and chlorpromazine. 1. Clin. Psychiatry,
43:148-150.
Links P.S., Steiner M., Boiago I. et al. (1990) Lithium therapy for borderline patients: preliminary findings.
Journal of Personality Disorders, 4:173-18l.
Lion IR (1979) Benzodiazepine in the treatment of aggressive patients. 1. Clin. Psychiatry, 40:70-109.
Marder S.R, Midha K.K., Vanputten T. et al. (1991) Plasma Levels of Fluphenazine in Patients Receiving
Fluphenazine Decanoate-Relationship to Clinical Response. Br. J. Psychiatry, 158:658-665.
Markovitz P.I, Calabrese IR, Schulz S.c. et al. (1991) Fluoxetine in the Treatment of Borderline and
Schizotypal Personality Disorders. Am. 1. Psychiatry, 148:1064--1067.
Markovitz P.I and Wagner S.c. (1995) Venlafaxine in the treatment of borderline personality disorder.
Psychopharmacol. Bull., 31:773-777.
Mattes IA. (1984) Carbamazepine for uncontrolled rage outbursts. Lancet, ii:1164--1165.
Mcevoy IP., Hogarty G.E., and Steingard S. (1991) Optimal Dose of Neuroleptic in Acute Schizophrenia-
A Controlled Study of the Neuroleptic Threshold and Higher Haloperidol Dose. Arch. Gen. Psychi-
atry,48:739-745.
Moleman P. (1992) Praktische Psychopharmacologie. 2nd: Houten: Bohn Stafleu van Loghem.
Montgomery S.A. and Montgomery D. (1982) Pharmacolgical prevention of suicidal behaviour. J. Affective
Disord.,4:291-298.
Murphy S.M. and lYrer P.A (1991) Double-Blind Comparison of the Effects of Gradual Withdrawal of
Lorazepam, Diazepam, and Bromazepam in Benzodiazepine Dependence. Br. 1 Psychiatry,
158:511-516.
New AS., Trestman RL., and Siever L.J. (1994) The pharmacotherapy of borderline personality disorder.
CNS Drugs, 2:347-354.
Norden M.J. (1989) Fluoxetine in borderline personality disorder. Progr. Neuro-Psychopharmacology &
Biological Psychiatry, 13:885-893.
Nyberg S., Nordstrom AL., Halldin C. et al. (1995) Positron emission tomography studies on D-2 dopamine
receptor occupancy and plasma antipsychotic drug levels in man. Int. Clin. Psychopharmacol.,
10:81-85.
Quitkin EM., Stewart J.w., Mcgrath P.J. et al. (1993) Columbia Atypical Depression-A Subgroup of
Depressives with Better Response to MAOI Than to Tricyclic Antidepressants or Placebo. Br. 1
Reyntjens AM. (1972) A series of multicentre trials with pimozide in psychiatric practice. Acta Psychiatrica
Belgica,72:30-36.
Rifkin A, Quitkin E, and Carillo C. (1972) Lithium carbonate in emotionally unstable character disorder.
Arch. Gen. Psychiatry, 27:519-523.
Rosenberg R (1994) Borderline States-Pharmacotherapy and Psychobiology of Personality-A Discussion
of Soloffs Article. Acta Psychiatr. Scand., 89:56-60.
Salzman C., Wolfson AN., Schatzberg A et al. (1995) Effect of fluoxetine on anger in symptomatic
volunteers with borderline personality disorder. 1 Clin. Psychopharmacol., 15:23-29.
Serban G. and Siegel S. (1984) Response of borderline and schizotypal patients to small doses of thithixene
and haloperidol. Am. 1 Psychiatry, 141:1455-1458.
Sheard M.H., Marini J.L., Bridges C.I. et al. (1976) The effect of lithium on impulsive aggressive behaviour
in man. Am. 1 Psychiatry, 133:1409--1413.
Soloff P.H., George A, Nathan S. et al. (1986) Progress in pharmacotherapy of borderline disorders; a double
blind study of amitriptyline, haloperidol, and placebo. Arch. Gen. Psychiatry, 43:691--697.
Soloff P.H., Anselm G., Nathan S. et al. (1986) Paradoxical effects of amitriptyline on borderline patients.
Am. 1 Psychiatry, 143:1603-1605.
Soloff P.H., George A, Nathan S. et al. (1989) Amitriptyline versus haloperidol in borderlines: final outcomes
and predictors of response. 1 Clin. Psychopharmacol., 9:238--246.
Soloff P.H., Cornelius J., George A et al. (1993) Efficacy of Phenelzine and Haloperidol in Borderline
Personality Disorder. Arch. Gen. Psychiatry, 50:377-385.
Stein G. (1992) Drug Treatment of the Personality Disorders. Br.l Psychiatry, 161:167-184.
Thase M.E., Trivedi M.H., and Rush AJ. (1995) MAOIs in the contemporary treatment of depression.
Neuropsychopharmacology,12:185-219.
Thiel A, Nau R, Lehmann K. et al. (1993) Intoxication in Manic Patients Following Chaotic Self
Administration of Lithium. Acta Psychiatr. Scand., 88:289--291.
Tunks E.R and Dermer S.w. (1977) Carbamazepine in the dyscontrol syndrome associated with limbic
dysfunction. 1 Nerv. Ment. Dis., 164:56-63.
Tupin IP., Smith D.B., Clanon T.L. et al. (1973) The long-term use of lithium in aggressive prisoners.
Comprehensive Psychiatry, 14:311-317.
Vanputten T., Marder S.R, and Mintz J. (1990) A Controlled Dose Comparison of Haloperidol in Newly
Admitted Schizophrenic Patients. Arch. Gen. Psychiatry, 47:754-758.
Vanputten T., Marder S.R, Mintz J. et al. (1992) Haloperidol Plasma Levels and Clinical Response-A
Therapeutic Window Relationship. Am. 1 Psychiatry, 149:500-505.
Wickham E.A and Reed J.V. (1987) Lithium for the control of aggressive and self-mutilating behaviour. Int.
Clin. Psychopharmacol., 2:181-190.
16
NEW DRUGS IN THE TREATMENT OF

BORDERLINE PERSONALITY DISORDER
F. Benedetti, C. Colombo, L. Sforzini, C. Maffei, and E. Smeraldi
Istituto Scientifico Ospedale San Raffaele

University of Milan, School of Medicine
Via Prinetti 29-20127 Milano-Italy
1. THE BORDERLINE DIAGNOSIS
From both a clinical and a neurobiological point of view, Borderline Personality

Disorder (BPD) as defined by DSM-IV (A.P.A., 1994) must be considered an hetero-
geneous condition. Clinical studies on BPD patients showed that BPD diagnosis is
rarely the only Axis II diagnosis present in a personality disordered patient,
and broadly overlaps with other diagnoses pertaining to the three DSM category
clusters (e.g. Nurnberg et aI., 1991). Moreover, comorbidity with Axis I pathologies is
often present: as reported in DSM-IV itself, heterogenous Axis I pathologies (such
as mood disorders, substance-related disorders, eating disorders, post-traumatic stress
disorders, and attention-deficit disorder) are commonly diagnosed in patients affected
by BPD.
This heterogeneity may account for the traditional split, in psychiatric literature,
between two schools of thought which underline the relationship of BPD with either
schizophrenia or mood disorders (e.g. Cowdry, 1987), and may also account for the con-
trasting results reported by psychobiologic research on BPD patients (see review by
Coccaro and Siever, 1995). A categorially defined association of basic impulsivity and
affective instability with transient cognitive and perceptual disturbances could then be
the common clinical feature of different neurobiologic substrates.
In the clinical practice, the lack of strictly defined diagnostic boundaries and of
reliable neurobiologic alterations in BPD hampers the choice of drug therapies for
BPD patients; up to now, no clear-cut guidelines for pharmacological therapy in BPD
Tel: 0039/2/26433229, Fax: 0039/2/26433265, e-mail: [email protected]
Treatment of Personality Disorders, edited by Derksen et aI.

Kluwer Academic 1 Plenum Publishers, New York, 1999. 229
230 F. Benedetti et al.
may be drawn from literature. Given the above, psychotherapeutic interventions must
still be considered the basis of any clinical approach to borderline personality pathol-
ogy. Nevertheless, the intensity of symptomatology experienced by patients affected by
severe BPD leads them to make an extensive use of health services, and many of them
can only be managed with the use of psychotropic medication (Skodal et aI., 1983;
Zanarini et aI., 1988; Stein, 1992; Soloff, 1994).
In the presence of Axis I comorbidity, the pharmacological treatment must
focus on the Axis I pathology (e.g. Stone, 1989). In this respect it should be noted that
most studies report a worse outcome for Axis I disorders in the presence of BPD
comorbidity (e.g. Gitlin, 1993; Casey et aI., 1996).
In the absence of Axis I diagnoses, psychiatric studies focused on the operational
definition of therapeutic targets for drug therapy. In clinical practice, targets are
selected basing on the clinical evaluation of the relative relevance of the different symp-
tomatological features which, from among the polymorphous and fluctuating BPD
symptomatology, bring the patient to the psychopharmacologist's attention: this means
that the pharmacological treatment of BPD must narrowly focus on specific sympto-
matological domains which characterize the clinical presentation of patients (Soloff,
1994; Hirshfeld, 1997; Coccaro, 1998). From a methodological point of view, this
approach could be useful for a pharmacological dissection of BPD as a diagnostic cat-
egory, with the identification of homogenous subgroups of patients biologically defined
on the basis of a common response to the same psychotropic drug (for a discussion on
these points, see Soloff, 1994; Rosenberg, 1994; Coccaro and Siever, 1995).
Following this approach, three main symptomatological clusters, corresponding
to different modalities of clinical emergence, have been empirically isolated as targets
for drug therapy in BPD patients: impulsivity, affective-related symptoms, and
psychotic-like symptoms.
2. IMPULSIVITY
Early clinical perpectives on the treatment of impulsivity in BPD patients based

on the analogy between the sometimes dramatic outbursts of behavioral dyscontrol
shown by BPD patients and the periodic, explosive behavioral abnormalities observed
in patients affected by complex partial seizures.
The hypothesis of an "epileptoid overactivity" of limbic system structures
common to BPD and epileptic patients (Cowdry et aI., 1980; Andrulonis et aI., 1981)
lead Gardner and Cowdry (1986) to try carbamazepine in BPD patients. In a double-
blind placebo controlled crossover trial, carbamazepine produced a decrease in
episodes of behavioral dyscontrol, angry outbursts, and suicidality. It should be noted
that in this study the 16 female subjects were selected for having extensive histories of
behavioral dyscontrol in the absence of abnormal EEG activity, and that 50% of the
sample had a lifetime history of mood disorders, a diagnostic category which has been
shown to positevely resent of the mood stabilizing effects of carbamazepine (e.g.
Watkins et aI., 1987).
More recently, Stein et ai. (1995) conducted a preliminary open label trial of val-
proate in 11 BPD patients, showing a broad amelioration of BPD symptomatology in
half of the subjects, without specific effects on the different symptom domains; and
Wilcox (1995) reported a positive effect on agitation and anxiety of open label adminis-
tered divalproex sodium in BPD patients. Moreover, several anticonvulsivants (carba-
New Drugs in the Treatment of Borderline Personality Disorder 231
mazepine, valproate, and phenytoin) have been shown to share therapeutic effects in the
control of behavioral impulsivity in different psychopatological conditions (Mattes, 1990;
Giakas et aI., 1990; Hasan et aI., 1990; Barratt, 1991; Keck et aI., 1992).
However, the efficacy of anticonvulsivants in controlling BPD impulsivity must
still be considered an area of debate, since a recent double-blind parallel placebo-
controlled trial on 20 BPD patients failed to find any therapeutic effect of carba-
mazepine (De la Fuente et aI., 1994). The same research group reported a 40% inci-
dence of electroencephalographic abnormal diffuse slow activity in a sample of 20 BPD
patients, which was not influenced by carbamazepine (De la Fuente et aI., 1998).
More recent researches focused on the role of brain serotonin in the control of
impulsive behaviors. Animal research shows that a reduction in the activity of brain
serotoninergic pathways results in an increase in aggressive behavior, sexual behavior,
and ethanol intake, while the administration of drugs that enhance serotonergic func-
tion results in an inhibition of these behaviors (e.g. Coccaro, 1989). Remarkably, the
behavioral dishinibition which follows a reduction in serotonergic activity does not con-
figure itself as a purposeless hyperactivity, but as an overreactivity to enviromental
stimuli. A neurophysiological role of behavioral inhibition has been then proposed for
brain serotoninergic pathways, with a direct relationship between the level of brain
serotonin activity and the threshold for reaction to enviromental stimuli.
Following this perspective, it has been hypothesized that the basic impulsivity
shown by BPD patients could be due to a dysregulation of central serotonergic func-
tion. Several neurobiologic findings have been reported in agreement with this hypoth-
esis. A blunted prolactine response to the non specific serotonin release/reuptake
blocker fenfluramine was inversely correlated with impulsive aggression and with an
history of suicide attempts in a sample of 8 BPD patients (Coccaro et aI., 1989a). The
same research group reported an inverse correlation between a blunted prolactine
response to the 5HT1a agonist buspirone and irritability in a sample of personally dis-
ordered patients (Coccaro et aI., 1989b). Correlations between platelet serotonergic
function (serotonin content, MAO activity, and paroxetine binding) and impUlsivity and
affective instability in BPD have been recently observed (Verkes et aI., 1998); the same
group reported association between platelet serotonergic function and recurrence of
suicidal behavior in BPD (Verkes et aI., 1997). Moreover, the acute administration of
the partial serotonin agonist m-chlorophenylpiperazine caused a decrease in anger
and fear in 12 patients affected by BPD (Hollander et aI., 1994), and a blunted pro-
lactine response with an increased cortisol response to m-chlorophenylpiperazine were
correlated with assaultiveness in antisocial patients (Moss et aI., 1990).
Basing on this preclinical and clinical findings, the administration of serotoniner-
gic drugs has been investigated in the treatment of BPD impUlsivity. The admin
istration of the selective serotonin reuptake inhibitor (SSRI) fluoxetine has been
reported to ameliorate impulsivity and global psychopathology in case reports
(Coccaro et aI., 1990; Hull et aI., 1993), in open label trials (Norden, 1989; Cornelius et
aI., 1990; Markovitz et aI., 1991), and in a double-blind placebo controlled-trial
(Salzman et aI., 1995). Remarkably, in the controlled study the amelioration in impul-
sivity after fluoxetine was found to be independent from changes in depressive symp-
tomatology. The SSRI sertraline has been investigated in an open label trial, and
reported to improve impulsive-aggressive behaviors and irritability in BPD patients
(Kavoussi et aI., 1994).
Contrasting results have been reported following the administration of less
specific drugs. The MAOI phenelzine has been shown to acutely ameliorate anger and
232 E Benedetti et al.
hostility, but chronically worsen excitement and reactivity in BPD patients (Soloff et
aI., 1993; Cornelius et aI., 1993). The MAOI tranylcypromine produced a decrease in
rage which followed amelioration in affective instability (Cowdry and Gardner, 1988).
Lithium salts showed a therapeutic effect on behavioral dyscontrol, possibly because
of their hypothesized proserotoninergic and anticatecholaminergic action (Shader
et aI., 1974; Sheard, 1975; Links et aI., 1990).
Finally, impulsivity in BPD patients has been shown to be worsened by alprazo-
lam administration in a double-blind trial (Cowdry and Gardner, 1988), but not in
anedoctal case reports (Faltus, 1984).
3. AFFECTIVE SYMPTOMS
Affective-related traits are a core diagnostic feature of BPD, according to DSM

classification. Affective instability is sometimes so important in characterizing the clin-
ical emergence of BPD, that it offers a diagnostic challenge and raises the issue of the
definition of diagnostic boundaries between BPD and mood disorders. In clinical prac-
tice, these problems in differential diagnosis often generate considerable difficulties in
the management of these patients (e.g. Bolton and Gunderson, 1996).
Following the perspective of an important theoretical paper by Gunderson and
Phillips (1991), the relationship between depression and BPD could be either defined
with depression or BPD as primary clinical pictures which produce the symptoms
referred to the second one, or with the two clinical pictures unrelated or related in a
nonspecific fashion.
Epidemiological and genetic studies do not sustain the hypothesis of a relation-
ship between BPD and mood disorders. Comorbidity studies failed to find a specific
relationship between the two disorders (Zanarini et aI., 1989), with Axis II diagnoses
other than BPD being equally associated with mood disorders (Shea et aI., 1987; Pfohl
et aI., 1984). Familial relationships have been noted between BPD and mood disorders
(Akiskal et aI., 1985; Gasperini et aI, 1991): recent data, however, suggest that this rela-
tionship may be secondary to comorbid major depression in the BPD probands
(Silvermann et aI., 1991). The higher frequency of affective disorders in the relatives of
patients with BPD could be accounted for by concurrent affective disorder in the
probands (Pope et aI., 1983; Zanarini et aI., 1988; Gunderson and Phillips, 1991), with
BPD alone in the proband being associated with greater frequencies of BPD, substance
abuse, and antisocial behavior among the relatives (Loranger et aI., 1982; Links et aI.,
1988). Conversely, the relatives of patients affected by bipolar disorder show higher
frequencies of anxious cluster personality disorders, but not of BPD (Coryell and
Zimmerman, 1989). Moreover, two studies found the same prevalence of affective dis-
orders among the relatives of BPD and schizotypal patients (Soloff et aI., 1983; Schultz
et aI., 1986).
The findings on biological studies about the "markers" of depression (dexam-
ethasone suppression test, thyrotropin-releasing hormone stimulation test, and sleep
EEG) failed to confirm a relationship between BPD and mood disorders (see review
by Lahmeyer et aI., 1989; Gunderson and Phillips, 1991), except when the studied
sample included BPD patients with an high familial morbid risk for mood disorders
(Battaglia et aI., 1993; Battaglia et aI., 1998).
From a phenomenological point of view, the nature of depression in borderline
patients seems to be qualitatively distinct from mood disorders. Even in the presence
of Axis I codiagnosis with major depression, BPD depressed patients show less specific
symptoms and the persistence of their maladaptive personologic tracts (Southwick et
aI., 1995; Rogers et aI., 1995), and discriminant analyses on symptomatological scales
suggest that the depressive episodes of BPD patients are qualitatively different from
those of non-BPD patients (Bellodi et aI., 1992).
Despite the above mentioned evidences of a lack of association between BPD
and mood disorders, the similarities between the depressive complaints presented by
BPD and the depressive symptomatology pertaining to mood disorders has led to test
antidepressant drugs in the treatment of this symptomatological domain. The first
methodologic problem in reviewing the literature on this matter is given by the neces-
sity to verify the exclusion of Axis I comorbidity for mood disorders.
Only one study is available on the efficacy of tricyclic antidepressants in BPD
patients diagnosed according to DSM criteria. Soloff et ai. (1986a, 1989) performed a
randomized trial of haloperidol (mean dose 7.2mg), amitriptyline (mean dose 147.6),
and placebo on 90 BPD patients: haloperidol was equal to amitriptyline on depressive
symptomatology, but produced a broader amelioration of BPD symptomatology. More-
over, in a double-blind parallel placebo-controlled trial, a paradoxical worsening of
suicide threats, paranoid ideation, and demanding behaviors was observed in 15 BPD
patients treated with amitriptyline (Soloff et aI., 1986b): the authors explained this
paradoxical effect as a dishinibition of impulsive behaviors which occurred indepen-
dently of the antidepressant effect of amitriptyline. One study of lithium in BPD
patients found no differences between lithium and placebo or desipramine on
depressive symptoms (Links et aI., 1990).
In a placebo-controlled crossover study with alprazolam, carbamazepine, triflu-
operazine, and the MAOI tranylcypromine, Cowdry and Gardner (1988) found that
the greatest improvement in depressive symptoms was obtained with the administra-
tion of tranylcypromine. However, it should be noted that also the administration of
trifluoperazine and carbamazepine was followed by positive effects on depressive
symptomatology. In a double-blind placebo-controlled trial with haloperidol (mean
dose 3.93mg/day) and phenelzine (mean dose 60.45 mg/day) , phenelzine, but not
haloperidol, was reported to acutely ameliorate depressive symptoms (Soloff et aI.,
1993); the continuation of the same therapy resulted in a loss of efficacy of phenelzine
and in an antitherapeutic effect of haloperidol on affective symptomatology (Cornelius
et aI., 1993).
The efficacy of neuroleptics in treating borderline depression must, however, still
be considered an area of debate, since the open label administration of flupenthixol
(3mg/day) to BPD patients has been reported to produce an improvement in all symp-
tomatological areas (Kutcher et aI., 1995), and flupenthixol decanoate has been
reported to prevent recurrent suicidal acts (Montgomery and Montgomery, 1982;
Montgomery et aI., 1985).
Recent studies on the effect of selective serotonin reuptake inhibitors focused on
the effect of these drugs on patients selected for the clinical predominance of impul-
sive and aggressive behaviors, and then do not consent a reliable evaluation of the
effects of SSRI on mood symptoms (see section on impulsivity).
Finally, the involvement of the cholinergic system has been recently proposed in
the regulation of affect in BPD: BPD patients (but not other personality disordered
patients) showed a greter depressive response to physostigmine than normal subjects,
and this response correlated with traits related to affective instability but not with
impulsivity (Steinberg et aI., 1997). It should be noted, however, that the presence of
234 F. Benedetti et al.
an adrenergic/cholinergic imbalance in never-depressed BPD had been proposed

basing on abnormalities in cardiac autonomic function (Battaglia et aI., 1995).
4. PSYCHOTIC· LIKE SYMPTOMS
"Psychotic-like" symptoms in BPD include paranoid ideation, magical and refer-

ential thinking, perceptual distortions, dissociative symptoms, hypnagogic phenomena.
A major problem in the evaluation of this symptomatological dimension comes from
the difficult definition of its boundaries.
Clear cut psychotic symptoms (such as delusions and hallucinations) are
extremely rare in BPD patients, and, if present, often due to substance abuses. Broader
definitions of "psychotic-like" symptoms in the literature, however, have led to the
inclusion of extremely aspecific symptoms, such as depersonalization and derealization,
which have been reported to occur also in non-psychotic diagnostic categories and in
normal subjects (for a discussion on this subject, see Jonas and Pope, 1984; Pope et aI.,
1985; Chopra and Beatson, 1986; George and Soloff, 1986). The extreme heterogeneity
of this psychopathological dimension causes considerable problems in the evaluation
of the available psychopharmacological trials.
Moreover, the presence of this symptomatology raises the issue of the differen-
tial diagnosis between BPD and Cluster A personality disorders. Clinical studies show
that schizotypal symptoms are often present in BPD and that borderline and schizo-
typal personality disorders are often codiagnosed. Following the DSM-IV perspective,
the feature which distinguishes the two disorders in respect to this symptomatological
cluster is the transience and the relationship to interpersonal stresses which character-
izes the emergence of psychotic-like symptoms in BPD. Nevertheless, the schizo typal
dimension in BPD must still be considered an area of debate since the published studies
on the pharmacological treatment of this area have been often performed with samples
including both borderline, schizotypal and mixed-diagnoses patients.
In analogy with the dopaminergic hypothesis of schizophrenia, psychotic-like
symptoms in personality disorders were hypothesized to be due to an abnormal
functioning of brain dopaminergic pathways, without sharing then the substrate of sero-
tonergic dysregulation proposed for the other symptomatological clusters. Several psy-
chobiological findings in agreement with this hypothesis have been reported. Patients
affected by schizotypal personality disorder were found to have higher plasma and spinal
fluid concentrations of homovanillic acid, which positively correlated with the intensity
of psychotic-like symptomatology (Siever et aI., 1991; Siever et aI., 1993). The adminis-
tration of amphetamine, which is known exacerbate schizophrenic psychosis, elicited
transient worsening of psychotic like symptoms in patients with a co-diagnosis of
schizotypal and BPD, but not in patients affected by BPD alone.
Brinkley et aI. (1979) showed that the administration of neuroleptics, given at a
dosage below the usual clinical effective dose for psychotic disorders, had positive
effects in BPD patients with transient psychotic symptoms: the administration of either
thiothixene (10 mglday) , or thioridazine (25 mg/day) , or perphenazine (16mglday)
resulted in broad spectrum amelioration of BPD symptomatology. Similar results were
obtained with the open-label administration of loxapine (13.5-14.5mglday) or chlor-
promazine (105-120mglday) (Leone, 1982). In a double-blind comparison of thiotix-
ene (mean dose 9.4mg/day) and haloperidol (mean dose 3.0mglday), 84% of treated
patients showed some degree of improvement, covering all psychopathological areas
New Drugs in the Treatment of Borderline Personality Disorder 23S
(cognitive disturbances, derealization referential thinking, anxiety, depression), with

marginal differences between the two treatments (Serban and Siegel, 1984). The posi-
tive effect of thiothixene was then confirmed in a double-blind placebo controlled trial
(mean daily dose 8.67 mg): interestingly, better results were obtained in patients with
worse baseline referential thinking, paranoid ideation, perceptual disturbances, anxiety,
but not depression (Goldberg et aI., 1986).
Finally, it must be remembered that the already cited double-blind trials of
Soloff et ai. (1986a, 1986b, 1989, 1993) and of Cornelius et ai. (1993) showed contrast-
ing results for the administration of low-dose haloperidol: in a first study, haloperidol
was shown to produce a broad improvement in BPD symptomatology, with the
improvement being predicted by the presence of psychotic-like symptoms at baseline;
this result, however, was not confirmed by subsequent studies, which failed to replicate
previous reports of efficacy. In the same studies, the MAOI phenelzine was shown to
share the same efficacy of haloperidol and of placebo in the treatment of psychotic-
like symptoms.
These clinical trials seem to confirm the hypothesis of a neurobiologic distinction
between psychotic-like symptomatology and other symptomatological dimensions
in BPD. However, striking issues against this perspective come from recent reports
on the efficacy of SSRls: the open-label administration of fluoxetine has been shown
to substantially improve psychotic-like symptoms both in borderline and schizotypal
patients (Markovitz et aI., 1991), thus suggesting a role for serotonergic dysfunction
even in the pathogenesis of this symptomatological dimension. Up to now, only one
double-blind study is available on the efficacy of fluoxetine in borderline patients,
but it was conducted in patients with mild symptomatology (Salzman et aI., 1995):
further researches are then needed to confirm these findings. Finally, the administra-
tion of the partial serotonergic agonist m-chlorophenylpiperazine has been reported
to cause, in BPD patients, the acute occurrence of pleasant depersonalization/-
derealization experiences, but not of other "psychotic-like" symptoms, thus confirming
the extreme heterogeneity of this symptomatological dimension (Hollander et aI.,
1994).
5. NEW ANTIPSYCHOTIC DRUGS

