Professional Documents
Culture Documents
Agricola Rintje
Agricola Rintje
Rintje Agricola
The Rise and Fall of the Hip
Rintje Agricola
The printing of this thesis was financially supported by:
ISBN: 978-94-91487-19-4
Proefschrift
door
Rintje Agricola
geboren te Heerenveen
Promotiecommissie
Chapter 12 Cam impingement of the hip: a risk factor for hip 189
osteoarthritis
Nature Reviews Rheumatology 2013;9(10):630-634
References 239
PhD portfolio 261
List of publications 267
Curriculum vitae 271
Acknowledgements / Dankwoord 275
Chapter 1
Introduction
11
Chapter 1
Figure 1: Coronal section of a specimen of a nine year old boy, with unknown cause
of death. From the anatomical department (prof G.J. Kleinrensink) of Erasmus Medical
Centre, Rotterdam, the Netherlands.
A coordinated growth of these three growth fronts determines the size and
shape of the adult proximal femur. A disturbance in any of these growth fronts
will lead to alterations in the final shape of the proximal femur.6
There are indications that certain shape variants of the hip lead to a higher risk of
development of hip OA. It is therefore of great relevance to study if such shape
variants are genetically imprinted and thus not modifiable, or whether they are
influenced by external (and possibly modifiable) factors. A prerequisite for the
latter is that bone can be affected by external factors. It is known that bone is
not an indolent tissue, but that it has the capacity to constantly renew itself. This
is realized by osteoclasts and osteoblasts working together in structural basic
multicellular units (BMUs). Osteoclasts are cells that resorb old bone, which are
followed by osteoblasts that deposit the new bone matrix. The possibility for
self renewal, or remodeling, is essential for repairing micro-damage in the bone
12
Chapter 1
General introduction
tissue and fracture healing. During growth, remodeling takes also place to adapt
its structure to changing mechanical loads, and is then referred to as modeling.
More than a century ago, Roux and Wolff already observed that bone is capable
to adapt itself to its mechanical environment.7,8 This has been well illustrated
by the trabecular structure of the proximal femur, which is aligned along the
principle loading directions.9,10 This has led to the idea that external forces as a
result of sporting activities, specifically during the period of skeletal growth, can
influence bone architecture and bone morphology.
Trabecular bone architecture is often expressed in bone mineral density
(BMD) as measured by DEXA scans. Multiple studies using DEXA have shown
that athletic activities during growth, and especially those with high impact,
stimulate bone formation and lead to a higher BMD. This has been illustrated
by studies showing that the dominant arm of professional tennis players has
a much higher bone mass than the contra lateral arm.11-13 In addition, many
studies showed a higher bone mass at specific locations in the proximal femur of
athletes who participated in high impact sports such as soccer, basketball, and
jumping sports.14-17 Thus, by increasing external loading of the hip, internal bone
formation occurs as an adaptive process.9 Conversely, when the loads applied
to bone decrease, an increase in bone resorption and a subsequent decrease
in bone formation takes place as in cosmonauts during a space flight or during
prolonged bed rest.18,19
Figure 2. Five weeks after rotationplasty (left) and 7 years postoperatively (right), with
increasing centering of what now appears as the femoral head. Note the orientation of the
medial part of the growth plate towards the vector of force.
13
Chapter 1
Bone morphology refers to the surface shape of bones and might be influenced
by sporting activities in the same manner as bone architecture, though the
actual shape of bones is much less studied. This is primarily due to the radiation
load of the radiographs or Computed Tomography (CT) scans needed to study
bone morphology in adolescents. Bone is a dynamic tissue and external loads
applied to it might largely determine its shape, which is strikingly demonstrated
during follow-up after a hip rotationplasty (figure 2).20 During hip rotationplasty,
the femur is completely resected (e.g. due to a malignant tumor) and the
proximal tibia is placed into the acetabulum to avoid prosthetic replacement or
amputation at the level of the hip. Figure 2 shows that years after this procedure
in a 5 years old girl, the bone shape has adapted towards the shape of a proximal
femur, despite it was genetically imprinted to become a proximal tibia.
Assuming that the shape of the hip can change resulting from the loads applied
to it during skeletal maturation, an abnormal hip morphology might be a result
of high impact sporting activities during skeletal growth. This phenomenon has
already been demonstrated in other joints: in shoulders of elite baseball players
there is more proximal humeral retrotorsion and glenoid retroversion in their
throwing than in their nonthrowing arm.21 Playing soccer during childhood and
adolescence is associated with development of genu varum (bowlegs) in the
knees.22 For the hip, there is only one study available that investigates whether
different loading conditions, especially those experienced by athletic activities,
can influence the shape of the hip.23
In 1965 Murray hypothesized a ‘tilt deformity’ to be a causative factor in hip OA
and described the tilt deformity as a slight SCFE.24 In 1971 it was consequently
studied whether this tilt deformity –nowadays known as a cam deformity-
might have been a result of chronic stress during the years of adolescence. He
investigated three groups of young mature males between 17 and 21 years of
age with different athletic backgrounds. The first group were boys at a boarding
school well known both for its intellectual standards and athletic successes;
the second group attended an identical school, though athletic activities were
more voluntary and practiced less; and finally, a third group comprised boys
from the same age working in industry, of which many had left state schools
at the age of 15, though most had continued with some athletic activity. The
prevalence of a tilt deformity on AP radiographs differed significantly between
the groups, being 24% in the first group, 9% in the second group, and 15% in
the third group. It was suggested that certain forms of sports, especially when
undertaken compulsorily during adolescence, might precipitate the abnormality.
Another indication that a cam deformity might develop during skeletal growth
14
Chapter 1
General introduction
was provided by Siebenrock et al., who showed that a cam deformity in adults
was associated with an unusual extension of the epiphyseal scar onto the femoral
neck.25 However, prospective studies in adolescents are not available. As recent
studies indicate that certain shape variants of the hip pose a higher risk of
hip OA, a better knowledge on how these shape variants of the hip develop is
required.
Figure 3. A. A healthy hip without signs of osteoarthritis B. A hip with radiographic signs
of advanced osteoarthritis including large osteophytes, joint space narrowing (indirect
measure of cartilage loss), subchondral sclerosis and joint deformity.
15
Chapter 1
OA in the Netherlands have been estimated to be €722 per patient per month.31
A precise definition of the disease has been difficult to determine and the
prevalence of OA is therefore difficult to express in one number. There is often a
discrepancy between the clinical presentation and the radiographic evidence of
OA. In research, the commonly used definitions of hip OA include ‘symptomatic
OA’ as quantified by the ACR criteria32, ‘radiographic OA’ as quantified by
for example the Kellgren and Lawrence scale33, or it is defined by total joint
replacement as a result of OA.
OA can affect any synovial joint, but joints most affected by OA are in order
of decreasing prevalence the spine (cervical and lumbar), phalanges (distal
interphalangeal joint and the first carpometacarpal joint), the knee, and the
hip.34 Besides the definition used, the reported prevalence of hip OA depends
largely on the characteristics of the study population such as age, gender, and
ethnicity. In the US, the prevalence of radiographic hip OA (K&L>1) in population
based studies ranges from 11.9% to 27.6%.35,36 In the Netherlands, a population
based study in the region of Zoetermeer reported a prevalence of around 10%
having a K&L grade >1.34 Another estimation of the incidence of hip OA is
provided by the first annual report of the Dutch registry of orthopaedic implants
(LROI). In 2011, an incidence of approximately 26.000 total hip replacements
due to osteoarthritis was reported.37 This number is expected to increase in the
coming years and there is a trend towards a younger age of patients needing a
primary hip replacement.
16
Chapter 1
General introduction
not a clear risk factor in contrast to other types of OA, which are generally more
prevalent in females than in males.
The influence of obesity on the risk of developing hip OA is not clear. In contrast
to the knee in which BMI is a definite risk factor for OA, obesity is not associated
with radiographic hip OA and only moderately associated (OR around 2) with
symptomatic hip OA.42,43 It has therefore even been suggested that obesity can
be regarded as a negligible risk factor for hip OA.44
There is consistent evidence that occupational factors lead to hip OA.45 A
systematic review reported ORs between 1.1 and 13.8 for heavy physical
workload and development of hip OA, when compared with light physical work.46
Typical examples of people with heavy workload are farmers, dockers and
bricklayers.
Sporting activities confer a moderately increased risk for development of hip
OA (OR around 2) and a clear dose-response relationship was found.47 A higher
level or frequency of sporting activities raised the OR substantially, and in this
subgroup, hip OA also developed at a younger age.48 A recent Swedish cohort
study of over 2000 individuals compared former elite athletes with matched
controls.48 They found that the risk of hip OA was doubled and the risk of total
hip replacement (THR) 2.5 times higher amongst the former elite athletes. The
OR of both knee OA and total knee replacement (TKR) was 1.6 among the
former athletes. Interestingly, when adjusted for soft tissue knee injury, the
OR for knee OA and TKR became non significant whereas the OR for hip OA
and THR became most pronounced among athletes that had participated in
high impact sports (OR around 3-4 for soccer, handball, and ice hockey). This
suggests that the association between high level sporting activities and knee
OA is primarily explained by a higher prevalence of anterior cruciate ligament
injury and meniscal tears, whereas local biomechanical factors probably explain
the association between high level (and high impact) sporting activities and hip
OA. A possible pathophysiological explanation might be that of a non-optimal
shape of the hip.
17
Chapter 1
slipped capital femoral epiphysis (SCFE), and congenital hip dysplasia pose a
high risk for early development of hip OA.51,52 OA as a result of these conditions
is therefore also called ‘secondary’ OA in contrast to primary OA in which the
cause is unknown. However, these childhood hip disease are responsible for
less than 10% of all primary THR cases.28 Half a century ago, it was suggested
that many cases of ‘primary’ or ‘idiopathic’ hip OA are actually the result of
previously unrecognized, more subtle shape deformities. This is illustrated by
the fact that also mild non-clinical forms of acetabular dysplasia have been
associated with development of OA.53 Furthermore, a non-spherical femoral
head was recognized as a potential risk factor for hip OA and first described in
the United Kingdom as a ‘tilt deformity’ by Murray and later in the US as a ‘pistol
grip deformity’ by Stulberg et al.54 It was not until a decade ago that a non-
spherical femoral head became of great interest when the pathomechanism by
which it could lead to hip OA was hypothesized by Ganz and colleagues.55
Femoroacetabular impingement
A pathophysiological explanation on how subtle shape deformities of the hip
might lead to OA was provided by the Swiss group from Ganz and colleagues. In
2001, they published a technique for a safe surgical dislocation of the hip, which
allows an almost complete visualisation of the femoral head, and described
the observation that acetabular chondrolabral damage often co-existed with a
nonspherical shaped femoral head.56 In 2003 they proposed the mechanism
of femoroacetabular impingement (FAI), which can, in turn, cause motion-
dependent soft-tissue damage.55 FAI became of great interest in recent years
and is the main topic of the current thesis.
FAI is a condition of a motion dependent abnormal contact between the femoral
head-neck junction and the acetabulum, due to a bone shape abnormality
on either the femoral or acetabular side. Based on the location of the shape
abnormality, two types of FAI can be distinguished: cam impingement (femoral
side) and pincer impingement (acetabular side) (figure 4).57
Pincer impingement is caused by over coverage of the acetabulum relative to
the femoral head, known as a pincer deformity. The hypothesis proposed by
Ganz et al. states that the femoral neck causes an abnormal linear contact
against the pincer deformity during terminal motion of the hip. It has classically
been described to occur primarily in middle-aged women, though no studies are
available to confirm this. Initially, labral damage is the main characteristic as
the cartilaginous labrum might be crushed between the acetabular bony rim and
femoral bony neck. When there are repetitive episodes of pincer impingement,
18
Chapter 1
General introduction
19
Chapter 1
20
Chapter 1
General introduction
The three parts of this thesis will cover the spectrum from the rise (skeletal
development) to the fall (osteoarthritis) of the hip. In order to present the aims
of this thesis in a sequence of general morphology to more specific FAI related
morphology, the relationship between hip shape and OA will be presented first
followed by the definition of FAI and how FAI morphology develops during
skeletal maturation.
21
Chapter 1
focusing on the femoral side of the hip joint. The association between a cam
deformity and cam impingement with hip OA was studied: the first prospective
cohort study on this topic in literature.
22
PART 1
Osteoarthritis:
Is it possible to diagnose and
treat early disease?
Lancet accepted
Chapter 2
ABSTRACT
Globally, osteoarthritis is a major source of pain, disability, and socioeconomic
cost. The epidemiology of osteoarthritis is complex and multifactorial, with
genetic, biological, and biomechanical components. Aetiological factors are
also joint specific. Joint replacement is an effective treatment for symptomatic
end-stage disease, however, there can be adverse functional outcomes and the
lifespan of prostheses is limited. Consequently, there is an increasing shift in
focus to disease prevention and the treatment of early osteoarthritis. This is a
challenging task given conventional imaging techniques are only able to detect
relatively advanced disease and there is a poor correlation between pain and
structural degeneration. Nevertheless, recent advances in both imaging and
biochemical markers offer potential as diagnostic tools and outcome measures
for novel treatments. Joint-preserving interventions under development include
lifestyle modification, pharmaceutical and surgical modalities. Some show
significant potential, however, at present few demonstrate a proven ability to
arrest or delay disease progression.
26
Osteoarthritis: an overview
Chapter 2
INTRODUCTION
Osteoarthritis is the most common joint disease world-wide affecting an
estimated 10% of males and 18% of females over 60 years of age.68 The
resultant pain and loss of function can be debilitating and in developed countries
represents a large socioeconomic burden, costing between 1% and 2.5% of
gross domestic product.69 Osteoarthritis treatment traditionally comprises pain
management with joint replacement for end-stage disease.70-72 This approach
does not address the morbidity associated with early disease or the limitations
of arthroplasty surgery, which include the possibility of adverse outcomes and
the finite lifespan of prostheses. An improved understanding of osteoarthritis
pathogenesis combined with superior assays of disease activity is facilitating a
shift in focus to the prevention and treatment of early osteoarthritis. Furthermore,
identification of different disease phenotypes may allow personalised patient
care. This seminar provides an update of recent developments in the prevention
and treatment of early disease.
EPIDEMIOLOGY
The identification of risk factors is central to understanding disease aetiology
and selecting targets for osteoarthritis prevention and treatment. Longitudinal
studies of large population cohorts have provided important insights and there
is an increasing appreciation that osteoarthritis develops through the action of
hostile biomechanics upon a susceptible joint. Biological pathways within a joint
are mechanosensitive,73 and biomechanical factors may be modifiable and offer
a potential means of intervention.
Joint biomechanics are dictated by anatomical and functional factors. Anatomical
factors include joint morphology. Hip dysplasia, where there is reduced
acetabular coverage of the femoral head, is a long established risk factor for
osteoarthritis.74 Femoroacetabular impingement, where there is abnormal
contact between the proximal femur and acetabulum can confer up to a 10-
fold increased risk of developing end-stage hip osteoarthritis within five years
(figure 1 and 2). However, depending on the characteristics of the cohort and
definition of abnormal morphology, the positive predictive value ranges from
6-25%, whereas the negative predictive value ranges from 98-99%.75 Similarly,
tibial and femoral bone morphology can predict the development of knee
osteoarthritis.76 Limb alignment also appears to be critical and recent studies
provide further evidence that varus and valgus knee alignment increases the
risk of osteoarthritis development and progression in the more loaded region of
27
Chapter 2
Figure 1: Anteroposterior and Lateral Radiographs of a normal hip (A and C) and a hip
with cam lesion femoroacetabular impingement morphology (B and D). In a normal
hip, the concavity of the femoral head-neck junction allows an extensive range of hip
movement without impingement of the femur against the acetabular rim. In cam lesion
femoroactebular impingement, the loss of this concavity at the anterior and superior head-
neck junction results in impaction of the femur against the acetabular rim when the hip
moves into flexion and internal rotation. This results in damage to the labrum, and may
progress to involve the acetabular cartilage with development of osteoarthritis. Surgery to
excise bone and reproduce a head-neck concavity is proposed as a means of preventing
the development and progression of hip osteoarthritis.
28
Osteoarthritis: an overview
Chapter 2
by systemic factors. Age remains the strongest risk factor for osteoarthritis82 and
this may reflect a reduction in regenerative capacity and accumulation of risk
factors. Osteoarthritis is also more common in females and although the role of
oestrogens has been extensively investigated, the mechanism remains unclear.
The material properties of bone may infer some susceptibility. Higher systemic
bone mineral density appears to increase the risk of incident osteoarthritis but
not disease progression.83
Injury may cause bone or cartilage damage that makes the joint more susceptible
to further insult, and damage to ligaments or meniscus may adversely affect
joint biomechanics. Knee injury confers a greater than four-fold increased
risk of developing knee osteoarthritis.84 Obesity increases the load on weight-
bearing joints, but may also increase joint susceptibility through the action of
inflammatory adipokines.85 It confers a three-fold increased risk of developing
knee osteoarthritis86 and accelerates disease progression.87 It remains a mystery
why obesity confers a far smaller risk of developing hip osteoarthritis.88 Given the
increasing prevalence of obesity, this risk factor is responsible for a significant
disease burden.
Labrum
Subchondral Acetabular
Bone Cartilage
Figure 2: Arthroscopic appearance of the hip of a patient with cam lesion femoroacetabular
impingement. The aspherical femoral head enters the acetabulum on hip flexion and
internal rotation leading to delamination of the acetabular cartilage from the underlying
subchondral bone and the development of osteoarthritis.
The strong genetic basis for osteoarthritis has been recognised for many years
through family-based studies. Genome-wide association studies, such as that
performed by the Arthritis Research UK Osteoarthritis Genetics (arcOGEN)
Consortium,89 have now identified 11 loci associated with osteoarthritis. The
29
Chapter 2
effect sizes are small (odds ratio 1.11 – 1.21),89 but consistent with those for
other similar complex traits. It is unlikely that genomics alone will reliably identify
individuals who will develop disease, but may reveal new biological insights
into disease pathogenesis for individual joints. Single-nucleotide polymorphisms
have been associated with a number of known risk factors including hip shape,
body mass index and bone mineral density.90
PATHOGENESIS
Osteoarthritis was once viewed as a disease of purely mechanical cartilage
degradation, however, it is now appreciated to be a complex condition involving
the whole joint in which the activation of specific matrix proteases plays a pivotal
role (figure 3). The possibility that diverse risk factors give rise to osteoarthritis
through a common end pathway offers exciting therapeutic potential. It is likely
that cartilage, subchondral bone and synovium all play key roles in disease
pathogenesis and there may also be an association with systemic inflammation.
Cartilage
The main structural protein of cartilage is type II collagen. This provides a
meshwork that receives stabilisation from other collagen types and non-
30
Osteoarthritis: an overview
Chapter 2
collagenous proteins, such as cartilage oligomeric matrix protein (COMP), and
provides cartilage with tensile strength. Aggrecan and other proteoglycans are
embedded within this framework, which draw water into the cartilage, providing
compressive resistance. Cartilage architecture and biochemical composition are
strictly regulated by chondrocytes in response to changes in their chemical and
mechanical environment.91 On activation, they produce a number of inflammatory
response proteins such as cytokines, including IL-1β, IL-6 and TNF-α, and matrix
degrading enzymes including the metalloproteinases (MMPs) and a disintegrin
and metalloproteinase with thrombospondin-like motifs (ADAMTSes). Some of
these appear to have important pathogenic effects, such as the collagenases
(MMP-1, 3 and 13) and aggrecan-degrading-degrading enzymes (ADAMTS-4 and
5). Others may have beneficial matrix remodelling roles in healthy cartilage.73
Proteases, including ADAMTS-5, are upregulated in a highly mechanosensitive
fashion in mice following surgical joint destabilisation, and are downregulated
on joint immobilisation.92 This supports a potential role for addressing hostile
joint biomechanics as a preventative strategy.
The innate immune system is activated in osteoarthritis. Chondrocytes
express multiple Toll-Like Receptors (TLRs)93 activated by damage-associated
molecular patterns (DAMPs). In osteoarthritis, DAMPS comprise extracellular
matrix molecules that include the glycosaminoglycan hyaluronan.94 Calcium
pyrophosphate and sodium urate crystals also bind chondrocyte TLR receptors
and may therefore play a role in the aetiology of osteoarthritis.95 The finding
that the expression and activation of complement is abnormally high in human
osteoarthritic joints96 is intriguing. COMP is a potent activator of the alternative
complement pathway97 while proteoglycans such as fibromodulin target the
classical pathway.98 Chondrocytes also express receptors that bind advanced
glycation end-products (AGEs)99 which accumulate in aging tissues. This results
in a phenotypic shift to catabolism100 and may help explain the increasing
prevalence of osteoarthritis with age.
Rather than initiating osteoarthritis, responses to extracellular matrix
components may simply reflect amplification of established cartilage degradation.
Chondrocytes may first be activated by inflammatory signals originating from
other joint structures such as synovium or subchondral bone. This warrants
elucidation given therapeutic interventions are more likely to be effective when
acting further upstream.
31
Chapter 2
Subchondral Bone
Subchondral cortical bone forms an interface between the calcified cartilage
below the ‘tidemark’ and the underlying trabecular bone. Marked changes are
seen in the structure and composition of both the cortical plate and trabecular
bone in osteoarthritis.101,102 In fact, aspects of endochondral ossification
are reinitiated in osteoarthritis and the tidemark advances with associated
vascular penetration. This is accompanied by the formation of osteophytes and
subchondral cysts. Advances in imaging now allow bone marrow lesions to be
identified on MRI that correlate with a number of histological changes including
microfractures at different stages of healing.103 These lesions localise to areas
with the most severe cartilage damage. Whether these pre-date changes in
cartilage or occur as a consequence of cartilage damage is unclear. Some
studies suggest that changes in subchondral bone and osteophyte formation
precede cartilage degeneration,104,105 but such studies are always biased by the
sensitivity of the detection methodology.106 More recent work demonstrates
that like chondrocytes, osteoblasts respond to mechanical stimulation with the
expression of inflammatory cytokines and degradative enzymes.107 These may
act directly on cartilage, or changes in the mechanical properties of subchondral
bone may have deleterious effects on overlying cartilage. Conversely, subchondral
bone remodelling may result from increased loading through loss of cartilage
integrity. Being highly innervated, subchondral bone probably contributes to the
generation of pain in disease.