The use of new categories of antipsychotic drugs was recently proposed in the
treatment of patients with symptomatological patterns which respond to traditional
neuroleptics. Clozapine is an antipsychotic agent with proven efficacy in the treatment
of severe illness like schizophrenia, and which lacks the neurological side effects
induced by classical neuroleptics (Kane et aI., 1988; Tamminga et aI., 1994). In BPD
patients clozapine has been proposed to use at dosages lower than those administered
in schizophrenics. Promising results have been reported in a sample of 15 BPD patients
with psychotic (due to comorbid Axis I pathologies) or psychotic-like symptoms (mean
dose 253.3 mgldie) (Frankenburg and Zanarini, 1993): BPD patients showed a signifi-
cant improvement in both psychotic symptomatology and overall function. Moreover,
a single case BPD patient with self-destructive behaviors treated with clozapine (300
mgldie) (Chengappa et aI., 1995) showed a dramatic reduction of self-mutilating behav-
ior, and a considerable improvement was reported in a single case BPD patient
with Axis I Obsessive-Compulsive Disorder with previous nonresponse to standard
antidepressant and neuroleptic treatments (Steinert et aI., 1996).
236 F. Benedetti et aL
Given the above, our research group open-label investigated the effect of low-
dose clozapine in the treatment of severe BPD (Benedetti et al., 1998).12 patients were
selected for the presence of severe psychotic-like symptomatology in the absence of
Axis I codiagnosis. Due to the already discussed problems in the definition of this symp-
tomatological dimension, the presence of depersonalization/derealization experiences
alone was not sufficient for inclusion, and all subjects showed marked (but not
delusional or hallucinatory) cognitive and perceptual disturbances.
In this experimental sample, the four month administration of clozapine given
at a mean daily dose of 43.8mg was followed by a dramatic decrease in psychotic-like
symptoms within the first two weeks of treatment, and by a progressive and broad ame-
lioration in overall course of pathology, with a substantial improvement of both impul-
sivity and affective instability without any worsening in depressive symptomatology.
6. CONCLUSIONS
"The strategy of pharmacologic dissection of patients with BPD in categorical

subtypes has not proved productive" (Soloff et al., 1993). From a methodological
point of view, this remark by the Pittsburgh group must still be hold true: apart from
the relative unspecificity of several tested drugs, which precludes a reliable psychobio-
logic evaluation of their effects, the most striking issue against the pharmaco-biologi-
cal dissection of BPD comes from the fact that in drug responders the administration
of the drug which triggers the response is followed by a broad aspecific improvement
in all symptomatological domains. Further researches, using more specific drugs and
more refined definitions of the symptomatological domains, are needed to clarify the
topic. In particular, longitudinal long-term follow up studies could lead, from a clinical
point of view, to a more reliable subdivision of BPD patients in homogeneous sub-
groups (e.g.: does the same patient always present symptoms pertaining to the same
psychopathological dimension, or do shifts between symptomatological clusters
occur?).
In the absence of psychobiological definition in this pathological area, in clinical
practice the choice of the first treatment must rely on the empirical definition of a target
symptomatology as above discussed, keeping in mind that the absence of a specific cor-
relation between symptoms and drugs precludes the definition of a reliable treatment
algorithm, and that each patient could modify his clinical picture during the life time
course of the disorder.
REFERENCES
Akiskal H.S., Chen S.E., Davis nc., Puzantian Y.R., Kashgarian M., and Bolinger J.M. (1985) Borderline: an
adjective in search of a noun. J Clin Psychiatry 46:41-48.
American Psychiatric Association. Diagnostic and statistic manual of mental disorders. 4rd ed. Washington,
DC: AP.A. Press; 1994.
Andrulonis D.A, Glueck B.c., Stroebel c.F. et al. (1981) Organic brain dysfunction in the borderline syn-
drome. Psychiatr Clin North Am 4:47-66.
Barratt E.S., Kent T.A, Bryant S.G. et al. (1991) A controlled trial of phenytoine in impulsive aggression
(letter) J Clin Psychopharmacolll:388-389.
Battaglia M., Ferini Strambi L., Smirne S., Bernardeschi L., and Bellodi L. (1993) Ambulatory polysomnog-
raphy of never-depressed borderline subjects: a high-risk approach to rapid eye movement latency.
Bioi Psychiatry 33:326-334.
Battaglia M., Ferini Strambi L., Smirne S., Bernardeschi L., Oldani L., and Bellodi L. (1995) Cardiac auto-
nomic fonction during sleep of never-depressed bordeline subjects: a pilot study. J Psychiatric Res
29:397-405.
Battaglia M., Ferini Strambi L., Bertella S., Bajo S., and Bellodi L. (1998) First-cycle REM density in
never-depressed subjects with borderline personality disorder. Bioi Psychiatry, in press.
Bellodi L., Battaglia M., Gasperini M., Scherillo P., and Brancato V. (1992) The nature of depression in
borderline depressed patients. Compr Psychiatry 33:128-133.
Benedetti E, Sforzini L., Colombo c., Maffei c., and Smeraldi E. (1998) Low-dose clozapine in acute and
continuation treatment of severe borderline personality disorder. J Clin Psychiatry 59:103-107.
Bolton S. and Gunderson IG. (1996) Distinguishing bordeline personality disorder from bipolar disorder:
differential diagnosis and implications. Am J Psychiatry 153:1202-1207.
Brinkley J.R, Beitman B.D., and Friedel RO. (1979) Low-dose neuroleptic regims in the treatment of
borderline patients. Arch Gen Psychiatry 36:319-326.
Casey P., Meagher D., and Butler E. (1996) Personality, functioning, and recovery from major depression. J
Nerv Ment Dis 184(Suppl 4):240-245.
Chengappa K.N.R, Baker RW., and Sirri C. (1995) The successful use of c10zapine in ameliorating severe
self mutilation in a patient with borderline personality disorder. J Personality Disord 9(1):76-82.
Chopra H.D. and Beatson IA (1986) Psychotic Symptoms in Borderline Personality Disorder. Am J Psy-
chiatry 143:12.
Coccaro E.E (1989) Central serotonin and impulsive aggression. Br J Psychiatry 155:52-62.
Coccaro E.E, Siever L.I, Klar H.M., Maurer G., Cochrane K., CooperT.B., Mohs Re., and Davis K.L. (1989a)
Serotonergic studies in patients with affective and personality disorders: correlates with suicidal and
impulsive aggressive behavior. Arch Gen Psychiatry 46:587-599.
Coccaro E.E, Siever L.I, Kavoussi R, and Davis K.L. (1989b) Impulsive aggression in personality disorder:
evidence for involvment of 5HTl receptors. Bioi Psychiatry (Suppl) 25(7A):86A
Coccaro E.E, Astill IL., Herbert J.L., and Schut Al (1990) Fluoxetine treatment of impulsive aggression in
DSM-IIIR personality disorder patients. J Clin PsychopharmacoI1O:373-375.
Coccaro E.E and Siever L.I (1995) The neuropsychopharmacology of personality disorders. In: Bloom EE.
and Kupfer D.I, eds. Psychopharmacology: the fourth generation of progress. New York: Raven Press;
pp.1567-1578.
Coccaro E.E (1998) Clinical outcome of psychopharmacologic treatment of borderline and schizo typal per-
sonality disordered SUbjects. J Clin Psychiatry 59(Suppl 1):30-35.
Cornelius IR, Soloff P.H., Perel 1M., and Ulrich RE (1990) Fluoxetine trial in borderline personality dis-
order. Psychopharmacol Bull 26:151-154.
Cornelius IR, Soloff P.H., Perel 1M., and Ulrich RE (1993) Continuation pharmacotherapy of borderline
personality disorder with haloperidol and phenelzine. Am J Psychiatry 150:1843-1848.
Coryell W.H. and Zimmerman M. (1989) Personality disorder in the families of depressed, schizophrenic and
never-ill probands. Am J Psychiatry 146:496-502.
Cowdry RW., Pickard D., and Davies R (1980) Limbic dysfunction in the borderline syndrome. In: New
research abstracts, 133rd annual meeting of american psychiatric association. Washington, DC, A.P.A.
Press.
Cowdry RW. (1987) Psychopharmacology of Borderline Personality Disorder: A Review. J Clin Psychiatry
48(Suppl 8):15-22.
Cowdry RW. and Gardner D.L. (1988) Pharmacotherapy of borderline personality disorder: alprazolam, car-
bamazepine, trifluoperazine, tranylcypromine. Arch Gen Psychiatry 45:111-119.
De la Fuente 1M. and Lotstra E (1994) A trial of carbamazepine in borderline personality disorder. Euro-
pean Neuropsychopharmacol 4(4):479-486.
De la Fuente J.M., Tugendhaft P., and Mavroudakis N. (1998) Electroencephalographic abnormalities in bor-
derline personality disorder. Psychiatry Res 77:131-138.
Faltus EI (1984) The positive effect of alprazolam in the treatment of three patients with borderline per-
sonality disorder. Am J Psychiatry 141:802-803.
Frankenburg ER and Zanarini M.e. (1993) Clozapine treatment of borderline patients: a preliminary study.
Comprenhens Psychiatry 34(6):402-405.
Gardner D.L. and Cowdry RW. (1986) Positive effects of carbamazepine on behavioral dyscontrol in bor-
derline personality disorder. Am J Psychiatry 143:519-522.
Gasperini M., Battaglia M., Scherillo P., Sciuto G., Diaferia G., and Bellodi L. (1991) Morbidity risk for mood
disorders in the families of borderline patients. J Affective Disord 21:265-272.
George A and Soloff P.H. (1986) Schizotypal symptoms in patients with Borderline Personality Disorders.
Am J Psychiatry 143:2.
238 F. Benedetti et at.
Giakas w.J., Seibyl J.P., and Mazure C.M. (1990) Valproate in the treatment of temper outbursts (letter).
J Clin Psychiatry 51:525.
Gitlin M.l (1993) Pharmacotherapy of Personality Disorders: conceptual framework and clinical strategies.
J Clin Psychopharmacol 13:343-353.
Goldberg S.c., Schulz c., Schulz P.M., Resnick RJ., Hamer RM., and Friedel Ro. (1986) Borderline and
schizo typal personality disorders treated with low-dose thiothixene versus placebo. Arch Gen
Psychiatry 43:680--686.
Gunderson J.G. and Phillips K.A (1991) A current view of the interface between borderline personality dis-
order and depression. Am J Psichiatry 148:8967-975.
Hasan M.Y., Sewell RD.E., and Nicholls P.J. (1990) Does the anticonvulsant agent sodium valproate display
behaviorally selective anti-offensive activity? (abstract) J Pharm PharmacoI42(Suppl):185.
Hirschfeld RM. (1997) Pharmacotherapy of borderline personality disorder. J Clin Psychiatry 58(Suppl
14):48-52.
Hollander E., Stein D.J., DeCaria C.M., Cohen L., Saoud J.B., Skodol AE., Kellman D., Rosnick L., and
Oldham 1M. (1994) Serotonergic sensitivity in borderline personality disorder: preliminary findings.
Am J Psychiatry 151:2 277-280.
Hull lW., Clarkin J.F., and Alexopoulos G.S. (1993) Time series analysis of intervention effects. Fluoxetine
therapy as a case illustration. J Nerv Ment Dis 181:48-53.
Jonas J.M. and Pope H.G. (1984) Psychosis in borderline personality disorder. Psychiatr Dev 4:295-308.
Kane J., Honigfeld G., Singer J., and Meltzer H. (1988) Clozapine for the treatment-resistant schizophrenia.
Arch Gen Psychiatry 45:789-796.
Kavoussi RJ., Liu J., and Coccaro E.F. (1994) An open trial of sertraline in personality disordered patients
with impulsive aggression. J Clin Psychiatry 55:137-14l.
Keck P.E., McElroy S.L., and Friedman L.M. (1992) Valproate and carbamazepine in the treatment of panic
and post-traumatic stress disorders, withdrawal states, and behavioral dyscontrol syndromes. J Clin
Psychopharmacol12(Suppl 1):36-4l.
Kutcher S., Papatheodorou G., Reiter S., and Gardner D. (1995) The successful pharmacological treatment
of adolescents and young adults with borderline personality disorders: a preliminary open trial of flu-
penthixol. J Psychiat Neurosci 20:113-118.
Lahmeyer H.W., Reynolds c.F., Kupfer DJ., and King R (1989) Biologic markers in borderline personality
disorder: a review. J Clin Psychiatry 50:217-225.
Leone F.N. (1982) Response of borderline patients to loxapine and chlorpromazine. J Clin Psychiatry
43:148-150.
Links P.S., Steiner M., and Huxley G. (1988) The occurence of borderline personality disorder in the fami-
lies of borderline patients. J Personality Disord 2:14-20.
Links P.S., Steiner M., Boiago I., and Irwin D. (1990) Lithium therapy for borderline patients: preliminary
findings. J Personality Disord 4:173-181.
Loranger AW., Oldham J.M., and Tulis E.H. (1982) Familial transmission of borderline personality. Arch
Gen Psychiatry 39:795-802.
Markowitz P.J., Calabrese J.R, Schulz S.c., and Meltzer H.Y. (1991) Fluoxetine in the treatment of border-
line and schizotypal personality disorders. Am J Psychiatry 148:1064-1067.
Mattes J. (1990) Comparative effectiveness of carbamazepine and propanolol for rage outbursts. J Neu-
ropsychiatry Ctin Neurosci 2:159-164.
Montgomery S.A and Montgomery D. (1982) Pharmacological prevention of suicidal behavior. J Affect
Disord 4:291-298.
Montgomery S.A, Roy D., and Montgomery D.B. (1987) The prevention of recurrent suicidal acts. Br J Ctin
PharmacoI15:1835-1885.
Moss H.B., Yao J.K., and Panzak L. (1990) Serotonergic responsivity and behavioral dimensions in antiso-
cial personality disorder with substance abuse. Bioi Psychiatry 28:325-338.
Norden M.J. (1989) Fluoxetine in borderline personality disorder. Progr Neuropsychopharmacol Bioi Psy-
chiatry 13:885-893.
Nurnberg H.G., Raskin M., Levine P.E., Pollack S., Siegel 0., and Prince R (1991) The comorbidity of bor-
derline personality disorder and other DSM-III-R Axis II personality disorders. Am J Psychiatry
148:10 1371-1377.
Pfohl B., Stangl D., and Zimmerman M. (1984) The implications of DSM III personality disorders for patients
with major depression. J Affective Disord 7:309-318.
Pope H.G., Jonas J.M., Hudson J.I., Cohen B.M., and Gunderson J.c. (1983) The validity of DSM-III
borderline personality disorder. Arch Gen Psychiatry 40:23-30.
Pope H.G., Jonas J.M., Hudson J.I., Cohen B.M., and Tohen M. (1985) An empirical study of psychosis in
borderline personality disorder. Am J Psychiatry 142:1285-1290.
Rogers J.H., Wiediger T.A, and Krupp A (1995) Aspects of depression associated with borderline person-
ality disorder. Am J Psychiatry 152:268-270.
Rosenberg R (1994) Borderline states: pharmacotherapy in psychobiology of personality. Acta Psychiatr
Scand 89(Suppl 379):56--60.
Salzman c., Wolfson AN., Schatzberg A, Looper J., Henke R, Albanese M., Schwartz J., and Miyawaki E.
(1995) Effect of fluoxetine on anger in symptomatic volunteers with borderline personality disorder.
J Clin PsychopharmacoI15:23-29.
Schulz P.M., Kelly T., Di Franco Ret al. (1986) Borderline personality disorder: family histories, in CME
Syllabus and scientific Proceedings in Summary Form, 139th Annual Meeting of the American
Psychiatric Association. Washington, DC, APA
Serban G. and Siegel S. (1984) Response of borderline and schizotypal patients to small doses of thiothix-
ene and haloperidol. Am J Psychiatry 141:1455-1458.
Shader RI., Jackson AH., and Dodes L.M. (1974) The antiaggressive effect of lithium in man. Psychophar-
macology 40:17-24.
Shea M.T., Glass D.R, Pilkonis P.A et al. (1987) Frequency and implications of personality disorders in a
sample of depressed outpatients. J Personality Disord 1:27-42.
Sheard M.H. (1975) Lithium in the treatment of aggression. J Nerv Ment Dis 160:108-118.
Skodal A, Buckley P., and Charles E. (1983) Is there a characteristic pattern to the treatment history of clin-
ical outpatients with borderline personality? J Nerv Ment Dis 171:405-410.
Siever L.J., Amin E, Coccaro E.E, Bernstein D., Kavoussi RJ., Kalus 0., Horvath T.B., Warne P., Davidson
M., and Davis K.L. (1991) Plasma homovanillic acid in schizotypal personality disorder. Am J
Psychiatry 148:1246-1248.
Siever L.J., Amin E, Coccaro E.E, Trestmann R, Silvermann J., Horvath T.B., Mahon T.R, Knott P., Altstiel
L., Davidson M., and Davis K.L. (1993) CSF homovanillic acid in schizotypal personality disorder.
Am J Psychiatry 150:149-151.
Silverman J.M., Pinkhan L., Horvath T.B., Coccaro E.E, Klar H., Shear S., Apter S., Davidson M., Mohs RC.,
and Siever L.J. (1991) Affective and impulsive personality disorder traits in the relatives of border-
line personality disorder. Am J Psychiatry 148:1378-1385.
Soloff P.H. and Millward J.w. (1983) Psychiatric disorders in the families of borderline patients. Arch Gen
Soloff P.H., George A, Nathan RS., Schulz P.M., Ulrich RE, and Perel J.M. (1986a) Progress in pharma-
cotherapy of borderline disorders. A double blind study of amitriptyline, haloperidol, and placebo.
Arch Gen Psychiatry 43:691--697.
Soloff P.H., George A, Nathan RS., Schulz P.M., and Perel J.M. (1986b) Paradoxical effects of amitriptyline
on borderline patients. Am J Psychiatry 143:12 1603-1605.
Soloff P.H., George A, Nathan RS., Schulz P.M., Cornelius J.R, Herring J., and Perel J.M. (1989) Amitryp-
tiline versus haloperidol in borderlines: final outcome and predictors of response. J Clin Psychophar-
macol 9:238-246.
Soloff P.H., Cornelius J., George A, Nathan S., Perel J.M., and Ulrich RE (1993) Efficacy of phenelzine and
haloperidol in borderline personality disorder. Arch Gen Psychiatry 50:377-385.
Soloff P.H. (1994) Is there any drug treatment of choice for the borderline patient? Acta Psychiatr Scand 89
(Suppl 379):50-55.
Southwick S.M., Yehuda R, and Giller E.L. (1995) Psychological dimensions of depression in borderline per-
sonality disorder. Am J Psychiatry 152:789-791.
Stein G. (1992) Drug treatment of the personality disorders. Br J Psychiatry 161:167-184.
Stein M.R, Simeon D., Frenkel M., Islam M.N., and Hollander E. (1995) An open trial of valproate in bor-
derline personality disorder. J Clin Psychiatry 56:506-510.
Steinert T., Schmidt-Michel P.O., and Kaschka w.P. (1996) Considerable improvement in a case of obsessive-
compulsive disorder in an emotionally unstable personality disorder, borderline type, under treatment
with clozapine. Pharmacopsychiatry 29:111-114.
Stone M.H. (1989) The role of pharmacotherapy in the treatment of patients with bordeline personality dis-
order. Psychopharmacol Bull 25:564-571.
Tamminga C.A, Thaker G.K., Moran M., Kakigi T., and Gao X.M. (1994) Clozapine in tardive dyskinesia:
observations from human and animal model studies. J Clin Psychiatry 55:102-106.
Verkes RJ., Fekkes D., Zwinderman AH., Hengeveld M.W., Van der Maast RC., Tuyl J.P., Kerkhof AJ., and
Van Kempen G.M. (1997) Platelet serotonin and [3Hlparoxetine binding correlate with recurrence of
suicidal behavior. Psychopharmacology 132:89-94.
Verkes RJ., Van der Maast RC., Kerkhof AJ., Fekkes D., HengeVeld M.W., Tuyl J.P., and Van Kempen G.M.
(1998) Platelet serotonin, monoamine oxidase activity, and [3Hlparoxetine binding related to impul-
sive suicide attempts and borderline personality disorder. Bioi Psychiatry 43:740-746.
240 E Benedetti et aL
Watkins S.E., Callender K., Thomas D.R., TIdmarsh S.E, and Shaw D.M. (1987) The effect of carbamazepine
and lithium on remission from affective illness. Br J Psychiatry 150:180-182.
Wilcox J.A. (1995) Divalproex sodium as a treatment for borderline personality disorder. Ann Clin Psychi-
atry 7:33-37.
Zanarini M.e., Gunderson J.G., Marino M.E et ai. (1988) DSM-III disorders in the families of
borderline outpatients. J Personality Disord 2:4292-4302.
Zanarini M.e., Gunderson J.G., and Frankenburg ER. (1989) Axis I phenomenology of borderline
personality disorder. Compr Psychiatry 30:149-156.
17
THE NARCISSISTIC PERSONALITY

DISORDER AND ADDICTION
PerVaglum
Department of Behavioural Sciences in Medicine

University of Oslo
Oslo, Norway
1. INTRODUCTION
It is a well established fact that substance abusers often also have one or more
personality disorders (PD). Verheul et al. (1995) recently reviewed the empirical liter-
ature on this topic, and found a usual co-occurrence of 40% among alcoholics and 79%
among drug abusers in treatment. The most frequent personality disorder found
is usually the antisocial and borderline personality disorder, partly biased by the used
questionaires and research populations.
Narcissistic personality disorders (NPD) as well as high levels of self-reported
narcissistic traits may occur among addicts. During the last two decades, psychody-
namically oriented clinicians and researchers have underlined the importance of nar-
cissistic disturbances among substance abusers (Kohut, 1971; Wurmser, 1974; Kernberg,
1985; Treece and Khantzian, 1986).
However, the co-occurrence of two different disorders or syndromes, here NPD
and substance abuse, in the same person does not imply that they are related (Wittchen
1996). There may be several possible explanations for this co-occurrence: (a) it may be
merely due to chance, (b) there may be a selection bias in clinical samples, (c) NPD
may be a predisposing factor to addiction, (d) NPD may be a secondary consequence
of the use of substances, (e) NPD and addiction may be spuriously linked together by
a third factor and it is also possible that (f) NPD and substance abuse develop in a rec-
iprocal relationship, each increasing the further development of the other.
In clinical work, it is important to know whether the presence of narcissistic
disturbances influences the response to different treatment modalities and/or course of
the substance abuse.
In this chapter, each of these six points will be discussed on the basis of the clin-
ical psychodynamic literature and the empirical studies of the DSM-IIIIR NPD disor-
der. Finally some clinical implications will be discussed.
242 P. Vaglum
2. PREVALENCE OF NPD IN THE COMMUNITY AND AMONG

SUBSTANCE ABUSERS
To be able to rule out the possibility that the occurrence of NPD is merely due
to a chance overlap with substance abuse, we need to know the prevalence of NPD
in the general population and among substance abusers separately. Unfortunately, we
still lack studies of nationwide, representative samples of the population, so that we
have to do with some studies from limited parts of the community and of non-patients.
Table 1 summarizes seven of such samples, the mean prevalence rate being 0.15%. Since
the sample sizes are to low to be used as epidemiological data, the true prevalence may
be different. However, this prevalence does not mean that the prevalence is low in
all subgroups of the popUlation. In fact, there are several studies of students that
indicate a higher prevalence of NPD among those who pursue higher education Drake
and Vaillant (1988) found that 4% of their 369 college men had NPD, and Bodlund
et al. (1993) identified with SCID-screen 7.3% of the men and 3.7% of the women
in their mainly student sample as having NPD. Maffei et al. (1995) found that 17% of
317 first year medical students fulfilled their PDQ criteria of NPD. Richman et al.
(1996) have recently shown that narcissistic vulnerability increased the level of alcohol
drinking among young doctors, also by increasing the risk of experienced workplace
harassment.
The prevalence of NPD in 14 samples of substance abusers in treatment is shown
in table 2. Mean prevalence is 6.8%. If we set the prevalence in the community study
to 0.5%, this is a 13 times higher prevalence than in the community. We can therefore
rule out the possibility that the co-occurrence of addiction is merely due to chance. To
explore whether it may be due to an increased help-seeking behaviour among NPD
persons who also misuse substances (Bergson's fallacy), we should know the preva-
lence of NPD among non-help-seeking abusers, the majority of the substance abusers.
Unfortunately, such data are still lacking.
However, if we compare the prevalence of NPD among general samples of non-
abusing psychiatric patients with the prevalence among treated addicts, we may get a
better impression of whether there is a specific relationship between NPD and sub-
stance use, and in this way control for-help seeking behaviour. The prevalence of NPD
in seven heterogeneous samples of psychiatric patients with all types of disorders
(including a few substance abusers) is shown in table 3. The mean prevalence is 3.6%,
which is nearly half the prevalence we found in the substance-abusing sample.
Table 1. Prevalence of NPD in non patient/community samples

Author Subjects N= Method NPD rate/100
Rich et al. (1989) Community 235 PDQ 0.4
Zimmerman & Coryell (1990) Relatives of patients and 697 PDQ 0.4
controls SIDP 0.0
Maier et al. (1990) Relatives of control group 447 SCID 0.0
Samuels et al. (1994) Community sample 762 DSM-interview 0.0
Moldin et al. (1994) Normal controls 302 PDE 0.3
Black et al. (1995) Relatives of normal controls 188 PDQ 0.0
SIDP
The Narcissistic Personality Disorder and Addiction 243
Table 2. Prevalence of NPD in samples of substance abusers

Author Substances Sample size Method NPD ratel100
Khantzian and Treece (1985) Polydrugs 133 ClinicallDSM -III 3.6
Vaglum and Vaglum (1985) Alcoholic women 65 SADSIDSM -III 9.3
Kosten and Rounsaville (1986) Opioid 384 SADSIDSM-III 2.4
Poldugro and Forti (1988) Alcoholic men 404 ? 0.2
Yates et al. (1989) Cocaine 59 PDQ 32.0
Alcoholics 47 PDQ 6.4
Clerici et al. (1989) Opioid 226 ClinicallDSM -III 6.1
Nace et al. (1991) Alcoholic 100 SCID 4.0
De Jong et al. (1993) Alcoholic 187 SIPD 6.7
Polydrug 86 SIDP 13.0
Rousar et al. (1994) Opioid 167 SCID 0.6
Cacciola et al. (1996) Opioid 210 SIDP-R 5.6
Brooner et al. (1996) Opioid 716 SCID 0.8
Barber et al. (1996) Cocaine 289 SCID 5.5
However, given the low rates in both samples, this difference is too small to completely
rule out the possibility that the increased prevalence of NPD among substance abusers
may be due to a selection bias. Those persons with NPD who also have substance abuse
may more often come to treatment. This may partly also be due to the fact that those
psychiatric disorders most often co-occurring with NPD in general (depression,
anxiety), also more often co-occur with addiction.
Summarizing: these studies show that one may expect to find about 7% of treat-
ment seeking substance abusers to fulfil the DSM criteria of NPD. The presence of sub-
threshold NPD traits may, however, be frequent among addicts, and the variation in
such traits may also have clinical importance. As shown in table 4, narcissistic traits are
more prevalent among substance abusers than in the sample of depressive outpatients,
studied by Stancovic et al. (1992). Studies using the Millon Clinical Multiaxial Inven-
tory (MCMI) usually report the highest scores on the narcissistic and antisocial scales.
Several studies on the factor structure of the NPD criteria for DSM-III or of the MCMI
narcissistic scales, have shown that there may be clear subgroups of patients fulfilling
Table 3. Prevalence of NPD in nonpsychiatric patients

Author Subjects N= Method NPD ratel100
Koenigsberg et al. (1985) in- and outpatients 2462 DSM-III 0.9
Pfohl et al. (1986) inpatients 131 SIDP 3.7
Alnres and outpatients from 298 SIDP 5.0 (men: 11)
Torgersen (1988) sector (women: 2)
Nakao et al. (1992) outpatients 149 DSM-II1 3.0
Loranger et al. (1994) in- and outpatients 716 IDPE 1.3
First et al. (1995) in- and outpatients 103 SCID 6.5
Mattanah et al.(1995) Adolescent inpatients 70 PDE 5.0
244 P. Vaglum
Table 4. Prevalence of MCMI-NPD-traits

Author Subjects MCMI-version Mean
McMahon et al. (1985) 96 VA alcoholics MCMI 55.3
33 VA drug abusers 57.7
Stark and Campbell (1988) 100 drug abusers MCMI 60.0
Marsh et al. (1988) 163 methadone users MCMI 68.1
Ravndal and Vaglum (1991) 144 polydrug abusers MCMI 55.0
Brown (1992) 50 alcoholic/drug abusing inpatients MCMI 65.1
Simonsen et al. (1992) 90 alcoholic outpatients MCMI 62.4
Stancovic et al. (1992) 43 depressive outpatients MCMI-II 38.2
different combinations of the criteria (Di Guiseppe et al. 1995, Choca et al. 1996). We
do not know whether substance-abusing NPD patients differ in this way compared with
non- substance-abusing NPD patients.
3. IS NPD PREDISPOSING FOR OR SECONDARY TO