Synovium
Synovitis is a frequent feature of osteoarthritis, even in early disease. In
established osteoarthritis there is proliferation of resident synoviocytes and
tissue hypertrophy, with increased vascularity.108 Synoviocytes synthesise
lubricants such as hyaluronic acid109 and lubricin.110 These contribute to optimal
joint function but display reduced lubricating capacity in some patients with
osteoarthritis.109,111 Like chondrocytes and osteoblasts, synoviocytes also
release inflammatory mediators and degradative enzymes. Activation is likely
secondary to inflammatory mediators and cartilage matrix molecules released
during an initial insult to the joint, after which synovial tissue appears to drive
progressive joint degeneration in a positive feedback cycle.108 Synovitis predicts
the development and progression of symptoms (odds ratio 9.2, 95% CI 3.2 to
26.3)112 and possibly cartilage loss (odds ratio 2.7, 95% CI 1.4 – 5.1),113 although
the relationship with structural change is less consistent. It is often difficult to
compare study findings due to different patient populations and varied methods
32
Osteoarthritis: an overview
Chapter 2
of diagnosing synovitis, however, synovitis may represent a rational target for
intervention.
Systemic Inflammation
Osteoarthritis is primarily seen as a local disease confined to the joint and studies
investigating the relationship with systemic markers of inflammation have
yielded conflicting results. A recent systematic review suggests that serum CRP
is associated with symptoms rather than radiographic osteoarthritis114 and pain
may represent a marker of systemic inflammation.115 It is not understood why
obesity remains a risk factor for osteoarthritis in non-weight-bearing joints.116
Adipokines released from adipose tissue have been proposed as mediators of this
effect, however, their role is speculative and not borne out in clinical studies.117
DIAGNOSIS
The clinical diagnosis of osteoarthritis requires that patients have symptoms,
and the prevention or alleviation of these is the goal of any intervention.
Indeed, it is symptoms that prompt patients to seek medical attention outside of
screening or research programmes. The difficulty using symptoms to define the
presence of osteoarthritis is that they may develop only once there is advanced
and likely irreversible disease. This may follow a period of earlier subclinical
structural change. With disease modification in mind, symptoms therefore have
limited value diagnosing early osteoarthritis when intervention is more likely
to be successful. Further limitations are that symptoms fluctuate significantly
over time, and are influenced by concurrent pathology and pain pathway
modulation.118
In this review, we shall define structural osteoarthritis to be present when
there is evidence of cartilage loss in the absence of inflammatory or crystal
arthropathy, irrespective of whether a patient has symptoms. This definition
aims to describe osteoarthritis at an earlier stage. Although cartilage changes
might be preceded by changes within synovium and bone, cartilage degeneration
appears to be the common endpoint of all osteoarthritis phenotypes. As our
understanding of disease pathogenesis improves, measures relating to other
joints structures are likely to gain validity. The greatest limitation of addressing
structural osteoarthritis is our inability to predict whether it will progress to
clinical osteoarthritis.
Interventions employed when patients remain relatively asymptomatic must
carry a low risk profile alongside proven efficacy to be ethically acceptable.
33
Chapter 2
Figure 4: Biochemical markers and imaging modalities offer the ability to evaluate
osteoarthritis at different stages of disease, as assessed using the Osteoarthritis Research
Society International (OARSI) histological grade. Modified and reproduced with permission
and copyright © of the British Editorial Society of Bone and Joint Surgery.120
IMAGING
Radiography
Osteoarthritis is traditionally diagnosed using plain film radiography where
features include narrowing of the joint space width (JSW), osteophyte formation,
and the development of subchondral sclerosis and cysts. Scoring systems include
those proposed by Kellgren and Lawrence33 and Osteoarthritis Research Society
International (OARSI)],121 however, JSW alone is more sensitive and reliable.122,123
34
Osteoarthritis: an overview
Chapter 2
JSW remains the only structural end-point accepted by the European Medicines
Agency and Food and Drug Administration to prove efficacy of disease-modifying
osteoarthritis drugs (DMOADs),124 yet this measure has many limitations.125 JSW
lacks sensitivity and is unable to detect localised cartilage damage125,126 making
it unsuitable for the detection of early osteoarthritis. It also lacks specificity,
and in addition to cartilage thickness, JSW in the knee is dependent on the
structural integrity of the meniscus and whether it is extruded from the joint
space.127 Standardisation of image acquisition is essential given JSW is strongly
influenced by joint positioning.128 The usually slow progression of osteoarthritis
and the limited responsiveness to change means that when JSW is used as
an outcome measure, large cohorts are required and ideally patient follow-up
should exceed two years,129 however, this must be balanced against the risk of
reduced participant retention. Despite these limitations, radiography is relatively
inexpensive and readily available and continues to play a role in both clinical and
research settings.
35
Chapter 2
Figure 5: Sagittal dGEMRIC of the hip of a patient with cam lesion femoroacetabular
impingement. Although there was no evidence of degenerative change on radiographs,
the white arrow highlights a region of glycosaminoglycan depletion within the acetabular
cartilage which may indicate early osteoarthritis at the site of impingement.
36
Osteoarthritis: an overview
Chapter 2
such as bone marrow lesions, meniscal status and synovitis are also taken
into account.145 Scoring systems have been developed for knee, hip and hand
osteoarthritis (table 1), which show good reliability and responsiveness in clinical
trials.146
Table 1: Validated semi-quantitative MRI scoring systems for knee, hip and hand
osteoarthritis that assess morphological features of osteoarthritis.146
Knee Whole-Organ Magnetic Resonance Imaging Score (WORMS)
Biochemical Markers
Biochemical markers can be effector molecules, such a cytokines and enzymes,
or extracellular matrix constituents, such as precursors or degradation products
of collagen and proteoglycan. Their concentrations reflect tissue metabolism and
can be measured in blood, urine, or serum. The BIPED classification stratifies
biomarkers into ‘burden of disease’, ‘investigative’, ‘prognostic’, ‘efficacy of
intervention’, or ‘diagnostic’.149 A plethora of biochemical markers have been
proposed, but at present none are sufficiently well validated for use in clinical
practice. CTX-II (C-terminal telopeptide of collagen type II) and COMP are
markers of tissue degradation and remain the most widely investigated and best
performing biochemical markers across all BIPED categories.149
‘Diagnostic’ biomarkers aim to identify patients with pathology. Urinary CTX-
II and serum COMP are both raised in patients with osteoarthritis compared
with healthy controls.150,151 Sensitivity and specificity are poor for all biochemical
markers, and inferior to imaging measures. In knee osteoarthritis diagnosed using
the Kellgren-Lawrence Score, the area under the curve (AUC) has been reported
as 0.70 (95% CI 0.57 to 0.81) for urinary CTX-II, 0.73 (95% CI 0.58 to 0.86)
37
Chapter 2
for radiographic JSW, and 0.82 (95% CI 0.72 to 0.91) for MRI measurements.
Combining CTX-II with MRI measurements gives an AUC of 0.84 (95% CI 0.77
to 0.92).152 When measured systemically, biochemical markers may originate
from any site, hence their value is limited unless disease is confined to the
specific joint under investigation, which in practice is rarely the case. Synovial
fluid assays overcome this problem, but are limited by acceptability to patients
and the potential absence of an effusion. Recent developments suggest that
post-translational protein modification may be joint-specific and this warrants
further investigation.153
The potential value of ‘prognostic’ biochemical markers is large and may allow
the identification of patients most likely to benefit from intervention. Urinary
CTX-II and serum COMP perform reasonably well in predicting the incidence and
progression of radiographic hip and knee osteoarthritis in longitudinal cohort
studies.154-156 The predictive value of urinary CTX-II (odds ratio 3.2) is greater
than JSW (odds ratio 1.4), but lower than MRI measurements (odds ratio 4.8).
Combining CTX-II with MRI measurements has the greatest prognostic value for
structural knee osteoarthritis progression (odds ratio 5.8).152 Biochemical markers
have only shown a limited ability to predict symptoms.156 The interpretation
of assay results is limited by an incomplete understanding of the biological
activity they signify and whether it is relevant to clinical osteoarthritis.157 Further
validation is essential given biochemical markers are already used extensively
as outcome measures in clinical studies to assess ‘efficacy of intervention’.158
The number of ‘investigative’ biomarkers has increased rapidly with expansion
of the field referred to as ‘omics’ where biological molecules are characterised
and quantified. Interesting findings include a proteomic study of cartilage that
identified biomarkers that appear to be joint-specific.159
The future of biochemical markers is likely to comprise broad-spectrum panels of
assays that allow the assessment of osteoarthritis with disease phenotyping to
determine the appropriate therapy. Performance may be enhanced if combined
with imaging and genotyping. At present, clinical application remains a fairly
distant prospect and many challenges remain. Sampling technique is critical
and biochemical marker concentrations are influenced by factors including diet,
physical activity, and systemic metabolism.
TREATMENT
An improved understanding of disease pathogenesis and advances in the field
of biomarkers makes it increasingly possible to identify patients at greatest
38
Osteoarthritis: an overview
Chapter 2
risk of disease, diagnose the earliest osteoarthritis, and measure treatment
efficacy within a short timeframe. Consequently, a plethora of novel therapeutic
strategies have been proposed and tested in clinical trials. Thus far none
have been approved by regulatory bodies, which require concurrent structural
modification and symptom improvement.131
Lifestyle Modification
Many aetiological factors of osteoarthritis are amenable to lifestyle changes.
Weight loss in obese patients reduces the risk of developing symptomatic
osteoarthritis160 and improves symptoms once there is evidence of disease.161
Radiographic structure modification has not been demonstrated, although
benefits are evident with dGEMRIC162 and several biochemical markers.163 The
effects of exercise require further elucidation, but activities focusing on improved
muscle strength and aerobic capacity improve symptoms (effect size >0.8)164
and confer cardiovascular and all-cause mortality benefits.
Surgery
i) Correction of Aetiological Factors
Some aetiological factors may be amenable to surgery. The progression of
osteoarthritis secondary to hip dysplasia is successfully delayed by reorientating
the acetabulum. In addition to sustained symptomatic improvement, hip survival
rates exceed 80% at 10 years.165 More recently, arthroscopic hip surgery to
recontour the proximal femur (figure 6) and prevent femoroacetabular
impingement has shown symptomatic benefit beyond five years and may modify
the long-term risk of developing osteoarthritis,166 however, evidence to date
is confined to small cohort studies. Knee alignment predicts the development
of osteoarthritis in the compartment of greatest loading, hence unloading this
compartment offers therapeutic potential. In an interesting study, temporary
surgical joint distraction produced symptomatic and structural improvement
in end-stage knee osteoarthritis and suggests reparative potential remains in
osteoarthritis.167 Periarticular osteotomies to correct the mechanical axis of
the knee show promise, and prospective studies demonstrate symptomatic
improvement extending beyond 10 years.168 However, in general, evidence for
the efficacy of these interventions is limited. Randomised controlled trials with
long-term follow-up are required to determine whether these joint-preserving
operations prevent clinical and structural osteoarthritis progression.169
39
Chapter 2
Acetabular Acetabular
Rim Rim
Femoral Femoral
Head Head
Cam Lesion Femoral Neck
40
Osteoarthritis: an overview
Chapter 2
into the cartilage defect several weeks later. The technique has undergone
several iterations, but to date it remains unclear whether ACI confers improved
clinical outcomes over more simple techniques such as microfracture.171 Latest
developments comprise alternative sources of cell, including mesenchymal and
embryonic stem cells, utilisation of growth factors, and the implantation of cells
into three-dimensional scaffolds or matrices that support growth, differentiation,
and maintenance of a chondrogenic phenotype (figure 7).172
A B
There is little evidence that the above techniques modify the development of
osteoarthritis. Cartilage repair is unlikely to be successful if the joint environment
remains biologically or mechanically hostile, but may provide an important
adjunct to the correction of aetiological factors.
Pharmaceuticals
Many patients who develop osteoarthritis do not have identifiable risk factors
amenable to intervention. Furthermore, it is not known whether the correction of
risk factors is sufficient to reverse a catabolic tissue phenotype. Pharmaceutical
agents already play a key role in symptom control, particularly paracetamol
and non-steroidal anti-inflammatories,172,173 but an increasing number of drugs
are also under investigation as DMOADs. None to date have been approved by
regulatory bodies.
41
Chapter 2
i) Supplementation
Chondroitin and glucosamine have demonstrated anti-inflammatory and anti-
catabolic properties in vitro174 and their ability to improve symptoms or delay
structural progression of osteoarthritis has been extensively investigated in
clinical trials. Results have been conflicting, purportedly due to different study
designs and patient populations, investigator bias, or the use of different
drug formulations. Studies tend to report more positive findings when using
glucosamine sulphate rather than glucosamine hydrochloride, however,
assuming glucosamine is the active ingredient, there is no available rationale to
explain this effect.175 Overall the literature does not indicate that chondroitin or
glucosamine confer a clinically relevant benefit176 and they are not recommended
in guidelines published by international bodies.176-178 However, it is noteworthy
that both have safety profiles comparable with placebo.
Hyaluronic acid is a glycosaminoglycan found in synovial fluid that acts as a
lubricant, but concentrations decline in osteoarthritis.109 Hyaluronic acids have
been widely used in osteoarthritis as viscosupplementation administered via
intra-articular injections, however, debate over efficacy and safety continues. A
recent meta-analysis concludes that there is no clinically relevant benefit with
respect to pain or function,178 neither is there is convincing evidence of structural
benefit. Lubricin is a glycoprotein that acts synergistically with hyaluronic acid
to provide lubrication,179 but exhibits reduced lubricating capacity in a subset of
patients with osteoarthritis.111 Supplementation restores normal joint lubrication
and may be chondroprotective,111,180 offering potential therapeutic benefit.
42
Osteoarthritis: an overview
Chapter 2
osteoarthritis where a reduction in urinary CTX-II was demonstrated compared
with placebo, but without a reduction in JSW narrowing over two years.184,185
Furthermore, the symptomatic improvement seen in one cohort184 was not
reproduced in a larger multi-national study.185 More recently, a single dose of
zoledronic acid was shown to improve pain and the size of bone marrow lesions
at six months.186 Strontium ranelate, in addition to osteoclast inhibition and
osteoblast stimulation, enhances chondrocyte matrix production in vitro.187 A
randomised controlled trial showed that strontium ranelate therapy for three
years reduces radiographic joint space narrowing over placebo, accompanied
by a modest improvement in symptoms and a reduction in urinary CTX-II.188
Further studies of strontium ranelate may be indicated, however, side effect
profiles are likely to limit its clinical utility in osteoarthritis.
43
Chapter 2
44
Osteoarthritis: an overview
Chapter 2
v) Stem Cell Therapy
Stem cell therapy may have an application in osteoarthritis and multipotent
mesenchymal stem cells are found in healthy and diseased cartilage. Kartogenin
is a molecule that promotes differentiation of chondrocytes. In animal models,
kartogenin reverses established osteoarthritis by promoting repair of cartilage
lesions.198 Whether this or similar molecules will translate to the clinical arena
remains to be seen.
DISCUSSION
An improved understanding of osteoarthritis aetiology and pathogenesis has
yielded an increasing array of potential targets for the prevention of disease
development and progression. Furthermore, advances in the fields of imaging
and biochemical markers have facilitated the diagnosis of earlier disease, and
may provide sensitive assays for treatment efficacy. But despite these advances,
effective preventative strategies have not been readily forthcoming.
Of all the interventions investigated to date, lifestyle modifications show
the greatest benefit. Promoting the maintenance of an optimum weight and
participation in regular exercise are cost-effective and also reduce all-cause
mortality. Results are eagerly awaited from ongoing trials investigating the
effects of surgically correcting adverse joint biomechanics.169 Interestingly, the
disease-modifying effect of doxycycline was negated in knees that were varus
aligned.199 Interventions to modify risk factors may on their own prove inadequate
if joint tissues have already shifted to a catabolic phenotype. Combining surgical
intervention with pharmaceutical agents may represent an optimum strategy.
Key challenges are to define and standardise outcome measures, and to
understand why there is a poor correlation between structural degradation and
symptoms.119 A better understanding of peripheral and central pain pathways,
aided by tools such as functional MRI, may help to unravel this enigma.200 The
limitations of targeting pain alone are highlighted by trials targeting nerve
growth factor (NGF). Tanezumab and fulranumab are monoclonal antibodies to
NGF, which in randomised controlled trials showed impressive improvements
in pain and function compared with placebo.191 However, a small minority of
patients developed rapidly progressive osteoarthritis200 raising the concern that
increased joint loading permitted by improved analgesia worsens disease.
Osteoarthritis has several disease phenotypes201 and identifying and specifically
targeting these is likely to prove critical for the successful development of novel
therapies. Clinical trials investigating the efficacy of an intervention that targets
45
Chapter 2
46
Chapter 3
ABSTRACT
Objective To investigate the association between baseline hip shape and both
clinical hip osteoarthritis (OA) and total hip replacement (THR) at 5-year follow-
up.
Design Individuals from the Cohort Hip and Cohort Knee (CHECK) study, with
early symptomatic OA, having standardized anteroposterior pelvic radiographs
at baseline and 5-year follow-up (n=723) were included. Hip shape on the
radiographs was assessed using statistical shape modelling (SSM). Hips fulfilling
the American College of Rheumatology (ACR) criteria at follow-up were classified
as clinical OA. The association between each mode of shape variation and both
outcome measures was calculated by Generalized Estimating Equations (GEE).
Results The included individuals comprised 575 females and 148 males (mean
age 55.9 ± 5.2 years). At baseline, 8% fulfilled the ACR criteria, 76% had no
radiographic hip OA [Kellgren & Lawrence (K&L) = 0] and 24% had doubtful OA
(K&L = 1). At follow-up, 147 hips (10.4%) fulfilled the ACR criteria and 35 hips
(2.5%) had received THR. Five shape variants (modes) at baseline associated
significantly with THR within 5 years. When combined in one GEE model, these
shape variants resulted in a predictive power indicated by an area under the
curve of 0.81. No shape variants associated with the presence of clinical OA at
follow-up.
Conclusion The shape of the hip as quantified by an SSM has a good predictive
value for THR, whereas variation in shape cannot predict clinical OA. Minor
shape variants may be used as a radiographic biomarker to predict the future
risk of THR.
48
Hip shape predicts THR
INTRODUCTION
Osteoarthritis (OA) is often present in multiple joints, but hip OA frequently occurs
Chapter 3
in isolation, suggesting that local factors are important in its development.202
There is growing evidence that morphology of the hip joint is one such a risk
factor. Morphological abnormalities of the hip probably predispose to OA by an
altered biomechanical behavior of the hip.55 This seems plausible for hips with
an evident non-optimal shape as seen in (congenital) hip dysplasia, Perthes
disease, and slipped capital femoral epiphysis.51,53 Recently however, the more
prevalent cam-type deformities have also been recognized as a causative factor
for end-stage OA, with a positive predictive value as high as 52%.203 Thus, the
morphology of the hip appears promising for prediction of hip OA before the
actual onset of OA.101,204
Obvious shape abnormalities are usually quantified by predefined measures
such as the centre-edge angle for dysplasia and the alpha angle for cam-type
deformity. However, subtle morphological variation might also play an important
role, but these are difficult to capture by predefined measures.
By using statistical shape modeling (SSM), a sophisticated technique which
identifies independent shape variants, it is possible to quantitatively describe
the total morphology of the hip.205,206 An SSM describes all variation in shape
that exists in the study population, and is therefore a method which can identify
shapes ‘at risk’ for OA without any assumptions.
Hip OA is usually defined by clinical symptoms such as pain and decreased
function, or radiographically by structural alterations as seen on radiographs.
However, a poor association between clinical and radiographic definitions for hip
OA has been reported.207 Previously, it has been shown in cross-sectional and
case-control studies that subtle shape variants of the proximal femur associate
with radiographic OA.63,101,206,208,209 However, it is unknown whether hip shape
associates with OA as defined by clinical criteria. Possibly, ‘at risk’ shapes are
different for both definitions, as it has been shown that those shape variants that
associated with radiographic hip OA were different from those that associated
with pain.209
We investigated whether minor shape variants of hips without definite
radiographic signs of OA at baseline, can be predictive in people with first onset
hip or knee pain for the development of hip OA after 5 years, as classified either
by the American College of Rheumatology (ACR) criteria for clinical OA or by
total hip replacement (THR).
49
Chapter 3
METHODS
Study cohort
All individuals were participants of the Cohort Hip and Cohort Knee (CHECK)
cohort. CHECK is a nationwide prospective cohort study of 1002 individuals with
early symptomatic OA of knee or hip. On entry, all participants had pain or
stiffness of knee or hip and were aged 45-65 years; they had not yet consulted
their general practitioner (GP) for these symptoms, or the first consultation was
within 6 months before entry. Participants with a pathological condition other
than early OA that could explain the symptoms were not included in the cohort
(for hip: trauma, rheumatoid arthritis, congenital dysplasia, Perthes disease,
subluxation, osteochondritis dissecans, fracture, septic arthritis, Kellgren &
Lawrence (K&L) grade 4 or total hip replacement, previous hip surgery, and
individuals having only symptoms of bursitis or tendinitis).210
Radiographs, serum samples, and clinical examination were obtained from
eleven (general and university) hospitals at baseline and at five year follow-
up. Individuals were recruited either by GPs who were invited to refer eligible
persons to one of those centres and by advertisements in local newspapers. The
723 of the 1002 individuals who had anteroposterior (AP) pelvic radiographs
of sufficient quality obtained both at baseline and at five year follow-up were
included [the mean (sd) follow-up was 5.06 (0.17) years. Of the initial 1002
individuals, 137 subjects did not have pelvic radiographs at both baseline and
follow-up, of the remaining individuals, 124 subjects had AP hip instead of AP
pelvic radiographs at baseline, and 18 subjects did not have radiographs of
sufficient quality at baseline to add them to the SSM. Excluded individuals did
not differ on any baseline characteristic from the included individuals. The study
was approved by the medical ethics committees of all participating centres, and
written informed consent was obtained from all participants.