SUBSTANCE ABUSE?
Theoretically, use of substances during adolescence and young adulthood may

impair the normal development of the self (including self-efficacy, self-esteem, self-
image, identity, self-ideal). Disturbances in the development of the self, may, on the
other hand, lead to a need for substances-either as a way to cope with intolerable
feelings, or as a way to bolster an overvalued self. Ideally, the development of the self
related to substance abuse, should be studied through prospective longitudinal studies
of narcissistic features starting at an age before the risk of substance abuse. Such studies
are still lacking. There is also a lack of systematic empirical studies that explicitly have
studied whether narcisstic features come first. The most extensive knowledge we have
about this question therefore, comes from psychoanalytically oriented clinicians who
have treated and/or evaluated addicted patients themselves, and have been in charge
of treatment programmes for substance abusers.
Through the history of psychoanalysis, one has passed through different stages
concerning the view of addiction. One may say that one has moved from seeing addic-
tion as an oral longing, to a sign of a narcissistic despair. I will very shortly review some
of the most important contributions to this development, relying heavily on Treece and
Khantzian (1986), Donnovan (1986), and Miller (1990).
Freud and his first followers mainly saw substance abuse as an impulse neurosis,
characterized by a longing for regressive, pleasurable states, without inhibitions and
prohibitions (id-psychology) (Freud, 1887; Abraham, 1908; Freud, 1930; Rado, 1933;
Fenichel, 1944). Chein et al. (1964) introduced the ego-psychological point of view; they
described addiction as an attempt to cope with painful feeling states, as well as with
overwhelming developmental tasks and responsibilities in the outside world. As a rep-
resentative of the object-relational school, Balint (1968) proposed that alcoholics might
be especially vulnerable to rejection and aggressive feelings in close relationships
(because of a "basic fault"): "in alcoholics .......... object relationships, though
usually fairly intense, are shaky and unstable. These people are easily thrown off their
The Narcissistic Personality Disorder and Addiction 24S
balance; the most common cause being a clash between themselves and one important
love object. This clash easily appears to them so overwhelming that they feel utterly
unable to remedy the situation ....... and start to drink. The first effect of intoxica-
tion is invariably the establishment of a feeling that everything is now well between
them and the environment. ... In this state of harmony, there are no people or objects
of love and hate, especially no demanding people and objects". (Balint 1968).
Kernberg (1985) connected the use of substances directly to narcissistic distur-
bances when he wrote: "In the case of narcissistic personalities, alcohol or drug intake
may constitute a predominant mechanism to 'refuel' the pathological grandiose self
and assure its omnipotence and protection against a potentially frustrating and hostile
environment in which gratification and admiration are not forthcoming." He also
suggested that the prognosis for treatment of addictive states in narcissistic personal-
ities would be much worse than in the case of borderline and depressive personalities
(Kernberg 1975).
Kohut (1971, 1977, 1978), in his development of psychoanalytic self-psychology,
underlined that the potential explanatory power of this psychology was especially great
in the area of addictions. Subsequently, he mentioned addictive behaviour as a major
symptom of narcissistic disorders. In short, his self psychological-theory of narcissism
and addictions can be summarized as follows:
- The addicted person has stopped in the normal development of the self at
the stage of the archaic, nuclear self comprising of the grandiose self and the
idealized self-object.
- Consequently, the addict has inadequately internalized important self-object
functions: tension regulation, self-soothing, self-esteem regulation, a stimulus
barrier.
- Addiction is a futile attempt to compensate for this failure in internalization.
It is the lack of self esteem of the unmirrored self, the uncertainty of the very
existence of the self, the dreadful feeling of fragmentation of the self, that the
addict tries to counteract by his addictive behaviour.
- The substances, however, serve not as a substitute for love or loving objects,
or for a relationship with them, but as a replacement for a defect in the psy-
chological (self) structure.
Have these theoretical models of narcissism and substance abuse got any support
from clinical research? In the 1970s and 80s, several studies supported the notion that
many substance abusers may use substances as an expression of, or an attempt to cope
with, narcissistic disturbances (Krystal and Raskin, 1970; Hendin, 1973; Wurmser, 1974;
Blatt, 1984; Treece and Khantzian, 1986; Dodes, 1990). These and other studies are often
based upon relatively large clinical samples. Treece and Khantzian (1986) reviewed
former findings from substance abusers in different forms of treatment (psychother-
apy, therapeutic community, methadone maintenance), and concluded that they seem
to be characterized by:
- Difficulties with management of affects: poor tolerance for affects, all-or-
nothing quality of affects, easily overwhelmed and flooded by affects,
problems with affect modulation and a vulnerability for self-criticism. Intense
affective arousal may constitute a psychic trauma.
- Severe narcissistic vulnerability: inability to turn to others for comfort, fear
of closeness and aggression in relation to others, intolerance of rejection,
246 P. Vaglum
disappointment or loss may lead to rage in close relationships, reliance on

others and denial of dependency at the same time. Omnipotence and sense
of entitlement permit the criminal and/or dangerous self destructive behav-
iour. Failure of self-care leads to ignorance about the dangerous nature of
substances. They do not anticipate harm and danger, do not recognize warning
signals. They do not care for themselves.
- Substances are often used as self-treatment: patients often prefer substances
which help them to cope with the particular affects that trouble them most.
Opiates, alcohol, and barbiturates may calm intense feelings of rage, shame,
loneliness, emptiness, anxiety, and depression. Psychedelics may counteract
boredom, emptiness and meaninglessness, and induce the illusion that the self
is mystically boundless and grandiose. Central stimulants provide a sense of
mastery, invincibility, and grandeur.
Acquiring knowledge about the addict's most preferred substance may give
important clues about which types of affects are most troublesome and/or whether it
is a feeling of omnipotence that is sought. This may also clarify which emotional prob-
lems the abuser will have to face in a drug-free treatment programme, and be able to
determine whether methadone will give sufficient psychological benefits, or whether
the risk for additional use of other substances is high.
Summarizing: psychodynamic studies document that many addicts have distur-
bances in the development and function of the self. Why some narcissistically disturbed
persons abuse substances while others don't, is a question which needs further explo-
ration, and biologic, social, cultural, and additional psychopathological factors should
be considered. Experiences from psychotherapy indicate that, for many addicts, the nar-
cissistic disturbances come first. Sometimes, the start of substance abuse can be con-
nected to a "narcissistic crisis" in which narcissistic defences are seriously threatened
(Wurmser, 1974). On the other hand, one cannot rule out the possibility that for some
addicts, use of substances may weaken the more major defences, and thereby increase
the domination of narcissistic disturbances. Substance abuse will thereby increase
the level of narcissistic disturbances, and drug use and narcissistic disturbances may
interact in a reciprocal way over time. Regardless of the primary/secondary question,
psychodynamic knowledge should have important implications for the psychosocial
treatment of addicts.
4. CAN A "THIRD VARIABLE" EXPLAIN THE CO-

OCCURRENCE OF NPD AND SUBSTANCE ABUSE?
Theoretically, the co-occurrence of NPD and substance abuse could be due to

biologic disturbances in affect regulation, disinhibition, and impulse control. This would
make the individual susceptible both for a disturbed development of the self functions,
and for a special sensitivity or reactivity to alcohol and other substances. This biologic
vulnerability may be genetically or environmentally induced. Unfortunately, genetic
and prospective longitudinal studies, giving answer to this question in children are still
lacking. The first genetic study concerning narcissistic traits has been published,
showing a heritability estimate of 0.53 (Jang et aI., 1996). The clinically frequent co-
occurrence of antisocial personality disorder (ASPD), which is partially genetically
determined (Cadoret et aI., 1995), may also indicate a possible co- occurent genetic/
biologic vulnerability for these disorders. Environmental factors may then explain why
some persons develop NPD and not ASPD, or vice versa, while others have the com-
bination. Research on psychopaths (Hart and Hare, 1989) has found that if psychopa-
thy is divided into two factors, it is the antisocial factor and not the narcissistic factor
which is related to drug abuse among psychopaths. One should explore whether sub-
stance abusing-persons with NPD more often have a co-morbid ASPD than non-
substance-abusing persons with NPD.
5. IS NPD RELATED TO DROPOUT FROM TREATMENT AND

TO THE CLINICAL COURSE?
Table 5 shows six studies that have explored the relationship between NPD and
attrition from treatment and/or clinical outcome. Craig (1984), Stark and Campbell
(1988), and Cacciola et al. (1996) found no relationship between NPD and short-term
dropout from treatment. Ravndal and Vaglum (1991a) found the opposite of what we
expected: Patients who scored high on the MCMI narcissistic scale completed more
often than the others the drug-free inpatient year in the Phoenix House therapeutic
community (while those on the MCMI antisocial scale did not). Kernberg (1982) has
previously proposed that persons with narcissistic disturbances would tend to stay
longer in therapeutic communities, because this treatment model might specially gratify
their narcissistic needs. Our findings support this hypothesis, especially because we
found that the MCMI narcissistic scores increased significantly during the inpatient
year (as did antisocial, histrionic, and compulsive scores) (Ravndal and Vaglum, 1991b).
This increase in scores may, however, alternatively mean that the patients had devel-
oped a stronger self-feeling and a better self-esteem during treatment. We are explor-
ing these possibilities in the five-year follow-up of these clients.
Studies on the influence of NPD on the further clinical course, is also very few
(table 5). Kosten et al. (1989) found that the presence of a self disorder was related to
the 2.5 year outcome of the medical condition of methadone patients, but not to the
substance abuse or to the criminal activity. In our prospective, five year follow up study
Table 5. NPD-and course and outcome of addiction

Author Subjects Observation time Method Results
Craig (1984) 100 opiate MCMI No relationship
Stark and Campbell (1988) 100 polydrug 2 months MCMI No relationship
Kosten et al. (1989) 150opioid 2.5 years DSM-I1I More medical problems,
no relation to substance
abuse
Ravndal and Vaglum (1991) 144 polydrug 18 months MCMI High NPD score-low
dropout from Phoenix
House
Cacciola et al. (1996) 197 opioid men 7 months SCID-II No relationship
Ravndal and Vaglum (in press) 200 polydrug 5 years MCMI OR: 4.8 for death if score
>75. No relationship to
substance abuse
248 P. Vaglum
of 200 substance abusers, being a case «BR: 75) on the MCMI narcissistic scale
increased the risk of death in the following year significantly (OR: 4.8), and more
strongly than being an ASPD case. The NPD scores were, however, not related to the
further level of substance abuse five years later (Ravndal and Vaglum, in press).
Summarizing: these studies are still very few, and need replication both in drug
free and methadone maintenance programmes. So far, they suggest that narcissistic
problems may be important for involving addicts in therapeutic community models, but
they may also indicate a serious of lack of self-care that may increase the risk of death
or medical disorders. Even if NPD may influence the recruitment to substance abuse
among those who have become addicts, NPD does not seem to influence the further
course of the addiction itself. NPD addicts do not seem harder to treat than other
addicts, but we lack studies that have explored whether a significant reduction of nar-
cissistic problems through treatment may have a positive effect on the further course
of substance abuse. As mentioned earlier, several studies have shown that the nar-
cissistic items on the DSM-III or MCMI narcissistic scale can be divided into sev-
eral subsets like self-centred arrogance, extroversion, and independent superiority (Di
Guiseppe et aI., 1995; Choca et aI., 1996). Such different subsets may have different
importance for treatment completion and for clinical outcome, and this should be
further explored in forthcoming research (Ronningstam et aI., 1995).
6. CLINICAL IMPLICATIONS
Treatment of narcissistic disturbances in general is outside the scope of this

chapter, but it is very well discussed in the books and papers by Kohut and his follow-
ers, and by Kernberg (Kohut, 1971, 1974, 1977; Kernberg, 1975, 1984). Here we will only
focus on what may be especially important concerning narcissistic persons with addic-
tive disorders.
So far, clinical research and experiences show that a relatively small portion of
substance abusers who come to treatment fulfil the criteria of a DSM-III NPD disor-
der, while many addictive patients report a high prevalence of narcissistic traits when
they are given self-report instruments. Using the DSM criteria and self-report instru-
ments, one should be able to carefully diagnose narcissistic disturbances and disorders
in the substance-abusing patient population. Self-report instruments should also be
useful in monitoring the development of the level of self reported narcissistic traits
during the participation of the treatment programme.
For clients whose use of substances is strongly motivated by a need to prevent a
fragmentation of the self, or to bolster or increase pathological omnipotence and
grandiosity, psychotherapeutic treatment models which aim at supporting and strength-
ening the self should be of special interest to therapists in the addiction field. Self-
psychology offers such a model for both individual and group psychotherapy. However,
Ulman and Paul (1989) rightly noticed that except for a few textbooks (Adams, 1978;
Forrest, 1983; Levin, 1987), self-psychologists in general have been surprisingly reluc-
tant to apply a self-psychological conceptualization of narcissism to understanding
and treating addictions. Ulman and Paul (1989, 1990) themselves proposed a self-
psychological model which is based on the view that the addict is addicted to a self-
object experience of fantasy and mood, which are triggered biochemically, physiologi-
cally, or psychologically by any substance (including food) (or behaviour or person)
with whom the addict is obsessively attached. The archaic self object fantasies are
usually accompanied by moods of narcissistic bliss, consisting of an intensely pleasur-

able feeling of grandiosity, invulnerability; tranquillity, serenity, and numbness. These
activations of archaic self object fantasies and arousal of mood of narcissistic bliss tem-
porarily buffer against, anaesthetize, and provide dissociation from (1) painful and
chronic states of self-fragmentation and anxious feelings of falling apart, going to pieces
and disintegrating, and (2) painful and chronic states of self-collapse and depressive
feelings of emptiness, depletion, and deadness. In their two papers, Ulman and Paul
(1989,1990) describe their treatment of individual patients by means of a self-psycho-
logical method, showing how the fantasized relationship to the therapist may replace
the self-object experience that previously was sought by chemical means. Controlled
studies of psychotherapy with substance abusers, based on a self-psychological model,
should be conducted to see whether they can add something significantly new to the
existing service models we have.
Concerning the AA model, Bell and Khantzian (1991) have pointed out that this
model of self-help is designed in the way one would design a helping programme for
people with narcissistic disturbances. They underline that the twelve steps of AA can
be viewed as a way of transforming narcissistic disturbances to a better functioning of
the self. The philosophy of AA provides a new edition of the self-experience that allows
tension regulation through: (1) an unconditioned acceptance, (2) affirmation, (3)
mirroring, (4) opportunities for idealization, and (5) for twinships. Recovery in AA
involves a gradual surrender of one's grandiosity, and a gradual acceptance of oneself
as imperfect, limited and therefore human. This may be illustrated by Step one: "we
admit we were powerless ....", Step two: "we have come to believe that a power greater
than ourselves could restore us ...... ", Step three: "we made a decision to turn our
will and lives over to the care of God ...... ", and so forth, reducing the illusion of
the self-sufficiency. Since AA, and possibly also other religious programmes, seem to
have better results than many professional programmes, this may support the other
clinical observations that the restoration of the self should be a central topic in treat-
ment modalities.
Concerning the specific therapeutic community models, our findings (Ravndal
and Vaglum, 1994) indicate that many abusers with a high degree of narcissistic dis-
turbances may "survive" in this, so called, Phoenix House model. The problem is that
they do not change enough to be more able to manage without substances when they
leave the inpatient programme. Therapeutic communities have in many ways the same
idea as the AA: First an initial phase of powerlessness, followed by phases in which a
new self should manifest itself. We have, however, previously shown (Ravndal and
Vaglum, 1994) that the Phoenix House model includes several elements that may
undermine the clients' possibility for identification, idealization, mirroring, twinship
relationships, especially in the early phases. The confrontative aggressive style in many
therapeutic communities is also very different from the supportive, anxiety reducing
atmosphere at the AA meetings. It is possible that the therapeutic community models
should be revised with a more careful look at how one should relate more specifically
to the narcissistic disturbances of the clients.
7. CONCLUSIONS
The overlap of NPD and substance abuse is not merely due to chance. Though
there is a selection bias in clinical samples, there are some indications that NPD may
250 P. Vaglum
increase the risk of becoming an addict, at least of becoming a psychiatric help-seeking

patient.
NPD and narcissistic traits seem to be unrelated to the further course of
addiction per se, but may increase the risk of medical disorders and death among
addicts.
Substance abuse may increase the domination of narcissistic disturbances and vica
versa.
There are, up to now, no studies showing the genetic/biological prove that NPD
and substance abuse are related.
Whether treatment of narcissistic disturbances leads to a good outcome with
regard to addictive behaviour is still an open question because of lack of research. Psy-
chodynamic experiences from clinical work and research have increased our under-
standing of the importance of narcissistic disturbances for the individual's psychological
vulnerability for abusing substances.
Some treatment modalities (not AA) should possibly be somewhat revised on the
basis of the knowledge from self-psychology and object-relational theory, about the
relationship between narcissistic disturbances and substance abuse.
Narcissistic disturbances may keep addicts in a therapeutic community model, but
the model does not change the narcissistic disturbances enough to reduce the need of
substances later on.
REFERENCES
Abraham K.: The psychological relations between sexuality and alcoholism (1908), in Selected Papers on
Psychoanalysis. London, Hogarth Press, 1973.
Adams J.W.: Psychoanalysis of drug dependence. The understanding and treatment of a particular form of
pathological narcissism. New York/San Francisco/London. Grune & Stratton. 1978.
Alnres R and Torgersen S.: DSM-III symptom disorders (axis II) in an outpatient population. Acta Psychiatr
Scand 1988,78,348-355.
Balint M.: The basic fault. London, Tavistock, 1968.
Barber J.P., Frank A., Weiss RD., Blaine J., Siqueland, Maras K., Calvo N., Chittanes J., Mercer D., and
Salloum LM.: Prevalence and correlates of personality disorder diagnoses among cocaine dependent
outpatients. J Personality Dis 1996, 10,297-311.
Bell CM. and Khantzian E.J.: Contemporary psychodynamic perspectives and the disease concept of addic-
tion: complementary or competing models? Psychiatric Annals 1991,21,273-281.
Black D.W., Goldstein RE., Blum N., and Noyes R: Personality characteristics in 60 subjects with psycho-
sexual dysfunction: a non patient sample. J Personality Dis 1995,9,275-285.
Blatt S.J., Berman W., Bloom-Feshback S., Sugarman A., Wilber C, and Kleber H.D.: Psychological asses-
sment of psychopathology in opiate addicts. J Nerv Ment Dis 1984, 172, 156-165.
Bodlund 0., Ekselius L., and Lindstrpm E.: Personality traits and disorders among psychiatric outpatients and
normal subjects on the basis of the SCID screen questionnaire. Nord J Psychiatry 1993,47,425-433.
Brooner RK., King Y.L., Kidorf M., Schmidt CW., and Bigelow G.E.: Psychiatric and substance use comor-
bidity among treatment-seeking opioid abusers. Arch Gen Psychiatry 1997,54,71-80.
Brown H.P.: Substance abuse and the disorders of the self Examining the relationship. Alcoholism Treatment
Quarterly 1992,9,1-27.
Cacciola J.S., Rutherford M.J., Alterman A. I. , McKay J.R, and Snider E.C: Personality disorders and
treatment outcome in methadone maintenance patients. J Nerv Ment Dis 1996,184,234-239.
Cadoret RJ., Yates W.R, Troughton E., Woodworth G., and Stewart M.: Adoption study demonstrating to
genetic pathways to drug abuse. Arch Gen Psychiatry 1995,52,42-52.
Chein I., Gerard D.L., Lee RS., and Rosenfeld E.: The road to H. New York, Basic Books, 1964.
Choca 1., Retzlaff 1'., Strack S., Mouton A., and van Denburg E.: Factorial elements in Millon's Personality
Theory. J Personality Dis 1996,10,377-383.
Clerici M. and Cazzullo c.L.: Substance abuse and psychopathology. A diagnostic screening of Italian narcotic
addicts. Social Psychiatric Epidemiology 1989,24,219-226.
Craig RJ.: Can personality tests predict treatment dropouts? Int J Addictions 1984, 19,665-674.
De Jong C.A.J, van den Brink W., Harteveld F.M., and van der Wielen E.G.M.: Personality disorders in
alcoholics and drug addicts. Comprehensive Psychiatry 1993,34,87-94.
Di Guiseppe R, Robin M., Szeszko P.R, and Primavera L.H.: Cluster analysis of narcissistic personality
disorders on the MCMI-Il. J Personality Dis 1995,9,304-317.
Dodes L.M.: Addiction, helplessness and narcissistic rage. Psychoanal Q. 1990, 109, 398-419.
Donovan J.M.: An etiological model of alcoholism. Am J Psychiatry 1986, 143, I-II.
Drake RE. and Vaillant G.E.: Predicting alcoholism and personality disorder in a 33-year longitudinal study
of children of alcoholics. Br J Addiction 1988,83,799-807.
Fenichel 0.: The psychoanalytical theory of neurosis. London, Routledge and Kegan Paul, 1946.
First M.B., Spitzer RL., Gibbon M., Williams J.B., Davies M., Bones J., Howes M.J., Kane J., Pope H.G., and
Rounsaville B.: The structured clinical interview for DSM-III-R personality disorders (SCID-II) Part
II: Multi-site test-retest reliability study. J Personality Dis 1995, 9, 92-104.
Forrest G.G.: Alcoholism, narcissism and psychopathology. Springfield, Charles C. Thomas, 1983.
Freud S.: A reply to criticism of my paper on anxiety neurosis. Standard Edition, 1895,3,123-139. London,
Hogarth Press, 1962.
Freud S.: Civilazation and its discontents. Standard Edition, 1930, 21, 59-145. London, Hogarth Press, 196I.
Golomb M., Fava M., Abraham M., and Rosenbaum J.F.: Gender differences in personality disorders. Am J
Psychiatry 1995, 152,579-582.
Hart S.D. and Hare RD.: Discriminant validity of the Psychopathic Checklist in a forensic psychiatric
population. Psychological Assessment: A journal of Consulting and Clinical Psychology. 1989, 1,
211-218.
Hendin H.: Marijuana abuse among college students. J Nerv Ment Dis 1973, 156,259-270.
Jang K.L., Livesley W.J., Vernon P.A, and Jackson D.N.: Heritability of personality disorder traits: a twin study.
Acta Psychiatr Scand 1996, 94, 438-444.
Kernberg 0.: Borderline conditions and pathological narcissism. New York, Jason Aaronson, 1975.
Kernberg 0.: Advantages and liabilities of the therapeutic community in: Pines M., Rafaelsen L. (eds): The
individual and the group. Boundaries and interrelations. New York, Plenum press, 1982.
Kernberg 0.: Severe personality disorders. Psychotherapeutic strategies. New Haven, Yale University Press,
1984.
Khantzian E.J.: Opiate addiction: A critique of theory and some implications for treatment. Amer J Psy-
chotherapy 1974, 28, 59-70.
Khantzian E.J. and Treece c.: DSM-III psychiatric diagnosis of narcotic addicts. Arch Gen Psychiatry 1985,
42,1067-1071.
Khantzian E.J. and Mack J.E.: Alcoholism Anonymous and contemporary psychodynamic theory in Galanter
M (ed): Recent developments in alcoholism, vol 7. New York, Plenum Press, 1989.
Koenigsberg H.W., Kaplan RD., Gilmore M.M., and Cooper A.M.: The relationship between syndrome
and personality disorder in DSM-lII: Experience with 2462 patients. Am J Psychiatry 1985, 142,
207-212.
Kohut H.: The analysis of the self. New York, International Universities Press, 197I.
Kohut H.: The restoration of the self. New York, International Universities Press, 1977.
Kohut H. and Wolf E.: The disorders of the self and their treatment: an outline. Int J Psycho-Anal 1978, 59,
413-425.
Kosten T.A, Kosten T.R, and Rounsaville B.1.: Personality disorders in opiate addicts show prognostic
specificity. J Substance Abuse Treatment 1989,6,163-168.
Krystal H. and Raskin H.A.: Drug dependence: aspects of ego functions. Detroit, Wayne State University
Press, 1970.
Levin J.D.: Treatment of alcoholism and other addictions. A self psychological approach. London, Jason
Aronson 1987.
Loranger AW., Sartorius N., Andreoli A, Berger P., Dieckstra R, Fergusson B., Jacobsberg L.B., Mombour
W., Pull C.h., Ono Y., and Regier D.A.: The international personality disorder examination. Arch Gen
Psychiatry 1994,51,215-224.
Maffei c., Fossati A, Lingiardi v., Madeddu F., Borellini c., and Petrachi M.: Personality maladjustment,
defences, and psychopathological symptoms in non clinical subjects. J Personality Dis 1995,9,330-345.
Maier w., Lichterman D., Klingler T., and Heun R: Prevalences of personality disorders (DSM-III-R) in the
community. J Personality Dis 1990,6,187-196.
252 P. Vaglum
Marsh D.T., Stile S.A, Stoughton N.L., and Trout-Landen B.L.: Psychopathology of opiate addiction:
comparative data from the MMPI and MCMI. Am J Drug Alchol Abuse 1988,14,17-27.
Mattarah IIF., Becker D.F., Levy KN., Edell W.S., and McGlashan T.H.: Diagnostic stability in adolescents
followed up 2 years after hospitalization. Am J Psychiatry 1995, 152, 889-894.
McMahon Re., Flynn P.M., and Davidson RS.: The personality and symptoms scales of the Millon Clinical
Multixial Inventory: sensitivity to posttreatment outcomes. J Clinical Psychology 1984, 4, 862-866.
Miller L.: Neuropsychodynamics of alcoholism and addiction: Personality, psychopathology, and cognitive
style. J Substance Abuse Treatment 1990,7,31-49.
Moldin S., Rice IP., Erlenmeyer-Kimling L., and Squires-Wheeler: Latent structure of DSM-III-R axis II.
Psychopathology in a normal sample. J Abnorm Psychology 1994,103,259-266.
Nace E.P., Davis e.w., and Gaspari IP.: Axis II comorbidity in substance abusers. Am J Psychiatry 1991, 148,
118-120.
Nakao K, Gunderson IG., Phillips KA, Tanaka N., Yorifuji K., Takaishi I, and Nishimura T.: Functional
impairment in personality disorders, 1992, 6, 24-33.
Pfohl B., Coryell W., Zimmerman M., and Stangl D.: DSM-llI Personality disorders: Diagnostic overlap and
internal consistency of individual DSM-III criteria. Comprehensive Psychiatry 1986,27,21-34.
Poldugro F. and Forti B.: Personality disorders and alcoholism treatment outcome. Drug Alcohol Dependence
1988,21,171-176.
Rado S.: The psychoanalysis of pharmacothymia. Psychoanal Q 1933,2, 1-23.
Ravndal E. and Vaglum P.: Psychopathology and substance abuse as predictors of program completion
in a therapeutic community for drug abusers: a prospective study. Acta Psychiatr Scand 1991, 83,
217-222.
Ravndal E. and Vaglum P.: Changes in antisocial aggressiveness during treatment in a hierarchical
therapeutic community. A prospective study of personality changes. Acta Psychiatr Scand 1991, 84,
524-530.
Ravndal E. and Vaglum P.: Why do drug abusers leave the therapeutic community? Nord Psychiatr J 1994,48,
Suppl. 33.
Ravndal E. and Vaglum P.: Psychopathology, treatment completion and 5 year outcome: a prospective study
of drug abusers. J Substance Abuse Treatment, in press.
Reich I, Yates W., and Nduaguba M.: Prevalence of DSM-III personality disorders in the community. Social
Psychiatry and Psychiatric Epidemiology 1989,24,12-16.
Richman IA., Flaherty lA, and Rospenda KM.: Perceived workplace harassment experiences and
problem drinking among physicians: broadening the stress/alignment paradigm. Addiction 1996,91,
391-403.
Ronningstam E., Gunderson IG., and Lyons M.: Changes in pathological narcissism. Am J Psychiatry 1995,
152,253-257.
Rousar E., Brooner RK., Regier M.W., and Bigelow G.E.: Psychiatric distress in antisocial drug abusers:
relation to other personality disorders. Drug Alcohol Dependence 1994,34,149-154.
Samuels IF., Nestadt G., Romanovski A.I, Folstein M.F., and McHugh P.R: DSM-Ill personality disorders
in the community. Am J Psychiatry 1994,151,1055-1062.
Simonsen E., Haslund I, Larsen A, and Bprup e.: Personality pattern in first-time-admitted alcoholics. Nord
Psychiatric J 1992,46,175-180.
Stancovic S.R, Libb IW., Freeman AM., and Roseman 1M.: Posttreatment stability of the MCMI-II
personality scales in depressed outpatients. J Personality Dis 1992,6,82-89.
Stark M.I and Campbell B.K: Personality, drug use and early attrition from substance abuse treatment. Am
J Drug Alcohol Abuse 1988, 14,475-485.
Treece e. and Khantzian E.I: Psychodynamic factors in the development of drug dependence. Psychiatric
Clinics of North America 1986,9,399-412.
Ulman RB. and Paul H.: A self psychological theory and approach to treating abuse disorders: The
"intersubjective absorption" hypothesis. in Goldberg A (ed): Dimensions of self experience. Progress
in self psychology, vol 5. Hillsdale, The Analytic Press, 1989, 121-141.
Ulman RB. and Paul H.: The addictive personality and "addictive trigger mechanisms" (ATMs): The self
psychology of addiction and its treatment. in Goldberg A (ed): The realities of tranference. Progress in
self psychology, vol 6. Hillsdale, The Analytic Press, 1990, 129-156.
Vaglum S. and Vaglum P.: Borderline and other mental disorders in alcoholic female psychiatric patients: a case
control study. Psychopathology 1985,18,50--60.
Verheul R, van den Brink W., and Hartgers e.: Prevalence of personality disorders among alcoholics and
drug addicts: an overview. Eur Addict Res 1995, 1, 166-177.
Wittchen H.W.: Critical issues in the evaluation of comorbidity of psychiatric disorders. Br J Psychiatry 1996,
168, Suppl. 30, 9-16.
Wurmser L.: Psychoanalytic considerations of the etiology of compulsive drug use. J American Psychoana-
lytic Ass 1974,22,820-843.
Yates W.R., Fulton A.I., Gabel J.M., and Brass C.T.: Personality risk factors for cocaine abuse. AM J Public
Health 1989,79,891-892.
Zimmerman M. and Coryell W.: DSM-III personality disorder diagnoses in a nonpatient sample. Arch Gen
Psychiatry 1989, 46, 682--689.
18
PHARMACOTHERAPY FOR PATIENTS WITH

Experiences from a Group Analytic Treatment Program
S. Friis, T. Wilberg, T. Dammen, and 0. Urnes
Department of Psychiatry
Ulleven University Hospital
N-0407 Oslo, Norway
1. INTRODUCTION
There is an increasing interest in the pharmacological treatment of personality

disorders. Despite the experience that psychotropic drugs often have only a modest
effect on patients with personality disorders, medication has been suggested as an
important adjunct to the overall treatment of these patients (Gitlin, 1993; Soloff, 1994).
According to various conceptualisations of personality pathology, one may apply dif-
ferent strategies in the pharmacological treatment of personality disordered patients
(Gitlin, 1993). Treatment may be directed towards various personality disorder cate-
gories, or specific symptom clusters within or across disorders, some of which have been
suggested to have a biological substrate (Tuinier and Verhoeven, 1995; Soloff, 1994).
The symptom clusters most frequently cited are affective instability, impulsivity,
and aggressiveness, anxiety, and inhibition, and cognition and perception (DeBattista
and Glick, 1995). A third alternative is to treat associated axis I disorders. However, in
the presence of personality pathology the response to pharmacological treatment of
axis I disorders is less pronounced (Reich and Vasile, 1993; Shea, Widiger, and Klein,
1992).
The empirical support for the different pharmacological treatment strategies is
still sparse (Stein, 1992). Pharmacological studies of patients with personality disorders
are few, involve a small number of patients with a restricted range of personality dis-
orders, and are suffering from limiting methodological problems (DeBattista and Glick,
1995). Furthermore, many of the relevant drug groups have been reported to have non-
specific effects in patients with personality disorders (Rosenberg, 1994). On the basis
256 S. Friis et al.
of pharmacological research it has not yet been possible to delineate clinical subtypes
of for instance borderline personality disorder (Soloff, 1994).
For personality disorders of the cluster C type, characterised by anxiety and inhi-
bition, one hypothesis has been that they would respond to anxiolytic and antidepres-
sant treatment. However, to our knowledge there are no studies of pharmacotherapy
of adult patients with cluster C disorders. Of the cluster A disorders, only schizotypal
personality disorder has been focused. A few open trials suggest that neuroleptics may
have a positive effect on psychotic-like symptoms, anxiety, and depression in schizo-
typal patients (Hymowitz, Frances, Jacobsberg, Sickles, and Hoyt, 1986; Serban and
Siegel,1984).
Among the cluster B personality disorders, associated with affective instability
and impulsive and aggressive behaviour, most studies focus on borderline personality
disorder (BPD). The effect of neuroleptics on patients with BPD is controversial.
Whereas three of five placebo-controlled studies report positive response on a broad
range of symptoms including depression, anger, and hostility, impulsiveness, schizotypal
features, and suicidal behaviour (Goldberg et aI., 1986; Montgomery and Montgomery,
1982; Soloff et aI., 1989), two studies do not report marked effect of neuroleptic
treatment (Cornelius, Soloff, Perel, and Ulrich, 1993; Soloff et aI., 1993). Nevertheless,
small doses of neuroleptics are recommended for brief psychotic episodes and episo-
des of severe behavioural dyscontrol in patients with BPD (Stein, 1992). Furthermore,
carbamazepine has been found effective on behavioural dyscontrol in one of two
placebo-controlled studies (Tuinier and Verhoeven, 1995).
Concerning antidepressant treatment, tricyclics have not been found very useful,
with a significant risk of paradoxical worsening of behavioural dyscontrol (Stein, 1992;
Soloff, 1994). Investigators have suggested a role for monoamine oxidase inhibitors in
the treatment of BPD, in particular the atypical depressive features. However, the
results are not consistent across studies (Tuinier and Verhoeven, 1995; DeBattista
and Glick, 1995). Promising reports have been presented concerning selective serotonin
reuptake inhibitors (SSRIs), which have been used to target the affective instability
and impulsivity. Studies have reported positive response on depressed mood, lability,
rejection sensitivity, impulsive behaviour, self-mutilation, psychoticism, and hostility,
as well as increased level of global functioning. However, five of the six existing studies
are open trials and comprise a small number of patients, mainly recruited from out-
patient settings or through advertisement (Nordern, 1989; Cornelius, Soloff, Perel,
and Ulrich, 1990; Coccaro, Astill, Herbert, and Scut, 1990; Markovitz, Calabrese, Schulz,
and Meltzer, 1991; Kavoussi, Liu, and Coccaro, 1994; Salzman et aI., 1995). There is a
need for placebo controlled studies of SSRIs in patients with a wider spectre of
personality disorders or personality dimensions, and in a broader range of treatment
settings.
Despite the recommendation of adjunct pharmacotherapy for many personality
disordered patients, we do not know much about the actual use of psychotropic drugs
in various treatment settings. Furthermore, there is a lack of studies from psychothera-
peutic treatment programmes comparing outcome for patients receiving additional
drug therapy with outcome for those who do not receive such additional therapy.
The present paper is a preliminary study, presenting experiences with pharma-
cotherapy in a psychodynamic group therapy programme, specialised in the treatment
of patients with personality disorders. Emphasis will be put on antidepressants in per-
sonality disordered patients with comorbid mood disorders.
The aims of the study are to explore the following questions: 1. What is the phar-
macological treatment practice in a clinical psychotherapeutic setting for personality
Pharmacotherapy for Patients with Personality Disorders 257
disorders? 2. What is the relationship between axis-I and axis-II disorders, and phar-
macotherapy in such a sample? 3. What is the relationship between use of antidepres-
sant medication, depressive symptoms, and outcome?
2. METHODS
The study was a naturalistic one, conducted at the Day Unit, Psychiatric Depart-
ment, Ullev:'H University Hospital, Oslo, Norway. The Day Unit is a specialised unit for
treatment of personality disorders.
Assessments
Axis I and axis II disorders were assessed by SCID-I and SCID-II interviews. The
SCID-I interviews were made by two research residents (TW and TD). The SCID II
interviews were conducted by the clinical staff. The diagnostic interviews for both axis
I and axis II disorders were followed by a team discussion both at admission and dis-
charge. Additional information obtained during the stay was considered, and final con-
sensus diagnoses were made. Furthermore, SCID I interviews were audiotaped, and
were independently rated by the other interviewer, blind to the result of the former
interviewer, for the presence/absence of the following disorders: Dysthymia, Major
Depression, Bipolar Disorder, Social Phobia, Panic Disorder, Obsessive Compulsive
Disorder, Simple Phobia, Eating Disorder, Somatoform Disorder, and Substance Use
Disorder. The median kappa was 0.93 (range 0.63 to 1.00).
At admission and discharge the SCL-90R (Derogatis, 1977) was filled in by the
patients and a GAF (Global Assessment of Functioning Scale, 1994) score was made
for each patient by the therapists. All GAF-scores were made by team consensus esti-
mates. Based on rating of clinical vignettes, the team members' inter-rater reliability
on GAF scores has been found to be high.
Patients
During the period March 94 through December 95, 111 patients were consecu-
tively admitted to the unit. All patients were considered eligible for study inclusion.
However, nine patients were excluded, as they dropped out during the two first weeks.
The remaining 102 patients formed the sample of the study.
Seventy-six (75%) of the patients were female, 26 (25%) were male. Mean age at
admission was 33.3 years (SD 7.7). Mean years of education was 12.9 (SD 3.0). Fifty-
three (52%) were single, 31 (30%) were married or stable cohabiting, 16 (16%) were
separated or divorced, and 2 (2%) were widowed. At admission 34 (33%) patients were
on sick leave, 35 (34%) were on rehabilitation, one (1 %) on disability pension, and 12
(12%) out of work. Among the remaining 20 patients 15 (15%) were students, three
(3%) were housewives, and two (2%) were in paid work.
Sixty (59%) of the patients had previously had psychotherapy, and 46 (45%) had
previously been admitted to a psychiatric unit.
At admission the GSI-score of the SCL-90R was 1.50 (SD 0.68) and the mean
GAF score was 45.0 (SD 5.4).
The mean number of axis-II disorders for all patients was 1.5 (SD 1.1). For the
85 patients with at least one PD, the mean number of axis-II disorders was 1.74 (SD
0.98). In table I we have used a hierarchy, so that patients meeting the criteria for a
258 S. Friis et al.
Table 1. Distribution of Axis-I disorders by Axis-II disorders

CIA CIB CIC Mix No Total
Anxiety disorders:
Panic disorder (no agoraphobia) 4 2 2 10
Panic disorder with agoraphobia 5 7 5 2 3 22
Social phobia 3 11 13 3 0 30
Obs/Comp disorder 3 3 0 0 7
Mood disorders:
Major affective disorder 10 20 14 13 9 66
Other mood disorder 0 8 5 0 6 19
Other disorders:
Any substance use disorder 5 7 5 2 20
Any eating disorder 2 4 4 3 3 16
Any somatoform disorder 5 4 4 4 5 22
Total number of patients 13 31 26 15 17 102
Cluster A diagnosis was assigned to this group irrespective of comorbidity with other
axis-II disorders. The next step in the hierarchy was Cluster B. As a consequence, those
with more severe personality disorders, had the highest number of axis-II disorders.
(The mean numbers were: Cluster A 2.7; Cluster B 2.0, Cluster C 1.3).
All patients had at least one axis-I disorder. The mean number of axis-I disorders
was 2.4 (SD 1.2). The distribution of disorders is given in table 1. As seen from the
table, mood disorders were the most prevalent group. Eighty-three percent of the
patients met the criteria for a mood disorder. The table reveals a considerable comor-
bidity between axis-land axis-II disorders. Over all, the more severe the personality
disorder, the higher the number of axis-I disorders. (Mean number: Cluster A: 3.1;
Cluster B 2.5; Cluster C 2.0, Mixed PD 2.0, and No PD 2.0). A main reason for this
difference was that the more severe the personality disorder, the greater the chance
for one or more comorbid anxiety disorders.
Treatment
The Day Unit treatment programme and outcome results have been described in
several papers elsewhere (Vaglum et ai., 1990; Vaglum, Friis, Vaglum, and Larsen, 1989;
Mehlum et ai., 1991; Karterud et ai., 1992; Mehlum, Friis, Vaglum, and Karterud, 1994;
Karterud et ai., 1995; Karterud et ai., 1998; Wilberg et ai., 1998). From 1994 the treat-
ment programme was changed from a therapeutic community treatment with combined
individual- and group therapy, to a programme exclusively based on different kinds of
group therapies. All patients were offered a 18 weeks programme. An overview of the
programme is given in Fig. 1. The Day Unit programme was part of a long-term psy-
chotherapy programme with an offer of long-term outpatient group therapy after
discharge.
Pharmacotherapy was administered through a medication group, led bya psy-
chiatric resident. At admission, all patients were referred to this group for evaluation
of medication. Decisions regarding pharmacotherapy were made partly in this group,
partly in the treatment team. Patients were recommended to continue medication, or
were prescribed medication based on a clinical evaluation. For some patients showing
Small group Cognitive

psychotherapy
Management
group
Large group
Ca. 35 participants
1 hour daily Cognitive
behavioral
group (anxiety)
Body
awareness
Figure 1. The treatment program.
depressive symptoms, a MADRS-rating (Montgomery and Asberg, 1979) was made

when in doubt whether or not antidepressant was to be considered appropriate treat-
ment. A MADRS-score > 20 was taken as an indication that antidepressive medication
should be prescribed. SSRIs were the drugs of choice for depression, mainly paroxe-
tine or fluvoxamine. Antidepressants (basically SSRIs) were also recommended or pre-
scribed for Panic Disorder, Obsessive Compulsive Disorder, Social Phobia, and Eating
Disorders. Neuroleptics were given for psychotic symptoms, or symptoms bordering on
psychosis. Sedatives were suggested for patients suffering from sleeping disturbances.
The patients were free to refuse to take medication, but most of them accepted
medication when suggested. Patients who received pharmacotherapy continued to
attend the medication group for adjustment of doses and assessment of side effects. No
objective measure of treatment compliance, such as serum analysis of medication/
metabolites or tablet counting, were applied.
3. RESULTS
3.1. Medication
At admission, nineteen (19%) patients were treated with neuroleptics, 44 (43 %)
with antidepressants, and eight (8%) with some other medication (sedatives or anxi-
olytics). During the stay, 16 patients (16%) used neuroleptics,58 (57%) antidepressants,
and six patients (6%) some other medication. As antidepressants was the only group
of drugs given to a substantial number of patients, the further analyses are limited to
antidepressants. The patients given antidepressants are termed the Ad-group, while
those not given are termed the NoAd-group. Among the 58 patients in the Ad-group,
260 S. Friis et al.
Table 2. Number and percentage of patients given medication

Mood Disorder No Mood Disorder Total
Ad-group:
Ad only 41 3 44
Ad+N 8 2 10
Ad+O 3 0 3
Ad+N +0 0 1
Sum 53 5 58
NoAd-group
N only 2 2 4
o only 1 0
N+O 0
No medication 28 10 38
Sum 32 12 44
Ad: Antidepressants.
N: Neuroleptics.
0: Other psychotropic medication.
17 started antidepressive medication after admission, while 41 continued outpatient

antidepressive treatment. Some patients used more than one type of medication.
Details are given in table 2.
3.2. The Relationship between Antidepressants and Axis I and

Axis II Disorders
Table 3 shows the percentage of Ad patients in various diagnostic groups.
As seen from the table there was no difference among groups of personality dis-
orders concerning use of antidepressants, except for a somewhat lower percentage
among patients with a cluster C disorder. The distribution was much more uneven
among patients with different axis I disorders. Among the patients with Obsessive Com-
pUlsive disorder all used antidepressants, while among those with Panic Disorder, only
30% did. The table also shows the mean maximum dosage of medication for the Ad
patients in diagnostic subgroups. Defined Daily Dosage (DDD) was used as unit. One
DDD is the equivalent of 20mg Paroxetine or 100mg Imipramine. As seen from the
table, the mean dosage was about 1.5 DDD for most disorders.
3.3. Outcome
Tables 4-6 show the admission and discharge score of global health (GAF), global
symptoms, and depressive symptoms of the SCL-90R in the subsample of patients with
mood disorders.
As seen from the tables there were no major over all differences in global health
or symptom level between the Ad and the NoAd groups neither at admission nor at
discharge. However, the NoAd-group had a significantly greater improvement than the
Ad-group concerning GSI (t = 2.54, df = 76, P = 0.013) and depressive symptoms (t =
2.91, df = 76, P = 0.005). The levels of improvement for the various groups of person-
ality disorders are shown in Figs. 2-4. There was a statistical interaction concerning
improvement in GAF, GSI, and the depression subscale of SCL-90R: Among those with
a Severe Personality Disorder (SevPD = Cluster A + B) the No Ad patients had clearly
Table 3. Number and percentage of patients given antidepressive medication, and mean
dosage in DDD for those who were given medication
Patients with
All patients mood disorder
(N = 102) (N = 85)
DDD DDD
n % Mean SD n % Mean SD
Total 58 57 1.4 0.7 58 62 1.4 0.7
Personality disorders:
Cluster A 7 54 1.6 0.8 6 60 1.6 0.8
Cluster B 20 65 1.3 0.7 17 61 1.4 0.8
Cluster C 11 42 1.3 0.6 11 58 1.3 0.6
Mixed/Atypical PD 10 66 1.4 0.8 10 77 1.4 0.8
NoPD 10 59 1.3 0.5 9 60 1.4 0.4
Axis-I disorders:
Anxiety disorders:
Panic disorder (no agoraphobia) 3 30 2.1 1.6 3 38 2.1 1.6
Panic disorder with agoraphobia 16 73 1.1 0.5 14 82 1.2 0.6
Social phobia 16 53 1.2 0.5 15 63 1.2 0.5
Obs/Comp disorder 7 100 1.6 0.5 6 100 1.5 0.5
Any mood disorder 53 62 1.4 0.7 53 62 1.4 0.7

Any substance use disorder: 13 65 1.2 0.6 12 75 1.2 0.6
Any eating disorder 11 69 1.4 0.7 10 71 1.5 0.7
Any somatoform disorder 10 45 1.5 0.7 10 50 1.5 0.7
Table 4. Mean GAF scores by types of personality disorders. Subsample of patients with
mood disorders
Patients given no antidepressants
Admission Discharge Improvement
n Mean SD n Mean SD n Mean SD
All 32 45.0 5.2 32 51.8 7.2 32 6.8 6.0
CIAPD 4 42.3 3.1 4 46.5 6.2 4 4.3 6.6
CIBPD 11 42.9 5.3 11 50.7 6.7 11 7.8 5.3
CICPD 8 47.5 5.3 8 54.4 6.8 8 6.9 5.0
OPD 3 43.3 3.2 3 49.7 9.3 3 6.3 11.0
NoPD 6 48.3 4.6 6 54.8 7.7 6 6.5 7.0
Patients given antidepressants

All 53 45.4 5.3 53 50.0 9.2 53 4.6 6.8
ClAPD 6 38.2 3.1 6 40.3 6.9 6 2.2 6.6
CIB 17 44.4 4.1 17 46.5 8.4 17 2.1 7.0
ClCPD 11 45.0 5.1 11 52.4 5.4 11 7.4 5.2
OPD 10 47.3 4.0 10 51.3 6.0 10 4.0 7.2
NoPD 9 50.4 4.4 9 58.9 10.1 9 8.4 6.3
262 S. Friis et aL
Table 5. Mean GSI scores by types of personality disorders. Subsample of patients with
mood disorders
All 32 1.67 0.62 29 0.97 0.60 29 0.68 0.70
CIAPD 4 1.97 0.95 3 0.98 1.16 3 1.35 0.45
CIBPD 11 1.78 0.70 10 1.00 0.48 10 .0.64 0.75
CICPD 8 1.68 0.43 8 1.14 0.65 8 0.55 0.76
OPD 3 1.59 0.45 3 0.77 0.67 3 0.82 1.07
NoPD 6 1.28 0.45 6 0.75 0.45 6 0.47 0.23

All 53 1.51 0.70 49 1.22 0.78 49 0.28 0.63
CIAPD 6 2.42 0.72 6 1.90 0.91 6 0.52 1.03
CIBPD 17 1.58 0.61 16 1.36 0.68 16 0.22 0.57
CICPD 11 1.63 0.48 11 1.20 0.78 11 0.42 0.56
OPD 10 1.01 0.63 9 1.02 0.68 9 -0.13 0.61
NoPD 9 1.17 0.54 7 0.56 0.53 7 0.52 0.27
Table 6. Mean score on the depression subscale of the SCL-90R by types of personality
disorders. Subsample of patients with mood disorders
All 32 2.29 0.70 29 1.32 0.85 29 0.95 1.04
CIAPD 4 2.69 0.78 3 0.97 1.37 3 2.07 1.00
CIBPD 11 2.45 0.81 10 1.26 0.76 10 1.08 1.15
CICPD 8 2.41 0.42 8 1.71 0.90 8 0.70 0.94
OPD 3 2.00 0.50 3 1.00 0.91 3 1.00 1.33
NoPD 6 1.71 0.60 5 1.18 0.74 5 0.38 0.24

All 53 2.02 0.88 49 1.67 0.97 49 0.33 0.82
CIAPD 6 2.63 0.72 6 2.14 0.91 6 0.49 1.12
CIBPD 17 2.20 0.88 16 1.86 0.96 16 0.34 0.54
CICPD 11 2.25 0.76 11 1.76 0.95 11 0.49 0.71
OPD 10 1.41 0.88 9 1.59 0.98 9 -0.32 1.12
NoPD 9 1.71 0.58 7 0.84 0.84 7 0.76 0.47
Improvement
in GAF score
9
5 • NoAd
4
DAd
3
2
Figure 2. Improvement in global health for clus-
ters of personality disorders for patients with 1
(Ad) and without (NoAd) antidepressive med-
ication. Subsample of patients with comorbid o
mood disorder. CIA CIB CIC No PO
stronger improvement than the Ad patients, while it tended to be the other way
round among the patients with No Severe Personality Disorder (NoSev = CIC PD +
No PD).
To evaluate the specific contribution of SevPD, prescribed antidepressive medi-
cation and their interaction, we made a series of analyses of variance. As dependent
variable we used the discharge score of either GAF, GSI or depressive symptoms, with
the admission score of the dependent variable as covariate. For GAF we found that
the presence of a severe PD gave a close to significant contribution to explained vari-
ance (F = 2.97; df = 1/67; P = 0.09), while the interaction between antidepressive medi-
Improvement
in GSI score
1.4
1.2
0.8
• NoAd
0.6 DAd
0.4
Figure 3. Improvement in SCL-90R global
symptom level for clusters of personality
disorders for patients with (Ad) and 0.2
without (NoAd) antidepressive medica-
tion. Subsample of patients with comorbid o
Mood disorder. CIA CIB CIC No PO
264 S. Friis et aL
Improvement
in symptom score
2.5
1.5
.NoAd
DAd
Figure 4. Improvement in SCL-90R

0.5 depressive symptom level for clusters of
personality disorders for patients with
(Ad) and without (NoAd) antidepressive
o medication. Subsample of patients with
CIA CIB CIC NoPD comorbid mood disorder.
cation and SevPD was statistically significant (F = 4.06; df = 1/67; p = 0.048). For the
global symptom level, only antidepressive medication approached a significant contri-
bution to explained variance (F = 3.02; df = 1/61; P = 0.09). For the depressive symp-
toms, the medication gave a significant contribution to explained variance (F = 4.40;
df = 1/61; P = 0.04), but more importantly the interaction between medication and
SevPD was also significant (F = 3.25; df = 1/61; p = 0.022).
We recalculated all the analyses to see whether there was any difference
between two subgroups of patients given antidepressants: Those who just continued
their medication from before admission, and those who started medication at the
day unit. We found no clear differences between these two subgroups, and the dis-
tinction gave no significant contribution to explained variance. In an additional set
of reanalyses we found that additional neuroleptic medication gave no significant
contribution to explained variance, neither alone nor in combination with
antidepressants.
4. DISCUSSION
The main findings in this study are: 1. In this setting a majority (63 %) of the
patients received some kind of pharmacotherapy. Antidepressants, mainly SSRIs, was
the most frequently used group of drugs. 2. All patients with personality disorders had
comorbid axis I disorders. 3. In the subsample of patients with a comorbid axis I mood
disorder, only 62% received antidepressants. 4. Among patients with mood disorders
those who received antidepressive medication tended to have poorer symptomatic
improvement than those who did not. 5. More important than the overall difference
between Ad and NoAd patients was the finding of a statistical interaction: The more
severe the personality disorder, the poorer the outcome among those who received
medication, compared to those who did not.
To our knowledge this is the first published study of pharmacotherapy in day unit
treatinent of personality disorders. However, Stein (1992) reviewed two studies which
reported that pharmacotherapy was given to 53% and 84-87% of inpatients with BPD.
SSRIs was the most frequently used group of drugs in the Day Unit. This is in line with
the preliminary, but promising research on these drugs. As psychotic features was an
explicit indication for neuroleptic treatment in this programme, the small number of
patients receiving this kind of medication indicates that brief psychotic episodes were
not frequent.
The fact that not every patient with an axis I mood disorder were treated with
antidepressant, may have several reasons. The often modest response to antidepres-
sants in patients with personality disorders may imply that coexistent axis I mood dis-
order is not an imperative indication for pharmacotherapy. The patients may also have
had previous experiences of poor response to antidepressants (Shea et aI., 1992). Fur-
thermore, the reported mood disorders were mainly major affective disorder, but also
comprised other disorders, like dysthymic disorder and non-specific mood disorder, for
which the indication for antidepressants is unclear in the presence of characterological
problems (Ravindran, Bialik; and Lapierre, 1994). However, there was no difference in
the distribution of various mood disorders among the Ad patients compared to the No
Ad patients.
The lack of over all difference in outcome between the Ad patients and the NoAd
patients certainly has to be interpreted with care. Firstly, patients were not randomly
assigned to medication/no medication treatment, and no standardisation of medication
was made. Most of the Ad-patients had received medication before admission. As they
had been admitted to the programme in spite of the medication, they may represent a
group of non-responders to drug therapy and possibly also to the group psychother-
apy. However, nearly all the patients in the present study had tried some sort of psy-
chiatric treatment prior to admission. The fact that they were referred to this specialised
unit may indicate that they were all hard to treat. This fact, however, does not exclude
the possibility of unrecognised differences between the groups. One possibility may be
that those who were not given medication represent a group of patients with a high
level of psychological mindedness, who wanted to find solutions to their problems
without medication. Psychological mindedness has been reported to be associated with
a favourable outcome in a group oriented day programme (Piper, Joyce, Rosie, and
Azim, 1994).
As pharmacotherapy usually is administered for acute symptoms, the patients
who received medication before admission may have experienced some treatment
response prior to admission, enough to make them able to benefit from this treatment
programme. Soloff (1994) states that for patients with BPD, pharmacological control
of various symptom clusters may make the patients more amenable to psychosocial
treatment. An important point, however, is that pharmacological studies of patients
with personality disorders have mainly been short-term, from weeks to a few months.
Today, no study has documented a beneficial effect of long-term pharmacotherapy
of personality disorders (Cornelius et aI., 1993). In our sample those who started
medication during the stay may have been a group of poor responders to treatment.
Several of them seemed to respond poorly to the group treatment alone, and additional
medication did not seem to help them substantially. They had just as poor outcome as
those who just continued medication from before admission.
As mentioned above, SSRls are reported to have effect on various symptoms
besides level of depression, like affect instability or impulsivity. In the present
266 s. Friis et al.
study, there was no registration of actual indication for pharmacotherapy and the Ad
patients may have had pharmacological responses on dimensions not assessed by our
measures.
The poorer outcome among the Ad patients was basically limited to the patients
with cluster A or cluster B personality disorders. This may be in line with the theory
that in patients with severe personality disorders, depression is of a different kind,
perhaps less biologically based. An alternative or additional hypothesis, is that patients
with personality disorders have personality traits that complicate response to treatment
(Shea et al., 1992). The various characteristics of mood disorders or complicating per-
sonality traits, may be associated with various personality disorders or clusters. Shea et
al. (1990) did not find any differences between the clusters, regarding treatment
outcome of depression, across different treatments, including antidepressants, but
ignored overlap between clusters. Sato, Sakado, and Sato (1993) reported that both the
presence of a cluster A disorder and the number of personality disorders were associ-
ated with worse short-term outcome of antidepressant therapy. In the present study,
with hierarchic cluster categories, patients with cluster A and B disorders had also the
highest number of personality disorders.
Patients with severe personality disorders are known to have poorer compliance.
We had no measures on compliance with medication. However, the Ad patients with
severe personality disorders did not have a specially high percentage of drop out, and
they had about the same percentage continuing with outpatient group psychotherapy
as the NoAd patients. Non the less it might be that some of the reason for the fairly
poor outcome for the Ad-patients with severe personality disorders might be that some
of these patients simply did not take their medication.
Our results show that patients with severe personality disorders may respond
fairly well to a eighteen weeks psychotherapy programme without medication, in spite
of considerable comorbidity with axis-I disorders. Even if preliminary findings from
other studies indicate that pharmacotherapy is useful in relieving various symptoms in
personality disordered patients, the role of such treatment is still unclear. The present
study suffers froin obvious methodological short-comings, but underlines the impor-
tance of comparing combined and single interventions for these patient groups. Hope-
fully, it will stimulate to research to clarify the relative role of different treatment
strategies, aiming at giving these patients more effective treatment.
REFERENCES
American -Psychiatric Association, 1994, Global assessment of functioning scale (GAP Scale). Diagnostic
and Statistical Manual of Mental Disorders, 4th ed., Washington D.C.
Coccaro, E.F., Astill, 1.L., Herbert, 1.L., and Scut, A.G., 1990, Fluoxetine treatment of impulsive aggression in
DSM-III-R personality disorder patients, 1. Clin. Psychopharmacol. 10:373-375.
Cornelius, 1.R, Soloff, P.H., Perel, 1.M., and Ulrich, RF., 1990, Fluoxetine trial in borderline personality dis-
order, Psychopharmacol. Bull. 26:151-154.
Cornelius, 1.R, Soloff, P.H., Perel, 1.M., and Ulrich, RF., 1993, Continuation pharmacotherapy of borderline
personality disorder with haloperidol and phenelzine, Am. 1. Psychiatry 150:1843-1848.
DeBattista, C. and Glick, I.D., 1995, Pharmacotherapy of personality disorders, Current Opinion in Psychi-
atry 8:102-105.
Derogatis, L.R, 1977, The SCL-90: administration, scoring, and procedures manual-II for the revised version,
<?linical Psychemetric Research, Baltimore.
Gitlin, M.1., 1993, Pharmacotherapy of personality disorders: conceptual framework and clinical strategies, 1.
Clin. Psychopharmacol. 13:343-353.
Goldberg, S.c., Schulz, c., Schulz, P.M., Resnick, R1., Hamer, RM., and Friedel, RO., 1986, Borderline and
schizo typal personality disorders treated with low-dose thiothixene vs. Placebo, Arch. Gen. Psychia-
try 43:680-686.
Hymowitz, P., Frances, A., Jacobsberg, L.B., Sickles, M., and Hoyt, R, 1986, Neuroleptic treatment of schizo-
typal personality disorders, Compr. Psychiatry 27:267-271.
Karterud, S., Friis, S., Irion, T., Mehlum, L., Vaglum, P., and Vaglum, S., 1995, A SCL-90-R derived index of
the severity of personality disorders, 1. Pers. Dis. 9:112-123.
Karterud, S., Pedersen, G., Friis, S., Urnes, 0., Irion, T., Brabrand, 1., Falkum, L.R, and Leirvag, H., 1998, The
Norwegian Network of psychotherapeutic day hospitals. Ther. Comm. 19:15-28.
Karterud, S., Vaglum, S., Friis, S., Irion, T., Johns, S., and Vaglum, P., 1992, Day hospital therapeutic commu-
nity treatment for patients with personality disorders. An empirical evaluation of the containment
function, 1. Nerv. Ment. Dis. 180:238-243.
Kavoussi, R1., Liu, 1., and Coccaro, E.F., 1994, An open trial of sertraline in personality disordered patients
with impulsive aggression, 1. Clin. Psychiatry 55:137-141.
Markovitz, P.1., Calabrese, 1.R, Schulz, S.c., and Meltzer, H.Y., 1991, Fluoxetine in the treatment of border-
line and schizotypal personality disorders, Am. 1. Psychiatry 148:1064-1067.
Mehlum, L., Friis, S., Irion, T., Johns, S., Karterud, S., Vaglum, P., and Vaglum, S., 1991, Personality disorders
2-5 years after treatment: a prospective follow-up study, Acta Psychiatr. Scand. 84:72-77.
Mehlum, L., Friis, S., Vaglum, P., and Karterud, S., 1994, The longitudinal pattern of suicidal behaviour in
borderline personality disorder: a prospective follow-up study, Acta Psychiatr. Scand. 90:124-130.
Montgomery, S.A. and Montgomery, D., 1982, Pharmacological prevention of suicidal behavior, 1. Affect.
Disord.4:291-298.
Montgomery, S. and Asberg, M., 1979, A new depression scale designed to be sensitive to change, Br. 1.
Nordern, M.1., 1989, Fluoxetine in borderline personality disorder, Prog. Neuro-Psychopharmacol. BioI.
Piper, W.E., Joyce, A.S., Rosie, 1.S., and Azim, H.F.A., 1994, Psychological mindedness, work and outcome in
day treatment, Int. 1. Group Psychotherapy 44:291-311.
Ravindran, A.V., Bialik, R1., and Lapierre, Y.D., 1994, Therapeutic efficacy of specific serotonin reuptake
inhibitors (SSRI) in dysthymia, Can. 1. Psychiatry 39:21-26.
Reich, 1.H. and Vasile, RG., 1993, Effect of personality disorders on the treatment outcome of axis I condi-
tions: an update, 1. Nerv. Ment. Dis. 181:475-484.
Rosenberg, R, 1994, Borderline states: pharmacotherapy and psychobiology of personality. A discussion of
Soloffs article, Acta Psychiatr. Scand. 89(Suppl. 379):56-60.
Salzman, c., Wolfson, A.N., Schatsberg, A., Looper, 1., Henke, R, Albanese, M., Schwartz, 1., and Miyawaki,
E., 1995, Effect of fluoxetine on anger in symptomatic volunteers with borderline personality disor-
der,1. Clin. Pharmacol. 15:23-29.
Sato, T., Sakado, K., and Sato, S., 1993, Is there any specific personality disorder or personality disorder cluster
that worsens the short-term treatment outcome of major depression? Acta Psychiatr. Scand.
88:342-349.
Serban, G. and Siegel, S., 1984, Response of borderline and schizotypal patients to small doses of thiothix-
ene and haloperidol, Am. 1. Psychiatry 141:1455-1458.
Shea, M.T., Pilkonis, P., Beckham, E., Collins, 1.F., Elkin, I., Scotsky, S.M., and Docherty, 1.P., 1990, Personal-
ity disorders and treatment outcome in the NIMH treatment of depression collaborative program,
Am. 1. Psychiatry 147:711-718.
Shea, M.T., Widiger, T.A., and Klein, M.H., 1992, Comorbidity of personality disorders and depression: impli-
cations for treatment, 1. Consult. Clin. Psychol. 60:857-868.
Soloff, P.H., 1994, Is there any drug treatment of choice for the borderline patient? Acta Psychiatr. Scand.
89(Suppl. 379):50-55.
Soloff, P.H., Cornelius, 1., George, A., Nathan, S., Perel, 1.M., and Ulrich, RF., 1993, Efficacy of phenelzine
and haloperidol in borderline personality disorder, Arch. Gen. Psychiatry 50:377-385.
Soloff, P.H., George, A., Nathan, S., Schulz, P.M., Cornelius, 1.R, Herring, 1., and Perel, 1.M., 1989, Amitipty-
line versus haloperidol in borderliners: final outcomes and predictors of response, 1. Clin. Psy-
chopharmacol. 9:238-246.
Stein, G., 1992, Drug treatment of the personality disorders, Br. 1. Psychiatry 161:167-184.
268 s. Friis et al.
'fuinier, S. and Verhoeven, W.M.A., 1995, Dimensional classification and behavioral pharmacology of per-
sonality disorders; a review and hypothesis, European Neuropsychopharmacol. 5:135-146.
Vaglum, P., Friis, S., Irion, T., Johns, S., Karterud, S., Larsen, F., and Vaglum, S., 1990, Treatment response of
severe and non severe personality disorders in a therapeutic community day unit, J. Pers. Dis.
4:161-172.
Vaglum, P., Friis, S., Vaglum, S., and Larsen, F., 1989, A comparison between personality disorder diagnoses
in DSM-III and DSM-IIIR: Reliability, diagnostic overlap, predictive validity, Psychopathol.
22:309-314. .
Wilberg, T., Karterud, S., Urnes, 0., Pedersen, G., and Friis, S., 1998, Outcomes of Poorly Functioning Patients
With Personality Disorders in a Day Treatment Program, Psychiatr. Servo 49:1462-1467.
19
A NEW INTERPERSONAL THEORY AND