50
Hip shape predicts THR
points that were positioned along the surface of the bone in the image by three
investigators, who were unaware of any clinical or radiographic outcomes. Each
point is always positioned on the same anatomical landmark (e.g., most lateral
Chapter 3
point of greater trochanter, most distal point of ischial bone etc.) of the outline,
to allow comparison between the shapes (Figure 1).
Principal component analysis was used to transform the set of points into an
SSM, which consists of a number of modes that together describe the total
variation in shape in the study population. Shape variants which are correlated
are captured in one mode such that each single mode represents independent
shape variants. Each mode is quantitatively described as the mean, which
corresponds with 0, and the positive or negative deviation from the mean as
expressed in the number of SDs.205
To examine the inter-observer reliability of the modes obtained, the point set
was positioned by each investigator in 24 randomly selected radiographs. Intra-
observer reliability was tested in 10 randomly selected radiographs with an
interval of 2 months.
51
Chapter 3
Outcome measures
The primary outcome measures were meeting the clinical ACR criteria for hip
OA at 5-year follow-up, and hips having received a THR within 5 years.32 In
short, a prerequisite for meeting the ACR criteria is hip pain, either together
with internal hip rotation <15° and an erythrocyte sedimentation rate ≤45
mm/h, or together with hip internal rotation ≥15°, and pain on internal rotation,
and morning stiffness of the hip ≤60 minutes, and age >50 years. Secondary
outcome measures were two items of the ACR criteria separately; amount of hip
pain and decreased internal hip rotation (<15°) at 5-year follow-up. The severity
of pain in the previous 48 h was assessed per hip using the Visual Analog Scale
(VAS). This scale runs from 0 to 10, where 0 equals no pain and 10 very intense
pain. Internal hip rotation was measured according to a standardized protocol
in sitting position by a goniometer in 90° of flexion, which previously showed
satisfactory reliability212.
Statistical analysis
Reliability of positioning the point set was assessed using intra-class correlation
coefficient (ICC). Univariable differences in baseline characteristics between
hips that developed OA and normal hips were evaluated by the Mann-Whitney
test for continuous variables, by chi-square test for sex, and by Generalized
Estimating Equations (GEE) for K&L score.
To analyze whether a mode was predictive for the various outcome measures,
regression models using GEE were constructed. All modes were corrected for
age, gender, and BMI. In order to account for the many modes (24) tested, an
effect was considered significant at a P-value smaller than 0.002 (p=0.05 / 24
modes). From these predictive models, odds ratios (OR) were calculated for
each mode to describe the strength for each independent predictive mode. The
predictive power of the GEE model including all significant modes was tested by
the area under the ROC curve (AUC). All statistical analyses were performed in
SPSS version 17.0.
52
Hip shape predicts THR
RESULTS
Participants
Chapter 3
Of the 723 individuals (1411 hips), 575 were women and 148 were men with
a mean age of 55.9 years (± 5.2 years). At baseline, 8% of the included hips
fulfilled the ACR criteria whereas 92% did not meet the clinical criteria of hip OA.
Radiographically, 76% of the included individuals had no signs of radiographic hip
OA (K&L = 0) and 24% had doubtful radiographic hip OA (K&L = 1). Additional
baseline characteristics are presented in Table 1, stratified for the presence or
absence of THR and clinical OA at follow-up
Outcome measures
A total of 147 (10.4%) hips fulfilled the ACR criteria for clinical OA at 5-year
follow-up and 35 hips (2.48%) underwent THR within 5-year follow-up. At
follow-up, 23 hips (1.63%) had internal hip rotation less than 15°.
Predictive modes
A total of 24 modes were extracted from the SSM, which together explained
90% of the total variance in shape. We could not identify any mode at baseline,
which was predictive for OA at 5-year follow-up (p<0.05) as defined by the ACR
criteria. When corrected for age, sex, and BMI, five modes (modes 7, 11, 12,
15, and 22) independent of each other associated significantly with THR within
5 years. The p-values, OR, and ICC scores of these modes are summarized in
table 2. Although a mode does not represent only one single aspect of variation
in shape, but is a combination of various correlated aspects of variation in shape,
we described the most obvious patterns in shape variation that the predictive
modes represent (Figure 2). Modes with a p-value less than 0.05, but greater
than 0.002 are also presented in table 2 and illustrated in figure 3. Combining
the five significant modes in the GEE model for calculating the area under the
ROC curve, resulted in a predictive value of 0.81.
Although no modes were found to be significantly predictive for the ACR criteria
at follow-up, we found modes, which could predict severity of pain and limited
internal rotation at follow-up when analyzed separately. For pain, mode 9 was
nearly significantly associated with the VAS scores (p-value of 0.007). Higher
values of mode 7 were almost significantly predictive for internal rotation <15°
(p-value of 0.003). The association between all modes and the secondary
outcome measures is given in table 3.
53
54
Chapter 3
Table 1. Baseline characteristics of the participants stratified by the absence or presence of clinical OA and THR at 5-year follow-up.
P-value p-value
Absence of Presence of Absence of Presence of
Total Absence vs Absence vs
THR THR clinical OA clinical OA
(1411 hips) presence of presence of
(1376 hips) (35 hips) (1264 hips) (147 hips)
THR clinical OA
Age in years: mean (SD) 55.9 (5.20) 55.9 (5.2) 57.7 (4.1) 0.052 56.0 (5.2) 55.1 (5.4) 0.041
Women, No. (%) 1120 (79) 1097 (80) 23 (66) 0.030 1001 (79) 119 (81) 0.67
BMI, kg/m²: mean (SD) 26.1 (4.1) 26.1 (4.2) 25.7 (4.1) 0.56 26.1 (4.2) 26.3 (3.6) 0.16
Height in cm: mean (SD) 169.9 (8.2) 169.8 (8.2) 170.2 (8.5) 0.88 169.8 (8.2) 170.4 (7.9) 0.42
Weight in kg: mean (SD) 75.3 (13.6) 75.3 (13.7) 74.6 (13.3) 0.99 75.1 (13.7) 76.6 (13.4) 0.19
Grade 0, No (%) 1058 (76) 1048 (77) 10 (29) 964 (77) 94 (65)
Grade 1, No (%) 331 (24) 306 (23) 25 (71) 280 (23) 51 (35)
55
Hip shape predicts THR
Chapter 3
Table 3. Strength of the relation between all modes and secondary outcome measures pain and decreased internal rotation
56
Relation with decreased internal rotation
Relation with Pain (VAS) at follow-up
(<15°) at follow-up
Chapter 3
Modes OR (95% CI) P-value aOR (95% CI) P-value OR (95% CI) P-value aOR (95% CI) P-value
Mode 0 0.85 (0.74-0.98) 0.02 0.90 (0.78-1.04) 0.15 1.87 (1.19-2.94) 0.007 1.68 (0.97-2.92) 0.06
Mode 1 1.12 (0.96-1.32) 0.16 1.12 (0.95-1.32) 0.19 1.26 (0.59-2.71) 0.55 1.71 (0.80-3.68) 0.17
Mode 2 0.99 (0.87-1.13) 0.90 1.09 (0.95-1.25) 0.24 0.87 (0.46-1.64) 0.67 0.72 (0.40-1.29) 0.27
Mode 3 1.17 (1.01-1.35) 0.04 1.08 (0.93-1.27) 0.32 0.87 (0.52-1.48) 0.61 1.03 (0.51-2.05) 0.94
Mode 4 0.96 (0.83-1.12) 0.62 0.99 (0.86-1.15) 0.90 1.00 (0.66-1.50) 0.98 1.00 (0.65-1.54) 0.99
Mode 5 1.02 (0.89-1.16) 0.80 1.05 (0.93-1.20) 0.44 1.20 (0.80-1.80) 0.39 1.24 (0.81-1.90) 0.33
Mode 6 1.03 (0.91-1.17) 0.63 1.06 (0.94-1.21) 0.35 0.77 (0.45-1.32) 0.34 0.79 (0.44-1.45) 0.79
Mode 7 0.92 (0.80-1.07) 0.30 0.96 (0.82-1.11) 0.55 1.92 (1.26-2.93) 0.002 1.77 (1.22-2.56) 0.003
Mode 8 0.97 (0.86-1.11) 0.97 0.97 (0.85-1.11) 0.65 0.66 (0.39-1.10) 0.11 0.77 (0.48-1.26) 0.30
Mode 9 1.17 (1.03-1.33) 0.01 1.19 (1.05-1.36) 0.007 1.03 (0.69-1.53) 0.90 1.00 (0.64-1.55) 0.98
Mode 10 0.86 (0.76-0.97) 0.01 0.87 (0.76-0.98) 0.022 1.76 (1.00-3.08) 0.05 1.62 (0.94-2.82) 0.08
Mode 11 0.90 (0.79-1.02) 0.10 0.89 (0.78-1.01) 0.07 0.98 (0.65-1.46) 0.91 1.11 (0.79-1.57) 0.54
Mode 12 0.93 (0.81-1.07) 0..32 0.92 (0.80-1.06) 0.23 1.63 (1.12-2.39) 0.011 1.59 (1.05-2.41) 0.028
Mode 13 1.02 (0.89-1.17) 0.75 1.04 (0.91-1.19) 0.56 0.96 (0.62-1.51) 0.87 0.93 (0.61-1.42) 0.75
Mode 14 0.96 (0.84-1.09) 0.49 0.98 (0.86-1.12) 0.73 1.08 (0.75-1.58) 0.67 1.11 (0.75-1.65) 0.60
Mode 15 1.06 (0.92-1.22) 0.43 1.03 (0.89-1.18) 0.74 1.47 (1.08-2.00) 0.015 1.46 (1.09-1.94) 0.011
Mode 16 1.08 (0.94-1.23) 0.30 1.07 (0.93-1.22) 0.35 1.10 (0.71-1.70) 0.68 0.59 (0.71-1.82) 0.59
Mode 17 0.94 (0.83-1.08) 0.38 0.95 (0.83-1.08) 0.41 0.61 (0.40-0.92) 0.019 0.57 (0.36-0.88) 0.012
Mode 18 1.03 (0.90-1.18) 0.65 1.01 (0.88-1.16) 0.89 1.29 (0.76-2.17) 0.35 1.25 (0.73-2.14) 0.42
Mode 19 0.98 (0.86-1.11) 0.73 0.99 (0.87-1.12) 0.85 1.04 (0.61-1.78) 0.88 1.17 (0.69-2.00) 0.56
Mode 20 1.00 (0.88-1.15) 0.97 1.03 (0.90-1.17) 0.71 0.78 (0.53-1.14) 0.20 0.78 (0.52-1.16) 0.22
Mode 21 0.96 (0.84-1.09) 0.53 0.98 (0.86-1.11) 0.72 1.40 (0.96-2.04) 0.08 1.46 (0.99-2.16) 0.06
Mode 22 1.04 (0.92-1.18) 0.51 1.02 (0.90-1.15) 0.80 0.99 (0.67-1.46) 0.96 0.99 (0.65-1.51) 0.98
Mode 23 0.93 (0.83-1.05) 0.24 0.92 (0.82-1.04) 0.17 0.66 (0.43-1.02) 0.06 0.64 (0.41-1.00) 0.05
aOR were adjusted for age, sex, and body mass index at baseline. The presented OR represent every increase in one SD.
(a)OR; (adjusted) odds ratio, CI; Confidence interval, VAS; Visual Analog Scale, as a continuous measure (0-10).
Hip shape predicts THR
Chapter 3
Figure 2. Modes which were significantly predictive for THR are shown. The shapes
corresponding with the -4 and +4 standard deviations of the mean are illustrated on the
left and right side respectively. The middle column shows the overlapping shapes of the
-4 and +4 standard deviations; the extremes which are predictive for THR are shown in
red. Mode 7 represents variation in the length of the femoral neck, mode 11 represents
variation in the concavity of the superior head-neck junction together with variation of the
posterior wall, mode 12 represents variation in the superior joint space width together
with the femoral head coverage by the lateral acetabular rim, and mode 15 represents
variation in the femoral neck width, together with the resulting variation in head-neck
offset.
57
Chapter 3
Figure 3. Modes which predicted THR at a p-value level <0.05, but >0.002 are shown. The
shapes corresponding with the -3 and +3 standard deviations of the mean are illustrated
on the left and right side respectively. The middle column shows the overlapping shapes of
the -3 and +3 standard deviations; the extremes which are predictive for THR are shown
in red.
58
Hip shape predicts THR
DISCUSSION
In this prospective study we showed that the shape of the hip at baseline, as
Chapter 3
quantified using SSM, can predict THR after 5 years but not clinical OA after 5
years as defined by the ACR criteria. By using SSM in individuals that consulted
their GP for the first time with knee or hip pain, it was possible to identify
shape variants that increased the risk of requiring THR, before the actual
radiographic onset of OA. At baseline, especially a broad and short femoral
neck, and a retroverted acetabulum together with a non-spherical femoral head
were predictive of fast progressing OA. In addition, hip shape at baseline might
predict hip pain and decreased internal rotation at 5-year follow-up.
This study confirmed the important role of hip shape on development of OA.
In previous cross-sectional and case control studies, the importance of hip
morphology as a risk factor of radiographic hip OA was already shown.63,208 In
these studies, shape variants of the femoral head and femoral neck appeared
to pose the highest risk for end-stage OA. Recently, Barr et al. retrospectively
studied hip morphology of individuals, which presented with hip pain 5 years
before, also using THR as an outcome measure213. Consistent with their findings,
we found a predictive mode which represented femoral head flattening and
superior neck broadening (mode 15, figure 2). The predictive role of bone shape
on clinical OA remains poorly studied, as most other studies defined the presence
of OA by radiographic criteria. Only one study by Waarsing et al., using dual-
energy X-ray absorptiometry (DXA) images, indicated that subtle shape aspects
of the proximal femur not captured by common radiological measures contain
information about clinical status.209 However, in the present study, we found that
baseline hip shape could not predict development of clinical OA as determined
by the ACR criteria. This might be explained by the fact that in this study, the
ACR clinical criteria were not stable in participants with early symptomatic OA.
For instance, 103 out of 116 (89%) individuals which fulfilled the ACR criteria
at baseline did not have OA anymore at 5-year follow-up when determined by
the same criteria. An explanation of the discrepancy between hips fulfilling the
ACR criteria at baseline and at follow-up might be the presence or absence of
pain at both time points. For example, if an individual met the ACR criteria at
baseline, he or she might not necessarily experience hip pain during the follow-
up visit because the presence and severity of pain in hip OA is highly variable,
especially in the early stage.214 When analyzing those hips that fulfilled the ACR
criteria either at baseline or at follow-up as an outcome measure, no predictive
modes were found either. Another explanation of this discrepancy might be the
variability in the measured internal hip rotation.215
59
Chapter 3
Although hip shape was not predictive for OA as determined by the combination
of the ACR criteria, it could predict two clinical criteria independently; hip pain
and decreased internal rotation. For pain, higher values of mode 9 at baseline
associated with more pain at 5-year follow-up. The thicker and shorter femoral
neck as represented by higher values of mode 9 shows striking similarities with
mode 3 of the study by Waarsing et al., who also found a broad and short
femoral neck to be the most significant predictor for VAS pain score.209 For
internal rotation, a higher value of mode 7, which correspond with a straight
and longer femoral neck (figure 2), was predictive for internal rotation <15° at 5
year follow-up. Remarkable in this respect is that the opposite of mode 7 (lower
values), representing a short femoral neck, was predictive for THR.
As decreased internal rotation is a clinical sign for cam impingement, we assumed
a mode describing a non-spherical femoral head to be predictive for decreased
internal rotation.216 However, we did not find such a mode when applying a
threshold value of 15°, but for limited internal rotation of 20° or less, mode 4,
describing a non-spherical femoral head together with a shallow acetabulum,
became highly significant (figure 3). Interestingly, this mode was a predictor for
THR as well (p-value of 0.003), but did not remain significant when corrected for
multiple testing (p-value threshold of 0.002).
A larger statistical shape model might be more powerful for predicting OA.213
Previous studies using SSM mostly included the proximal femur only, and despite
the importance of the interaction between the proximal femur and acetabulum,
only two studies additionally included the acetabular roof or a portion of the
pelvis.206,209,213,217 In order to quantify this interaction, we created a shape model
of the complete hip joint by including both the proximal femur and the pelvis.
The advantage of this model is that it can describe both the position of the
proximal femur relative to the pelvis, and it can describe morphological variation
of the femur, which is correlated with morphological variation of the pelvis. The
importance of the interaction between proximal femur and pelvis for predicting
OA was reflected in the significant modes. For example, higher values of mode
11 describe a flat head-neck junction, resulting in a broad femoral neck.
Interestingly, the same mode also described a retroverted acetabulum as seen
by a posterior wall located medially with respect to the centre of the femoral
head (see figure 2). Acetabular retroversion has previously been described as a
risk factor for hip OA, but the evidence is conflicting.61,218,219 Our results from the
SSM indicate that acetabular retroversion only when combined with a flattened
head poses a higher risk for THR.
60
Hip shape predicts THR
Strengths of this study are the large number of hips assuring a robust statistical
shape model, the prospective design, and the large statistical shape model. Also,
the shape of the hips at baseline was not influenced by the arthritic process, as
Chapter 3
no hips showed definite radiographic OA at baseline. This was confirmed when
the analysis were corrected for K&L grade at baseline. The same modes became
significant without change in OR, assuring that the found shapes were true OA
predisposing shapes. There are however some limitations. The hip joint is a
complex three-dimensional structure and variants of shape might not be visible
on the AP radiographs. Still, SSM is able to quantify variation in orientation of
bone structures from the projection of the radiographs. Another issue concerns
variation in orientation of the bones, which will influence the projected shape.
Variation in position was minimized by using a standardized protocol. Since
remaining positional variation is often dictated by variation in anatomy, the
effect on the projected shape cannot be separated from true anatomical shape
variants. Still, both effects might contain valuable information with regards to
OA development. We aimed to use clinical outcomes of OA, although THR could
be considered not to be a pure clinical outcome, but is rather a combination
between radiographic signs of OA and symptoms. The significant modes for THR
might therefore also be predictive for radiographic OA. However, when analyzing
the five predictive THR modes for those hips with radiographic OA as defined by
a K&L grade of 2, 3, or 4 at follow-up (n=64), only mode 15 (a broader femoral
neck) could predict radiographic OA. Further, THR is a validated and clinically
relevant outcome measure.220
In conclusion, the morphology of the hip at baseline could not predict which hips
fulfilled the ACR criteria at 5-year follow-up, probably due to the instability of
those criteria in these participants, likely related to variability in pain. However,
receiving a THR within 5 years was predicted well by the shape of the hip. In
particular, before the presence of definite radiographic OA as defined by the K&L
score, shape variants can be identified that pose a higher risk for THR during
the 5-year follow-up. Hip morphology might therefore be used as a radiographic
biomarker to predict the future risk of THR.
61
Chapter 4
ABSTRACT
Objectives: To prospectively investigate whether hip shape variants at baseline
are associated with the need for future total hip replacement (THR) in women
and to validate the resulting associated shape variants of the CHECK cohort in
the Chingford cohort.
Results: In the CHECK and Chingford cohorts, the respective mean(SD) age
was 55.8(±5.1) and 53.6(±5.4), and BMI 26.14(±4.3) and 25.7(±3.3). Multiple
shape variants of the hip were significantly (p<0.05) associated with future THR
in both the CHECK (modes 4,11,15,17, and 22) and Chingford (modes 2 and 17)
cohorts. Mode 17, representing a flattened head-neck junction and flat greater
trochanter, could be validated in the Chingford cohort. Mode 4 and 15 of the
CHECK cohort also showed non-significant trends in the Chingford cohort.
Conclusion: Several baseline shape variants are associated with the future
need for THR within a cohort. Despite differences in participant characteristics,
radiographic protocol, and follow-up time, we could validate at least one shape
variant, suggesting that SSM is reasonably transferable between cohorts.
64
Validation of SSM to predict THR
INTRODUCTION
Osteoarthritis (OA) is a common disease which accounts for a detrimental impact
on quality of life and a considerable economic burden.27,68 Due to the aging
population, the need for total hip replacement (THR) is expected to grow 174%
to more than half a million primary THRs per year by 2030 in the United States
only.28
Chapter 4
An increasing amount of evidence suggests that the shape of the hip plays
a causative role in the development of hip OA and the subsequent need for
THR.63,74,203,206,221 A cam deformity, characterised by extra bone formation at
the anterolateral head neck junction, and acetabular dysplasia are consistently
found to be associated with hip OA in prospective cohort studies.53,63,74,203,222
Interestingly, because these shape abnormalities are present before the actual
onset of OA, relatively simple measurements (alpha angle223,224, centre-edge
angle225) can strongly predict the future risk of THR.63,203,226,227
Many other non-optimal shape variants might predict the development of hip OA
as well, but these are difficult to capture by linear measurements (i.e. lengths
and angles) because of the complex overall hip shape. Statistical shape modeling
(SSM) is a technique which recognises all independent shape variants in a given
population, and describes them quantitatively. Using this technique, it is possible
to identify shape variants at risk for developing OA without any predefined
hypothesis. SSM has therefore considerably increased in popularity as a tool to
study the association between hip shape and the risk of OA. As a result, multiple
subtle shape variants of the proximal femur and/or pelvis have been reported
to be associated with OA.206,209,213,217,221,228-231 SSM might therefore be a potential
radiographic biomarker which can be used in research and clinical practice.
However, all published studies on SSM and OA used different shape models (point
sets of femoral head-neck junction or entire proximal femur with or without
acetabulum), which makes the results difficult to compare and interpret. Certain
shape variants have been associated with hip OA within a single cohort, but their
generalisability remains unknown. To this end we have used one identical shape
model for anteroposterior (AP) pelvis radiographs in two cohorts.
The aim of this study was to investigate which shape variants at baseline in
OA free female hips are associated with the need for THR at follow-up in two
prospective cohorts (CHECK and Chingford). Since both cohorts now have an
identical model in our study, a comparison can be made between the findings
of the shape variants in the two cohorts that are associated with THR at a later
time point. Thereby we can test how consistent the associated shape variants for
development of OA are in various populations.