THE TREATMENT OF DEPENDENT
PERSONALITY DISORDER
John Birtchnell* and Giuseppe Borgherini**
* Institute of Psychiatry
de Crespigny Park
London SE5 8AF, United Kingdom
** Department of Neurological and Psychiatric Sciences
University of Padova
Via Guistiniani 2,35128
Padova, Italy
1. INTRODUCTION
The method of treatment of dependent personality disorder that will be described

in this chapter requires an understanding of the setting of this disorder within the entire
range of personality disorders. Such an understanding is provided by a classificatory
system, called the interpersonal octagon (Fig. 2) within which the ten DSM-IV, Axis II
(American Psychiatric Association, 1994) personality disorders have been accommo-
dated (Birtchnell, 1997a). The system is based upon a new interpersonal theory that
has been developed by the senior author (Birtchnell, 1993/6). Within it, personality dis-
orders are categorised according to the individual's relating characteristics. The therapy,
based upon the theory, involves minimising what will be called negative relating char-
acteristics and maximising what will be called positive ones. The chapter will be divided
into five parts. In the first, the version of interpersonal theory to be used will be
described; in the second, dependent personality disorder will be defined within the
context of this theory; in the third, the therapy of dependent personality disorder, based
upon this definition, will be explained; in the fourth, a method of measuring change in
therapy, based upon the theory, will be introduced and, by this method, the respon-
siveness of dependent patients to therapy will be demonstrated; and in the fifth, the
general themes of the chapter will be brought together. Although it is acknowledged
270 J. Birtcbnell and G. Borgberini
that dependent personality disorder is more prevalent in women, for ease of reading
only, the male gender will be used throughout the chapter.
2. THE ORIGINAL AND THE NEW INTERPERSONAL THEORY

The first version of the original interpersonal theory was published by Freedman,
Leary, Ossorio, and Coffee in the Journal of Personality in 1951, but throughout the
fifties and sixties, a number of variations of it appeared in the literature (Wiggins, 1982).
It's essential feature is the proposition that relating takes place along two orthogonal
axes: a horizontal one, which represents a dichotomy such as warm versus cold, loving
versus hating, or friendly versus hostile, and a vertical one which represents a
dichotomy such as active versus passive, powerful versus weak, or dominating versus
sUbmitting. Between the resulting, four polar positions, are inserted intermediate posi-
tions, which represent a blending of the characteristics of the horizontal and vertical
positions to either side of them. The entire system, comprising eight segments (with
one intermediate position per quadrant) or sixteen segments (with three intermediate
positions), is called the interpersonal circle.
The version of interpersonal theory that forms the basis of this chapter was devel-
oped independently of this original interpersonal tradition. Interestingly, from the point
of view of this chapter, it grew out of a preoccupation with the nature of psychologi-
cal dependence. The proposal was made (Birtchnell, 1987), which will be central to the
arguments developed, that dependence is compounded of the two motives of seeking
closeness to another person and needing to be the recipient of something that is offered
by that other person. The closeness-seeking motive was called attachment and the need-
fulness motive was called receptiveness. From this, it was but a small step to proposing
a biaxial theoretical system that had attachment versus detachment as the horizontal
axis and directiveness versus receptiveness as the vertical axis. It soon became appar-
ent that the naming of these axes had to be changed. Attachment, as Bowlby (1977)
defined it, like dependence, incorporates both a proximity-seeking and a receptive com-
ponent. Therefore, closeness versus distance was selected as a purer distinction for the
horizontal axis. Directiveness already had been defined by Ray (1976) as "the desire
or tendency to impose one's will on others." (page 314). This was considered too restric-
tive a definition for what ought to be something like relating from a position of rela-
tive strength. Similarly, the term "the receptive orientation" already had been used by
Fromm (1947) to mean the tendency to receive what one wants from an outside source.
This was considered too restrictive a term for what ought to be something like relat-
ing from a position of relative weakness. Therefore, upperness versus lowerness was
selected as a purer, and more generic distinction for the vertical axis.
Once the new terminology was in place, a more comprehensive theoretical system
was developed (Birtchnell, 1990). In the pages that follow, only the briefest account of
this system can be presented. For a more complete account see Birtchnell, 1993/1996.
As with the original interpersonal system, four intermediate positions were selected,
which were called upper closeness, lower closeness, upper distance, and lower distance.
The entire system, in order to distinguish it from the interpersonal circle, was called the
interpersonal octagon (Birtchnell, 1994) (see Fig. 1). Each of the eight octants was con-
sidered to include a range of related, interpersonal characteristics. For closeness they
were: interacting, communicating, sharing, cooperating, self-disclosing, and showing an
interest in others. For distance they were: separating, setting up personal boundaries,
maintaining privacy, establishing a personal identity, being restrained, and being origi-
Treatment of Dependent Personality Disorder 271
NO NC
LN
UN
Figure 1. The interpersonal

octagon: The upper diagram, the
positive version, indicating for
each octant, positive forms of
relating; the lower diagram, the
negative version, indicting for NO NC
each octant, negative forms of
relating. C, D, U, L, and N stand
for close, distant, upper, lower,
and neutral. From: Birtchnell, J.
"The interpersonal octagon: An
alternative to the interpersonal
circle." Human Relations. 47,
pages 518 and 524. Copyright
The Tavistock Institute, 1994.
Reprinted with the permission
of the publisher. LN
nal. Upperness was considered to range from upper distance, through an arbitrary
neutral position called upper neutral, to upper closeness; and lowerness was considered
to range from lower distance, through an arbitrary neutral position called lower neutral,
to lower closeness. The more distant forms of upperness included such characteristics
as management, control and judgement, and the closer forms, such characteristics as
caring for, supporting and protecting. The more distant forms of lowerness included
such characteristics as respect, loyalty and obedience, and the closer forms, such char-
acteristics as being cared for, supported and protected.
272 J. Birtchnell and G. Borgherini
2.1. An Evolutionary Contribution to Interpersonal Theory

Of the four authors of the first paper on interpersonal theory, Leary (1957), in his
seminal book "Interpersonal Diagnosis of Personality," carried the theory furthest
forward, but an important difference between the Leary version of the theory and
the present version concerns the assumed motivation for interpersonal behavior;
Influenced by Horney (1937) and Sullivan (1953), Leary considered interpersonal
behavior to be directed toward reducing anxiety. In contrast, in the present version, it
is considered to be directed toward attaining and maintaining what will be called relat-
ing objectives, these objectives being closeness, distance, upperness, and lowe mess.
Gilbert (1989) observed that humans come into the world in a state of pre-
paredness, that is, they have certain inborn dispositions, namely their relating objec-
tives. Such objectives, have evolved in the same way that physical attributes have
evolved. They have evolved because they are advantageous to us. If it is assumed that
we are born with the innate tendencies of closeness, distance, upperness, and lowerness,
then what we must do, during the course of psychological maturation, is develop
interpersonal skills and competencies, that enable us to relate effectively in these four
directions.
2.2. The Distinction between Positive and Negative Relating

Most versions of interpersonal theory draw a distinction between what is some-
times called adaptive and maladaptive relating. These are not good terms because, in
an evolutionary sense, all relating is adaptive. However, it is generally understood that
some forms of relating are more indicative of good adjustment than others. What then
is the difference between relating that is so indicative and relating that is not? Leary
(1957), again under the influence of Sullivan (1953), who believed in a continuum from
normality to abnormality, considered maladaptive relating to be simply an extreme
form of adaptive relating. This is clearly not the case, for it is possible to be extremely
adaptive without becoming maladaptive. Take dependence for example, a patient on a
life support machine is extremely dependent, yet his relating to either the machine or
the doctors and nurses is not maladaptive.
In the present theory, the terms positive and negative are used to distinguish
between relating that is and is not indicative of good adjustment. Positive relating is
the relating of someone who has adequate interpersonal skills and competencies and
negative relating is the relating of someone who has not. The terms positive and nega-
tive are used here categorically; and there is no implication of a range from one to the
other. A totally well adjusted person is someone who relates positively, i.e. with com-
petence, in all four of the main areas of relating. As and when it is required, such a
person can be close, distant, upper or lower. Such a person is described as versatile. It
is not possible to locate a versatile person (i.e. someone who does not have a person-
ality disorder) in anyone place within the octagon.
Negative relating is the relating of someone who lacks certain interpersonal skills.
A person may relate negatively in one or more of the eight octants of the interpersonal
octagon (see Fig. 1). A number of forms of negative relating has been proposed and
these will briefly be described. A person who lacks the skills in one zone of the octagon
may be inclined to avoid (be fearful of) that zone and may operate predominantly in
the opposite zone. Thus, avoidance of a zone is one form of negative relating. A person
who cannot be close operates predominantly towards the distant end of the horizon-
tal axis; a person who cannot be distant operates predominantly towards the close end;
a person who cannot be upper operates predominantly towards the lower end of the
vertical axis; and a person who cannot be lower operates predominantly towards the
upper end. Where a universally positive relater, i.e. a versatile relater, is able to move
freely around the octagon, a negative relater has a more restricted range of relating,
and operates only in those zones in which he is competent. It is for this reason that per-
sonality disorders can be located within particular zones of the octagon (Fig. 2).
A person who remains predominantly in one zone is afraid of being displaced
from that particular zone. This becomes manifest as holding on insecurely to that zone.
Thus, insecurity is a second form of negative relating. An insecurely close person clings
excessively to people; an insecurely distant person is afraid of being encroached upon
or invaded by people, and draws back and creates barriers against them; an insecurely
upper person is afraid of losing her/his position of influence, and acts boastfully and
bombastically and pushes people down; and an insecurely lower person is afraid of the
exploitation, abuse or abandonment by those upon whom s/he relies, and seeks reas-
surance from them and is careful not to offend them.
A person who is not confident of his competence in a particular position is likely
to resort to desperate or unscrupulous means of attaining it. This involves showing little
concern for the effect that his relating behavior is having upon the person with whom
he relates. Thus desperation is a third form of negative relating. A desperately close
person will force himself upon others or adopt measures that keep others close to him;
a desperately distant person will withdraw from, shun and reject others; a desperately
upper person will insult, intimidate, suppress, and dominate others; and a desperately
lower person will feign illness, weakness or helplessness and plead and beg others to
help and care for him. Figure 1 provides examples of positive and negatives forms of
each of the octants of the interpersonal octagon.
UN
UO UC
oOJ
(f)
Figure 2. The ten disorders of the m

(f) (f)
DSM-IV, Axis II accommodated ~ ~ (f)
within the interpersonal octagon. -

N < 0
NO m :r:
C, D, U, L, and N stand for close, ~
-< 0 N NC
00
distant, upper, lower, and neutral.
From Birtchnell, 1. "Personality set
;t; ~ 0
r C
within an octagonal model of relat- r
(f)
ing." In Plutchik, R. and Conte, H.R. <:
m
(Eds.) Circumples Models of Per-
sonality and Emotions. American
Psychological Association Press:
LO
Washington, D.C. Copyright 1997
American Psychological Associa-
tion. Reprinted with the permission
of the publisher. LN
2.3. Transient and Enduring Forms of Negative Relating

Everyone relates negatively at certain times and under certain circumstances.
Negative relating can also be a response to the behavior of others. When others
threaten to leave us, we cling tighter; when they threaten to encroach upon us, we push
them away; when they challenge our authority, we put them down; and when they
threaten to abandon us, we beg them not to. Negative relating can also be induced by
certain affective states. The relationship between pathological dependence and depres-
sion is a case in point (Birtchnell, 1984). A number of studies have shown that people
are more pathologically dependent when they are depressed (Hirschfeld, Klerman,
Clayton, Keller, McDonald-Scott, and Larkin, 1983), and that this dependence subsides
when the depression is lifted (Hirschfeld, Klerman, Lavori, Keller, Griffiths, and
Coryell, 1989; Birtchnell, Deahl, and Falkowski, 1991). The pathological dependence
that accompanies depression is called state dependence, to distinguish it from trait
dependence, which refers to the more enduring, dependent personality disorder. The
treatment of state dependence (Birtchnell, 1996a) is much easier and much quicker
than the treatment of trait dependence.
If negative relating is not simply a consequence of an affective state, if it occurs
under most circumstances, in relation to most people, and is persistent over time, it is
classifiable as a personality disorder. If the interpersonal octagon purports to be a com-
prehensive system, it should be possible to locate the ten DSM-IV, Axis II, personality
disorders within its eight octants (Birtchnell, 1997a). On the basis of the published
descriptions, this was found to be the case, and their locations are presented in Fig. 2.
It will be seen that (1) only four disorders can be located in a single octant, (2) each
of the remaining six disorders straddle at least three octants, (3) some disorders share
the same location within the octagon, and (4) there is a preponderance of disorders on
the distant side of the octagon. It was found that those disorders that share the same
location can be differentiated by their intrapersonal, rather than their interpersonal
characteristics. An account of such differentiation for the distant disorders is provided
in Birtchnell (1996b). Since the DSM-IV personality disorders are simply listed, without
any indication of their possible interconnectedness, a time may come when the octag-
onal descriptions of negative relating characteristics will be seen as providing a more
rational basis for their classification.
It is pertinent to enquire whether there might be a connection between the
present theory and the widely adopted five factor model of personality (Costa and
McCrae, 1986). Wiggins and Pincus (1989) have considered it likely that personality
disorders would involve maladaptive and inflexible expressions of some or all of the
model's five dimensions. However, of these five dimensions, extraversion, neuroticism,
openness to novel experience, agreeableness, and conscientiousness, only the first is
interpersonal. It would seem then that the overlap between the present theory and the
five factor model is minimal. Whilst Trapnel and Wiggins (1990) have produced a
measure which combines their own interpersonal dimensions and the Big Five dimen-
sions (the IASR-B5), they have also found it necessary to retain their separate inter-
personal measure (the IASR).
2.4. Interpersonal Psychotherapy

Interpersonal psychotherapy was introduced by Sullivan (1954). In contrast to the
psychoanalytic tradition of his time, he laid great stress upon the process of inter-
personal learning. Instead of the detached, silent stance of psychoanalysis, he actively

shared impressions with the client. By subtly challenging the client's expectational set,
he opened the possibility of the client's reappraisal of interpersonal situations and the
development of more effective interpersonal behavior. Sullivan died in 1949, two years
before the publication of the first version of interpersonal theory, and so was not aware
of the biaxial classification of relating behavior. Those who followed after him applied
his methods within the framework of the interpersonal circle (Anchin and Kiesler,
1987). Many drew upon the theory of complementarity, first introduced by Leary (1957)
and later developed by Carson (1969), which states that each form of relating evokes
a complementary response, which reinforces the original mode of relating; for example,
dominance evokes submissiveness, which in turn, reinforces the original dominance. By
providing an anticomplementary response the therapist aims to extinguish the original
action (see Kiesler, 1986). Orford (1986) demonstrated that the rules of complemen-
tarity do not always hold, and although the approach has some validity, it might be
considered rather simplistic and essentially behavioral. Benjamin (1984) developed the
more flexible approach of adopting, at the outset of therapy, a position that is comple-
mentary to the patient, and then moving toward other desired parts of interpersonal
space in small, progressive steps. Another school of interpersonal psychotherapy
(Klerman, Weissman, and Rounsaville, 1984), is limited in that it is directed specifically
at the resolution of depression. It bypasses the approaches based upon the inter-
personal circle and draws inspiration directly from Sullivan's writing. It pays little
attention to the effects of past experiences or unseen, intrapsychic processes and con-
centrates upon actual experiences in the here and now. It is a highly pragmatic approach
and aims to help clients within four specific problem areas: grief, interpersonal deficits,
interpersonal disputes, and role transitions.
The form of interpersonal psychotherapy adopted in this chapter is in line with
Sullivan's original approach, but it is set within the framework of the interpersonal
octagon. Unlike many of the more recent interpersonal developments, it does seek to
relate present relating deficits to early experiences, but it does not draw upon the theory
of complementarity. Benjamin (1987) laid down the following useful principles: (1)
create and maintain a working, collaborative relationship with the patient; (2) help the
patient recognise his interactive patterns; (3) help the patient discover how he learned
the patterns, including underlying reinforcements; (4) help the patient decide if he
wants to give up those old reinforcements and learn new ways of interacting; and (5)
provide support as he feels the stress of major change.
The present approach places a strong emphasis upon distinguishing between
positive and negative relating, a distinction largely ignored by the anticomplementarity
school. Since positive relating is indicative of good adjustment and negative relating is
indicative of bad adjustment, its overall objective is to maximise positive relating and
minimise negative relating. Since negative relating is largely explicable in terms of lack
of interpersonal skills, the primary task of the therapy is to improve interpersonal skills
and increase interpersonal competence. It is also concerned with the fact that clients
may be induced to relate negatively by the negative relating of others toward them,
both in the past and in the present. An important feature of therapy therefore, is to
make clients more aware of the effect that certain others have had, and are continuing
to have, upon them and to help them avoid slipping into complementary responses to
the negative relating of others. Whilst this bears some resemblance to Kiesler's (1986)
anticomplementarity approach, it is directed more toward the client's relating outside
of the therapy session. Of course, during the session itself, clients relate negatively to
the therapist, and it is important to point this out to them. It is also important for the
therapist not to reinforce clients' negative relating by responding in a complementary
manner to it; but simply adopting an anticomplementary response to it is not sufficient.
3. DEPENDENT PERSONALITY DISORDER
The thinking of the first author (Birtchnell) about the nature of dependence has
changed over recent years. Since during some of the stages of this thinking, the con-
cepts of closeness and lowerness were not clearly formulated, they have been added in
brackets. In the first account of this thinking (Birtchnell, 1984) it was proposed that
there are three types of dependence called affectional (the need for closeness), onto-
logical (a poorly defined, separate identity), and deferential (lowerness). In the second
account (Birtchnell, 1988) these were expanded to five: failure to separate (closeness),
failure to establish a separate identity (closeness), lack of competence (lowerness) lack
of self-worth (lowerness), and failure to feel deserving of the status of adult (lower-
ness). In the third account (Birtchnell, 1991b), they had been expanded to nine: close-
ness seeking, poor identity (closeness), easily influenced, suggestible (lower closeness),
needing to seek advice (lowerness), care eliciting (lower closeness), inclined to look up
to others (lowerness), approval seeking (lower closeness), fear of rejection (lower close-
ness), and self-judging (lowerness). With the formulating of the new interpersonal
theory (Birtchnell, 1993/96) the definition of dependence became greatly simplified,
because it could be condensed into various subdivisions of closeness and lowerness.
3.1. The Two Components of Dependence

As far back as 1987 the idea was beginning to form that dependence has two
major components, which are now called closeness and lowerness. Around 1990 it was
concluded that Bowlby's (1969) concept of attachment is also made up of the same two
components, and recently Birtchnell (1997b) proposed that dependence is the adult
equivalent of what in childhood is called attachment. In Fig. 2 the dependent person-
ality disorder, as defined within the DSM-IV Axis II, was shown to span the three
octants neutral close, lower close, and lower neutral. The official descriptions of depen-
dent personality disorder in the DSM-IV (American Psychiatric Association, 1994), the
leD 10 (World Health Association, 1989), and certain other published descriptions
(Livesley, Schroeder, and Jackson, 1990; Pilkonis, 1988), can be accommodated
extremely well within these three categories of relating (See Fig. 3). It seems that the
purest form of dependence is located within the lower close octant and that less pure
forms spread outwards into the neighbouring lower neutral and neutral close octants.
3.2. Positive and Negative Dependence

Around 1991, the distinction was first made between what was then called normal
(positive) and pathological (negative) dependence. This greatly clarified the definition.
Although dependence is commonly considered normal in childhood (Parens and Saul,
1971), there can be pathologically (negatively) dependent children (Berg, 1974). There-
fore it is not just a matter of maturity. It was important to appreciate that sometimes
dependence can be quite normal (positive) in adulthood (Birtchnell, 1991a). The person
on a life support machine has already been mentioned. Similar examples are the pilot
CLOSE
Experiences devastation and helplessness when a close relationship ends (lCD 10)
Preoccupied with the fear of being abandoned and needs constant reassurance against this
(lCD 10)
Feeling very uncomfortable when alone (lCD 10)
Insecure for fear of losing an important relationship or person (pilkonis, 1988)
Secure base effect, proximity seeking, need for affection, feared loss, separation protest
(Livesley et aI., 1990)
LOWER CLOSE
Is unrealistically preoccupied with fears of being left to take care of herlhimself (DSM-
IV)
Goes to excessive lengths to obtain nurturance and support from others to the point of
volunteering to do things that are unpleasant (DSM-IV)
Feels uncomfortable or helpless when alone because of exaggerated fears of being unable
to care for herlhimself (DSM-IV)
Has difficulty expressing disagreement with others because of fear of loss of support or
approval (DSM-IV)
Need for care and support, need for approval (Livesley et aI., 1990)
LOWER
Encourages or allows others to assume responsibility for major areas in herlhis life (lCD
10)
Tends to react to adversity by transferring responsibility to others (lCD 10)
Subordinates herlhis own needs to those of others on whom slhe is dependent, and unduly
complies with their wishes (lCD 10)
Has difficulty making everyday decisions without an excessive amount of advice and
reassurance from others (DSM-IV)
Gives up control to others and underestimates abilities and resources for coping (pilkonis,
1988)
Figure 3. Examples from the ICD 10, the DSM-IV, Livesley et aI., 1990 and Pilkonis, 1988, of close, lower
close and lower forms of dependence.
who is dependent upon radar and the blind person who is dependent upon a guide dog.
In terms of interpersonal relating, a trainee is dependent upon his teacher (positive
lowerness) and people who are deeply in love are dependent upon each other (posi-
tive closeness). In fact, a person can have a pathological inability to be dependent, that
is, he can be fearfully avoidant of closeness and/or lowerness.
Central to what might be called positive dependence is the issue of trust. In fact,
without trust, positive dependence is not possible. Trust is a preparedness to place
oneself in a position of vulnerability in relation to another. Trust applies both to close-
ness and to lowerness. In terms of closeness, lovers trust each other not to break away,
knowing that, if either does, it will hurt. In terms of lowerness, the trainee trusts the
teacher not to mislead him, knowing that if the teacher does, he will become incom-
petent. The pathologically (negatively) dependent child does not trust his mother to
return when she goes away, or to act responsibly towards him, which brings Berg's
278 J. BirtchneU and G. Borgherini
(1974) pathological dependence in childhood close to what Bowlby (1973) called

anxious attachment. It follows that an important feature of dependent personality dis-
order is lack of trust, and part of the therapy of such a disorder is the building up of
this trust.
3.3. Ontological Dependence

Ontological dependence is not normally listed among the characteristics of the
dependent personality (Fig. 3), but there is no doubt that it is an important feature of
it. The term was first coined by Laing (1965). It is a consequence of the avoidance of
distance. Distance is not only the capacity to be alone (Winnicott, 1958), it is also the
capacity to create a secure and separate identity. The ontologically dependent individ-
ual has no clearly defined sense of self. He either avoids people for fear of what Laing
called engulfment, or becomes so involved with people that he becomes parasitic upon
them, i.e. assuming their identity.
3.4. The Self-Centeredness of the Dependent Individual

Perhaps as a consequence of the lack of a secure identity, the dependent individ-
ual sucks up closeness from others, but is incapable of giving closeness to others. Thus
he wants others to be interested in him and concerned about him but he takes little
interest in and shows little concern for others. This corresponds with what Fromm
(1947) called the receptive orientation (described earlier in the chapter). It is an
example of that form of negative relating called desperation.
3.5. Explaining the Apparent Overlap between Dependent and

Borderline Personality Disorder
In Fig. 2 it will be seen that the dependent and borderline personality disorders
share the same three octants. How then do they differ? When Pilkonis (1988) carried
out a cluster analysis of items relating to dependence, he extracted two higher order
constructs which he called "excessive dependence" and "borderline features." The latter
construct included the items: reacts to real or perceived deprivation with anger or rage,
and alternates rapidly between intense positive and negative emotions. The DSM-IV
criteria for borderline disorder include such items as impulsivity, marked reactivity of
mood, frequent displays of temper, alternating between extremes of idealisation and
devaluation. Borderline personality disorder is more difficult to treat than dependent
personality disorder because, in addition to the basic close, lower close and lower char-
acteristics, there are also these tendencies to impulsivity and mood swings. These are
intrapersonal rather than interpersonal, which is why they do not affect the location of
borderline disorder within the octagon. They are likely to be genetically rather than
environmentally determined and to seriously impair response to psychotherapy.
3.6. Is There a Genetic Input to Dependent Personality Disorder

Livesley (1996) has argued that genetically determined symptoms are resistant to
psychotherapy. If this is true, it is important to know the level of genetic input of the symp-
toms of dependent personality disorder. Livesley, Jang, Jackson, and Vernon (1993) con-
cluded from a popUlation study of monozygotic and dizygotic twins that the heritability
of the closeness and lowerness components of the condition are 36% and 25% respec-
tively, which, compared with other DSM-IV personality disorders, is relatively low.
4. PSYCHOTHERAPY OF DEPENDENT PERSONALITY

DISORDER
The psychotherapy of any personality disorder is a long and difficult exercise. This
is because: (1) Patterns of relating that have become established and repeatedly prac-
tised over many years require therapy extending over a number of years to correct.
This is the more so because, outside of therapy, clients are involved in relationships,
over which therapists have no direct control, in which these established relating pat-
terns continue to be reworked and reinforced. (2) People develop internal representa-
tions of past interpersonal experiences. Important others from their past become
internalised. They also have internal representations of themselves and, in their inner
world, the internal representations of past others continue to relate to the internal rep-
resentations of themselves and their internal selves continue to relate to internalised
past others (Jacobson, 1964). It is useful to imagine that people have what might be
called a generalised, internal, close other, which is an amalgam of various, past, impor-
tant, close others; and a generalised, internal, upper other, which is an amalgam of
various, past, important, upper others. When considering the dependent personality dis-
order, the close other keeps them bound within a kind of symbiotic capsule, from which
they cannot extricate themselves; and the upper other keeps them subdued within a
kind of obedient, submissive, abased, self-blaming, and apologetic state, out of which
they cannot lift themselves. Their current involvements with others become a re-
enactment of these internalised relationships. Over time, the therapist may becomes a
part of this internal world and able to influence the nature of the internally close and
the internally upper other.
People tend to seek therapy not because of their personality disorder but because
of the interpersonal difficulties that result from such disorder. This is particularly so of
dependent personality disorder. Millon (1981) graphically described how the demands
that dependent clients make upon the important others in their lives frequently lead,
over time, to the rejection (loss of closeness or loss of lowerness) that their behavior
(imperfectly as it turns out) is directed toward preventing, and that this frequently pre-
cipitates them into severe depressions, which may need treatment in their own right.
Most often, therefore, it is these depressions that bring dependent clients into treat-
ment. However, the depressions cannot adequately be treated unless the underlying
relationship failure has been addressed.
4.1. Interactions of Client and Therapist

Whilst it is possible to treat dependent personality disorder by either individual
or group therapy, the focus in this chapter will be on individual therapy. The general
principles however would be the same for group therapy. Considering first the close-
ness component of dependence, generally, dependent clients find therapy pleasurable.
They speak freely and enjoy being listened to. They speak openly about themselves
and, if they are brave enough, they ask therapists about themselves. Initially therapists
find such willingness to enter into, and openly indulge in, therapy gratifying, but soon
they may find it intrusive. Any effort therapists may make to create space for them-
selves, or erect barriers between themselves and their clients, are likely to be met with
by further efforts by the clients to get closer. Clients may try to prolong sessions by
introducing new and interesting material toward the end and even have difficulty
leaving. They may try to make contact between sessions or ask for additional sessions.
In order to create a stable working environment, throughout therapy, therapists should
remain constantly and unvaryingly, reasonably close, and clients should be told that the
length and frequency of sessions will remain the same, whatever the clients may say or
do.
Considering now the lowerness component, clients will also find it pleasurable
to be in the presence of someone they consider to be of high status. At the distant end
of lowerness, they will act respectfully and obediently and at the close end they will
be flattering and admiring. Some therapists may find this a source of narcissistic
gratification, but in time, as with the clients' excessive closeness, they may begin to find
it irritating. One of the biggest obstacles to therapeutic progress with dependent clients
is their passivity. They are inclined to offer themselves up to therapists to be worked
on, or to expect therapists to tell them what they have to do to get better. They rapidly
learn to say and do the things that please their therapist and repeat their therapists'
words back at them. This can create the impression that they are improving, but what
may really be happening is that they are pretending to be the kind of person that they
believe their therapists would like them to be.
By their very nature, many therapists, particularly medically qualified ones, derive
pleasure from caring for weak, ill and helpless people, that is, they have a markedly
upper close tendency. Since people with a dependent personality have a markedly lower
close tendency, therapists, and clients readily complement each other. This can be a bar
to therapeutic progress because therapist and client can become locked in a mutually
satisfying interaction. Dependent clients learn that the weaker, iller or more helpless
they appear to be, the more responsive their therapists are to them. Therefore they are
tempted to feign weakness, illness and helplessness. Therapists need to be aware of this
tendency in themselves. From an early stage, they must convey to their clients that they
will get neither more nor less attention if they adopt such care eliciting behavior.
In a similar vein, clients learn how to alarm their therapists by creating crises.
Therefore they are tempted to damage themselves, gamble or drink excessively and
accumulate debts. The most alarming and effective crisis is a suicide bid. Therapists,
particular medical therapists, and particularly in the U.S.A., are aware of the danger of
legal reprisal should their client die. Therefore they are inclined to respond to suicide
bids by imposing restraints upon the client which only serves to reinforce the depen-
dent position. Schwartz, Flinn, and Slawson (1974), writing at a time when the legal
pressures were less stringent, warned against the dangers of allowing dependent clients
to gain such control over their therapists.
4.2. Tapping into Clients' Anxieties

People with personality disorders hold on nervously to certain states of related-
ness and do not dare venture toward the opposite states of relatedness. Therefore they
are afraid of two things: losing the state or states of relatedness that they have, or are
locked into, and venturing toward the opposite state or states of relatedness, the one
or ones they have always avoided. The first set of fears has to be tackled first, because,
until clients feel secure in their present states of relatedness, they will not have the
confidence to venture toward the opposite states. Thus, firstly, dependent personalities
anxiously hold on to, and are afraid of losing both their closeness and their lowerness,
and secondly, they are afraid of moving toward both distance and upperness.
Their fear of losing closeness manifests itself in their clinging behavior toward
both their therapist and the significant others in their lives. They are afraid that people
do not want to stay close to them and that when people go away from them they will
not come back. They are nervous when they are alone and try combat their isolation
by playing music, listening to the radio, watching television, and telephoning people.
They try to make themselves attractive to others and frequently say to them, "You will
never leave me will you?" They are frightened when those who are close to them show
an interest in others, or are shown an interest in by others. They try to dissuade those
who are close to them from having any involvement with others, and to control their
comings and goings. This makes others feel tied down or shut in, and when inevitably,
these others try to create some space for themselves, they respond by being even more
clinging and restricting. They are frightened of separation which is why they are reluc-
tant to terminate the session and to leave the room. This is an area where there is a
clear overlap between dependent and borderline personality disorder.
Their fear of losing lowerness manifests itself in (1) an attitude of helplessness
and needfulness and (2) an excessive deference and respectfulness toward both the
therapist and the significant others in their lives. They fear that those who assume
responsibility for them, do things for them, tell them what to do, protect them, or care
for them will either desert them, or exploit or abuse them. They feel that they have no
negotiating power. They tolerate insult or abuse from those upper to them because
they fear that, if they object, they may lose whatever lowerness they have. They are
excessively grateful to upper people for what they do for them. They frequently
ask, "Do you approve of me? Have I offended you?" They are excessively apologetic
for any small indiscretions and are overinclined to accept blame for anything that
may go wrong. This kind of behavior can be extremely irritating to other people,
but when these people show their irritation, they become even more apologetic and
self-blaming.
. A particularly intense and destructive form of lowerness is shame. Kaufman
(1996) has defined shame as the emotion of inferiority. The negatively lower person
thinks badly of himself and considers himself unworthy of the attention or praise of
others. This is a firmly held belief which is a consequence of what Kaufman has called
the shaming behavior of others toward the client. In therapy the client must be encour-
aged to return to and relive early shaming experiences.
4.3. Revision of the Client's Relating Pattern

Therapists may be coerced into responding reassuringly to clients questions by
replying, "No I will not leave you." or "No you have not offended me." and whilst this
will have some temporary effect on quelling clients' fears, the questions are bound to
recur with tedious regularity. More importantly, therapists need to identify the origin
of the fears by encouraging clients to free associate about them. The object of this is
to enable them to connect up with those early experiences from which the fears orig-
inated. When clients become connected up they often show strong emotion and the
therapist needs to stay reassuringly with them whilst the emotion is expressed. It is
important for both therapist and client to be able to make meaningful connections
between the early experiences and the fears. No two people's stories are the same, so
that it is not possible to lay down hard and fast rules about the kinds of experience that
282 J. BirtchneU and G. Borgherini
give rise to particular fears; but essentially they have never been enabled to feel
securely close or to feel securely lower.
Normally the client's story centres upon one or both parents whose behavior
toward the client appears to have given rise to the fears, though other people, like teach-
ers, may have reinforced them. Within the security of the session, it is necessary for the
client to renegotiate his relationship with these people, in order that he may break out
of the deadlock that he has remained in to the present day. This requires him to be
both critical of, and sympathetic toward, them, for usually there is an explanation for
why these other people behaved toward the client in this way. Next it is necessary
to help the client see how he has tended to relate to everyone as though they were
replicas of these people.
Inevitably, the therapist becomes both a representative of and an alternative to
them. The client alternates between behaving as though the therapist were one of them
and acknowledging that the therapist is a different person who treats him differently.
As the months go by, the client comes to see the therapist less and less as a represen-
tative of these other people and more and more as the person he really is, and to relate
to the therapist differently from the way he related to these other people. This means
that he feels more securely close and more securely lower in relation to the therapist,
and then hopefully toward other people.
4.4. Overcoming the Client's Fears of Moving toward the Opposite Pole
of Each Axis
On the horizontal axis, from the position of secure closeness, the client must be
encouraged to make excursions into distance. In Mahler's (1963) terminology, this
includes both separation and individuation. In terms of separation, his parents proba-
bly never encouraged him to go places on his own and overemphasised the dangers of
straying too far from the home base. The therapist must belatedly provide this encour-
agement. The client must feel more confident about leaving people, going places by
himself, being left alone and letting people leave him and have friends other than
himself.
In terms of individuation, his parents probably never encouraged him to think
things out for himself, and have ideas and preferences of his own, and never asked him
what he thOUght or what he liked or wanted to do. For them, his opinions did not count
for anything. Belatedly he must be encouraged to do this. He must come to experience
himself as a separate individual with ideas, values and interests of his own. To this end,
the therapist must repeatedly ask him what does he think about this or that, what does
he like and dislike and why, what does he prefer to do and what does he definitely not
like doing, and to convey to him that his opinions do matter. He must come to recog-
nise what it feels like to prefer one thing to another. He must be encouraged to go into
shops and choose things to buy, decide where to go on holiday and experience what it
is like to have preferences.
On the vertical axis, from the position of secure lowerness, the client must be
encouraged to make excursions into upperness. There are two aspects to this: (1)
increasing his sense of upperness, and (2) becoming confident of acting in an upper way
toward others. In terms of increasing a sense of upperness, his parents would have been
over inclined to do everything for him or to convey to him that he was hopeless and
incompetent and that he always made a mess of things, so he might as well let them do
things for him. Belatedly, the therapist must assure him that he is capable of doing
things for himself, and select tasks that he has never have the confidence to do on his
own, encourage him to try to do them, and praise him for any successes. In terms of
acting in an upper way toward others, his parents were probably overkeen to keep him
in the one-down position, never accepting him into the world of adults. At this late
stage, the therapist should make a point of treating him as an equal, or push him into
upperness by asking his advice on things. He should be encouraged to find opportuni-
ties of assuming responsibility for others, showing other people how to do things and
being protective and caring toward others. Caring for his own or other people's chil-
dren might be a step in this direction.
4.5. The Client's Relating to Others Outside of Therapy

The person the client identified as being most responsible for his dependent relat-
ing style (most usually a mother or a father) may still feature prominently in his life.
This person, more than any other, is capable of obstructing therapeutic progress or
reversing any changes that the therapy has brought about. This, in a sense, is the person
with whom the therapist has had to compete throughout therapy and with who he has
to continue to compete. Because this person has not been in therapy, he has not
changed, and can only continue to relate to the client the way he always has done.
When, as the result of therapy, the client tries to relate to this person differently, the
person cannot reciprocate and persists in relating to the client the way he always has.
The client feels himself being forced back into the dependent position. After repeated,
unsuccessful efforts to change the nature of the relationship, the client may reluctantly
conclude that the only solution is to reduce contact with this person or to break off
contact completely from him.
4.6. Inviting Significant Others into Therapy

Inviting a parent into therapy is rarely effective. If it is attempted, it should only
be after a long period of therapy during which time the client has had the opportunity
to build up some resistance to the parent's seductive behavior. The parent has usually
so much invested in maintaining the status quo, that he blinds himself to the problem.
He perceives the therapist's remarks as accusatory, and adopts an attitude of innocence.
He denies that he has ever held the client back or tried to keep the client helpless. He
is likely to be hostile toward the therapist and may even try to persuade the client to
stop attending.
Inviting a marital partner into therapy can be a more rewarding experience. It is
easier to explain to both the client and the partner the cyclical nature of the interac-
tion between them. On the horizontal axis, when the dependent client becomes cling-
ing, the partner becomes withdrawing, which makes the client cling even more. If the
clinger clings less, the withdrawer withdraws less, and vice versa. The partners try this
and find that it works. On the vertical axis, when the dependent client behaves help-
lessly, the partner moves in to help, and this makes the client even more helpless. If the
client is less helpless, the partner is less helpful, and vice versa. In Birtchnell (1986), two
dependent individuals, one a husband and one a wife, were successfully treated during
the course of marital therapy. In both these marriages, the other partner appeared to
do a great deal to maintain the individual's dependent behavior. In the first, the wife
complained that her husband was weak and ineffectual, yet she spent much of her time
giving him instructions. He, for his part, was afraid of acting on his own initiative lest
what he did would not meet with her approval. In the second, the wife complained that
her husband dominated her and would not give her the chance to make any decisions
for herself. The husband maintained that she was so indecisive that it was easier if he
made the decisions for her. When the wife insisted on being allowed to make some of
the decisions, the husband became confused, because his role in the marriage was being
taken from him. In both marriages, the main therapeutic objective was to persuade the
dominant partner to allow the dependent one to be more responsible. In marriages of
this kind, the dependent partner has often become a dull, almost atrophic figure. If the
other can be persuaded to relax the pressure upon the dependent one he can be sur-
prised by how much more interesting and lively the dependent one becomes.
5. DEMONSTRATING THE EFFECTIVENESS OF

PSYCHOTHERAPY
An instrument is being developed for measuring a person's relating characteris-

tics within the framework of the interpersonal octagon (Birtchnell, Falkowski, and
Steffert, 1992). It is called the Person's Relating to Others Questionnaire (PROQ). It
comprises 96 items, twelve for each of the eight octants of the octagon. In order to
detect forms of relating pathology, it is designed predominantly as a measure of nega-
tive relating. Therefore, aU but two of the twelve octant items are negative, and the
negative items are the only ones that are scored. The sixteen positive items (ten sets of
two), which are evenly distributed throughout the questionnaire, are included to relieve
its overall negative tone. A computer program produces a set of octant scores and also
presents them graphically as shaded areas of octants within an octagon. Administered
before therapy, it shows the client's main areas of negative relating, and administered
after therapy, it shows those areas in which there has been improvement (Fig. 4). The
mean pre-therapy scores of a series of psychotherapy clients were shown to be
significantly higher than those of a student sample, the most marked difference (a mean
of 21 versus a mean of 12) occurring in the lower close octant. Thus, dependence is a
prominent feature of many psychotherapy patients. The mean post-therapy scores were
shown to be significantly lower than the mean pre-therapy scores, the biggest drop in
scores occurring in the lower close octant (Table 1). Whilst most of this change is likely
to be in state dependence rather than trait dependence, it does indicate that negatively
dependent relating is readily modifiable by psychotherapy. This is in accord with
Millon's (1981) observations.
6. CONCLUSIONS
In this chapter a new version of interpersonal theory is presented. It is repre-

sented as a biaxial theoretical system called the interpersonal octagon. An important
feature of this version of the theory is the distinction that is made between positive and
negative relating, and the explanations and precise definitions that are provided for the
main forms of negative relating.
The treatment of any personality disorder is facilitated by defining its location
within the interpersonal octagon. This is so because the major part of any disorder is
explicable in terms of negative forms of relating. Whilst it is acknowledged that varying
Treatment of Dependent Personality Disorder 28S
Before After (30 sessions)
Figure 4. Use of the PROQ as a measure of change in psychotherapy. In these, before and after psy-
chotherapy, computer print-outs, negative scores are represented as shaded areas of octants. With this par-
ticular patient, the distant and lower distant scores have been reduced and the lower and lower close scores
have been almost eliminated. From Birtchnell, 1., Attachment in an interpersonal context. British Journal of
Medical Psychology, 70, 265-279.
proportions of the symptomatology of personality disorders can be accounted for by

genetic disposition, there are always those that cannot. These are the ones toward which
the main therapeutic thrust should be directed. It is necessary to identify those com-
ponents of the disorder that are attributable to relating on the horizontal axis and those
that are attributable to relating on the vertical axis, and to devise separate strategies
to combat them.
Table 1. Mean pre- and post-therapy PROQ scores in a series of 25 patients receiving psycho-
dynamic psychotherapy. Mean number of session = 27.1 (sd = 9.9). PROQ stands for Person's
Relating to Others Questionnaire (see text) . The letters UN, UC, NC, LC, LN, LD, ND, and UD
stand for Upper Neutral, Upper Close, Neutral Close, Lower Close, Lower Neutral, Lower
Distant, Neutral Distant, and Upper Distant, and represent the eight oct ants of the Interpersonal
Octagon. The letters TOT stand for Total, and refer to the score obtained by adding together the
scores of all the eight octants. The table shows that the total mean post-therapy score is sub-
stantially lower than the total mean pre-therapy score, and that there is a mean pre- to post-
therapy drop of 25 points. The biggest drop is registered for the LC octant, which corresponds
most closely with the construct of dependence
UN UC NC LC LN LD ND UD TOT
Pre-therapy mean octant scores
me 13.7 20.0 16.2 22.7 14.3 17.9 17.7 10.0 132.3
sd 6.9 6.6 8.3 6.0 7.0 5.4 7.0 5.4 24.0
Post-therapy mean octant scores
me 13.1 18.0 11.1 16.5 10.5 11.8 14.7 9.7 104.7
sd 6.7 7.7 7.1 7.9 7.7 7.2 6.9 4.3 31.7
Mean pre- to post-therapy octant score change
me -{).6 -1.9 -5.0 -6.2 -3.8 -4.9 -2.8 -{).3 -25.2
sd 4.5 4.2 6.5 7.4 7.4 7.6 6.4 2.6 32.2
When the horizontal and vertical components of a disorder have been defined a
likely aetiology in terms of early exposure to the negative relating of significant others
can be postulated. It is proposed that the relating of people with personality disorders
is restricted to a particular location within both the horizontal and the vertical axis.
They remain insecurely attached to one location and fearful of venturing toward the
opposite location. On each axis separately, the therapy involves first improving the
security of attachment to the one location, and second encouraging movement toward
the opposite location. An important component of therapy involves acknowledging the
part played by the negative relating of others in maintaining the client in a particular
form of negative relating, and enabling the client to resist responding to them.
Dependent personality disorder is defined in terms of insecure closeness and fear
of distance on the horizontal axis and insecure lowerness and fear of upperness on the
vertical axis. Compared with most of the other nine DSM-IV personality disorders, the
genetic input to both these forms of negative relating is relatively low (Livesley, Jang,
Jackson, and Vernon, 1993). This suggests that the condition would be relatively respon-
sive to therapy. The development of an instrument (the PROQ) to measure negative
relating within the interpersonal octagon is described. Administration of the instrument
before and after therapy confirms that lower close forms of negative relating are par-
ticularly responsive to therapy.
REFERENCES
American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders, 4th edn.
(DSM-IV). American Psychiatric Association Press, Washington, De.
Anchin, 1.e. and Kiesler, D.1. (1987) Handbook of interpersonal psychotherapy. Pergamon, New York.
Benjamin, L.S. (1984) Principles of prediction using structural analysis of social behavior. In Zucker, R.A.,
Aronoff, 1., and Rabin A.1. (Eds.) Personality and the Prediction of Behavior. Academic Press: New
York.
Benjamin, L.S. (1987) Use of the SASB dimensional model to develop treatment plans for personality dis-
orders. 1: Narcissism. Personality Disorders, 1,43-70.
Berg, I. (1974) A self-administered dependency questionnaire (SADQ) for use with the mothers of school
children. British Journal of Psychiatry, 124, 1-9.
Birtchnell,1. (1984) Dependence and its relationship to depression. British Journal of Medical Psychology,
57,215-225.
Birtchnell,1. (1986) The imperfect attainment of intimacy: a key concept in marital therapy. Journal of Marital
Therapy, 8, 153-172.
Birtchnell,1. (1987) Attachment-detachment, directiveness-receptiveness: A system for classifying interper-
sonal attitudes and behaviour. British Journal of Medical Psychology, 60,17-27.
Birtchnell,1. (1988) Defining dependence. British Journal of Medical Psychology, 61,111-123.
Birtchnell, 1. (1990) Interpersonal theory: Criticism, modification, and elaboration. Human Relations, 43,
1183-1201.
Birtchnell, 1. (1991a) Redefining dependence: A reply to Cadbury's critique. British Journal of medical
Psychology, 64, 253-261.
Birtchnell, J. (1991b) The measurement of dependence by questionnaire. Journal of Personality Disorders, 5,
281-295.
Birtchnell,1. (1993/1996) How humans relate: A new interpersonal theory. HB (1993) Praeger, Westport, CT.
PB (1996) Psychology Press, Hove, Sussex.
Birtchnell,1. (1994) The interpersonal octagon: An alternative to the interpersonal circle. Human Relations,
47,511-529.
Birtchnell,1. (1996a) The management and resolution of dependency in the depressed client. In Flach, F.
(Ed.) The Hatherleigh guide to managing depression. Hatherleigh Press, New York.
Birtchnell, 1. (1996b) Detachment. In Costello, e. (Ed.) Personality characteristics of the personality disor-
dered. Wiley, New York.
Birtchnell, I (1997a) Personality set within an octagonal model of relating. In Plutchik, R and Conte, H.R
(Eds.) Circumplex models ofpersonality and emotions. American Psychology Association Press, Wash-
ington, D.C.
Birtchnell, I (1997b) Attachment in an interpersonal context. British Journal of Medical Psychology, 70, (in
press).
Birtchnell, I, Deahl, M., and Falkowski, I (1991) Further exploration of the relationship between depression
and dependence. Journal of Affective Disorders, 22, 221-233.
Birtchnell, I, Falkowski, I, and Steffert, B. (1992) The negative relating of depressed patients: A new
approach. Journal of Affective Disorders, 24, 165-176.
Bowlby, I (1973) Attachment and Loss. Vol. 2: Separation, Anxiety and Anger. Hogarth Press/Institute of
Psychoanalysis, London.
Bowlby, I (1977) The making and breaking of affectional bonds, 1, Aetiology and psychopathology in the
light of attachment theory. British Journal of Psychiatry, 130,201-210.
Carson, RC. (1969) Interaction concepts of personality. Aldine, Chicago, Ill.
Costa, P.T. and McCrae, RR (1986) Personality stability and its implications for clinical psychology. Clini-
cal Psychology Review, 6, 407-432.
Freedman, M.B., Leary, T., Ossorio, AG., and Coffee, H.S. (1951) The interpersonal dimension of personal-
ity. Journal of Personality, 20,143-161.
Fromm, E. (1947) Man for himself. Rinehart, New York.
Gilbert, P. (1989) Human nature and suffering. Erlbaum, Hillside, New Jersey.
Hirschfeld, RM.A, Klerman, G.L., Clayton, P.I, Keller, M.B., McDonald-Scott, P., and Larkin, B.H. (1983)
Assessing personality: Effects of the depressive state on trait measurement. American Journal of
psychiatry, 140,695-699.
Hirschfeld, RM.A, Klerman, G.L., Lavori, P., Keller, M.B., Griffith, P., and Coryell, W. (1989) Pre-
morbid personality assessment of first onset major depression. Archives of General Psychiatry, 46,
345-350.
Horney, K. (1937) The neurotic personality of our time. Norton, New York.
Jacobson, E. (1964) The self and the object world. International Universities Press, New York.
Kaufman, G. (1996) The psychology of shame: Theory and treatment of shame-based syndromes. Springer,
New York.
Kiesler, D.I (1986) Interpersonal methods of diagnosis and treatment. In Cavenar, 10. (Ed.) Psychiatry, Vol.
1,1-23. Lippincott, Philadelphia.
Klerman, G.L., Weissman, M.M., Rounsaville, B.I et al. (1984) Interpersonal psychotherapy of depression.
Basic Books, New York.
Laing, RD. (1965) The divided self. Penguin: Harmondsworth, Middlesex.
Leary, T. (1957) Interpersonal diagnosis of personality. Ronald Press, New York.
Livesley, W.I (1996) The implication of recent research on the structure and etiology of personality disor-
ders for treatment. Paper read to the II European Congress on Personality Disorders, Milan, June
29th.
Livesley, W.I, Jang, K.L., Jackson, D.N., and Vernon, P.A. (1993) Genetic and environmental contributions to
dimensions of personality disorder. American Journal of Psychiatry, 150,1826-1831.
Livesley, W.I, Schroeder, M.L., and Jackson, D.N. (1990) Dependent personality disorder and attachment
problems. Journal of Personality Disorders, 4, 131-140.
Mahler, M. (1963) Thoughts about development and individuation. Psychoanalytic Study of the Child, 18,
307-324.
Orford, I (1986) The rules of interpersonal complementarity: Does hostility beget hostility and dominance,
submission? Psychological Review, 93, 365-377.
Parens, H. and Saul, L.I (1971) Dependence in man. International Universities Press, New York.
Millon, T. (1981) Dependent personality: The submissive pattern (Chapter 4). In Disorders of personality,
DSM-III, Axis II. Wiley, New York.
Pilkonis, P.A (1988) Personality prototypes among depressives: Themes of dependency and autonomy.
Ray, II (1976) Do authoritarians hold authoritarian attitudes? Human Relations, 29, 307-325.
Schwartz, D.A., Flinn, D.E., and Slawson, P.F. (1974) Treatment of the suicidal character. American Journal
of Psychotherapy, 28,194-207.
Sullivan, H.S. (1953) The interpersonal theory of psychiatry. Norton, New York.
Sullivan, H.S. (1954) The psychiatric interview. Norton, New York.
Trapnel, P.D. and Wiggins, IS. (1990) Extension of the Interpersonal Adjective Scales to include the Big Five
dimensions of personality. Journal of Personality and Social psychology, 59, 781-790.
Wiggins, IS. (1982) Circumplex models of interpersonal behavior in clinical psychology. In Kendall, P.e. and
Butcher, IN. (Eds.) Handbook of research methods in clinical psychology. Wiley, New York.
Wiggins, IS. and Pincus, A.L. (1989) Conceptions of personality disorders and dimensions of personality.
Psychological Assessment: A Journal of Consulting and Clinical psychology, 1,305-316.
Winnicott, D.W. (1958) The capacity to be alone. International Journal of Psycho-Analysis, 9, 420-438.
World Health Organisation (1989) International classification of diseases, 10th edn. (ICD-lO). World Health
Organisation, Division of Mental Health, Geneva.
20
HIV INFECTION, PERSONALITY

STRUCTURE, AND PSYCHOTHERAPEUTIC
TREATMENT
R. Visintini,t E. Campanini,t A. Ama,! R. Alcorn,2 s. Corbella,3

s. Gessler,3 D. Miller/ L. Nilsson Schonnesson,4 and F. Staracel5
1 Medical Psychology Unit
Scientific Institute Ospedale San Raffaele
Milan, Italy
2Psychoanalyst of SPI
(Societa Psicoanalitica Italiana)
and IPA (International Psychoanalytic Association)
Trainer and Supervisor APG
(Associazione Psicoterapia di Gruppo)
and member of American Group Psychotherapy Association
Milan, Italy
3Camden & Islington
Community Health Services NHS Trust
The Mortimer Market Centre
London, Great Britain
4Department of Social Work
University of Gothenburg
Sweden
5 Psychiatric Institute
School of Medicine
II' University of Naples
Italy