65
Chapter 4
METHODS
Study population
Subjects were selected from two prospective cohorts, the Dutch CHECK cohort
and the UK Chingford cohort. CHECK is a nationwide multicenter prospective
cohort study of 1002 individuals aged 45-65 years (mean 55.9) at baseline
with early symptoms of OA (pain) of the hip and/or knee. They had not yet
consulted their general practitioner for these symptoms, or the first consultation
was within 6 months before entry. An extensive description of the CHECK cohort
can be found elsewhere.210 For the current study, only females of the CHECK
cohort were included to allow comparison with the Chingford cohort. Of the 1002
individuals of the CHECK cohort, 791 (79%) were women and of these, 682 had
radiographs available on both baseline and 5 years follow-up. Of these 682
women, the first 93 women that had entered the cohort had AP hip radiographs
instead of AP pelvis radiographs obtained and were therefore excluded as a
different radiographic view might influence the morphological appearance of
the hip joint on the radiographs. Of the remaining 589 individuals who had AP
pelvis radiographs at both timepoints, 39 women had radiographs of insufficient
quality to apply the shape model, most often because the edge of the greater
trochanter was missing on the radiographs. Ultimately, 550 women (1100 hips)
were included. The 241 excluded women did not differ on baseline age (p=0.75)
or BMI (p=0.69) from the 550 included individuals.
The Chingford cohort is a population-based cohort of 1003 women aged 44-
67 years (mean 54.2 years) at baseline. These women were registered at a
single general practice in London and were invited to participate in a study
assessing musculoskeletal disease in the population. Yearly clinic visits included
morphometric, clinical, biological, and radiographic measurements. Subjects
who had radiographs obtained both at baseline and at 19-year follow-up were
included using a nested case control design. To allow comparison with the CHECK
cohort, only women without definite signs of radiographic OA (Kellgren&Lawrence
[K&L] score <2) at baseline were included.211 This selection resulted in 19 cases
who received THR within 19 years follow-up. For each case, 5 control hips were
matched based on age and BMI resulting in a total of 95 controls. Only one hip
per person was included. When a woman had received bilateral THR, the left
side was selected.
66
Validation of SSM to predict THR
Radiographs
In the CHECK study, weight-bearing AP pelvis radiographs were obtained from
the 11 participating research centers according to a standardized protocol,
taken at baseline and at 2 and 5 years follow-up. Feet were positioned such that
the medial side of the distal part of the first phalanx touched and a wedge was
used to assure 15° internal rotation. In the Chingford cohort, each woman had
Chapter 4
a standardized supine AP pelvis radiograph, taken at years 2, 8 and 20. A small
sand bag under the knees was used to minimize hip rotation so that the hips
were in neutral position.
In both the CHECK and Chingford cohorts, baseline AP pelvis radiographs were
graded using an atlas-based scoring method (K&L), with investigators blinded to
all clinical and demographic information.211,232 The K&L scores were independent
of the positioning of the SSM point set.
67
Chapter 4
Figure 1. The point set of the statistical shape model which consisted of 75 points
Outcome measure
Total hip replacement due to OA at follow-up (5 years follow-up in the CHECK
cohort and 19 years follow-up in the Chingford cohort) was used as an outcome
measure. THR was confirmed for all cases on the follow-up radiographs.
Statistical analysis
Differences between baseline characteristics within a cohort and between the
cohorts were calculated by independent samples T-test for normally distributed
continuous variables, by Chi squared test for sex and K&L score in the Chingford
cohort, and by generalized estimating equation model (GEE) for K&L score in
the CHECK cohort, to account for the correlation between the left and right
hips within an individual. To analyze whether a mode of shape variation was
associated with THR in the CHECK cohort, logistic regression with GEE was used.
The model was adjusted for age, BMI, and baseline K&L score. In the Chingford
cohort, these associations were calculated by conditional logistic regression
due to the matched nature of the data. From these models, odds ratios (ORs),
95% confidence intervals (95% CI) and p-values were calculated to describe
the strength of association for each independent mode within each cohort. To
68
Validation of SSM to predict THR
analyze the validity of the modes of shape variation for THR as an outcome, a
comparison was made between the findings of the shape variants in the two
cohorts that are associated with THR. An effect was considered significant at a
p-value <0.05. All statistical analysis were performed in SPSS version20.0.
RESULTS
Chapter 4
Participants
The baseline characteristics of the CHECK and Chingford cohort are presented
in table 1. At baseline, the 550 females (1100 hips) of the CHECK cohort had
a mean age of 55.8 years (range 44-66) and a mean BMI of 26.1 (range 17.5-
48.9). Of the 1100 hips, 80% had a K&L score of 0, and 20% had a K&L score of
1. Of these hips, 22 hips had received THR within the 5 years follow-up period.
In the Chingford cohort, the included 114 females (114 hips) had a mean age of
53.6 years (range 46-64) and a mean BMI of 25.7 (range 20.0-37.8) at baseline.
The K&L score was 0 in 95% of the hips and 1 in 5% of the hips. Between the
two cohorts, there were significant differences in baseline age (55.8 in CHECK vs
53.6 in Chingford, p<0.001) and baseline K&L=1 score (20% in CHECK vs 5% in
Chingford, p<0.001), but no difference in mean BMI was found (26.1 in CHECK
vs 25.7 in Chingford, p=0.26).
Table 1. Baseline characteristics of the individuals of the CHECK and Chingford cohorts
CHECK Chingford
69
Chapter 4
modes of shape variation had a significant positive (+) or negative (-) association
with THR at follow up in the CHECK cohort [mode 4 (-), 11 (+), 15 (+), 17 (-),
and 22 (+)] and two modes of shape variation in the Chingford cohort [mode 2
(+) and 17 (-)](figure 2). Negative values of mode 17, representing a flattened
head-neck junction, a flat major trochanter, and a prominent acetabular posterior
wall, was associated with future THR in both the CHECK and Chingford cohorts
(figure 2). Moreover, negative values of mode 4 and positive values of mode 15,
which were significantly associated with future THR in the CHECK cohort, had
the same effect in the prediction of THR in the Chingford cohort, though not
statistically significant (p=0.27 and p=0.16 respectively).
Table 2. the strength of association between all modes at baseline and THR at follow-up
(5 years in the CHECK cohort and 19 years in the Chingford cohort).
Check Chingford
The statistically significant modes (p<0.05) are highlighted in bold. aOR=adjusted OR; corrected for
age, BMI, and K&L grade at baseline. The presented ORs represent every increase in one SD of that
specific mode. CI=confidence interval.
70
Validation of SSM to predict THR
Chapter 4
Figure 2. Modes significantly predictive for THR are shown. The odds ratios presented in
table 2 represent every increase in SD, but for clarity the -3SD (left column) and +3 SD
(right column) from the mean (middle column) are presented. The extremes which are
associated with THR are indicated in red.
71
Chapter 4
DISCUSSION
Multiple shape aspects of the hip as quantified by SSM have previously been
shown to be associated with, or even predictive for radiographic OA206,209,228,229,
clinical criteria of OA209,221, and THR206,213,221. However, due to the application
of different shape models (different point sets) the results between studies
are difficult to compare. Consequently, the generalisability of the resulting
modes of shape variation is limited to the characteristics of each specific cohort
and therefore, the validity in the prediction of OA is unknown. In this study
we showed in two prospective cohorts that several shape aspects in female
hips without definite features of radiographic OA at baseline was significantly
associated with the need for THR at follow-up within each cohort. One shape
aspect was associated with THR in both the CHECK and Chingford cohorts. This
specific shape variant –describing a flattened head-neck junction together with
a flat major trochanter and a prominent acetabular posterior wall- might be a
generalisable shape aspect in the prediction of THR in women, even regardless
of other factors including cohort characteristics, follow-up time, or radiographic
protocol. In contrast, some other shape variants found to be associated with
THR might be co-dependent on those or other factors.
Gregory et al. were the first to quantify hip shape by SSM in order to study its
relationship with OA in a case-control design.206 Their point set of the shape
model involved the femoral head and neck (but excluded the acetabulum) and
was applied to 110 subjects (without radiographic OA) from the Rotterdam
study at baseline and 6 years follow-up. One shape variant that they identified
that could significantly predict OA at follow-up represented a less pronounced
curve from the upper femoral neck into the head, indicated by negative values
of this variant. This association was even more pronounced in the OA cases who
had received a THR. The variant values also decreased significantly over time
in the OA group, highlighting the importance to study predictive shape aspects
before the actual onset of radiographic OA. Furthermore, another variant they
identified could predict the need for THR but not the development of radiographic
OA, suggesting that baseline shape aspects may predict disease severity.206
The finding that hip shape was associated with the future risk of THR was
further supported by other studies. Previous research in the CHECK cohort for
both men and women showed five modes predictive of future THR in baseline
radiographic-free hips (K&L <2).221 The predictive modes primarily described a
broad and short femoral neck, and a retroverted acetabulum together with a
non-spherical femoral head. In the Rotterdam study, especially shape variants
related to the superior head-neck junction and the shape of the acetabular
72
Validation of SSM to predict THR
Chapter 4
as quantified by SSM have therefore been suggested as a potential radiographic
biomarker to predict the future risk of OA. But again, all studies on SSM and
OA used different shape models so that the resulting predictive modes of shape
variation cannot be directly compared.
Our results are in line with previously published articles showing that certain
shape variants of the hip can predict the risk of a THR within a cohort. More
importantly however, we found that one shape variant was predictive for a THR
in both the CHECK and Chingford cohorts. There are several explanations for the
fact that not all predictive shape variants of the CHECK cohort could be validated
in the Chingford cohort, which include differences in follow-up time, radiographic
protocol and participant characteristics. First, there was a difference in follow-up
time between both cohorts and the pathophysiology of fast progressing OA (THR
within 5 years in CHECK) might be different from slow progressing OA (THR
within 19 years in Chingford). This is supported by a study showing that the
predictive value of hip shape and other risk factors at baseline decreased when
using a longer follow-up time.235 An example is a cam deformity that results in
fast progressing hip OA203 which will be detected at 5 years but it is less likely to
be detected at 19 years of follow-up as other factors may dilute the predictability
of a cam deformity. Lower values of mode 4, representing a non-spherical
femoral head (together with a shallow acetabulum), and higher values of mode
15 might be modes that are stronger associated with fast progressing OA (figure
2). Though not as strong as in CHECK, the predictive effect of these modes were
still present in the Chingford cohort, though not statistically significant (p=0.27
for mode 4 and p=0.16 for mode 15). Secondly, the radiographic protocol was
different between both cohorts. in the CHECK cohort, AP pelvic weight-bearing
radiographic views with 15° internal hip rotation were obtained while in the
Chingford cohort AP pelvic supine radiographs were obtained in neutral position.
Consequently, there will be differences in how the outline of the bone, and
thus hip shape, appears on the radiograph. Thirdly, although no difference in
BMI between the cohorts was found and the significant difference of 2.2 years
in baseline mean age is probably negligible, differences in other participants
characteristics might play a role. In the CHECK cohort there is a larger proportion
73
Chapter 4
of females presenting with hip and/or knee pain (as first onset of knee and/or hip
pain was an inclusion criterion) whereas Chingford women were recruited from
the general population. This difference might influence the association between
hip shape and OA. Furthermore, there might be other (unknown) differences
between the two cohorts, for example genetics, which is known to affect both
hip shape and the relationship between hip shape and OA.217,230
It is unknown whether the shape variants found to be associated with OA are
causative factors for OA, or whether they represent early changes in bone
morphology as a result of OA. As the shape variants of our model describe the
overall hip shape, it is unlikely a result of the OA process. Furthermore, most
modes associated with THR show some variation in the head-neck junction,
which might reflect different subtypes of cam deformities, which result in OA
by a motion dependent process of cam impingement.75 Interestingly, a cam
deformity is known to be highly associated with OA, and develops during growth,
probably as a bone adaptation to high impact athletic activities.226,227,236 It is
unknown whether other predictive shape variants are also present before the
onset of OA.
Several potential limitations of this study need to be acknowledged. First, we
only included females and hence the results of this study might be different for
males, as there might be a gender difference in how much hip shape contributes
to the prediction of OA and also, the predictive shape variants might be different
between males and females.229 Second, rotation of bones might influence the
projected radiographic shape, which was however minimized by the use of a
standardized radiographic protocol. Thirdly, THR is a surrogate measure for OA
and as it is a surgical endpoint, questions might arise on the accessibility to
health care and THR. In this study, all THRs were a result of OA, a THR was
confirmed on follow-up radiographs, and it is a validated and clinically relevant
outcome measure.220 Also, both the Netherlands and the United Kingdom have a
public health system in which all people have equal access to THR, independent
of for example social status. Moreover, the definitions of radiographic OA and
clinical OA have their limitations as well.221
Predictive shape variants might be dependent on many factors including follow-
up time, inclusion criteria of the cohort, and radiographic protocol. Despite the
differences in these factors between the two cohorts, we could validate one
mode (mode 17), while two modes (modes 4 and 15) that were significantly
associated with THR in the CHECK cohort showed trends in the same direction
in the Chingford cohort, though not statistically significant. Therefore, SSM
might be a useful tool as a radiographic biomarker. For this purpose, the manual
74
Validation of SSM to predict THR
Chapter 4
radiographic biomarker was reasonably transferable between cohorts.
75
Chapter 5
ABSTRACT
Objective Determining the relation between acetabular coverage, especially
overcoverage which may lead to pincer impingement, and development of
osteoarthritis (OA) of the hip.
78
Pincer deformity and osteoarthritis
INTRODUCTION
The aetiology of osteoarthritis (OA) is mainly unclear, though both systemic
factors and local biomechanical factors are known to play a role.238 OA of the hip
often occurs without the presence of OA in other susceptible joints, indicating
that local biomechanical factors may predominate.202
Growing evidence supports the theory that these local factors are mainly explained
by bone shape variants of the hip, causing OA by an altered biomechanical loading
pattern.55,101,209,239 These bone shape variants can be located at the femoral
side, acetabular side, or both. An example of a femoral sided morphological
Chapter 5
abnormality is a non-spherical femoral head (cam deformity) which may lead
to a motion dependent abnormal contact between the femoral head and the
acetabulum, also known as cam-type FemoroAcetabular Impingement (FAI).55
A cam deformity is thought to develop during growth and is an important risk
factor for OA.203,226,240 An abnormal shape of the acetabulum may also lead to
OA by either acetabular undercoverage, also known as (mild) dysplasia, or
acetabular overcoverage, known as a pincer deformity.
In hips with mild acetabular dysplasia, a decreased contact area between femur
and acetabulum results in higher static loads on the anterosuperior acetabular
cartilage. In hips with a pincer deformity, the mechanism leading to OA is much
less understood. The proposed mechanism is that of a dynamic abnormal linear
contact between the overcovered acetabular rim and the femoral neck during
terminal motion of the hip, which is known as pincer-type FAI.55 When vigorous
hip motion causes repetitive impingement events, the soft tissue structures
within the hip joint might gradually damage, leading to hip OA.
This hypothesis is supported by intra-operative findings in symptomatic patients
with a pincer deformity, where acetabular cartilage damage was found throughout
the acetabulum in a small thin strip around the labrum.57 Also, cartilage damage
at the posterior-inferior site has been described as a result of a ‘countrecoup
lesion’ by the femoral head, due to the leverage effect of the neck when it abuts
against the anterior acetabular rim (Figure 1).
Evidence for the relation between mild dysplasia and development of OA provided
by cross-sectional or retrospective studies is inconsistent, but prospectively
designed studies generally show a moderate increased risk for hip OA.53,222,241-243
In contrast, the relation between pincer deformities and development of OA is
conflicting.59-61,63-66,244 However, these studies are often limited by a retrospective
or cross-sectional design, making it difficult to draw conclusions on causality. As
in mild dysplasia, prospective studies might identify an association, but no such
studies are available for pincer deformities.
79
Chapter 5
The aim of this study was to examine the relation between baseline anterior
and lateral acetabular coverage, specifically pincer deformities, and the risk
of developing OA after 5 years follow-up. We further investigated whether
acetabular coverage was associated with pain and decreased hip function.
METHODS
Study design and participants
Individuals were extracted from Cohort Hip and Cohort Knee (CHECK). CHECK
is a nationwide prospective cohort study of 1002 individuals with symptoms
of early OA of knee or hip. On entry, all participants had pain with or without
stiffness of knee or hip and were aged 45-65 years; they had not yet consulted
their general practitioner (GP) for these symptoms, or the first consultation was
within six months before entry. Participants with another pathological condition
that could explain the symptoms were excluded.210
Radiographs and clinical examination were obtained from eleven (general
and university) hospitals at baseline, and at 5-year follow-up. The mean (SD)
follow-up was 5.06 (0.17) years. Individuals were recruited either by GPs who
were invited to refer eligible persons to one of the participating centres and
by advertisements in local newspapers. The 720 individuals (1391 hips) of the
1002 participants who had both anteroposterior (AP) pelvic radiographs and
False Profile (FP) radiographs of sufficient quality obtained both at baseline and
at five year follow-up were included in the current study (figure 2). The study
was approved by the medical ethics committees of all participating centres, and
written informed consent was obtained from all participants.
Radiographs
Weight bearing AP pelvic radiographs and FP oblique view radiographs were
obtained according to a standardized protocol at baseline and five years follow-
up.203 For AP radiographs, feet were positioned in 15° internal rotation. For
FP radiographs, individuals stood sidewise with the hip of interest against the
radiographic table and with the second metatarsal phalanx of the same leg
parallel to the radiographic table. Then, the pelvis was rotated 25° backwards,
confirmed by a 65° wedge between the back and radiographic table, to profile
the anterosuperomedial edge of the acetabulum.245
80
Pincer deformity and osteoarthritis
Chapter 5
Figure 1. The mechanism of pincer impingement
A normal hip (A) and a hip with a pincer deformity (B) are shown. The anatomy of the normal hip
provides the hip a physiological ROM (C) whereas a pincer deformity (arrow) is proposed to lead to
an abnormal linear contact between the overcovered acetabular rim and the femoral neck during
terminal motion of the hip, which is known as pincer impingement (D). When vigorous hip motion
causes repetitive impingement events, the acetabular cartilage might gradually damage throughout
the acetabulum in a small thin strip around the labrum. Also, the leverage of the femoral head in the
acetabulum might lead to a contrecoup lesion posteroinferiorely (F).
81
Chapter 5
Figure 2. Flow of subjects from cohort inclusion to the final study population
Radiographic measurements
The shape of the hip on the AP and FP radiographs was outlined using the ASM
tool kit (Manchester University, Manchester, UK). The shape is given by a set
of landmark points that are positioned along the surface of the bone in the
radiograph.206,221 The lateral centre edge angle (LCEA) and anterior centre edge
angle (ACEA) were automatically calculated from the points outlined on the AP
and FP radiographs respectively, using Matlab (version 7.1.0, MathWorks Inc,
Natick, Massachusetts, USA).
The centre edge angle measures the amount of acetabular coverage with respect
to the centre of the femoral head. The centre of the femoral head is found by
drawing a best-fitted circle around the femoral head. From the centre of the
femoral head, a first line is drawn vertical and a second line is drawn to the
edge of the sourcil (the dense subchondral bone) of the acetabulum. The angle
between those lines is the centre edge angle (Figure 3).225,246 The vertical line
was drawn perpendicular to the horizontal line connecting both femoral heads
In the AP view, and perpendicular to the horizontal line of the radiographic film
in the FP view. We defined the presence of a pincer deformity by an LCEA or
82
Pincer deformity and osteoarthritis
ACEA >40° and the presence of mild acetabular dysplasia by an LCEA or ACEA
<25°.247,248
At baseline and 5 years follow-up, all AP pelvic radiographs were scored for OA
according to the Kellgren and Lawrence (K&L) classification (grade 0-4).211 In
addition, superior and inferior joint space was determined on FP radiographs
(grade 0-3).
Chapter 5
Figure 3. The LCEA and ACEA A. The lateral centre edge angle (LCEA) as measured on an
anteroposterior pelvic radiograph. B. The anterior centre edge angle (ACEA) as measured
on a false profile radiograph.
Clinical measurements
At baseline, all individuals were clinically examined by measuring the range
of motion (ROM) of the hip. Active ROM was measured by a goniometer in
flexion, internal rotation, external rotation, abduction, and adduction. Further,
the presence and severity of hip and knee pain was assessed using the Visual
Analog Scale (VAS) at baseline and follow-up. This scale runs from 0 (no pain)
to 10 (very intense pain) .249
83
Chapter 5
Outcome measures
Development of incident OA was defined by K&L grade 2, 3, 4, or a total
hip replacement (THR) at 5-year follow-up. This outcome measure was further
subdivided in mild OA (K&L=2) and end-stage OA (K&L=3, 4, or THR). As it
has been described that pincer impingement may result in specific acetabular
cartilage damage anterosuperiorely, adjacent to the point where the abnormal
contact occurs, and posteroinferiorely, as a result of the contrecoup lesion, both
superior and inferior joint space narrowing (JSN) (grade ≥1) on the FP view was
used as an outcome measure.250 Grade 1 JSN is indicative for mild, but definite
disease.
Statistical analysis
Reliability of the LCEA and ACEA was assessed using intra-class correlation
coefficient (ICC). Univariate differences in baseline characteristics between
included and excluded hips, and between hips that developed OA and normal
hips, were evaluated by the Mann-Whitney test for continuous variables, by chi-
square test for sex, and by generalized estimating equations (GEE) for K&L score.
The strength of the independent relationship between acetabular coverage at
baseline and development of OA was calculated using GEE, adjusted for baseline
K&L classification, age, sex, and body mass index (BMI), and expressed in terms
of odds ratios (ORs). Using GEE allowed to model the correlation between the
left and the right hip. Hips with either a pincer deformity or mild dysplasia were
compared with a
reference group of hips with a centre edge angle ≥25° and ≤40° in the
corresponding radiographic view, representing normal acetabular coverage.