290 R. Visintini et al.
1. INTRODUCTION
We observed that socially, continuous and specific campaigns to build an aware-
ness of the ways HIV infection can be transmitted, have not been able to effectively
control the transmission of the disease.
The high incidence of cluster B personality disorders (DSM IV) in HIV-positive
subjects found in clinical practice emphasized the need for deeper knowledge of the
clinical-theoretical characteristics typical of the relationship between these personality
disorders and HIV-infection and the psychological/psychotherapeutic methods used
with these patients.
Kernberg (1993) suggested that the borderline personality organization, include:
"a wide range of character pathologies or personality disorders which had in common
a lack of integration of the self-concept and of the concept of significant others", there-
fore the borderline organization model can be extended to several DSM personality
disorders: narcissistic, histrionic, borderline, and antisocial.
Kernberg (1996) includes two different levels of psychopathological functioning,
one moderately severe: the high level borderline personality organization and the other
extremely severe: the low level borderline personality organization. Borderline and
antisocial personality disorders, are placed at a low level of functioning, histrionic and
narcissistic at a high level.
The weak ego structure found in these personality disorders specifically reflects
the use of primitive/low level defense mechanisms: splitting, primitive idealization, pro-
jection, projective identification, denial, devaluation, omnipotence.
Our clinical experience has confirmed how the lack of a structured personality
implies a greater risk being exposed to infection.
All the subjects with these kind of personality organizations do not tolerate
anxiety, due to their weak ego structure, and as a result act-out and therefore increase
the risk of HIV infection. Their perception of the real risk of infection is limited by
their rigid use of low level defense mechanisms.
We suggest that, the different modalities of contracting HIV-infection, analo-
gously, reflect the level of functioning of the underlying personality structures and the
defenses primarily used by the subject. The way the disease is transmitted seems to
reflect either a personality organization with functioning characterized by narcissistic
defenses, which we call "narcissistic functioning" and which could be included in
Kernberg's "high level borderline personality organization", or functioning character-
ized by more primitive defenses, specific to Kernberg's "low borderline personality
organization", which we will call "borderline functioning".
Subjects with a narcissistic functioning structure use idealization to compensate
their weak ego structure and lack of self-esteem by perceiving the self as hypertrophic.
This compensating mechanism, represented by the subject's omnipotence and percep-
tion of others as a narcissistic extension (self-object), can be placed at a superior devel-
opmental level compared to borderline functioning, where the structuring of the self,
instead, is not consolidated in one grandiose and idealized identification, but remains
in one primary narcissistie position where the self has not been structured.
Acting-out can also have other underlying significance and differentiate narcissis-
tic functioning from borderline functioning. In subjects with narcissistic functioning
acting-out is characterized by an angry attempt to oppose the depressive experience tied
to the frustration of omnipotence. Instead, in borderline functioning acting-out has a
discharging function, the inability to tolerate anxiety is, concretely, the acting-out.
HIV Infection, Personality Structure, and Psychotherapeutic Treatment 291
In reference to the modality of transmission, we observed that:

- in borderline functioning subjects who have great difficulty in containing
anxiety due to deficient differentiation and structuring of the self, the conta-
gion represents extreme self-mutilating acting-out as well as drug dependency
behavior and promiscuous sexual behavior.
- in narcissistic functioning subjects, with greater structuring and differentia-
tion of the self, there is low self-esteem and the contagion is more tied to the
grandiose and omnipotent transposition of the self which reduces the thresh-
old of perceiving the risk of infection ("it will never happen to me"), and the
object as a potentially dangerous "object-other than the self". In fact since
narcissistic functioning subjects do not live the relationship in a bi-univocal
sense, they do not perceive the possibility that even the virus can be part of
the relationship and thus underestimate its transmissibility.
We think that the virus is concretely both the persecutor and punisher tied to
unconscious guilt feelings; the anxiety linked to the infection replaces existential
anxiety; frequent social isolation makes feelings of inadequacy and low self-esteem
that characterized the infected person's life become real. These patients have, in
most cases, suffered particularly inadequate intimate relationships in the micro-social
environment.
Psychodynamic oriented authors sustain that these pathological structures of per-
sonality originate in very early stages of development, due to the lack or loss of an ade-
quate holding environment, and thus leads to particular difficulties during processes of
separation-individuation. In this case the lack of object-constancy makes it impossible
to have an integrated relationship and as a result the object relations of these patients
are characterized by prevalent splitting, since the processes of differentiation could not
follow natural developmental phases. All of this produces a fragile-self, a subsequent
weakening of the capacity of the ego in carrying out its functions and an anachronis-
tic use of archaic defense mechanisms. All crucial moments of development then
put at stake this already unstable balance, and therefore increase unsolved conflicts.
We can conclude this brief digression by sustaining that these are all object relation
pathologies.
In the course of HIV disease progression, the individual is confronted not only
with the HIV-diagnosis as such but also with various physical, social, and sexual threats.
These threats, that are intimately connected to emotions of helplessness, powerlessness,
and anxieties of separation and annihilation, are more or less salient depending on
which medical phase of HIV infection has been entered. The individual also runs the
risk of facing the social individual implication of blame and personal responsibility for
his infection. Additionally, they face, with multiple losses and grief, the mourning
process related to infection. The latter include existential concerns of death, freedom,
existential isolation, and the meaning of life and their related existential anxiety. It can
be reinforced, colored, and inereased by unresolved conflicts originating in the first
years of life. Worries of annihilation can be affected in an unproportionate strong way
by early losses and non-healed grief.
Another way to view HIV-related threats is in terms of internal and external per-
secutors. The internal persecutor is the virus itself and its related potentiai physical,
psychological, sexual, and social limitations. The potential stigmatizing environment
may be perceived as an external persecutor. Persons withHIV can also be potential
persecutors, because they can infect others. The double facet-both being potentially
persecuted and being a potential persecutor-finds its counterpart in an internal per-

secution scenario. Within this scenario these persons are haunted by the destructive
power of the H1V infection, but may also be haunted by their own psychological per-
secutors such as guilt, bad conscience, self-blarne, degradation, and devaluation of
themselves. So the individual's self is heavily attacked by concrete and existential con-
cerns in combination with these internal persecutors. As a consequence this self and its
functions may be disrupted and de-stabilized and there may be loss of self-confidence,
self-esteem, self-respect, and even human dignity.
The H1V-infection may also re-activate unresolved intrapsychic conflicts related
to early traumatic experiences or various psychic threats related to symbiosis and
separation.
We observed that many of the H1V-positive women, who have been infected
through sexual intercourse with an active drug addicted partner or ex-drug addict, seem
to share common typical personality structures. This is also based on our observation
of the kind of relationships they have with their partner.
These women knew their partners were active drug addicts or ex-drug addicts,
furthermore these women were infected when there was already enough and valid
information available on the modality of transmission.
It should be highlighted that some of these drug addicts of ex-drug addicts, at first,
although they knew they were H1V-positive, did not inform their partners. Others
attempted to convince their partners and themselves of their presumed H1V-
negativity, producing statements like: "I have never exchanged. needles!".
Apart from ethical-moral issues regarding the behavior of these subjects, it seems
these women, knowing of the potential risk of infection, however still accepted the
manipulatory justifications put forth by their partners and had indiscriminately and
consequently unprotected sexual intercourse.
These women often justify themselves during interviews by saying: "I never con-
sidered the problem", "every now and then I thought about it, but then I always
believed his explanations". These statements seem to reflect the denial and splitting
between the knowledge of the objective risk of infection and adequate precautionary
behavior and personal protection.
What is peculiar about these patients is that they totally dedicated themselves to
their partners and his problems, at times up to the point of annulling themselves. They
spend most of their energy trying to redeem them from self-destructive behavior tied
to the drug addiction, hoping and expecting they will change. The logic that character-
izes this relationship is similar to: "I'll save you".
Another common characteristic in these women is the extremely negative opinion
they have of themselves, evident in the belittling and ruthless judgment they have
of themselves. This could reflect a deficient individuation process in their primary
fusional relationship with their mother which may have, consequently, led to inade-
quate personal recognition and therefore devaluation.
These women, in fact, are obsessed by the doubt of lacking something or not being
good enough. The ruthless judgment they have of themselves makes them highly sen-
sitive to others judgment and, particularly empathetic and able to sense the feelings
and needs of others, and look after them. This willingness to help others makes
them less important: personal problems are of secondary importance, or are annulled,
compared with those of others, especially if needy.
On the other hand the particular partner, the difficult relationship and evident
devotion to the partner give shades of uniqueness and omnipotence to the actions of
the patient. In fact we observed that these women were unconsciously pleased and had
a sense of narcissistic gratification in subjugating themselves to their love object and
proud of the mental representation of themselves as the greatest martyrs on earth: "I
will sacrifice myself for the great cause, so the greatness of the cause will fall back onto
me and I will become omnipotent" (Kernberg, 1993) what Fenichel (1945) called "pride
in suffering".
These relationships answer to these women's unconscious need to feel heroic
in their attempt to sustain the precariousness of their self-esteem. This heroic act,
therefore, is the result of a need for special recognition to answer to an extremely high
ego ideal, gratifying the tyranny of a rigid, severe and self-depreciating super-ego
(Jacobson, 1964), which, however, exposes these women to unquestionably frustrating
relationships and in our case dangerous. Their partners, in fact, either because they are
active drug addicts or because they are ill or because of their personality disorder
are not able to offer a truly gratifying and complete relationship. For these women
both the situation in which they found themselves, and the fact that they have been
infected can be considered, as Fenichel states (1945) "the undesired consequence of
something desired". These women inevitably put themselves in a situation that gener-
ates suffering or exploitation which then makes them feel mistreated, despised or
humiliated by their partner (Kernberg, 1993). Going back to Freud and Fenichel's
concept of "moral masochism" these dynamics reflect unconscious regression to pre-
oedipal dynamics.
The challenge with themselves is primarily to beat the partners heroin depen-
dency which seems similar to competing with the other woman (the mother) to take
away the loved one. Later another challenge is that of becoming indispensable so that
their partner in the end becomes dependent on them, in fact the ill partner is always
more needy, subjugated and not able to leave or choose another woman, that is: his
mother. Finally the omnipotent challenge, par excellence, is between the HIV-virus and
the idealized love, superior to all, eternal and unassailable, like that demonstrated with
deep suffering and frustration for their partner.
We see how the masochistic behavior intrinsically has an accusatory and black-
mailing tone towards the partner. It is evident how the narcisistic gratification described
up now, originates from having been unfairly treated, but however from a morally supe-
rior position compared to that of the "persecutor". The partner, therefore, is not only
lovingly looked after and super-protected but deeply devaluated and unknowingly
becomes the object of these patient's narcissistic need to increase their self-esteem.
In this way the self-esteem is maintained by the super-ego through approval, narcisis-
tically nourishing the ego.
"The aim of any self-destruction is to attempt to free oneself from the pressure
of the super-ego, especially evident in cases in which self-destruction is tied to a kind
of ascetic pride" (Fenichel, 1945).
From the anamnestic recostruction of their history it was possible to learn the
scarce recognition these women received in childhood consequently generated
extremely low level of self-esteem and an idealization of their father, experienced as
unattainable. The devaluation of the partner/father allows them to realize a relation-
ship which would otherwise have been unthinkable, which gratifies their extreme need
for recognition and love. These women probably lacked individual recognition in their
primary fusional mother/child relationship. Therefore the reactualization of a fusional
relationship with their father seems to be an extreme attempt, at an unconscious level,
to repair the lack of recognition during their pre-oedipal phase.
The process described seems to originate as a dynamic movement compensating

affective deficiency, frustration, personal disesteem, and lack of recognition, which
should have one aim: "self-conservation" of the subject; paradoxically this process can be
summed up as the risk of "self-destruction", made concrete by the HIV-infection.
2. SOME DIAGNOSTIC PROBLEMS USING DSM-IV CRITERIA IN

AT RISK SUBJECTS FOR HIV-INFECTION
We realized that the characteristics ofthe DSM IV borderline personality disorder

diagnosis appear very different when viewed in a gay context. It raises the question of
how pathologically one views symptoms which are acceptable mores in a different
culture. Thus, how do we rate sexual impulsivity in a gay bar where contact is likely to
move very fast to a sexual act, by common consent? Or substance misuse in a similar
context? Instability of self image is not unlikely in the wider gay community because of
the often difficult and tortuous process of coming out. Intensity of relationships, alternat-
ing between idealization and devaluation is also a common feature of the gay scene, to the
extent that some gay commentators are beginning to question the difficulty of the gay
community in sustaining longer term relationships (Lyttle, 1996).
According to DSM IV, antisocial personality disorder can be diagnosed if the
patient has had a conduct disorder since the age of 15 and at least three deviant (i.e.,
not conforming with social norms) kinds of behavior after the age of 18.
Antisocial personality disorder is a "pervasive pattern of disregard for, and vio-
lation of, the rights of others" (APA, 1994). The concept of antisocial personality dis-
order in DSM IV is quite different from the classical model of psychopathology, which
emphasizes the structural characteristics of a psychopathic personality, such as lack
of guilt feelings, the presence of manipUlative, and utilitarian behavior, the lack of
interest in long lasting affective relationships and the inability to tolerate frustration
(Cleckley, 1941).
At least three homogeneous diagnostic sub-groups of antisocial personality dis-
order can be identified among intravenous drug users (IDU's). The first includes many
adult IDU's who, although they have antisocial behavior, did not meet the required cri-
teria for childhood conduct disorder from the age of 15. These subjects are not differ-
ent from other addicts in terms of the severity of addiction, the treatment response or
the presence of at risk behavior for HIV infection (Brooner et aI., 1990; Brooner et aI.,
1992; CottIer et aI., 1995). In these cases, antisocial behavior could be reasonably
believed to be a consequence of substance abuse, with would act as continuative stress,
according to the secondary psychopathy model (Alexander, 1930; Bartemeir, 1930;
Whittlers, 1937; Karpman, 1941).
IDU's meeting DSM diagnostic criteria for antisocial personality disorder form
the second group. They show more frequent illegal behavior, an earlier onset of drug
abuse and a more severe addiction; however, they do not show a worse response to
treatment compared to IDU's who do not have antisocial personality disorder
(Cacciola et aI., 1994, 1995, 1996). A more precise clinical and psychopathological
assessment of subjects belonging to this diagnostic group show substantial hetero-
geneity in terms of vulnerability to stress and affective liability. The frequent comor-
bidity for several personality disorders is associated with a more severe addiction and
at risk behavior for HIV infection; however, improved compliance and a more
favourable prognosis have also been reported (Brooner et aI., 1993b).
mv Infection, Personality Structure, and Psychotherapeutic Treatment 295
The third group is composed of subjects meeting DSM criteria for antisocial per-
sonality disorder and showing psychopathic personality traits, according to appropri-
ate evaluation instruments, such as the California Psychological Inventory and the Hare
Psychopathy Checklist. These IDU's present more severe drug abuse and a very poor
response to treatment (Hare et aI., 1991). As noted above, there is evidence that IDU's
with a diagnosis of antisocial personality disorder according to DSM criteria, show a
more severe addiction and a higher frequency of at risk behavior for HIV infection
(Brooner et aI., 1990, 1993a).
And, in fact, IDU's with concomitant antisocial personality disorder diagnosis
showed a higher frequency of needle sharing and a higher number of partners with
whom the sharing occurred, compared to IDU's who do not have antisocial personal-
ity disorder. Interestingly, no difference was detected between the two groups in terms
of severity (Brooner et aI., 1990). In a more ent study (Brooner et aI., 1993a), the diag-
nosis of antisocial personality disorder was found to be significantly related to a more
severe addiction and to a higher frequency of needle sharing. Unfortunately, no infor-
mation was provided regarding the frequency of at risk sexual behavior. Moreover, the
high prevalence of subjects belonging to ethnic minorities in the selected sample ham-
pered the generalization of results.
A significant relationship between the diagnosis of personality disorder and HIV
infection has been reported by Jacobsberg et ai. (1995), in a study on subjects who vol-
untarily participated in an HIV testing program. The most frequent personality diag-
nosis among subjects with HIV infection was borderline personality disorder and
antisocial personality disorder. A higher prevalence of personality disorder in subjects
with HIV infection has also been reported by Perkins et ai. (1993), who found a close
relationship between antisocial and borderline personality disorder, and HIVinfection
risk. IDU's with antisocial personality disorder showed a severe risk of contracting and
spreading the disease.
Further studies seem necessary to define the role that other diagnoses of axis II
play in determining behavior at risk for HIV infection, and to analyze the specific
importance of some psychopathologic features, such as impulsiveness or irresponsibil-
ity, which are typically found in antisocial personality disorder.
A common assumption about antisocial personality disorder concerns its incur-
ability by means of psychotherapeutic approaches (Andreoli et aI., 1990). This assump-
tion is based on the patient's inability to establish significant affective and social
relationships, which are thought to be essential conditions for a good response to psy-
chotherapeutic treatments (Luborsky et aI., 1983; Strupp, 1980).
In addition, personality traits such as manipulative attitudes and basic deceitful-
ness in relationships, typically described in subjects with an antisocial personality, are
traditionally considered factors which negatively interfere in the establishment of a
therapeutic setting.
Unfortunately, only few controlled studies can be found in the vast literature on
treatment strategies for IDU's with a concomitant diagnosis of antisocial personality
disorder. Here we will review these investigations.
The outcome of different interventions (supportive-expressive therapy & drug
counselling; cognitive behavioral therapy & drug counselling; drug counselling alone)
has been evaluated by Woody et al. (1983, 1987), and Grestley et ai. (1989). At 6 and
12 months follow-up only subjects receiving brief psychotherapy interventions showed
stable improvements.. Supportive-expressive therapy seems to produce a better
response as regards the severity of addiction and the social and working adjustment.
296 R; Visintini et at.
The effectiveness of this psychodynamic oriented psychotherapy may well be related

to the establishment of a therapeutie alliance, that is known to be a reliable predictor
of treatment success, in the area of drug abuse and of social and working adjustment
(Luborsky et aI., 1975).
Counselling and psychosocial interventions are indispensable to support brief
psychotherapy. The counsellor, in this context, plays three essential roles: a) he co-
ordinates the different therapeutic interventions; b) he coordinates the psychosocial
supportive interventions; e) he offers the patient a structured and well-organized
behavioral model.
3. PSYCHODYNAMIC PSYCHOTHERAPY IN
HIV-SEROPOSITIVE PATIENTS WITH NARCISSISTIC AND
BORDERLINE FUNCTIONING
On the basis of what has been put forward, we feel that psychodynamic psy-
chotherapy is an important instrument for the treatment of patients with good life
expectations and with a deep need to improve the quality of their existence.
In the last few years the effectiveness of psychodynamic psychotherapy in clini-
cal work with HIV-positive patients has become even more clearly evident. In the 80's
many psychotherapists and psychoanalysts were sceptical about the possibility of using
the psychodynamic model with patients whose bodies were at risk in such an impor-
tant and above all concrete way. In the 90's this scepticism decidedly reduced and even
case reports appeared in literature; moreover, some collegues' initial resistance began
to diminish.
The efficacy of the psychodynamic psychotherapeutic approach with HIV-
positive patients can be demonstrated by highlighting the extent to which the HIV-pos-
itive condition is associated, in certain cases, not only with a particular psychological
condition directly caused by the infection, but also with the psychopathological condi-
tion present prior to the infection. This is seen in clinical practice as well as in some
studies recently published in internationaljournals.
Through psychotherapeutic work and the patient's relationship with his/her
psychotherapist these patients have the possibility of becoming aware of the underly-
ing dynamics in their lives, as well as the feeling being recognized, accepted, and
considered.
3.1. The Individual Psychodynamic Psychotherapy in mV-Seropositive

Patients with Narcissistic and Borderline Functioning
Symbolically the life of a person with HIV can be pictured in the following way:
the individual's well-known, familiar ground has changed into a new, unknown, and
chaotic one. The overall question is: "How am I to become familiar with this new
ground, to feel confident in it, and to make it a part of myself." The overall question is:
"How can I achieve at least temporarily psychological and sexual well-being and carry
on a satisfying and meaningful life?" The answers to these questions are unfolded in
the individual's subjective adaptation process. Usually adaptation processes are
described in terms of stage crisis models. They do not justify the complexity of psy-
chological and social mechanisms involved in the adaptation processes. To better appre-
DIV Infection, Personality Structure, and Psychotherapeutic Treatment 297
ciate these processes we would suggest a theoretical framework in which adaptation

processes are to be understood as a function of psychic metabolism. In other words:
metabolizing is necessary to achieve adaptation.
The concept of metabolizing is used here as a metaphor to describe an individual's
active internal process of coming to grip with a given stressor. Every time the person is
confronted with a stressor it may evoke psychic and cognitive chaos accompanied by a
range of, and often a combination of, emotions. The extent to which the individual's psy-
chological equilibrium and sense of self are threatened by the HIV stressor and mani-
fested in psychological distress depend on the person's life history, personality (including
personality structure), and current life context. By means of the individual's psychologi-
cal, social and cognitive tools the actual stressor and its concomitant emotions are cogni-
tively and mentally metabolized in terms of being incorporated, contained, and digested.
The aim of psychic metabolism is to bring order out of the chaos and in this way restore
the disrupted equilibrium and consequently minimize feelings of mental suffering.
However, the person's metabolizing is not always constructive to him in terms of reduc-
ing the chaotic scenario. Regardless of the success of the psychic metabolism, one has to
keep in mind that a given metabolizing makes sense and can be understood within the
context ofthe individual's life conditions.
Case 1: Andrew is 23 years old with no steady partner. He has no professional training but has
earned his living through various unskilled jobs. In 1987 Andrew was notified of his HIV-
seropositivity and since then he has been on sick leave on and off due to psychological stress.
He is ashamed of his homosexuality and his, as he claims, "immoral sexual life." He expresses
a wish for encounters and acquaintances. He stresses his longing to find trust and security in a
father figure. Andrew expresses some mistrust in others, but at the same time he is eager to
seek help and support in others. It is, however, difficult for him to remain within an intimate
relationship. There is only one man (who works in the HIV field) to whom he can come to and
confide. "Dad is a bastard and I would never like to get his support. But John is such a great
person and he gives me a lot of support. I can always count on him." Andrew knows that his
mother is concerned about him and he feels support from her. "But I don't want to bother her
with my problems."
Andrew's childhood was traumatic. His parents divorced when he was three. Between
the age of six and nine he was sexually abused by his mother's new partner. His internal
resources are limited; he has almost nonexistent self-esteem, is dissatisfied with himself, turns
aggressiveness inwards and often feels sad and anxious. He seems to have some problems with
writing and reading, difficulties in understanding certain words such as "self blame," "neglect."
Drugs and alcohol have become a resource to Andrew when he is confronted with problems.
The first year after the diagnosis, he withdrew socially and drank a lot. He feels like he was
persuaded to be tested for HIV while being an inpatient at a psychiatric clinic. Today he
strongly regrets that. Andrew feels that his life has "stopped, moved backwards" after the
HIV-diagnosis. Sometimes he has suicidal ideation. On the other hand he emphasizes that
he will not fulfill other people's expectations of him to commit suicide "I'm not a shit who
also commits suicide." Today he discloses his HIV status without any real discrimination which
puts him in a difficult situation (e.g. being battered, stigmatized). Andrew reports intrusive
HIV thoughts and feelings as well as strong AIDS ruminations. On the one hand he is very
afraid of becoming sick and on the other hand he has a sense of mastery as to the HIV-
progression. With respect to coping styles, Andrew displays a combination of self-assertiveness
and avoidance.
His internal world is characterized by chaos and emptiness. There is a splitting between
good and bad figures, between mistrust and idealization. Andrew has a fragile self image and
low self-esteem. He tries to compensate for his void by introjecting the idealized John who rep-
resents strength and goodness. However, this compensating strategy makes him vulnerable and
may manifest itself in anxiety attacks and psychosomatic symptoms.
At the bottom line Andrew experiences himself as nobody, but the HIV infection makes
him visible and provides him with a fragile, but still an identity. But there is a paradox here:
the identity that makes him visible today in the long run will make him invisible again and
destroy him. It is quite likely that the HIV infection reactivates the sexual abuse trauma. Both
represent the threat of extinction. Since Andrew has such low self-esteem, he introjects (by
means of John) a self-assertive attitude and in that way he acquires at least a pseudo self-
esteem. But this attitude also functions as a manic defense against the powerful HIV. His
avoidant coping style, just like drinking and other drugs can be viewed as another desperate
way to get away from HIV-related issues and mourning. Andrew's strong help seeking attitude
can be seen as a desperate hope that another person will fill his internal emptiness but also in
a magical way free him from the evil, the HIV infection.
People with HIV reflect various personality structures and they may experience
their current life situation as a recapitulation of earlier traumas, unresolved conflicts,
etc. It is thus important to provide the individual with psychological space to explore
potential recapitulations as well as to explore fantasies about disease and death, HIV-
related fears and worries. Hard as it may be, it is still important that clinicians do their
best to distinguish existential anxiety from neurotic anxiety. The latter is generated from
unconscious, internal, repressed conflicts whereas existential anxiety is a reflection of
insight into the inescapable limitations of existence and the individual's aloneness and
vulnerability (Wikstrom, 1990).
Working with HIV infected persons challenges various aspects of countertrans-
ferenee, such as attitudes on human sexuality, survival guilt, death wishes, omnipotence,
helplessness, and existential concerns. Another challenge is to stand the paradox of
empathy and hopelessness i.e. to share, contain, and to hold feelings of dispair and
hopelessness of the person with HIV. Taking these psychological challenges into con-
sideration, it is no wonder that the clinician's self may be attacked. Consequently,
her/his self-object needs have to be attended to, for example to feel recognized,
affirmed, and appreciated.
In the context of the devastating disruptions HIV brings, the therapist's mirroring,
alter ego, and idealizing functions are crucial as is her/his commitment for a long-term
continuity. To paraphrase Kohut, until the moment physical death arrives, it is vital to con-
tinually provide self-object functions to help maintain the dying person's sense of self. As
therapists we should strive to provide the person with HIV with a holding environment
that can bring hope, and a safe playground in which he can alleviate separation anxiety,
fear of loss,fear of death, and mourning. To quote Ethel Dreifuss-Kattan, we can "contain
what has been put into me by the patient and so become the equivalent of a good mother,
who provides a safe space, a framework, and a medium where the patient threatened by
the death can move freely between the illusion of union and the fact of separateness, as
happens in the transitional phase of infancy."
3.2. The Group Psychodynamic Psychotherapy in HIV-Seropositive

Patients with Narcissistic and Borderline Functioning
The object relation theory allows us to consider the individual and the group as
different points of a continuum. However, it is important to point out some specifics of
the group setting. In understanding group thematics it is useful to think of time as a
spiral (Corbella, 1996). The image of a spiral rotating around an axle permits us to syn-
thesize the plurality of dimensions and movement that constitute our temporal expe-
rience in the group. We can therefore go ahead or backwards, with the possibility of
returning to the same point in relation to the axis. This occurs at different levels, since
multiple levels of reality are present at the same time in each session and for each
individual. A fundamental element and one which characterizes group therapy is the
fusion-individuation dialectic movement underlying every session and therefore always
available. However, this concept of time as moving as a spiral means that fusionality in
the group is referred, not only to the possibility of symbolically re-enacting the symbi-
otic phase with the primary object, but also to the possibility of sharing other and more
advanced stages of fusion. This regression is precisely to that archaic fusional phase
(which Balint defines "of basic fault"). There is no distinction between the subject and
object and, this in fact characterizes the primitive fusional moment in the group, in
which fantasies of omnipotence emerge. This level of regression is potentially present
immediately at the beginning of the group's history and continues to be so for the whole
time the group exists. Being in the group requires the capacity of putting into play
mutual symbiotic zones and this is made possible by the particular permeability that
the boundaries of the ego take on in the group situation. The positive and transform-
ing aspect of this regression, in this case, is the possibility of going back in time to the
relationship with the primary object and thus enter into the area of the original fault
to repair the path of the "grandiose self" (a basis to develop the "true self"). As well
as, perhaps, for the first time, the possibility of forming reassuring fusional experiences
of holding within the group and then be able to resynthesize and integrate partial
objects into a whole object, thanks to group work and opportune therapeutic inter-
ventions. In this case the synchronicity, which should be an aspect of the mother-child
relationship, becomes the prototype for group interaction.
However, a precise example of this situation is not easy to give because the expe-
rience that this regressive movement makes possible is located at a preverbal level; in
this context, language loses conventional adult meaning and words are used as a sort
of transitional object. As such it is not possible to refer to the exact content of a session,
in which the positive aspect of this archaic level of regression has been experienced.
However it is possible to speak of the dominating emotional atmosphere which, shared
by all, is usually extremely intense and characterized by trust; everyone participates,
the therapist as well, in a sort of immersed serene fusion, an experience, which many
patients have never experienced. At first, the possibility of this phase, seems still super-
ficial, often the pronoun "we" is used; "me too" becomes a sort of password. From this
beginning of verbal fusion, present in the new group which has only just begun to feel
good together, even if in a confused way, the potential for a more authentic and deeper
fusion develops as the group process evolves.
When the possibility of regressing to this kind of archaic fusion appears in
the group it is important that the therapist does not make the mistake of interpreting
this situation. Moreover, it is not interpretable if not with a disturbing intrusive effect
on the intensity of the experience. Instead it should be allowed to evolve freely so
that the members and the group as a whole fully experience it, for as long as the ther-
apist feels it maintains a therapeutic function. Only after this experience is completely
lived out, it can be metabolized and transformed into thought. When the positive and
regenerating aspect of this experience begins to vanish and elements which disturb
therapeutic work begin to come out as well as the anxiety producing elements of the
fusion tied to the fear of losing one's identity, and along with these the anxiety of the
fear of fragmentation tied to emerging relationships with partial objects, only then
should the therapist reveal the dangers in continuing this state and keep it under
control.
300 R. Visintini et aL
Case 2: Apart from obvious individual particularities, the members of the groups with which
we worked seemed to have some common aspects regarding their personal life history. In par-
ticular, as the group work developed, as they communicated their past and present life history,
little by little, for each of these patients the fact emerged that there had been a message from
their environment of origin, not necessarily explicit, but precisely because of this even more
dangerous, of not having the right to a valued life, but at most of surviving. In fact, they were
either almost all unwanted children or had very absent and/or particularly narcissistic parental
figures. The latter consider children only a narcissistic extension of themselves, and therefore
fully accept them only and only if they answer to their expectations, otherwise they give signs
of intolerance or delusion. So these patients never felt wanted, and wanted for what they actu-
ally were, but had to take on suitable behavior or one falsely complacent and/or reactively
transgressive. Other members of the group, instead, had parents with severe pathologies, at
times even with specific disorders such as alcoholism or drug addiction. None of them had suf-
ficiently good family environments nor parental figures who could have acted as reliable and
valid models. In fact, the term environment of origin and not "parents" was implicit, because,
in some cases a real and true family of reference was actually lacking and instead there had
been more or less inadequate parental figures. These persons, even before having become HIV-
positive felt like marked persons and not in the banal social sense of the word but, as Zucca
Alessandrelli (1995) writes: "The mark is the primary terror of not being as a person, that is
of not being significant and valuable as the subject-object of a vital relationship. As the possi-
bility of facing fundamental transformations and thus separations and, the need to abandon
roles without limits. They battle with the fact that they actually have to accept, at a very vig-
orous age of life, the possible end of an "object" fundamental for everyone of us: our own life.
And yet, precisely with the approach of this incredible danger can the possibility emerge of
reproposing their request to be significant.
The group in fact in becoming at work, is considered by the members as an extension
of the self, and as a place where they can be and say, look, listen, and understand. The
feeling of belonging to the group is continuously nourished through the mutual sharing of
personal experiences, above all of anxiety and depression, which since they became communi-
cation, are acknowledged and possibly modified. The experience of belonging is particularly
important for this type of patient, since it is fundamental for the building of a sense of self,
as a person who has the right to live and to have affective space and one to be listened to,
as Neri (1995) also states, often takes on the fundamental function of self-object. An object
which makes the self of the individual emerge and maintains it and gives it significance. At
times it takes on the role of twin self-object which thanks to the warm and affective presence
of other persons gives an essential contribution to building-up the feeling of being: being a
human being among other human beings, and for these persons who so often feel different
this is very important.
In the group, in fact, just the fact of seeing each other and being many, makes the bodily
presence of others much more concrete and explicit, and stimulates the awareness of belong-
ing to an active and functional meeting. All of this, however, in particular with HIV-positive
patients, should not be interpreted but, as in positive fusional phases, should be allowed to be
experienced. One should keep in mind that, for this kind of patient, a facilitating and welcoming
environment represents even if positive, something new and unknown, and as such is fright-
ening and induces defensive reactions, even that of not attending the group sessions. But this
acting out that a patient may choose to express, is also valuable and important communication
for all the members of the group. In fact, when patients who have acted out their fear which
is also that of the other members, in return they are accepted in an environment, which con-
trary to the past, neither looks down on them or bans them, but instead gives them importance
by helping them to understand how they took on the role of "those who are afraid of the new"
for everyone in the group. It is thus possible to highlight that the fear of the new is a lived and
shared experience that thanks to this acting out becomes communication that permits aware-
ness and elaboration. As for the fundamental function of the group as self-object the group
takes on an ideal and omnipotent self-object role, as always occurs in a positive fusional phase.
This object is idealized but not distanced, on the contrary it is experienced as an extension of
the self and allows to experience being a whole with an ideal of calm and strength. It is clear
that from this phase the patients must later pass, and not only once (spiral time), to the healthy
phase of individuation.
The history in the therapeutic group allows patients to go beyond the splitting and
episodical parts of the ego, to share universal human experiences. It also allows patients to
achieve a positive synthesis between the synchronic and diachronic prospect and thus produce
a reverse movement but one complementary to the one towards individualization. Conse-
quently, it also provides the foundations to overcome the fear of separation and solitude, since
it ties the individual to others. Naturally all the above is also true for HIV-positive groups, but
for them history is much more important. The history assures them that they will leave a hered-
ity of affects in ari area of belonging. An area, where their right to a valued life has been
acknowledged and shared, and where consequently the need and right to the value of death
can also be acknowledged, as an indispensable event for all human beings who can be spoken
of, and not as the result of a personal fault. This aspect makes the therapeutic group setting
different from all other social settings.
Today there is the tendency to negate death in a maniacal way and to accuse
and ghettoize HIV-positive patients also because they represent a concrete "memento
mori".
In group work even the therapist must know how to face this reality in an authentic
and deep way and guarantee the patient that the group as a whole knows how to maintain
the memory and testimony of a valuable and significant existence.
It is not a coincidence that when new patients join an HIV-positive group or when
new groups are formed, which also include patients who have participated in groups
and for various reasons no longer had a sufficient number of patients, there is always
a patient who takes on the role of senior of the group. This person becomes spokesman
of the past history and revives the memory of persons who are no longer there and in
a certain way tests the therapist to check if these patients are still present and will
remain present. Moreover, it is not a coincidence that in these groups more than in
others, special events are celebrated that in the family had often never been adequately
celebrated (as christmas and carnival): in these occasions photographs are taken and
given to all the participants and therefore as well to the therapist and observer who
keep them for the group.
Frequently, a supervisor is necessary as a reference to conduct these groups. There
are many difficulties to face and the involvement requested from the therapist is
particularly deep and authentic. Therefore the regulating of emotional distance needed
to face certain themes that corne up from time to time in the best possible way is
problematic.
4. CONCLUSIONS
Based on what has been presented in this chapter, we can confirm that psy-
chological treatment for HIV-positive patients has progressed remarkably in these
past years.
From, an initial condition of scarce knowledge of HIV-infection, when a diagno-
sis was made at an advanced stage of the illness and psychological intervention could
only be the illness acceptation and the accompany towards death (Kubler-Ross, 1989;
Nichols, 1995), we have moved ahead to a condition in which a diagnosis can be made
earlier and both pharmacological and psychotherapeutic treatment can actually
increase expectations and improve the patient's quality of life. In our clinical experi-
302 R. Visintini et aL
ence, along with supportive psychological work to deal with experience tied to the
illness, we apply a specific intervention on the patient's personality traits; this inter-
vention permits to elaborate and/or to contain personality traits which have facilitated
contagion and which can still become self-mutilating or maladaptive.
The psychodynamic psychotherapy is chosen for subjects whose: a) life
expectancy is sufficient; b) need to face unconscious aspects of their disorder; c) insight
is good, and d) disorder is considered, by them, as present even before the illness.
As has been highlighted many times, intrasystem deficiencies (structural defects
of the self, lack of object-constancy, identity diffusion or lack of self-esteem, etc.) are
found in these patients, which also need an empathy-reconstructive psychotherapeutic
intervention. On the opposite pole of a continuum we put an interpretative model
of psychotherapy. These two different therapeutic approaches must keep in mind the
structural reality of the patient. An empathy-reconstructive psychotherapy primarily
deals with the defective personality structure which is typical of a borderline func-
tioning personality.
The empathy-reconstructive model should aim at helping the patient to structure
his/her identity and defective self through the construction of meanings with the use
of affirmative interventions.
The interpretative therapeutic treatment is useful for patients with prevalently
narcisistic functioning, who through the revealing of meaning, can become aware of the
unconscious dynamics that determined their existence.
REFERENCES
AAVV. (1985), La morte oggi, Feltrinelli, Milano.

Adler M.W. (1980), The terrible peril: a historical perspective on the venereal diseases. British Medical
Journal 281:206-211.
Alexander F. (1930), The neurotic character. Int. Psychoanall, 292-311.
Andreoli A, Gressot G., and Aapro N. (1990), Personality disorder as a predictor of outcome. 1 Pers. Disord.,
3:307-321.
American Psychological Association (1994), Diagnostic and Statistical Manual of Mental Disorders, Fourth
Edition, Washington, DC, American Psychological Association.
Ashbach C. and Schermer V. (1987), Object relations, the self and the group-London and New York,
Routledge and Kegan Paul.
Balint M.E. Ed. Or (1959), Thrills and regressions, trad. it (1983): La Regressione, Cortina, Milano.
Barteimer L.H. (1930), The neurotic character as a new psychoanalytic concept. Am. 1 Orthopsychiatry,
1:512-519.
Brandy K.T., Grice IE., and Dustan L. (1993), Gender differences in substance use disorders. Am. 1
Psychiatry, 150:1707-1711.
Brooner RK., Bigelow G.E., Strain E., and Schmidt c.w. (1990), Intravenous drug abusers with antisocial
personality disorder: Increased HIV risk behaviour. Drug Alcohol Dep., 26:3~4.
Brooner RK., Greenfield L., Schimidt C.W., and Bigelow G.E. (1993a), Antisocial personality disorder and
HIV infection among intravenous drug abusers. Am. 1 Psychiatry, 150:5358.
Brooner RK., Herbst IH., Schimidt c.w., Bigelow G.E., and Costa P.T. (1993b), Antisocial personality
disorder among drug abusers. Relations to other personality diagnosis and the Five-Factor model of
personality. 1 Nerv. Ment. Dis., 181:313-319.
Brooner RK., Schmidt c.w., Felch L.I, and Bigelow G.E. (1992), Antisocial behaviour of intravenous
drug abusers: implication for diagnosis of antisocial personality disorder. Am. 1 Psychiatry,
149:482-487.
Cacciola IS., Alterman AI., Rutherford, M.I, and Snider E.C. (1995), Treatment response of antisocial
substance abusers. 1 Nerv. Ment. Dis., 183:166-171.
Cacciola IS., Rutherford M.I, Alterman AI., and Snider E.C. (1994), An examination of the diagnostic
criteria for antisocial personality disorder in substance abusers. 1 Nerv. Ment. Dis., 182:517-523.
mv Infection, Personality Structure, and Psychotherapeutic Treatment 303
Cacciola 1.S., Rutherford M.1., Alterman AI., Mc Kay 1.R, and Snider E.C. (1996), Personality disorders and
treatment outcome in methadone maintenance patients. 1. Nerv. Ment. Dis., 184:234-239.
Carta I. and Clerici M. (1991), The treatment of psychiatric disorders in substance abusers; in Racagni G.
(ed) Biological Psychiatry. Amsterdam, Elsevier Science Publishers, 1991, vol. 2, pp. 51-54.
Clekley (1959), The mask of sanity: an attempt to clarify some issues about the socalled psychopathic
personality, St. Louis, Mosby.
Clerici M., Carta I., and Cazzullo c.L. (1989), Substance abuse and psychopathology. A diagnostic screening
of Italian narcotic addicts. Soc. Psychiatry Psychiatr Epiderniol, 24:219-226.
Compton W.M., Cottler L.B., Shillington AM., Prince RK., and Asmus G.A (1995), Is antisocial personal-
ity disorder associated with increased HIV risk behaviours in cocaine users? Drug Alcohol Depend,
in press.
Corbella S. (1988), La terapia di gruppo, in "Trattato di psicoanalisi" vol. I-Cortina, Milano.
Corbella S. (1996), Femrnine e maschi nel gruppo terapeutico: dagli stereotipi alIa relazione personale. In Gli
Argonauti n. 69, pp. 131-152, CIS Editore.
Cottier L.B., Prince RK., Compton W.M., and Mager D.E. (1995), Subtypes of adult antisocial behaviour
among drug abuser. 1. Nerv. Ment. Dis., 183:154-161.
Dreifuss-Kattan E. (1990), Cancer stories. Creativity and self-repair. Hillsdale, NJ: The Analytic Press.
Ellis D., Collis I., and King M. (1993), A controlled comparison of HIV and general medical referrals to a
liaison psychiatry service. AIDS Care.
Fell M., Newman S., Herns M., Durrance P., Manji H., Connolly S., McAllister R, Weller I., and Harrison M.
(1993); Mood and psychiatric disturbance in HIV & AIDS: Changes over time. Br. 1. Psychiat.,
162:604-610.
Foulkes S.H. Ed. Or. (1964), Trad. it. (1967), Therapeutic group analysis, trad. it.: Psicoterapia e analisi di
gruppo, Boringhieri, Torino.
Foulkes S.H. Ed. Or. (1975), Group-Analytic psychotherapy, trad. it. (1976): La psicoterapia gruppanalitica.
Astrolabio.
Gerstley L., Mc Lellan T.A, Alterman AI., Woody G.E., Luborsky L., and Prout M. (1989), Ability to form
an alliance with the therapist: a possible marker of prognosis for patients with antisocial personality
disorder. Am. 1. Psychiatry, 146:508-512.
Hare RD., Hart S.D., and Harpur T.1. (1991), Psychopathy and the DSM-IV criteria for antisocial personal-
ity disorder. 1. Abnorm Psychol, 100:391-398.
Jacobsberg L., Frances A, and Perry S. (1995), Axis II diagnoses among volunteers for HIV testing and coun-
selling. Am. 1. Psychiat., 192:1222-1224.
Karpman B. (1941), On the need for separating psychopathy into two distinct clinical types: symptomatic and
idiopathic. 1. Criminol. Psychopathol., 3:112-137.
Kernberg O. Ed. Or (1975), Borderline conditions and pathological narcissism, trad. it. (1978): Sindromi mar-
ginali e narcisismo patologieo, Boringhieri, Torino.
Kernberg 0. (1996), "A psychoanalytic theory of personality disorders", in Clarkin 1.E and Lenzenweger
M.E, Major theories of personality disorder, The Guilford Publications, New York.
Kohut H. (1982), The search for the self, trad. it.: La ricerca del st!, Boringhieri, Torino.
Kubler-Ross E. (1989), AIDS: the ultimate challenge, Raffacllo Cortina Editore, Milano.
Lazzari c., Campione E, and Chiodo E (1993), "Aspetti psicologici e sociali dell'epidernia di Aids" in Gli
Argonauti, n. 56, pp. 61-76. CIS Editore.
Luborsky L. and Singer B. (1975), Comparative studies of psychotherapies: Is it true that "everybody has
won and all must have prizes"? Arch. Gen. Psychiat., 32:995-1008.
Luborsky L., Crits-Cristoph P., and Alexander L. (1983), Two helping alliance methods for predicting
outcomes of psychoterapy: a counting signs vs. a global rating method. 1. Nerv. Ment. Dis., 171:480-492.
Lyttle 1. Column in the Independent newspaper, 28106196.
Maffei C. (1993), II disturbo borderline di personalita. Bollati Boringhieri, Torino.
Mariotti G. (1987), La morte annunciata. "Nuova unita sanitaria" n. 2, pp. 70-78. Lalli Editore.
Menzies I.E.P. (1961), The functioning of social systems as a defence against anxiety. Tavistock Pamphlet no
3, London: Tavistock Publications.
Neri C. (1995), Gruppo-Borla, Roma.
Paris 1. (1995), Borderline personality disorder. Cortina, Milano.
Perkins D.O., Davidson E.1., Leserman 1., Liao D., and Evans D.L. (1993), Personality disorder in
patients infected with HIV: a controlled study with implications for clinical care. Am. 1. Psychiat.,
150:309-315.
Robins L. (1966), Deviant children grow-up: a sociological and psychiatric study of sociopathic personality,
Baltimore, Williams and Wilkins.
Rounsaville B.1., Glazer W., and Wilber c.H. (1983), Shortterm interpersonal psychotherapy in methadone-
maintenance opiate addicts. Arch. Gen. Psychiat., 40:629-636.
Sini C. (1985), Irrazionalita e morte, in AA.VV., "La morte oggi", cit., pp. 69-76.
Strupp H.H. (1980), Success and failure in time-limited psychotherapy: Further Evidence. Arch. Gen. Psy-
chiat., 37:947-954.
Tempesta E., Pirronelli c., Laniri L., Clerici M., and Carta I. (1987), Appraisal of psichiatric disorders inci-
dence according to DSM-III in two populations of addicts. Acta. Med. Romana, 1:96-108.
Wanigarative S., Billington A., and Williams M. (1996), Initiating and maintaining safer sex: evaluation
of group work with gay men. The impacts of AIDS: psychological and social aspects of HIV infection,
in Sherr L., Catalan 1., and Hedge B. (eds) Harwood Academic Publishers; London.
Whittles 1. (1937), The Criminal Psychopath in the psychoanalytic system. Psychoanal Rev., 24:276-283.
Wikstrom O. (1990), Den outgrundliga manniskan. Livsfragor, psykoterapi och sjalavard. (The inscrutable
individual. Existential concerns, psychotherapy and cure of souls) Stockholm: Natur och Kultur.
Williams D., Gilbart v., and Williams 1. (1996), Behaviourial factors associated with recent HIV serocon-
version in gay men (manuscript in preparation).
Wilson 1.0. (1980), After Affluence: Economics to meet human needs. Harper & Row. New York.
Woody G.E., Luborsky L., Me Lellan T.A., O'Brien c.P., Beck A.T., Blane 1., Herman I., and Hole A. (1983),
Psychotherapy for opiate addicts. Arch. Gen. Psychiatry, 40:639-645
Woody G.E., Me LeHan T.A., Luborsky L., and O'Brien c.P. (1985), Sociopathy and psychotherapy outcome.
Arch. Gen. Psychiatry, 42:1081-1086.
Woody G.E., Mc Lellan T.A., Luborsky L., and O'Brien c.P. (1987), Twelve month follow-up of psycho-
therapy for opiate dependence. Am. 1. Psychiatry, 144:590-596.
Zucca Alessandrelli C. (1995), "II marchio" in Gli argonauti n. 66, pp. 231-244. CIS Editore.
INDEX
AA model, 249 Constitutional factor, 23

Acting out, 171 Containment, 59
Active-passivity, 191 Contingency-management, 199
ADHD, 11,90 Co-occurrence, 241
Adolescence, 77 Counter-identification, 83
Adoption, 12 Countertransference, 172
Agoraphobia, 139 Crisis behavior, 202
Alienation, 90 Cyclic pattern, 185
Anal conflict, 26
Anti-agressive, 216
Day Unit treatment, 265
Antipsychotics, 207, 224
Defense mechanism, 39
Anti-serotonergic,217
Defense style, \09
Approval seeking, 276
Defensive function, 47
Attachment, 54, 56, 63, 73, 278
Defined Daily Dosage, 260
Attenuation hypothesis, 144
Dependent child, 277
Avoidance, 145,203
Dependent personality disorder, 269, 286
Depersonalization, 234
Bad object, 71
Depressive disorder, 138
Basic dispositional traits, 35
Depressive transference, 180
Behavioral analysis, 161
Deprivation, 58
Behavioral deficit, 184
Designer treatments, 41
Behavioral dyscontrol, 256
Desynchrony, 192
Behavioral repertoire, 195
Developmental phase, 140
Benzodiazepines, 207, 225
Dexamethasone suppresion test, 232
Big Five, 4
Diagnostic homogeneity, 127
Biological marker, 1\0
Dialectical process, 97
Bio-psycho-social model, 72
Dimensional,147
Bipolar, 222
Dimensional evaluation, 91
Borderline adolescent, 86
Dimensional models, 120
Borderline functioning, 291
Dissociative episodes, 70
BORRTI,93
Distress tolerance, 197, 204
DMs profile, 46, 49
Carbamazepine, 224, 231
Drop out, 45, 50, 211, 213
Chaotic condensation, 181
Dynamic formulations, 42
Childhood abuse, 61
Dyscontrol, 212, 230
Chlorpromazine, 208
Dysfunctional levels, 112
Classical antidepressant, 220
Dysregulation, 231
Clinical characteristic, 98
Dysthymia, 218, 257
Cognitive-behavioral therapy, 108
Cognitive disorganization, 20
Commitment, 201 Early childhood, 69
Comorbidity, 129, 157 Early trauma, 298
Comorbidity model, 144 Eating disorder, 48
Conduct problem, 89 Egosyntonic, 34, 156
305
306 Index
Emotion dysregulation, 183, 184 Maternal sensitivity, 55

Etiology. 103 MCMI, 136, 243, 248
Evolution, 102, 103, 272 Medication group, 258
Extemal reality, 170 Mentalizing, 57, 64
Metabolizing, 296
Five Factor Model, 146,274 Mind representation, 73
Functional analysis, 163 MMPI-2,158
Functional level, 44 Multiple crisis, 193
Fusion, 300 Multi-problem. 196
GABA,221 Narcissistic despair, 244

GAF, 164,260,263 Needle sharing, 295
Gay community, 294 Negative relating, 272, 274
Genetics, 1, 246, 278 Negative symptoms, 209
Genotype, 3 Negative transference, 168
Gestalt, 106 Negotiating power, 281
Grandiose self, 245, 299 Neuroleptics, 259, 264
Grandiosity, 28 Neuroticism, 148
Grief,74. 194 Neurotransmitter, 21
Nonspecific tranquillizer, 215
Hallucination, 209 Novelty seeking, 6
Heritability, 9
High level, 290 Obsessive-compulsive, 145
HIV, 290, 297 Oedipal triangular, 176
HIV-positive, 292, 301 Omnipotence, 291
Orthogonality, 143
Identity diffusion, 62, 168 Overcompliant, 114
Identity problem, 8
IOU, 295 Panic disorder, 130
Image distorting, 50 Paradoxical, 233
Immune system, 100, 101 Paraphilias, 121
Impulse cluster, 113, 122 Parasuicide, 189
Impulsivity,230 Personality organization, 119
Incestuous desire, 87 Pervasive, 88
Instable relationship, 188 Phenotype, 2
Internal representation, 56, 279 Placebo, 210, 212, 214
Internal working models, 80 Polarity, 188
Interpersonal circle, 275 Post-eclecticism, 105
Interpersonal effectiveness, 205 Postive relating, 275
Interpersonal locus, 155 Posttraumatic, 229
Interpersonal octagon, 271, 273, 284 Pre-oedipal, 293
Interpersonal theory, 270 Primary caregiver, 53, 63
Intervention techniques, 99 Projective identification, 180
Intrapsychic processes, 275 Promiscuous behavior, 222
Intrapsychic reality, 51 Protocol, 162
Prototypical, 123
Lability, 27 Psychobiological structures, 33
Life-expectance, 302 Psychological dependence, 270
Life-threatening, 198 Psychological mindedness, 265
Lithium, 221, 224 Psychological self, 60
Long-term chronicity, 107 Psychotic regression, 175
Low level, 290 Psychotic transference, 178, 179
Lowerness, 272 Psychotic-like, 234, 236
Psychotropic, 255
Maladaptive functioning, 104
Maltreatment, 65 Rapprochement sub-phase, 79
MAO, 225 Reactivation, 176
Marital partner, 283 Reflective function scale, 64
Maturation. 31 Regression, 40, 173
Index 307
Schizoid,5 Suicide bid, 280

Schizotypal, 141,210 Superego, 177,293
Schizotypy, 10 Supportive technique, 167
SCID-II-R, 135 Supportive-expressive, 296
Self harming behavior, 32
Self identity, 30 Tardive dyskinesia, 209
Self system, 29 TCI,7,I59
Self-destructive behavior, 235 Temperamental features, 126
Self-disclosure, 200 Temper-tantrums, 15
Self-generated behavior, 186 Temporal stability, 93
Self-mutilation, 223, 291 Therapeutic alliance, 43
Self-sufficiency, 84 Therapeutic community, 62, 247
Self-validation, 190 Three-step procedure, 169
Separation-individuation, 80 Transgenerational, 81, 82
Serotonergic system, 14 Transition phase, 299
Sexual abuse, 85 Treatment policy, 160
Sexual conflict, 171 Twin study, 13
Side effects, 213
Social phobia, 22, 130, 142 Unidimensional model, 124, 126
Social learning theory, II Upperness, 272, 282
Splitting, 78, 86, 174
SSRI, 19,219,256 Vulnerability, 101,202, 161, 187,242
Substance abuse, 241, 250
Suicidal behavior, 212 Zen, 197

Treatment of Personality Disorders

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Treatment of Personality Disorders

Uploaded by

Copyright:

Available Formats

TREATMENT OF

Springer Science+Business Media, LLC

© 1999 Springer Science+Business Media New York

All rights reserved

Once we have established an accurate diagnosis, we can offer a range of treat-

Part I: Biological Foundations of Personality Disorders

1. Genetics of Personality Disorders ................................. 1

2. Psychobiology of Personality Disorders: Implications for the Clinic 17

3. The Implications of Recent Research on the Etiology and Stability of

Part II: Psychological Foundations of Personality Disorders

4. Psychodynamic Research Can Help Us to Improve Diagnosis and

5. Attachment, the Development of the Self, and Its Pathology in

6. Trauma and Personality .......................................... 69

7. Adolescence and Personality Disorders: Current Perspectives on a

Part III: Diagnostic Models of Personality Disorders

8. Integrative Perspectives on the Personality Disorders 97

9. A Multidimensional Approach to Personality Disorders and Their

10. The Structure of DSM-IV Borderline Personality Disorder and Its

11. The Relationship between Anxiety Disorders and Personality Disorders:

Part IV: Treatment of Personality Disorders

12. Psychodiagnostics and Indications for Treatment in Cases of Personality

13. The Psychotherapeutic Treatment of Borderline Patients 167

14. Functional Analysis of Borderline Personality Disorder Behavioral

15. Psychopharmacological Treatment of Personality Disorders: A Review 207

17. The Narcissistic Personality Disorder and Addiction. . . . . . . . . . . . . . . . . . 241

18. Pharmacotherapy for Patients with Personality Disorders: Experiences

19. A New Interpersonal Theory and the Treatment of Dependent

20. HIV Infection, Personality Structure, and Psychotherapeutic

GENETICS OF PERSONALITY DISORDERS

Svenn Torgersen 1 and Marco Battaglia2

Department of Neuropsychiatric Sciences

When we wonder about whether genetic factors influence the development of a

liability-or vulnerability-appears to interact with environmental factors in

In general, it should be remembered that when data show a continuous distribu-

3. COMMON PERSONALITY DIMENSIONS

Table 1. The genetic variance (G), the shared-in-families

1) Adjusted from Pedersen et aI., 1991; Bergeman et aI., 1993.

Not all personality dimensions embodied in the different available questionnaires

phenomenon may be understood in terms of interactions between independently inher-

Livesley et ai. (1993) have measured by means of a questionnaire personality

5. PERSONALITY DISORDERS EVALUATED THROUGH

6. ANTISOCIAL AND SCHIZOTYPAL

While it has been traditional to seek out environmentally based or psychologi-

cognitive/attentional dysfunction are not only associated with chronic schizophrenia

The capacity of an individual to organize their perceptions of the world around

A hallmark feature of many of the Cluster B or dramatic cluster personality

involved in other risky or self-destructive behaviors such as binge-eating, gambling, or

The cluster C personality disorder diagnoses are marked by persistent maladap-

THE IMPLICATIONS OF RECENT RESEARCH

Recent empirical research on normal and disordered personality suggests the

1. THE CAUSES OF PERSONALITY DISORDER

Table 1. Higher-order and basic dimensions of personality disorder

2. STABILITY AND CHANGE IN PERSONALITY

To find information on the stability of personality across the lifespan we need to

there is considerable variability across the different components of personality in sta-

Figure 1. Personality System.

3. CHANGE AND THE STRUCTURE OF TRAITS

4. STRATEGIES FOR PROMOTING MORE ADAPTIVE

Traits may be expressed through characteristic expressions that are maladaptive

4.1. Progressive Substitution of More Adaptive Expressions

4.2. Attenuating Trait Expression through the Acquisition of Strategies to

4.3. Increasing the Acceptance and Tolerance of Basic

The results of both etiological studies and investigations of personality change

treatment would involve the integration of diverse therapeutic interventions selected