When mild OA or JSN was used as an outcome measure, baseline hips that had
developed end-stage OA at 5-year follow-up were excluded from analysis. When
end-stage OA was used as an outcome measure, hips that had developed mild
OA were excluded. Differences in pain scores at follow-up and ROM at baseline
between hips with and without mild acetabular dysplasia or a pincer deformity
were calculated using GEE, adjusted for K&L grade at baseline. All statistical
analyses were performed in SPSS version 20.0.
84
Pincer deformity and osteoarthritis
RESULTS
Participants
Baseline demographic data is summarized in table 1. The 720 included individuals
did not differ on any baseline characteristic from the 282 individuals who were
excluded. At baseline, the 720 individuals (1391 hips) were included in the
CHECK cohort because they consulted the GP for the first time with mild pain
in their hip (n=127, mean ± SD VAS hip pain 2.73 ± 2.07), pain in their knee
(n=292, mean ± SD VAS knee pain 2.71 ± 2.04), or pain in both hip and knee
(n=301, mean ± SD highest VAS pain 3.29 ± 2.12).
Chapter 5
Table 1. Baseline characteristics of the participants, stratified by the absence or presence
of incident OA.
Absence of Presence of p-value
Total incident OA at incident OA Absence vs
n=720 follow-up at follow-up presence
(1391 hips) n=657 n=80 of incident OA
(1294 hips) (97 hips) at follow-up
Age in years: mean (±SD) 55.9 (5.21) 55.8 (5.2) 58.0 (4.7) <0.001
Women, No. (%) 572 (79) 540 (81) 54 (68) 0.001
BMI, kg/m²: mean (±SD) 26.1 (4.1) 26.2 (4.2) 25.8 (3.7) 0.67
Length in cm: mean (±SD) 169.9 (8.2) 169.7 (8.1) 172.6 (8.8) 0.003
Weight in kg: mean (±SD) 75.3 (13.7) 75.2 (13.7) 76.9 (12.8) 0.116
K&L grade 0, No (%) 1045 (76) 1029 (81) 16 (17) <0.001
K&L grade 1, No (%) 324 (24) 248 (19) 76 (83) <0.001
Osteoarthritis classification
At five years follow-up, 97 (7.0%) hips developed incident OA, of which 39 hips
(2.8%) end-stage OA and 58 hips (4.2%) mild OA. On the FP view, JSN was
present superiorly in 107 hips (7.7%) and inferiorly in 107 hips (7.7%).
At baseline, the K&L grade could be scored reliably in 1369 hips. Of these hips,
76% did not show radiographic evidence of OA (K&L grade 0) whereas 24% had
doubtful OA (K&L grade 1). On the FP view, superior joint space was scored as
grade 0 in 94%, grade 1 in 5%, and grade 2 in 1%, and inferior joint space as
grade 0 in 95% and grade 1 in 5%.
85
Chapter 5
86
Table 2. Association (n=1391 hips) between both pincer deformity and mild acetabular dysplasia at baseline and both osteoarthritis
and JSN at 5-year follow-up
Incident OA Mild OA End-stage OA Inferior JSN Superior JSN
(n=97 hips) (n=58 hips) (n=39 hips) (n=107 hips) (n=107 hips)
Unadjusted Adjusted Unadjusted Adjusted Unadjusted Adjusted Unadjusted Adjusted Unadjusted Adjusted
Radiographic
OR OR OR OR OR OR OR OR OR OR
View
(95%CI) (95%CI) (95%CI) (95%CI) (95%CI) (95%CI) (95%CI) (95%CI) (95%CI) (95%CI)
0.78 0.53 0.96 0.62 0.33 0.28 1.44 1.40 0.88 0.76
AP LCEA>40°
(0.38-1.57) (0.25-1.10) (0.41-2.24) (0.26-1.46) (0.08-1.37) (0.07-1.21) (0.85-2.43) (0.81-2.43) (0.48-1.59) (0.41-1.41)
(n=201 hips)
p=0.48 p=0.090 p=0.92 p=0.27 p=0.13 p=0.09 p=0.17 p=0.23 p=0.67 p=0.38
0.96 0.77 0.78 0.68 0.83 0.69 1.50 1.48 1.11 1.00
FP ACEA>40°
PINCER (0.58-1.59) (0.43-1.37) (0.41-1.48) (0.33-1.41) (0.40-1.72) (0.32-1.50) (0.97-2.31) (0.93-2.34) (0.73-1.70) (0.62-1.59)
(n=418 hips)
p=0.96 p=0.77 p=0.45 p=0.30 p=0.62 p=0.35 p=0.07 p=0.10 p=0.63 p=0.99
AP+FP 0.52 0.34 0.81 0.53 1.10 1.04 1.10 0.89
LCEA+ACEA>40° (0.19-1.43) (0.13-0.87) (0.26-2.49) (0.19-1.45) X X (0.56-2.17) (0.50-2.16) (0.59-2.07) (0.47-1.66)
(n=141 hips) p=0.21 p=0.025 p=0.71 p=0.22 p=0.78 p=0.93 p=0.77 p=0.70
2.50 2.83 1.98 2.33 3.42 3.80 0.97 0.95 1.15 1.13
AP LCEA<25°
(1.53-4.09) (1.54-5.20) (1.03-3.78) (1.03-5.27) (1.74-6.72) (1.84 - 7.84) (0.54-1.73) (0.51-1.80) (0.65-2.01) (0.58-2.20)
(n=176 hips)
p=0.000 p=0.001 p=0.040 p=0.042 p=0.000 p=0.000 p=0.90 p=0.88 p=0.63 p=0.73
2.61 2.62 1.78 2.02 4.51 4.34 1.49 1.59 0.98 0.91
MILD FP ACEA<25°
(1.55-4.41) (1.44-4.77) (0.83-3.80) (0.85-4.76) (2.12-9.62) (1.99-9.47) (0.83-2.67) (0.86-2.94) (0.47-2.08) (0.38-2.18)
DYSPLASIA (n=144 hips)
p=0.000 p=0.002 p=0.14 p=0.11 p=0.000 p=0.000 p=0.18 p=0.14 p=0.97 p=0.83
5.45 4.88 4.98
AP+FP 3.64 2.59 6.48 0.92 1.04 1.28 1.49
(2.40- (1.38- (2.17-
LCEA+ACEA<25° (1.91-6.99) (0.99-6.81) (2.69-15.59) (0.36-2.31) (0.39-2.78) (0.45-3.64) (0.41-5.43)
12.34) 17.26) 11.42)
(n=62 hips) p=0.000 p=0.053 p=0.000 p=0.85 p=0.94 p=0.64 p=0.54
p=0.000 p=0.014 p=0.000
Odds ratios were adjusted for K&L grade, age, BMI, and sex at baseline. X= no cases of end-stage OA.
Hips with pincer or dysplasia were compared with a reference group with a centre edge angle ≥25° and ≤40° in the corresponding radiographic
view. Hips with pincer or dysplasia in both radiographic views were compared with a reference group with a centre edge angle ≥25° and ≤40 in both
radiographic views. Significant associations (p < 0.05) are indicated in bold.
87
Pincer deformity and osteoarthritis
Chapter 5
Chapter 5
Table 3. Difference (n=1391 hips) in amount of internal rotation at baseline and VAS pain
score at follow-up between either hips with a pincer deformity or mild acetabular dysplasia
and normal hips.
at baseline at follow-up
Radiographic
view No pincer Pincer p-value No Pincer Pincer p-value
30.82 29.87 1.77 1.96
AP LCEA>40° 0.035 0.79
(8.68) (8.45) (2.39) (2.19)
30.85 30.29 1.76 1.87
PINCER FP ACEA>40° 0.10 0.74
(8.62) (8.73) (2.34) (2.40)
30.76 30.02 1.77 2.00
AP+FP 0.27 0.77
(8.70) (8.19) (2.37) (2.29)
No No
Dysplasia p-value Dysplasia p-value
dysplasia dysplasia
30.57 31.43 1.76 2.03
AP LCEA<25° 0.96 0.05
(8.57) (9.21) (2.34) (2.50)
30.68 30.67 1.82 1.60
Dysplasia FP LCEA<25° 0.87 0.22
(8.58) (9.25) (2.35) (2.40)
30.65 31.27 1.80 1.61
AP+FP 0.10 0.78
(8.61) (9.48) (2.36) (2.37)
IR: internal rotation in degrees
88
Pincer deformity and osteoarthritis
Odds ratios were adjusted for K&L grade, age, BMI, and sex at baseline. X=
no cases of end-stage OA. Hips with pincer or dysplasia were compared with a
reference group with a centre edge angle ≥25° and ≤40° in the corresponding
radiographic view. Hips with pincer or dysplasia in both radiographic views were
compared with a reference group with a centre edge angle ≥25° and ≤40 in
both radiographic views.
DISCUSSION
This prospective study of individuals with first onset of mild pain complaints
Chapter 5
without radiographic evidence of definite OA at baseline shows that a pincer
deformity does not increase the risk for development of OA whereas mild
acetabular dysplasia is associated with development of OA. The latter is in
agreement with previously reported prospective studies.
In retrospective or case control studies on pincer impingement, some authors
showed a moderately increased risk for OA in the presence of a pincer
deformity,60,61,66,244 where others did not find an association or even suggested
a potential protective effect of a pincer deformity for development of OA.59,63-65
Our data supports the latter studies and found surprisingly a protective effect
of a pincer deformity when present both anterior and lateral in one hip. This
was supported by the fact that none of the hips with a pincer deformity on both
radiographic views (n=141 hips) developed end-stage OA.
Interestingly, a nearly significant association was found between a pincer
deformity on the FP view and JSN posteroinferiorely, in the specific region
where the so called ‘contrecoup lesions’ may occur (figure 1).57,250 Most hips
with posteroinferior JSN at follow-up already showed JSN at baseline, indicating
that this indirect measure of cartilage loss is not progressive in hips with pincer
deformity. This could suggest that these cases do impinge and suffer from a
limited ROM. However, we did not find a limited flexion, neither in hips with
anterior and/or lateral overcoverage nor in those hips with overcoverage and
JSN posteroinferiorely at follow-up.
We did not find a positive association between a pincer deformity and OA, but
regarding its slow progression the 5-year follow-up might have been too short.240
However, Nicholls et al. could neither identify an association between a higher
LCEA and THR after 19 years follow-up.63 Moreover, the first structure that fails
in pincer impingement is the labrum, which is associated with pain.240,251,252
However, a pincer deformity was neither associated with higher pain scores at
follow-up. Second, pincer impingement might lead to OA rapidly, before the age
89
Chapter 5
of 45 years, so that those subjects were not included in this cohort. However,
this is unlikely as it has been described that complaints as a result of pincer
impingement starts from the age of 30-40 years.240
A pincer deformity is a description of a variety of acetabular morphological
or orientation abnormalities causing excessive coverage of the femoral head.
The overcoverage may result from either a generally overcovered acetabulum,
i.e., when the femoral head is positioned deep inside the acetabulum or when
there is an overgrowth of the acetabular rim, or it may result from focalized
overcoverage, as seen in acetabular retroversion.248 General overcoverage can
be quantified by the LCEA as seen on an AP radiograph. Coxa profunda has also
been used to define general overcoverage, but two recent studies showed that
coxa profunda should not be used to define a pincer deformity.253,254 Acetabular
retroversion, resulting in anterior overcoverage, can be quantified on AP
radiographs by the cross-over sign or posterior wall sign, though both measures
have their limitations.255 We measured anterior coverage on the FP radiograph
by the ACEA, by which we assume that hips with anterior overcoverage due to
retroversion are accurately identified.256
Although an LCEA and ACEA threshold for a pincer deformity of >40° is
subjective, it is most often used to quantify overcoverage. Gosvig et al. found
cross-sectionally a mild association between a deep acetabular socket (LCEA
>45°) and a joint space width of ≤2 mm, but we neither found a significant or
increased risk of OA when using a threshold of 45° (see Table 4).57.
FAI is a motion dependent abnormal contact between the proximal femur and
acetabular rim due to a non-optimal morphology of either the proximal femur
(cam-type) or the acetabulum (pincer-type). Both types of FAI are being treated
with increasing frequency by restoring the normal anatomy to prevent the
abnormal contact to occur. This appears to be justified for cam impingement
regarding its relation with OA and the promising short-term to mid-term surgical
outcomes.203 However, for surgical treatment (acetabular rim trimming) of pincer
impingement caution is necessary. Based on our results, a pincer deformity
might even have a protective effect. Obviously, some symptomatic pincer
impingement patients might benefit from a surgical procedure, but proper
patient selection is critical.
For mild dysplasia, prospective studies generally showed an association with
OA, though cross-sectional and retrospective studies have confused this issue.
Two well designed studies showed a moderate increased risk for development
of OA.53,222 Lane et al. conducted a prospective nested case-control study with 8
years follow up in 176 women (mean age 70 years) and showed an aOR of 3.3
90
Pincer deformity and osteoarthritis
(95% CI 1.1-10.1) for acetabular dysplasia (LCEA<30°) and OA. Reijman et al.
defined acetabular dysplasia in 835 subjects (mean age 65 years) by an LCEA
<25° and found an aOR of 2.4 (95% CI 1.2-4.7) for OA as defined by a K&L
grade ≥2 and an aOR of 4.3 (95% CI 1.8-4.5) for JSN ≥1mm. It was suggested
that the association between acetabular dysplasia and OA might be higher in
a younger population. In the current study of 720 individuals with a mean age
of 55 years, we confirmed above mentioned associations between dysplasia
and OA and showed aORs of roughly 3 to 4 for OA. Other studies did not use
an additional FP view, and we showed a clearly increased aOR when acetabular
dysplasia was present both anteriorely and laterally in one hip.
Chapter 5
There are some limitations that need to be acknowledged. Although both an
anteroposterior and lateral radiographic view per hip was obtained, acetabular
retroversion might have been missed in some cases. Secondly, as opposed to
the LCEA, we were not able to draw a horizontal reference line for calculating
the ACEA. Still, a FP view might be more sensitive for the diagnosis of dysplasia
and showed additional predictive value as compared with an AP view alone.246,257
Thirdly, a pincer deformity is a risk factor for pincer impingement, but the
occurrence and frequency of impingement depends on patient activity, which
is unknown in this cohort. However, the same holds true for cam deformities,
which were highly predictive for OA in this cohort.203 Moreover, the occurrence
of pincer impingement might also depend on femoral version and the presence
of a cam deformity. We could not correct for femoral version, but the prevalence
of cam deformities was similar in hips with and without pincer deformities,
implying that this risk factor did not influence the association between pincer
deformities and OA. Individuals with a known diagnosis of acetabular dysplasia
were excluded from the CHECK cohort, implicating that the study subjects only
represent a mild, subclinical subset of acetabular dysplasia. Finally, the reader
should bear in mind that all participants had mild pain in their hips, knees, or
both at baseline, which might represent a very early stage of symptomatic OA,
although no definite radiographic OA was present at baseline.
In conclusion, a significant association between both lateral and anterior
acetabular dysplasia and development of OA within 5 years was found. The
strength of association increased when dysplasia was present both anteriorly
and laterally in one hip. In contrast, when a pincer deformity was present both
anteriorely and laterally in one hip, a protective effect for development of OA
was found. No association between a pincer deformity and joint space narrowing
at the specific anatomical locations of pincer impingement was identified. These
results questions the hypothesis that pincer impingement leads to OA.
91
Chapter 6
ABSTRACT
Objective To determine the association between cam impingement, which is hip
incongruity by a non-spherical femoral head and development of osteoarthritis.
94
Cam impingement and osteoarthritis
INTRODUCTION
Osteoarthritis is the most frequently occurring chronic joint disease and its main
characteristics are pain and dysfunction.26 Symptomatic OA affects 9.6% of men
and 18% of women aged 60 years or older and, with an ageing population,
the prevalence of OA will increase substantially.68 Although osteoarthritis
implies an increasing economic burden and detrimental impact on quality of
life,27 no methods are available to prevent the disease or delay its progression.
If modifiable risk factors can be identified at an early stage of the disease,
preventive measures might be implemented.72
Many cases of idiopathic hip osteoarthritis may be secondary to minor variations
in the morphology of the proximal femur.24 These minor variations in shape
were described by Ganz et al., who introduced the concept of femoroacetabular
impingement (FAI).55 FAI is a condition of abnormal contact between the femoral
Chapter 6
head-neck junction and the acetabulum, due to a bone shape abnormality on
either the femoral or acetabular side. Two types of FAI can be distuingished,55
cam impingement and pincer impingement. Pincer impingement is caused
by overcoverage of the acetabulum relative to the femoral head, and cam
impingement is caused by extra bone formation (a cam-type deformity) in the
anterolateral head neck junction.67,258 Such a deformity can develop in response
to sporting activities during adolescence.226,227 These cam-type deformities may
cause impingement against the acetabular rim, especially during flexion and
internal rotation of the hip (figure 1). Subsequently, a cascade of structural
damage (including labral tears and cartilage delamination) may occur, which
might gradually lead to osteoarthritis of the hip.32 Therefore, cam impingement
has been proposed as a biomechanical risk factor for the development of hip
osteoarthritis.208,240
However, there are no cohorts combining radiographic and clinical parameters
of cam impingement to study the risk of the development of osteoarthritis.
Furthermore, due to the lack of prospective studies on this topic, no causal
relationship between cam impingement and the development of osteoarthritis
has been established.241 Interestingly, cam impingement is a potentially
modifiable risk factor of osteoarthritis, as a cam-type deformity can be treated
surgically and diagnosed before the onset of severe hip damage
Assuming an association between cam impingement and the development of
osteoarthritis, we examined the association between clinical and radiographic
parameters suggestive of cam impingement at baseline and the risk of developing
osteoarthritis within 5 years.
95
Chapter 6
METHODS
Study design and participants
All subjects were extracted from the Cohort Hip and Cohort Knee (CHECK)
cohort. CHECK is a nationwide prospective cohort study of 1002 individuals
with early symptomatic osteoarthritis of knee or hip. On entry, all participants
had pain or stiffness of knee or hip and were aged 45-65 years; they had
not yet consulted their general practitioner for these symptoms, or the first
consultation was within 6 months before entry. Participants with a pathological
condition that could explain the symptoms were excluded (for hip: trauma,
rheumatoid arthritis, dysplasia, Perthes disease, subluxation, osteochondritis
dissecans, fracture, septic arthritis, Kellgren & Lawrence (K&L) grade 4 or total
hip replacement, previous hip surgery, and individuals having only symptoms of
bursitis or tendinitis).210
96
Cam impingement and osteoarthritis
Radiographs
Weight-bearing anteroposterior pelvic radiographs and faux profil oblique view
radiographs were obtained according to a standardised protocol at baseline,
Chapter 6
and at 2 and 5 years follow-up.245 For anteroposterior radiographs, feet were
positioned such that the medial side of the distal part of the first phalanx touched
and a wedge was used to assure 15° internal rotation. The tube to film distance
was 100 cm and the beam was centred at the top of the pubic symphysis.
Radiographic measurements
The shape of the proximal femur and acetabulum on the anteroposterior
radiographs was outlined using statistical shape modelling (SSM) software
(ASM tool kit, Manchester University, Manchester, UK). The shape on the
anteroposterior radiographs is given by a set of landmark points that are
positioned along the surface of the bone in the image. Each point is always
placed on the same landmark of the outline, to allow comparison between the
shapes. The alpha angle was automatically calculated from the points set of the
SSM software (figure 2), using Matlab (V.7.1.0)
The alpha angle measures the extent to which the femoral head deviates from
spherical and is a measure mostly applied to quantify a cam-type deformity. It is
measured by first drawing the best fitting circle around the femoral head, then a
line through the centre of the neck and the centre of the head. From the centre of
the femoral head a second line is drawn to the point where the superior surface
of the head-neck junction first departs from the circle. The angle between these
two lines is the alpha angle.224 We defined the radiographic presence of a cam-
type deformity by an alpha angle greater than 60° and greater than 83°.259,260
At baseline, and at 2 and 5 years follow-up, all anteroposterior pelvic radiographs
were scored for osteoarthritis according to the K&L classification (grade 0-4).211
In addition, superior and inferior joint space was determined on faux profil
97
Chapter 6
Figure 2. Cam impingement associates with end-stage osteoarthritis of the hip. (A) A
spherical femoral head with an alpha angle of 43°; (B) a cam-type deformity resulting in
an alpha angle of 78°; and (C) scatterplot of baseline measurements of the alpha angle
and internal hip rotation in 90° of flexion (the red dots indicate hips that developed end-
stage osteoarthritis within 5 years).
Clinical measurements
All individuals were clinically examined by measuring the range of motion
(ROM) of the hip. The ROM was measured in flexion, internal rotation, external
rotation, abduction, and adduction. The presence of reduced internal rotation
(in 90° of flexion) is suggestive of impingement. We used a threshold value of
20° or less.216 Furthermore, the presence and severity of hip pain were assessed
using validated self-report questionnaires, including the Western Ontario and
McMaster University Osteoarthritis Index (WOMAC, pain subscale) and the
Visual Analog Scale.249,261
98
Cam impingement and osteoarthritis
reliability, the point set was positioned twice by each investigator in 25 randomly
selected radiographs with an interval of 2 months. Intra-observer reliability was
tested for each investigator in 10 randomly selected radiographs.
Statistical analysis
Reliability of the alpha angle was assessed using intraclass correlation coefficient
(ICC) (two-way random, absolute agreement). Univariable differences in
baseline characteristics between included and excluded hips, and between hips
that developed osteoarthritis and normal hips, were evaluated by the Mann-
Whitney test for continuous variables, by the chi-square test for sex and by
generalised estimating equations (GEE) for K&L score. The strength of the
independent relationship between cam impingement parameters at baseline and
the development of osteoarthritis was calculated using GEE and expressed in
Chapter 6
terms of OR. Using GEE allowed to model the correlation between the left and
the right hip. To adjust for confounders, in the GEE model, K&L classification
and sex were entered as a factor, and body mass index and age as a covariate.
Hips with end-stage osteoarthritis at 5-year follow-up were excluded from
analysis when incident osteoarthritis was used as an outcome measure. The
positive predictive value of cam impingement parameters for osteoarthritis
was also calculated. Differences in pain scores and ROM between individuals
with and without radiographic cam-type deformity were calculated using the
Mann-Whitney test and GEE, respectively, the latter adjusted for K&L grade at
baseline. All statistical analyses were performed in SPSS V.15.0.
RESULTS
Participants
Of the 1002 individuals in the CHECK cohort, 89 were lost to follow-up, 16 were
not able to visit the hospital during the 5-year follow-up, and 32 refused to
undergo radiographs or had missing radiographs. For this study, 865 individuals
(682 women, 183 men; aged 45-65 years) were included. At baseline, these
865 individuals (1730 hips) were included in the CHECK cohort because of
the first onset of pain in their hip (n=148), knee (n=354), or both (n=363).
Anteroposterior hip radiographs (instead of anteroposterior pelvic radiographs)
were obtained in the first 124 individuals of the CHECK cohort. These individuals
were excluded from baseline measurements because an anteroposterior hip view
is known to influence the alpha angle.248 Radiographs of insufficient quality were
also excluded (figure 3). To calculate the OR, all hips with an anteroposterior
99
Chapter 6
Figure 3. Flow of subjects from the start of the cohort to the final study population.
Osteoarthritis classification
Of the 1411 included hips, 4.25% (n=60) developed incident osteoarthritis and
2.76% (n=39) end-stage osteoarthritis within 5 years follow-up. THR patients
did not have other hip pathology than osteoarthritis nor other hip surgery during
follow-up.
At baseline, the K&L grade could be scored reliably in 1389 hips. Of these hips,
76% had no signs of radiographic hip osteoarthritis (K&L grade 0) and 24%
had doubtful radiographic hip osteoarthritis (K&L grade 1). On the faux profil
view, superior joint space narrowing was scored as grade 0 in 94%, grade 1 in
5%, and grade 2 in 1% (n=1379), and inferior joint space narrowing as grade
0 in 95% and grade 1 in 5% (n=1378). Of those hips having K&L grade 0 on
the anteroposterior view, 2% had a superior and 5% an inferior joint space
narrowing grade 1 on the faux profil view.
100
Table 1. Baseline characteristics of the participants, stratified by the absence or presence of end-stage osteoarthritis and incident
osteoarthritis.
Absence of Presence of p-value Absence of Presence of p-value
end-stage end-stage Absence vs incident incident Absence vs
Total
osteoarthritis osteoarthritis presence of osteoarthritis osteoarthritis presence of
n=723 (1411
at follow-up at follow-up end-stage at follow-up at follow-up incident
hips)
n=686 n=37 osteoarthritis n=669 n=52 osteoarthritis
(1372 hips) (39 hips) at follow-up (1312 hips) (60 hips) at follow-up
Age in years: mean (±SD) 55.9 (5.20) 55.9 (5.2) 57.7 (4.2) 0.032 55.8 (5.2) 58.2 (4.9) 0.001
Women, No. (%) 575 (80) 552 (80) 23 (62) 0.004 538 (80) 36 (69) 0.017
BMI, kg/m²: mean (±SD) 26.1 (4.1) 26.1 (4.2) 25.4 (4.0) 0.25 26.2 (4.2) 26.0 (3.5) 0.78
Length in cm: mean (±SD) 169.9 (8.2) 169.8 (8.2) 171.3 (9.3) 0.44 169.7 (8.1) 173.1 (8.5) 0.003
Weight in kg: mean (±SD) 75.3 (13.6) 75.3 (13.7) 74.6 (13.0) 0.97 75.2 (13.7) 78.0 (12.7) 0.064
K&L grade 0, No (%) 1058 (76) 1048 (78) 10 (26) <0.001 1042 (80) 6 (11) <0.001
K&L grade 1, No (%) 331 (24) 302 (22) 29 (74) <0.001 253 (20) 49 (89) <0.001
For the comparison between absence and presence of incident OA, hips which developed end-stage OA were excluded.
BMI, body mass index; K&L, Kellgren and Lawrence.
101
Cam impingement and osteoarthritis
Chapter 6
Chapter 6
Osteoarthritis classification
Of the 1411 included hips, 4.25% (n=60) developed incident osteoarthritis and
2.76% (n=39) end-stage osteoarthritis within 5 years follow-up. THR patients
did not have other hip pathology than osteoarthritis nor other hip surgery during
follow-up.
At baseline, the K&L grade could be scored reliably in 1389 hips. Of these hips,
76% had no signs of radiographic hip osteoarthritis (K&L grade 0) and 24%
had doubtful radiographic hip osteoarthritis (K&L grade 1). On the faux profil
view, superior joint space narrowing was scored as grade 0 in 94%, grade 1 in
5%, and grade 2 in 1% (n=1379), and inferior joint space narrowing as grade
0 in 95% and grade 1 in 5% (n=1378). Of those hips having K&L grade 0 on
the anteroposterior view, 2% had a superior and 5% an inferior joint space
narrowing grade 1 on the faux profil view.
102
Cam impingement and osteoarthritis
Table 2 Association (n=1372 hips) between cam impingement parameters and incident osteoarthritis
(n=60) at 5-year follow-up
Unadjusted Adjusted
Cam impingement parameter Incident No Incident OR OR†
(95% CI) (95% CI)
Radiographic Alpha angle >60º n = 19/139 n = 120/139 4.43 2.42
Referent group ≤60º n=41/1233 n=1192/1233 (2.36-8.32) (1.15-5.06)
p<0.0001 p=0.020
Alpha angle >83º n = 4/42 n = 38/42 2.25 1.09
Referent group ≤83º n=56/1330 1274/1330 (0.78-6.44) (0.35-3.38)
p=0.13 p=0.88
Clinical Hip internal rotation n =15/200 n = 185/200 2.27 1.47
≤20º n=45/1172 n=1127/1172 (1.19-4.33) (0.74-2.90)
Referent group >20º p=0.013 p=0.27
Combined Alpha angle>83º n = 1/9 n = 8/9 2.97 0.75
and hip internal n=59/1363 n=1304/1363 (0.49-18.17) (0.10-5.68)
rotation ≤20º p=0.24 p=0.78
Referent group: hips not
Chapter 6
having both conditions
† Adjusted OR was corrected for K&L grade, age, body mass index, and sex at baseline.
103
Chapter 6
Table 3 Association (n=1411 hips) between cam impingement parameters and end-stage
osteoarthritis (n=39) at 5-year follow-up
No
=End-stage Unadjusted Adjusted
end-stage
Cam impingement parameter osteoarthritis OR OR†
osteoarthritis
At follow-up (95% CI) (95% CI)
at follow-up
Radiographic Alpha angle >60º n = 139/156 n = 17/156 6.82 3.67
Referent group n=1233/1255 n=22/1255 (3.55-13.10) (1.68-8.01)
≤60º
Alpha angle >83º n = 42/56 n = 14/56 17.71 9.66
Referent group 1330/1355 n=25/1355 (8.71-36.02) (4.72-19.78)
≤83º
Clinical Hip internal n = 200/223 n =23/223 9.34 7.13
rotation ≤20º n=1172/1188 n=16/1188 (4.66-18.71) (3.38-15.04)
Referent group
>20º
Combined Alpha angle>83º n = 9/19 n = 10/19 53.79 25.21
and hip internal n=1363/1392 n=29/1392 (20.96- (7.89-80.58)
rotation ≤20º 138.08)
Referent group:
hips not having both
conditions
All p-values for OR are <0.0001
† Adjusted OR are corrected for K&L grade, age, body mass index, and sex at baseline.
Table 4 Range of motion and pain scores in all hips and in hips with a cam-type deformity.
Alpha angle >60° Alpha angle >83°
Total
Clinical Parameter (n=156) (n=56)
(n=1411*)
p-value† p-value‡
Range of Internal rotation 26.31 (8.61) 25.76 (9.59)
30.64 (8.67)
motion mean <0.001 <0.001
(±SD) External rotation 27.00 (8.62) 25.08 (7.53)
29.27 (9.88)
0.002 0.01
Flexion 117.10 (12.47) 117.27 (12.99)
118.64 (11.21)
0.18 0.31
Adduction 20.19 (8.61) 19.74 (9.24)
21.73 (8.40)
0.23 0.02
Abduction 31.56 (10.03) 30.56 (10.42)
33.81 (10.95)
0.72 0.30
Pain scores
23.56 (16.75) 23.39 (19.07)
mean (±SD) WOMAC pain subscale 24.43 (16.93)
0.45 0.33
Abbreviations: VAS, visual analog scale; WOMAC, Western Ontario and McMaster University
Osteoarthritis Index
*
n=1410 for internal rotation and flexion, n=920 for external rotation, adduction, and abduction
† Compared with hips with alpha angle ≤60°
‡Compared with hips with alpha angle ≤83°
104
Cam impingement and osteoarthritis
DISCUSSION
This first prospective study on cam impingement shows a strong association
between cam impingement in radiographically non-arthritic individuals presenting
with the first onset of pain complaints at baseline, and the development of
osteoarthritis within five years. The results indicate that hip incongruity by a
cam-type deformity is highly predictive of especially fast progression to end-
stage osteoarthritis.
The data suggest that the risk for end-stage osteoarthritis is higher when the
cam-type deformity is more pronounced. This is in line with a study reporting
a correlation between the magnitude of the alpha angle and the severity of
acetabular cartilage damage as measured with delayed gadolinium-enhanced
contrast MRI of cartilage.262 Another group examining the relationship between
morphological parameters and the risk of receiving THR within 19 years,
Chapter 6
concluded that the alpha angle was significantly associated with THR in women,
independent of the presence of radiographic hip osteoarthritis at baseline; they
reported an adjusted OR for receiving a THR of 1.052 for every degree increase
in the alpha angle.63
Cross-sectional studies have shown a positive association between cam
morphology and osteoarthritis. One study showed an estimated OR of 2.2 for
having a non-spherical femoral head and osteoarthritis as defined by a minimum
joint space width of 2 mm or less.60 Another group found an estimated OR
(adjusted for confounders) of 6.95 (95% CI 4.64-10.41) for the association of a
visually scored non-spherical femoral head with osteoarthritis in a case-control
study.208 This latter value lies within the range of our findings for a radiographic
cam-type deformity.
In those studies it was noticed that a non-spherical head could also partly be a
consequence of rather than a cause of osteoarthritis. However, in the present
study, as all hips had only doubtful or no signs of radiographic osteoarthritis at
baseline, we could exclude a cam-type deformity from being a consequence of
osteoarthritis. In addition, we found no increase in cam-type deformity over
time, as the alpha angle at 5 years follow-up did not differ from that at baseline
in either normal hips or hips that developed osteoarthritis. This indicates that
cam-type deformity is probably not a consequence of a dynamic remodelling
process in this adult cohort, but is a more or less static phenomenon. There
is some evidence that this deformity develops during puberty by (over)active
participation in certain sports.226,227
As only four hips had a K&L score of 3 or greater, most of the hips were judged
to be end-stage osteoarthritis based on having received a THR. Although THR
105
Chapter 6
106
Cam impingement and osteoarthritis
Chapter 6
for decreased internal rotation.216
Individuals with symptomatic cam impingement are eligible for treatment
of the deformity in an open or arthroscopic surgical procedure in which the
femoral head is shaved to its normal congruity.266 Such a treatment is currently
performed with increasing frequency in an attempt to relieve complaints and
to halt the destructive cascade. The short-term and mid-term results of this
treatment show a relief in pain and an increase in hip function; however, no
randomised (placebo) controlled trial is available and the long-term results on
the prevention of osteoarthritis are unknown.267 The findings of the present
study justify further study on the treatment of cam impingement, and indicate
that follow-up periods of a few years seem sufficient to show a potential effect
of treatment.
In conclusion, cam impingement is strongly related to the development of hip
osteoarthritis. Our results indicate that individuals who present with the first
onset of pain complaints, and have both a radiographic cam-type deformity and
reduced internal hip rotation, are at high risk of fast progression to end-stage
osteoarthritis. A cam-type deformity might be a modifiable risk factor that can
be diagnosed before severe hip damage is present, providing an opportunity to
prevent hip osteoarthritis.
107
PART 2
Definition of femoroacetabular
impingement
Chapter 7
Femoroacetabular impingement:
Defining the condition and its role in the
pathofysiology of osteoarthritis
ABSTRACT
Femoroacetabular impingement (FAI) is an increasingly recognized cause of hip
pain. It is best defined as a pathological mechanical process by which morphologic
abnormalities of the acetabulum and/or femur combined with vigorous hip
motion lead to repetitive collisions that damage the soft-tissue structures
within the joint itself. Based on cross-sectional studies in which FAI morphology
was studied before the presence of radiographic osteoarthritis (OA), and on
prevalence studies in younger, asymptomatic persons, it is clear that FAI and its
morphologic risk factors are common in young adult hips and predispose to the
later development of OA in certain patients. Longitudinal studies also support
the assertion that, in middle-aged adults, the presence of cam deformities at
baseline substantially increases the risk of developing OA and the need for total
hip arthroplasty. More long-term data are needed to better define the natural
history of pincer deformities as well as FAI in younger cohorts.
112
Definition of FAI
INTRODUCTION
Femoroacetabular impingement (FAI) is an increasingly recognized and treated
condition in patients presenting with pain about the hip, particularly active
young adults. In spite of the increased interest in this condition, the nature of
this clinical entity remains ill defined. Furthermore, the idea that FAI may be
a risk factor for the development of early hip osteoarthritis (OA) is viewed by
many as unproven. The primary purpose of this article is to define FAI, with
the goal of developing an operational definition for the purpose of clinical trial
planning. Secondary goals are to summarize the current knowledge about the
possible relationship between FAI and the later development of hip OA and to
define areas for future research.
Chapter 7
acetabulum and/or femur, combined with vigorous hip motion (especially at the
extremes), lead to repetitive collisions that damage the soft-tissue structures
within the joint itself. This definition contains five essential elements: (1)
abnormal morphology of the femur and/or acetabulum, (2) abnormal contact
between these two structures, (3) especially vigorous supraphysiologic motion
that results in such abnormal contact and collision, (4) repetitive motion resulting
in the continuous insult, and (5) the presence of soft-tissue damage. Although
FAI can be further divided into subcategories based on whether the primary
deformity is situated in the femur, acetabulum, or both, these five elements are
both necessary and sufficient for diagnosis.55
Although hip symptoms are not among the five essential elements of FAI
required for an operational definition to determine eligibility for clinical trials
investigating the management of FAI, individuals should be symptomatic in
the affected hip. Their clinical presentations, based on the history and physical
examination, should be consistent with the diagnosis of FAI.268,269 This generally
includes the gradual onset of hip pain, which occasionally may radiate laterally
toward the trochanteric region, medially into the adductor region, and rarely
into the buttocks or even down to the knee. In most patients, symptoms are
exacerbated by periods of hip flexion, such as prolonged sitting. Typical findings
include loss of motion, particularly with hip flexion and internal rotation, and
a positive impingement test (ie, pain with hip flexion to 90°, adduction, and
internal rotation).269
113
Chapter 7
114
Definition of FAI
Chapter 7
maturation.227,278
Jung et al.277 evaluated 838 asymptomatic hips in 419 randomly selected
patients aged 25 to 92 years who underwent abdominal or pelvic CT for other
medical diseases unrelated to the hip. AP scout views of CT scans, which provide
information similar to that of AP pelvic radiographs, were used to measure
asphericity of the femoral head-neck junction using the alpha angle as described
by Nötzli et al.224 Study participants were classified according to criteria defined
by Gosvig et al.260 based on the Copenhagen Osteoarthritis Study. For men, three
ranges of values were defined for the alpha angle: pathologic (≥83°), borderline
(69° to 82°), and normal (≤68°). For women, these values were pathologic
(≥57°), borderline (51° to 56°) and normal (≤50°).277 In 49 men aged ≤50
years, the mean alpha angle was 57.2°, with 14% demonstrating pathologic
alpha angles. A similar magnitude and distribution of alpha angles were found in
men older than 50 years. In 107 women younger than 50 years, the mean alpha
angle was 45.3°, with 7% demonstrating pathologic alpha angles. The authors
reported similar findings in 433 women older than 50 years.
115
Chapter 7
Reichenbach et al.272 General population 19.9 All male 244 (244) MRI
(18–24)
Hack et al.275 General population 29.4 44/56 200 (400) MRI
(21–50.6)
Laborie et al.276 General population 18.6 42/58 2,060 (4,120) AP and frog-
(young adults) (17.2–20.1) leg lateral
radiography
Kapron et al.278 Collegiate football 21 ± 1.9 All male 67 (134) AP and frog-
players leg lateral
radiography
a
Multiplanar imaging was used in three studies bPrevalence per joint c Visually scored (semiquantitatively).
AA = alpha angle, AI = acetabular index, AOR = anterior offset ratio, AVA = acetabular version angle,
FAI = femoroacetabular impingement, FHNO = femoral head-neck offset, FP = focal prominence, IR =
internal rotation, IS = impingement sign, LCEA = lateral center-edge angle, OA = osteoarthritis, PGD =
pistol grip deformity, PWS = posterior wall sign, TI = triangular index
116
Definition of FAI
Chapter 7
AA >55° in the 1:30 33.5% in ≥1 Males, 51.7%; females, 18.9%
(anterosuperior) position hip (13%
bilateral)
No Cam TI ≥0 mm All: 10.5% Males, 19.6%; females, 5.2%
Pincer LCEA ≥45° All: 17.8% Males, 15.2%; females, 19.4%
Combined All: 1.6% Males, 2.9%; females, 0.9%
No Cam PGD, FP of femoral neck, All: ≥1 Males, 35%; females, 10.2%
flattening of femoral head findingc in
20.6%
Pincer Crossover sign, PWS, excessive All: ≥1 Males, 34.3%; females, 16.6%
acetabular overcoverage findingc in
24.1%
No Cam Pathologic (males): AA ≥83° 14%b
Borderline (males): AA 69°–82° 14.9%b
Pathologic (females): AA ≥57° 5.6%b
Borderline (females): AA 51°–56° 6.1%b
No Cam AA >50° 72%b 78%b had ≥1 sign of cam
deformity
FHNO <8 mm 64%b
Pincer LCEA >40° 7%b 66%b had ≥1 sign of pincer
deformity
AI <0° 16%b
Crossover sign 61%b
Combined Mixed impingement 50%b 95%b had ≥1 sign of cam or
pincer deformity
a
Multiplanar imaging was used in three studies bPrevalence per joint c Visually scored (semiquantitatively).
AA = alpha angle, AI = acetabular index, AOR = anterior offset ratio, AVA = acetabular version angle,
FAI = femoroacetabular impingement, FHNO = femoral head-neck offset, FP = focal prominence, IR =
internal rotation, IS = impingement sign, LCEA = lateral center-edge angle, OA = osteoarthritis, PGD =
pistol grip deformity, PWS = posterior wall sign, TI = triangular index
117
Chapter 7
The Genetics of Osteoarthritis and Lifestyle (GOAL) study recruited subjects from
Nottingham, United Kingdom, and compared hip morphology in 566 persons
with unilateral hip OA with the contralateral asymptomatic hip as well as with
nonosteoarthritic control hips in 1,100 patients who had undergone intravenous
urography.208 The assumption was that hip morphology was symmetric and that
the contralateral hip represented the premorbid state. All patients were aged ≥45
years (mean, 66 years), and 48% of the subjects were women. Cam morphology
was defined as a ratio of head-to-neck diameter <1.27. Based on this criterion,
the investigators reported a 5.5% risk of developing OA in the contralateral
nonosteoarthritic hip, compared with a 3% risk in the control group. Similarly,
pistol grip deformity conferred an 8.3% risk of developing OA in the contralateral
nonosteoarthritic hip, compared with a 3.6% risk in the control group. Both
measures were associated strongly and significantly with an increased risk of
developing hip OA, after adjusting for age, body mass index (BMI), and sex.
Longitudinal studies of FAI are also critical to establish the temporal relationship
of cause to effect. By their nature, they ensure that the exposure to the risk
factor occurs before development of the disease by measuring proximal femoral
morphology before it is altered by OA. A prospective study of FAI in the Cohort
Hip and Cohort Knee (CHECK) study reported on a Dutch national sample of
723 patients presenting for the first time with recent onset of hip or knee pain,
thereby representing a sample of early potential OA.203 Of the cohort, 80% were
female, ranging in age from 45 to 65 years. Subjects had doubtful or no OA
at baseline, with a Kellgren-Lawrence osteoarthritic grade of 0 or 1 (76% and
24%, respectively). Initial AP pelvic radiographs were measured to determine
the alpha angle, and subjects were followed for 5 years to determine the risk of
developing end-stage OA. Hips with a baseline AP alpha angle >83° had a 25%
risk of developing end-stage OA within 5 years, compared with a <2% risk of
end-stage OA in hips with an alpha angle of <83° (OR, 9.7 [adjusted for age,
sex, BMI, and baseline Kellgren-Lawrence grade]). Interestingly, hips with both
an alpha angle >83° and decreased internal rotation ≤20° had a 53% risk of
developing end-stage OA within 5 years.
Gregory et al.206 performed a longitudinal study of hip OA in a population-based
sample drawn from residents of Rotterdam, the Netherlands. In the study, 835
subjects aged 55 to 80 years underwent an initial AP pelvic radiograph, with a
follow-up radiograph 6 years later. Of these, 57% had no definite evidence of OA
at baseline (Kellgren-Lawrence grade 0 or 1). Progression to significant OA at the
6-year follow-up was defined by the need for total hip arthroplasty (THA) or the
presence of Kellgren-Lawrence grade 3 or 4 OA. Statistical shape modeling was
used to characterize the main variants in shape of the proximal femur at baseline
118
Definition of FAI
and in changes over time. Baseline shape variants were compared between 55
patients with minimal arthritis (Kellgren-Lawrence grade 0 or 1) who developed
advanced arthritis or needed THA compared with 55 control subjects with no
arthritic progression. Statistical shape mode scores that were consistent with
flatter baseline femoral heads and reduced femoral head-neck offset in patients
without OA at baseline were associated significantly with an increased risk of
developing OA and requiring THA. For each standard deviation difference in this
femoral head shape mode score, the risk of developing advanced radiographic
OA of the hip increased 62% and the risk of THA increased 135%.
The Chingford cohort is another recent longitudinal study of the relationship
between hip morphology and the development of OA.63 The study enrolled 1,003
healthy women aged 44 to 67 years after they underwent baseline AP radiography
of the pelvis, which was measured for several morphologic characteristics,
including the presence of cam and pincer deformity. At 19-year follow-up, these
radiographs were repeated. The baseline hip morphology of the 25 hips that
went on to require THA was compared with a random sample of 243 hips that
did not require THA. The median AP alpha angle in those who required THA
Chapter 7
was 62.4° at baseline, compared with 45.8° in the control group (P = 0.001).
The odds ratio of needing a THA increased 1.05 for each 1° increase in initial
alpha angle. Additional results from the Chingford cohort were presented at the
2012 Osteoarthritis Research Society International meeting in Barcelona, Spain.
In women, hips with an AP alpha angle ≥65° at baseline had a higher risk of
incident radiographic hip OA than did those with an alpha angle of <65°. The
odds ratio was 2.7, adjusted for age, BMI, and baseline hip joint space width (P
<0.001). Although this and other longitudinal studies show strong associations
between cam deformities and later OA (table 2), it should be noted that, in this
study, as in the others, the overwhelming majority of hips with cam deformities
did not reach end points of THA or radiographic OA even up to approximately a
19-year follow-up.63
119
Chapter 7
Doherty et al.208 Case-control Subjects from GOAL cohort. C: C: 50/50 2,076
C: clinically severe OA; 67.7±7.1 Con: 54/46 (4,152
Con: persons from UK without OA. Con: [C:965/
64.3±8.4 Con:1,111])
Reichen bach Cross- Subjects from Sumiswald cohort, 19.9 All male 244
et al.264 sectional general population of young (18–24) (244)
adults
Gosvig et al.60 Cross- Subjects from Copenhagen Heart 60.5 37/63 3,620
sectional Study (22–90) (7,240)
Chung et al.66 Cross- Subjects from Korean study on 71.7 43/57 674
sectional health and aging (all ≥65 y) (1,378)
Ecker et al.244 Retrospective Contralateral hips of subjects who C: 57.5 C:73/26 119 (119)
case-control underwent THA. C: Tönnis grade (36-85) Con: C: 94
2; Con: Tönnis grade 0 or 1 Con: 5.8 52/48 Con:25
(60–82)
120
Definition of FAI
Chapter 7
Cartilage deformityc Cam hips: ~0.2 mm decreased
thickness cartilage thickness
AP radiography No JSW <2 mm Cam TI ≥0mm RR, 2.2 (95% CI, 1.7–2.8)
Pincer LCEA ≥45° RR, 2.4 (95% CI, 2.0–2.9)
AP radiography No Minimum Pincer LCEA ≥40° NS
JSW ≤2.0mm
Minimum OR, 2.3 (95% CI, 1.5–3.4)
JSW ≤2.5mm
AP radiography No Any subtle Cam PGDc or AA>68 1/17 hips developed OA
indication of (men), AA >50°
decreased (women), or
JSW and/or NSA <125°
osteophyte
Formation Pincer LCEA ≥35°, IA 7/34 hips developed OA
≤45°, crossover
sign, or PWS
Mixed Any 9/45 hips developed OA
combination of
cam and pincer
AP radiography No Tönnis grade 3 Cam MPFA <84 OR, 20.6 (95% CI, 3.4–34.8)
or THA (AA, NSA, LCEA: NS)
Pincer PWS OR, 10.2 (95% CI, 1.0–99.8)
(crossover sign, coxa profunda,
protrusio acetabuli: NS)
AP, cross- No THA Cam AA (LM) OR, 1.98 (95% CI, 1.33–2.95)
table, and for every 10° increase
false-profile FHR (LM) OR, 0.23 (95% CI, 0.07–0.82)
radiography for every 0.1 increase
Pincer LCEA (LM) OR, 0.87 (95% CI, 0.79–0.95)
CT, AP No Tönnis grade 2 Cam AA (LM) OR, 1.09 (95% CI, 1.03–1.15)
radiography Pincer LCEA (LM) OR, 1.14 (95% CI, 1.02–1.27)
121
Chapter 7
Based on the previously mentioned studies, in which FAI morphology was studied
before the presence of radiographic OA, and the prevalence studies in younger,
asymptomatic persons, it is clear that FAI and its morphologic risk factors are
common in young adult hips and predispose to the later development of OA in
certain patients. Longitudinal studies also support the assertion that, in both men
and women aged 45 to 65 years, the presence of cam deformities at baseline
substantially increases the risk of developing OA and the need for THA. In most
hips, however, the presence of a cam lesion is not sufficient in itself to lead to the
development of clinically significant and symptomatic OA. Currently, insufficient
evidence from longitudinal population studies exists to confirm a similar
association between the presence of a pincer deformity and the development
of clinical or radiographic OA. It also should be pointed out that many of the
longitudinal epidemiologic studies of FAI and hip OA are based on data from AP
pelvic radiographs, which are insensitive to milder cam deformities and those
located more anteriorly280 and may not assess pincer deformities accurately.281
Thus, these epidemiologic studies may underestimate the true association of
FAI-related deformities with the risk of developing hip OA. Evidence also is
insufficient to demonstrate a causal link between the presence of radiographic
FAI morphology, clinical findings of FAI, and FAI-associated tissue damage in
younger persons and the long-term risk of developing OA of the hip in mid to
late life.
FUTURE DIRECTIONS
Using existing longitudinal population studies, it will be important in the future
to determine which persons with FAI-related morphologic abnormalities are at
greatest risk of developing hip OA at a relatively young age and which are at
increased risk of hip OA in mid to late life. It also will be important to identify
the subject- and limb-specific risk factors of OA. We hypothesize that these risk
factors are likely to include the type, severity and combinations of the anatomic
abnormalities, genetic influences, activity levels, and clinical findings at baseline.
In addition, long-term follow-up of younger cohorts is needed to fully describe
the natural history of FAI and hip OA and to determine the relationship between
the type and severity of intra-articular damage and the longer-term risk of
developing clinically significant hip OA. Longitudinal data on the progression of
morphologic abnormalities in hips with FAI that are in the process of developing
osteoarthritic changes should be used to investigate whether the deformities
progress in a way that could increase the risk of hip impingement and worsen
the prognosis.
122
Chapter 8
ABSTRACT
Introduction Cam impingement is characterized by abnormal contact between
the proximal femur and acetabulum caused by a non-spherical femoral head,
known as a cam deformity. A cam deformity is usually quantified by the alpha
angle; greater alpha angles substantially increase the risk for osteoarthritis
(OA). However, there is no consensus on which alpha angle threshold to use to
define the presence of a cam deformity.
Aim To determine alpha angle thresholds that define the presence of a cam
deformity and a pathological cam deformity based on development of OA.
Methods Data from both the prospective CHECK cohort of 1002 individuals (45-
65 years) and the prospective population-based Chingford cohort of 1003 women
(45-64 years) with respective follow-up times of 5 and 19 years were combined.
The alpha angle was measured at baseline on anteroposterior radiographs, from
which a threshold for the presence of a cam deformity was determined based
on its distribution. Further, a pathological alpha angle threshold was determined
based on the highest discriminative ability for development of end-stage OA at
follow-up.
Results A definite bimodal distribution of the alpha angle was found in both
cohorts with a normal distribution up to 60°, indicating a clear distinction
between normal and abnormal alpha angles. A pathological threshold of 78°
resulted in the maximum area under the ROC curve.
124
Definition of a cam deformity
INTRODUCTION
Historically, the cause of most hip osteoarthritis (OA) has been defined as
‘idiopathic’, but recent evidence suggests that development of hip OA is largely
influenced by the presence of a cam deformity.55,203,208 A cam deformity is
characterized by extra bone formation at the anterolateral head-neck junction
resulting in a non-spherical cam-shaped deformity.67 It is forced into the
acetabulum during flexion and internal rotation of the hip, a process referred
to as cam impingement.75,240 In time and with repeated movement, the cam
deformity might damage the soft tissue structures of the hip, leading to pain,
decreased function, and subsequently OA of the hip.63,203,264 In the prospective
CHECK cohort, an odds ratio (OR) of 9.7 (95% CI 4.7-19.8) was found for a large
cam deformity (alpha angle >83°) at baseline and subsequent development
of end-stage OA after 5 years. Moreover, in a case control study within the
prospective Chingford cohort, an OR of 1.05 (95% CI 1.02-1.09) was found
for every degree increase in alpha angle at baseline and receiving total hip
replacement (THR) within 19 years follow-up.63,203
The presence of a radiographic cam deformity is a common finding with
prevalence numbers of roughly 15-25% in males and 5-15% in females.60,272,275
The wide range of prevalence reported is mainly due to the inconsistency in the
definition of what is a cam deformity. A cam deformity is commonly assessed by
the alpha angle, which measures the extent to which the femoral head deviates
Chapter 8
from spherical.224 Greater alpha angles increase the risk for development of
OA substantially.63,203,262,282,283 However, there is neither a validated alpha
angle threshold value to define the presence of a cam deformity, nor a
pathological threshold that indicates an increased risk for development of OA.
As a consequence, threshold values ranging from 50° to 83° have been used in
literature, which makes prevalence numbers and associations with subsequent
pathology difficult to compare and interpret.75,224,260
In order to determine alpha angle thresholds, large cohort studies are needed.
For that reason, we combined data of the CHECK cohort and Chingford cohort,
both with prospective follow-up. Using these data, the aim of this article is to
determine an alpha angle threshold for defining the presence of a cam-type
deformity, and to determine a pathological alpha angle threshold based on
development of OA at follow-up.
METHODS
Study population
The alpha angle threshold values were determined in the CHECK cohort with a
current follow-up of 5 years, and in the Chingford study with a follow-up of 19
years.
125
Chapter 8
Radiographs
In the CHECK study, weight-bearing Anterio-Posterior (AP) pelvis radiographs
were obtained from the 11 participating research centers according to a
standardised protocol, taken at baseline and at 2 and 5 years follow-up. Feet
were positioned such that the medial side of the distal part of the first phalanx
touched and a wedge was used to assure 15° internal rotation. In the Chingford
cohort, each woman had a standardised supine AP pelvis radiograph, taken at
years 2, 8 and 20. A small sand bag under the knees was used to minimize hip
rotation.
In both the CHECK and Chingford cohorts, AP pelvis radiographs were scored
atlas based and ‘blind’ to clinical details according to the method of Kellgren and
Lawrence (K&L) at baseline, and at 5 year follow-up in the CHECK cohort and at
year 20 in the Chingford cohort.211,232. End-stage OA was defined by K&L grade
3, 4, or total hip arthroplasty (THA) at follow-up.
Alpha angle
The alpha angle measures the extent to which the femoral head deviates from
spherical. It is measured by first drawing the best fitting circle around the
femoral head, then a line through the centre of the neck and the centre of the
head. From the centre of the femoral head, a second line is drawn to the point
where the superior surface of the head-neck junction first departs from the
circle. The angle between these two lines is the alpha angle (Figure 1).224
In both cohorts, the alpha angle was semi-automatically calculated. In the
CHECK study, the shape of the proximal femur was outlined by a set of points
that were positioned on anatomical landmarks using statistical shape modeling
126
Definition of a cam deformity
(SSM) software (ASM tool kit, Manchester University, Manchester, UK). From this
points set, the alpha angle was calculated using Matlab (V.7.1.0).203,221 In the
Chingford cohort, the alpha angle was also measured using a validated Matlab
based (Matlab R2009b; MathWorks) software package called Hip Morf 2.0.
Figure 1. The alpha angle quantifies the asphericity of the femoral head. A. a normal
alpha angle of 41° is shown representing a spherical femoral head B. an abnormal alpha
angle of 98° is shown representing a cam deformity.
Chapter 8
Reliability of the alpha angle was examined in both cohorts and between both
techniques. In the CHECK cohort, interobserver reproducibility was examined
by positioning the point set twice in 25 randomly selected hips by three
investigators. Intra-observer repeatability was tested for each investigator
in 10 randomly selected radiographs. In the Chingford cohort, intra-observer
repeatability was assessed by 1 investigator reading 10 randomly selected
blinded radiographs on 3 occasions. Interobserver reproducibility was assessed
by 2 further observers reading the same 10 radiographs.63 Finally, in order to
examine interobserver reliability between both techniques, the alpha angle was
calculated in 30 randomly selected hips using SSM software and Hipmorf 2.0 (14
hips of the CHECK cohort and 16 hips of the Chingford cohort).
Statistics
Reliability of the alpha angle as a continuous measure was assessed using
intraclass correlation coefficient (ICC) and Cohen’s kappa indicating agreement
for whether a hip was classified as having or not having a cam deformity. A Bland-
Altman plot was used to visualize agreement in the alpha angle measurements
between the two techniques (SSM and Hipmorf).284
127
Chapter 8
RESULTS
Participants CHECK and Chingford
Of the 1002 individuals in the CHECK cohort, 89 were lost to follow-up, 16
were not able to visit the hospital during the five year follow-up, 32 refused
to undergo radiographs or had missing radiographs, and 124 were excluded
because this first cluster of participants had AP hip radiographs (instead of
AP pelvic radiographs) obtained at baseline. Of the remaining 741 individuals
(1482 hips), another 71 hips (with a total of 18 subjects because of bilateral
involvement) were excluded because the radiographs were of insufficient quality
to reliably position the SSM points, leaving 1411 hips in 723 individuals for
inclusion with 80% (n=575) being women. Of the included hips, 76% had no
signs of radiographic hip OA (K&L=0) and 24% had doubtful radiographic hip
OA (K&L=1).
From the initial Chingford cohort of 1,003 individuals at baseline, 795 individuals
had AP pelvis radiographs obtained at year 2. Of those, 20 individuals were
excluded due to poor radiograph quality. Five hip joints were excluded because
they had a metalwork in situ, indicating previous femoral neck fracture. Seventy-
128
Definition of a cam deformity
two hip joints were excluded due to excessive tilt. A total of 119 hips in 61
individuals were excluded, leaving 1468 hips in 734 individuals for inclusion at
baseline. At baseline, 80% had a K&L score of 0, 6% had a K&L score of 1 and
14% had a K&L score of ≥2.
Chapter 8
Figure 2. Bland Altman plot. The difference vs average alpha angle showing the agreement
in alpha angle measurement of 30 hips between the SSM point set and Hipmorf technique.
129
Chapter 8
Figure 3. Presence of cam deformity: 60°. Histograms of the alpha angle as measured on
anteroposterior radiographs with kernel density plot show a clear bimodal distribution in
the CHECK cohort (A), the Chingford cohort (B), and the combined data of both cohorts
(C). A corresponding normal distribution of the alpha angle until a value of 60° is shown
in both men (D) and women (E) in the CHECK cohort. These histograms also illustrate the
higher prevalence of cam deformity in men than in women.
130
Definition of a cam deformity
– 69.7, figure 3c). The sensitivity analysis comprising one randomly selected hip
per person (n=1440 hips) yielded a similar value of 61.5°. Figure 3d-e depicts
the distribution of the alpha angle in men and women separately in the CHECK
cohort. A corresponding bimodal distribution of the alpha angle was found in
both men and women and the higher prevalence of cam deformities in men than
in women is clearly illustrated. When analyzing the hips with an alpha angle less
than 62.7° in the CHECK cohort, the mean (SD) alpha angle in men 45.0 (±
4.15) was significantly higher than in women 42.7 (± 3.9), p<0.001.
Chapter 8
Figure 4. ROC curve. A. the ROC curve of the alpha angle as a continuous measure for
end-stage OA is shown, with the pathological threshold of 78° indicated. B The area under
the ROC curve for each alpha angle threshold is shown.
131
Chapter 8
Pathological threshold
In the CHECK cohort, 2.76% (39 hips) developed end-stage OA within five years
follow up. THA patients did not have other hip pathology than OA nor other
hip surgery during follow-up. In the Chingford cohort, 7.05% (66 hips) had
developed end-stage OA at year 20.
A pathological threshold was determined based on the hips with cam deformity
that developed end-stage OA at follow-up. An alpha angle threshold of 78°
(95% CI 62° - 87°) resulted in the maximum area under the ROC curve, which
was 0.69 (95% CI 0.62-0.75) for end-stage OA (figure 4). It is illustrated by
figure 5 that the majority of hips with a cam deformity that developed OA had
an alpha angle greater than 78°. The sensitivity analysis based on the random
selection of one hip per person which contained 50 hips with end-stage OA at
follow-up, resulted in a maximum area under the ROC curve of 0.66 at an alpha
angle of 82°.
Figure 5. Pathological alpha angle threshold: 78°. Scatterplot showing the alpha angle on
the X-axis and a random variable on the Y-axis. The dots represent the combined hips of
the CHECK cohort (green) and Chingford cohort (blue) at baseline with a cam-deformity
that developed end-stage OA (K&L 3, 4, or THA). The pathological threshold of 78° is
indicated.
132
Definition of a cam deformity
DISCUSSION
In the present study that combined data of two large prospective cohorts,
comprising almost 3000 hips, an alpha angle threshold for defining the presence
of a cam deformity was determined based on the finding of a bimodal distribution
of the alpha angle. Also, a pathological threshold value was determined based
on development of end-stage OA at follow-up.
Previously, threshold values based on the distribution of the alpha angle in a
certain population have been proposed, and ranged from 50° to 83°.224,260,262,285
These studies defined the presence of a cam-type deformity either by the upper
limit of the 95% reference interval, or by the +1 standard deviation of the
mean. Although these values might give an indication, the reference interval
or standard deviation might for several reasons not be optimal to define an
alpha angle threshold. First, the assumption that only the upper 5% of a given
population has an abnormal alpha angle is incorrect. Many studies showed a
prevalence higher than 5% in asymptomatic individuals with various threshold
values for the alpha angle.226,258,273,275,276 Secondly, reference intervals assume
a normal distribution of the alpha angle, which was clearly not the case in
both cohorts. Also, different threshold values for men and women have been
proposed based on significant differences in mean alpha angle between men and
women.260,285 However, as the prevalence of cam deformities is higher in men
than in women, a greater mean alpha angle in men can be expected, but this
Chapter 8
does not imply that the alpha angle cut-off in women should be lower. Figure
3 d-e clearly illustrates that the higher mean alpha angle in men is actually
influenced by outliers, which are in fact the hips with a cam deformity.
To define the presence of a cam deformity, we propose a threshold of 60°. The
threshold of 62.7° was based on the minimum of the optimal fit through the
combined alpha angle data. However, the optimal fit actually consisted of three
distributions, as the distribution of the alpha angles <60° was slightly skewed
to the right. When we forced a fit of two distributions through the data, the
minimum was found at an alpha angle of 56.7°, which was clearly too low. The
benefit of a bimodal distribution is that there are only a few cases in between
those alpha angle thresholds, assuring that separation in cam versus non cam
cases is not too sensitive to small differences in threshold value (Figure 3).
We therefore regard a non gender specific threshold of 60° to be optimal to
distinguish between normal hips and hips with cam deformity.
When comparing the distribution of the alpha angle in the CHECK cohort with
the distribution in the Chingford cohort, a higher prevalence of cam deformities
in the Chingford cohort was found. As no systematic differences in measurement
between both techniques were found, this might be due to the differences of
inclusion or differences in radiographic protocol. The Chingford cohort is a
133
Chapter 8
population based cohort, where individuals might have had hip complaints for
a longer time. In contrast, an inclusion criterion of the CHECK cohort was that
individuals consulted the GP for the first time because of hip or knee pain.
Individuals suffering from pain because of cam impingement are likely to
present with complaints at a younger age than 45 years, which means that they
were not included in the CHECK cohort but those individuals might have been
included in the Chingford cohort. Secondly, radiographs in the CHECK cohort
were obtained in standing position, whereas in the Chingford cohort radiographs
were taken in supine position, which might have influenced alpha angles and
thereby partly explain the discrepancy in prevalence. Finally, radiographs in the
Chingford cohort were controlled for hip rotation in a lesser degree than in the
CHECK cohort, which might also have influenced alpha angles, as it is known
that alpha angles might become higher when the hip is more externally rotated
hips.260
As higher alpha angles are associated with more severe cartilage damage and
OA, Gosvig et al. proposed a pathological threshold based on the +2 standard
deviations of the mean which resulted in pathological thresholds of 83° for
men and 57° for women.260 However, in our opinion a pathological threshold
value should rather be based on ‘true’ pathology. Still, as opposed to the
threshold based on the bimodal distribution of the alpha angle, the definition
of a pathological threshold value might be more subjective. We determined
a pathological threshold value based on the highest discriminative ability for
development of end-stage OA during the follow-up period, using the maximum
area under the receiver operating characteristic (ROC) curve. Using this method,
we propose a non gender specific threshold of 78°. It can also be visualized in
figure 5 that most hips with cam deformity that developed end-stage OA had an
alpha angle higher than 78°.
The proposed threshold value for defining the presence of a cam deformity is
based on all hips (n=2879) and corresponds with the value obtained by the
sensitivity analysis in which only one hip per person was randomly selected
(n=1440). For the pathological threshold a slight discrepancy of 4° was found.
As the majority of the end-stage OA cases were unilateral, we regard the analysis
using all hips for the pathological threshold the most appropriate.
For upcoming clinical trials it is important to avoid misclassification of the
radiographic presence of a cam deformity. An alpha angle below 60° is often
used to quantify a cam deformity, but below this value we could not discriminate
between hips with and without a cam deformity. This might also explain
why a value of 60° has previously been shown to be optimal to discriminate
between symptomatic and asymptomatic cam deformities.259,286 However, a cam
deformity is highly prevalent in the general population and is only a prerequisite
for cam impingement. Whether someone with a cam deformity will become
134
Definition of a cam deformity
Chapter 8
techniques from 0.66 to 0.89. Except for these outliers, there was agreement
in categorizing hips with or without cam deformity in all hips. It appears that
drawing a best fitting circle is crucial for reliably determining the alpha angle.
In the CHECK cohort, 8 points on the femoral head were used to determine a
best fitting circle around the femoral head whereas 3 points were used in the
Chingford cohort. Those differences might explain the slight differences in best
fitting circle.
Several measures to quantify the cam deformity besides the alpha angle have
been proposed, such as the head-neck ratio, anterior offset ratio, and triangular
index.260 All those measures quantify the loss of concavity of the head-neck
junction and show good reproducibility. However, a nonspherical, flattened
femoral head might have a normal concavity of the head-neck junction. Such an
abnormal morphology might lead to impingement but is not captured by above
mentioned measures, except for the alpha angle. Further, in both the CHECK
and Chingford cohort, the alpha angle was most predictive for OA.63,203
This study has some limitations. First, the use of AP radiographs might not be
optimal for quantifying a cam deformity. Although the exact amount of cam
deformities missed on AP radiographs varies in different reports, we acknowledge
that the prevalence of cam deformities is underestimated.260,280,289 Though, when
135
Chapter 8
only AP radiographs are used, the alpha angle is still highly predictive for OA.
Second, it is unknown whether the proposed thresholds are applicable for other
radiographic views, or in other planes when 3D imaging techniques are used.
The proposed thresholds should therefore be validated in large cohorts using CT
or MRI.
In conclusion, the alpha angle in this study of combined epidemiological data
showed a bimodal distribution, indicating a clear distinction between hips with
and without cam deformity. An alpha angle threshold of 60° is proposed for
defining the presence of a cam deformity. In addition, a pathological threshold
of 78° is proposed based on development of OA at follow-up.
136
PART 3
ABSTRACT
Background Cam impingement is a well-recognized cause of hip pain and
might cause osteoarthritis of the hip. Clinically, cam impingement is mostly
observed in young, active males patients, but only few studies have focused on
the manifestation of cam-type deformities during skeletal development.
Results An alpha angle >60° was already found at the age of 12 years in
some soccer players and controls. A cam-type deformity defined by alpha angle
tended to be more prevalent in soccer players (26%) than in controls (17%,
p=0.31). In 13% of soccer players, a prominence was visible on radiographs
and was first seen at the age of 13 years. The anterosuperior flattening (56% vs
18%, p=0.0001) and prominence (13% vs 0%, p<0.03) were more prevalent in
soccer players than in controls.
Conclusion Cam-type deformities were recognizable and present from the age
of 13 years and were more prevalent in soccer players than in their non-athletic
peers. Cam-type deformity develops during adolescence and is likely to be
influenced by high-impact sports practice.
140
Development of cam deformity
INTRODUCTION
Some sporting activities have been associated with a higher incidence of
hip osteoarthritis (OA).47,290 This applies for soccer, especially in elite soccer
players.291 A possible explanation for the higher risk for hip OA in elite soccer
players is the existence of shape deformities of the femoral head.
From the 1960s until the 1980s, it was suggested that many cases of ‘primary’
or ‘idiopathic’ hip OA are actually secondary to shape deformities and thus do
not have a systemic origin.24,50,292,293 From the 1990s, research has indicated that
subsets of hip OA are indeed secondary to minor, previously unrecognized, or
subclinical shape deformities.55,64,206,244,294,295 The path from these minor shape
deformities to OA might follow the mechanism of femoroacetabular impingement
(FAI) proposed by Ganz et al.55
FAI is an abnormal contact between the femoral neck and the acetabulum. There
are two classic types.55 The first one, pincer impingement, is an overcoverage
of the femoral head by the acetabulum and is mostly seen in middle-aged
women. The second one, the subject of this article, is cam impingement, mostly
seen in young active males.240 It is caused by an abnormal morphology of the
proximal femur: a non-spherical head that is due to a flattening or prominence
on the anterosuperior part of the femoral head-neck junction.67 As a result of
this anatomical anomaly, the head-neck offset is smaller and squeezed in the
acetabulum. It has been suggested that this causes repetitive minor trauma
to the anterior acetabular margins, and thus causing labral tears and cartilage
damage.296,297 These injuries may eventually lead to hip OA.
Although the exact etiology of a cam-type deformity is not well understood,
suggested causes are bone remodelling,298-301 a growth abnormality of the
Chapter 9
epiphysis,25 or subclinical slipped capital femoral epiphysis (SCFE).294,302 The
general hypothesis based on research on young, active adult men is that the
deformity develops during adolescence after a long period of excessive sports
practice during growth.23 Recently, Siebenrock et al. showed that young basketball
players are at higher risk for having a cam-type deformity, which developed
especially around the time of physeal closure.227 Thus, high impact sports, such
as basketball, ice hockey, and soccer, might predispose to developing a cam-
type deformity because of the high shear stress applied to the femoral head.303
For this study, we investigated at what age a cam-type deformity develops
and tested the hypothesis that cam-type deformities are more prevalent in
adolescent preprofessional elite soccer players than in their nonathletic peers.
141
Chapter 9
METHODS
Participants
We included boys (12-19 years) who played in selection teams of Feyenoord
soccer club in Rotterdam, the Netherlands. They follow a particular school
program to train at the optimal intensity. An information letter was sent to
141 boys who met the inclusion criteria; 101 boys gave informed consent,
of whom 89 finally joined this study, while 12 boys did not come for clinical
and radiological evaluations. Exclusion criteria for soccer players were any hip
disorder. The study was approved by the Medical Ethical Committee of Erasmus
Medical Center (Rotterdam, the Netherlands).
All soccer players had anteroposterior (AP) pelvic and frog-leg lateral
hip (commonly known as the Lauenstein view) radiographs and a clinical
investigation. For controls, boys aged 12-19 years in whom both an AP and
a frog leg lateral radiograph of both hips were available were selected from
the radiology databases of Erasmus Medical Center and Reinier de Graaf Groep
Hospital. The exclusion criteria were any hip disorder as diagnosed clinically by a
pediatric orthopedic surgeon or radiographically by a musculoskeletal radiologist.
Further, 19 subjects were excluded because sports practice was reported by the
orthopaedic surgeon in their medical records. Subjects were not excluded when
a cam-type deformity without any other hip disease was diagnosed (n=2). The
selected controls have had radiographs of both hips in the diagnostic course of
complaints of hip (84%), knee (4%), back (5%), or other (7%). None of the
controls visited the hospital again because of their complaints within a minimum
of two years after the radiographs were obtained. There were 92 subjects in the
control group.
The soccer players were clinically examined to measure the range of motion
(ROM) of the hip, and an anterior impingement test was performed. The ROM
was measured by a goniometer in flexion, extension, abduction, adduction,
internal rotation and external rotation of the hip. A positive impingement test
was defined as a sharp pain in 90° of flexion with adduction and internal rotation.
The boys were asked to complete a questionnaire including items related to
groin pain, medical history and demographic data.
Radiology
Most cam-type deformities are located anterosuperior in the femoral head-neck
junction. To detect a cam-type deformity in this region, an AP view as well as
a lateral view of the hip is needed.248 Both radiographs were obtained, namely
an AP pelvic view and a frog-leg lateral view.304 The radiographs were made
according to a standardized protocol.
142
Development of cam deformity
For the AP radiograph the subject was positioned supine with his feet in a frame
to assure 15° internal rotation. This provides the best view of the head-neck
junction. The tube to film distance was 130 cm and the beam was centered 5
centimeters above the pubic symphysis.
The frog-leg lateral radiograph was performed in supine position with the leg of
interest abducted using a 45° wedge under the knees. The heel rested against
the medial aspect of the contralateral knee to position the leg in 90° of external
rotation. The tube to film distance was 115 cm, and the beam was centered
in the middle of the groin on the femoral head. Gonadal shields were worn to
reduce the radiation load.
Measures
We used the alpha angle as a quantitative outcome measure and a visual scoring
system as a semi-quantitative outcome measure for a cam-type deformity.
The shape of the proximal femur in the AP and frog leg lateral radiographs
was assessed using statistical shape modelling (SSM) software (ASM tool
kit, Manchester University, Manchester, UK). The shape is defined by a set of
landmark points that are positioned along the contour of the bone in the image.
Each point is always placed on the same landmark of the contour, to allow
comparison between the shapes. All radiographs of the left hip were mirrored to
appear as right hips in order to be able to use the same set of landmarks.
The alpha angle was automatically calculated in all radiographs from the point
sets of the SSM software using matlab (version 7.1.0, MathWorks Inc, Natick,
Massachusetts, USA) (figure 1). The alpha angle is measured by first drawing
the best fitting circle around the femoral head, then a line through the center of
Chapter 9
the neck and the center of the head. From the center of the head a second line
is drawn to the point where the superior surface of the head neck junction first
departs from the circle. The angle between these two lines is the alpha angle.224
We defined the presence of a cam-type deformity as an alpha angle >60°, as
osseous abnormalities below this threshold value have been suggested to pose
less of a risk for developing hip pain.259 In addition, radiographs were scored
semi-quantitatively by an experienced orthopedic surgeon and a musculoskeletal
radiologist based on consensus, using a three-point scoring system305 (figure 2).
The anterior head-neck junction was classified as
1. Normal: slight symmetric concavities of the anterior head-neck junction with
respect to the posterior head-neck junction.
2. Flattening of the head-neck junction: moderate decrease in the anterior
head-neck offset with respect to the posterior head-neck junction..
3. Presence of a prominence: a convexity in the anterior head-neck junction, as
opposed to a concavity.
143
Chapter 9
Figure 1. To measure the alpha angle we fitted a circle to the femoral head and placed
a line through the center of the head and the middle of the neck. The alpha angle is the
angle between the line through the neck and the point where the head-neck junction
deviates from the circle. (A) Normal alpha angle of 42° is shown in a 16-year-old soccer
player. (B) An abnormal alpha angle of 76° is shown in a 19-year-old soccer player.
Statistical analysis
A subject was classified as having a cam-type deformity when an alpha angle
> 60º was measured in at least one hip in either the AP or frog-leg lateral
view. A subject was classified as having a flattening or prominence when such
was visually scored in at least one hip in either the AP or frog-leg lateral view.
When a subject had both a flattening and a prominence, he was classified as
having a prominence. Differences in prevalence numbers between groups were
calculated using logistic regression, corrected for age. By correcting for age,
differences in prevalence between the groups were not influenced by the slight
difference in age between the groups. Differences in the ROM between cam-type
deformity cases and normal soccer players were calculated using generalized
estimating equations (GEE), corrected for age. By using a GEE regression
model, we could model the correlation between left and right hips. Differences
144
Development of cam deformity
between the Oxford score in soccer players and controls were calculated using
a univariate general linear model, corrected for age. Chi-squared test was used
to study groin pain, leg dominance, and impingement test results in relation to
the existence of a cam-type deformity.
Chapter 9
Figure 2. anteroposterior pelvic (left) and frog-leg lateral (right) radiographs showing
typical examples illustrating the three categories of the used visual scoring system.
Normal (top), flattening (middle, black arrow), and a prominence (bottom, white arrow)
were scored.
145
Chapter 9
RESULTS
The mean age of the 89 soccer players (178 hip joints) was 14.8 years (range,
12 – 19 years), and the mean body-mass index was 20.13 kg/m² with a standard
deviation (SD) of 2.25. The intensity of soccer practice had a mean ±SD of 7.96
± 1.77 h/wk and the years playing soccer ranged from 4 to 15 years, with
a mean of 8.97 years. None of the participants had a history of hip disease.
The mean age of the 92 (184 hip joints) controls was 13.8 years (range, 12
– 19 years), which was significantly lower than the soccer players (p<0.001).
From the age of 14 years, the Oxford score tended to be slightly higher in the
soccer players, although no statistically significant differences were measured in
skeletal maturity (p=0.07). Table 1 illustrates the mean per age.
In the soccer players, an alpha angle >60° was first found at the age of 12 years
on the frog-leg lateral view and at the age of 14 years on the AP view. In the
control group, an alpha angle >60° was found at the age of 12 years in both the
frog-leg lateral as well as the AP views. Hips having a visual prominence were
found only in the soccer players and were present from the age of 13 years.
Hips having a flattening were found both in soccer players and controls from the
age of 12 years.
In 26% of the soccer players and in 17% of the controls, an alpha angle >60°
was measured in at least one hip in either the AP view or the frog-leg lateral view,
although this difference was not significant when corrected for age (p=0.31).
Significantly more prominences in the head-neck junction were visually scored
in soccer players than in controls (13% vs 0%, p<0.033), with prevalence
increasing with age (p=0.003). A flattening of the head-neck junction was
significantly more frequently found in the soccer players as well (53% vs 19%,
p<0.001). Prevalence numbers per age are given in table 2 and figure 3.
146
Table 2. Prevalence of a cam-type deformity per age
12 25% (n=3/12) 0% (n=0/12) 33% (n=4/12) 16.7% (n=5/30) 0% (n=0/30) 10% (3/30)
13 7.7% (n=1/13) 7.7% (n=1/13) 23% (n=3/13) 15.4% (n=2/13) 0% (n=0/13) 0% (0/13)
14 30% (n=6/20) 5% (n=1/20) 65% (n=13/20) 17.4% (n=4/23) 0% (n=0/23) 17% (4/23)
15 25% (n=3/12) 8.3% (n=1/12) 83% (n=10/12) 15.4% (n=2/13) 0% (n=0/13) 38% (5/13)
16 27.3% (n=3/11) 18.2% (n=2/11) 45% (n=5/11) 20% (n=1/5) 0% (n=0/5) 40% (n=2/5)
17 27.3% (n=3/11) 27.3% (n=3/11) 73% (n=8/11) 25% (n=1/4) 0% (n=0/4) 50% (n=2/4)
≥18 40% (n=4/10) 40% (n=4/10) 40% (n=4/10) 25% (n=1/4) 0% (n=0/4) 25% (n=1/4)
total 26% (n=23/89) 13% (n=12/89) 53% (n=47/89) 17% (n=16/92) 0% (n=0/92) 18% (n=17/92)
147
Development of cam deformity
Chapter 9
Chapter 9
Figure 3. Prevalence of a cam-type deformity based on the alpha angle (A) and the
visual scores (B) stratified per age in years for soccer players and controls. The number of
participants in each group is given above the bars. note that cam-type deformities started
with a flattening, and with increasing age, more prominences were found in the young
soccer players.
148
Development of cam deformity
without (20.4 vs 29.8 p<0.001). No such differences in ROM were found when
having a flattening or prominence on the AP view. No significant association was
found between a positive impingement test result and the existence of a cam-
type deformity or between groin pain and the existence of a cam-type deformity.
DISCUSSION
Cam impingement has been defined as a cause of hip pain in young adults and
a potential cause for OA.268,307 Young, active males are thought to have a higher
risk for developing cam impingement. Some evidence exists that young active
males indeed have a higher prevalence and that a cam-type deformity develops
during closure of the growth plate.227 Our study supports this line of thinking.
The two main results of our study were that a cam-type deformity is present early
in adolescence, and seems to be more prevalent in soccer players. Furthermore,
prevalence of cam-type deformities based on visual scores increases with age.
The higher prevalence of cam-type deformities in soccer players has also be
shown by Murray et al., who reported more tilt deformities on AP radiographs
from males exposed to a more active regime in school.23 Recently, Siebenrock
et al.227 showed that high intensity basketball playing during adolescence was
associated with a higher risk of cam impingement. They reported a prevalence
of 89% cam-type deformities in basketball players after closure of the capital
growth plate, which is comparable with the visual scores for flattening and
prominence from the age of 14 years in our study. Other studies have shown
that the prevalence of such deformities in asymptomatic men ranges from 9%
to 25%.227,260,272,275 However, these percentages are difficult to compare, since
different measures were used to quantify a cam-type deformity. Nowadays, the
Chapter 9
alpha angle is the measure generally applied to quantify a cam-type deformity,
although a validated threshold value for an abnormal alpha angle remains to
be determined. Osseous abnormalities below the threshold value of 60° have
been suggested to pose less of a risk for developing hip pain,259 which is why we
applied this cut-off value as a quantitative outcome measure.
A remarkable finding was an alpha angle >60° in some 12-year-old boys.
However, in 12-year-old boys, the alpha angle might be a less accurate measure
for cam-type deformities. When the growth plate is not fused anterosuperiorally,
the contour of the femoral head appears more oval, resulting in large alpha
angles (Figure 4). In addition, we could not define a cam-type deformity purely
visually when the growth plate was incompletely closed. Once the growth plate
has almost closed, the typical shape of a cam-type deformity is more pronounced
and recognizable. An interesting observation was that in some 12 years old the
proximal femoral growth plate was more extended into the neck, which gave the
head-neck junction a flat appearance. A follow-up study is needed to find out
149
Chapter 9
whether the large alpha angles and an extended growth plate in the youngest
boys are precursors to the development of a classic cam-type deformity.
Figure 4. A high alpha angle in a frog-leg lateral radiograph of a 12-year-old boy when
the growth plate was not completely fused. On the left, the original radiograph which was
visually scored as normal and on the right, an alpha angle of 72° is shown.
Soccer players did not only have cam-type deformities more often, but these
deformities were also more pronounced at this age. In soccer players, most
hips with cam-type deformities that were identified with an abnormal alpha
angle were visually scored as having a flattening or as having a prominence.
In contrast, in the hips of controls with cam-type deformities identified by an
abnormal alpha angle, most hips were visually scored as normal and some as
having a flattening. In addition, in soccer players, 35% of the cam cases had
an alpha angle >60° in both the AP and frog-leg lateral radiographs, whereas
none of the controls had an alpha angle >60° in both radiographs. Dudda et al.
showed that those cases with an aspherical head-neck junction in both an AP and
a lateral radiograph had the most prominent asphericity on magnetic resonance
arthrography.280 Thus, obvious ‘bumps’ were not seen in the controls, whereas
these were present in soccer players and significantly increased in prevalence
with age, reaching a prevalence of 40% in soccer players older than 17. The
prevalence of a flattening did not increase with age in the soccer players, though
figure 3b suggests that a flattened head-neck junction can evolve towards a
prominence over time.
Our study showed some discordance between the presence of a cam-type
deformity defined by alpha angle >60° and the visual scores. In soccer players,
six of the 18 hips visually scored as having a prominence did not have an alpha
angle >60°. This was caused either by a convexity anterior in the head-neck
junction, which is visually detectable as a sclerotic region on the frog leg lateral
view, or by a convexity in the femoral neck instead of the head-neck junction.
150
Development of cam deformity
In both situations the alpha angle was normal because these abnormalities
do not affect the contour of the head-neck junction. Vice versa, in controls,
13 of the 17 hips with an alpha angle >60° were visually scored as normal.
Because all the radiographs of the controls were taken during normal clinical
routine, these were not as standardized as in the soccer players. As a result,
the alpha angles might have been influenced by the rotational position of the
femur.260 For instance, we observed that hips with an alpha angle >60° that
were visually scored as normal, often had more externally rotated legs in the
radiographs, which led to the higher alpha angle. In these cases, the head-neck
junction could still be visually assessed and a prominence or a flattened head-
neck junction would have been recognised. Finally, in both soccer players and
controls, the alpha angle might be less accurate when the growth plate is not
completely fused anterosuperiorely. Taken together, the alpha angle is important
as a quantitative outcome measure, but in this population we observed that the
alpha angle might have underestimated the prevalence in soccer players and
might have overestimated the prevalence in controls. Therefore, we applied a
visual scoring system which better represented the differences between soccer
players and controls.
In addition to an AP pelvic view, a lateral view is necessary to visualize cam-type
deformities.248 Meyer et al. showed both the Dunn view in 45° or 90° flexion,
and the cross-table projection to be optimal for identifying femoral head-neck
asphericity.263 However, they did not study the frog-leg lateral view. Clohisy et al.
concluded the frog-leg lateral view to be the most distinctive and reliable view
for detecting cam-type deformities.304 In our study as well, most prominences
were visually detectable in the frog leg lateral view.
Chapter 9
The anterior impingement test is most commonly used when FAI is suspected.
In our study, the outcome of the impingement test was not associated with the
existence of a cam-type deformity, for which we have two possible explanations.
The first is that the anterior impingement test was not sufficiently sensitive. The
second is that in these young soccer players, labral pathology, which causes
pain, was not yet present. The second explanation is more likely, since studies
with arthroscopically proven labral tears have reported high numbers of positive
impingement tests.308 In hip joints with a cam-type deformity defined either
by an alpha angle >60° or by a flattening and prominence, internal rotation
in 90° of flexion was significantly reduced. Thus, in young adolescents without
any significant labral or cartilage damage, reduced internal rotation might be a
more clinically relevant sign for a cam-type deformity than a positive anterior
impingement test. This point needs further study.
Cam-type deformities were more often bilateral in soccer players than in controls.
In the soccer players, prominences were bilateral in half of the cases, and cam-
type deformities defined by alpha angle were bilateral in 30% of the cases. On
151
Chapter 9
the other hand, cam-type deformities defined by alpha angle in the controls
were bilateral in only 6% of the cases. In addition, the higher prevalence and
the more pronounced cam-type deformities in soccer players, combined with
the fact that such deformities were already present when the growth plate had
closed, suggests that a cam-type deformity develops during skeletal maturation
and high impact sports practice. The longer lateral extension of the epiphysis
into the neck as described by Siebenrock et al. that we frequently noticed,
indicates that a growth abnormality of the epiphysis, due to the high shear
stresses applied on the proximal femoral epiphysis, might be a cause of a
cam-type deformity.25 This growth abnormality might be caused by a delayed
separation of the common physis between the femoral head and the greater
trochanter,309 or by a structural adaptation of the epiphysis in order to resist the
higher demands of the proximal femur. A subclinical SCFE has been hypothesized
to cause a cam-type deformity as well.294,302 However, all soccer players were
without complaints during weight-bearing, so a subclinical SCFE was not very
likely. The controls in our study did have complaints but because none of them
had follow-up radiographs within the first two years after the initial radiographs,
(subclinical) SCFE is unlikely. Radiographs of both soccer players and controls
had no signs of a preslip i.e. widening and irregularity of the physis.310 We
therefore have no indication that most cam-type deformities are caused by a
subclinical SCFE.
This study had some limitations. First, the prevalence of cam-type deformities
might have been underestimated because conventional radiographs were
used. Although radiographs are important for the initial diagnosis of cam
impingement, several studies have shown that 3-dimensional imaging such
as computed tomography or magnetic resonance imaging is slightly more
sensitive for diagnosing a cam-type deformity.280,289,311 Using 3-dimensional
imaging, the complete anterosuperior portion can be evaluated and maximum
alpha angles can be calculated.275 In contrast, when using AP and frog leg
lateral radiographs, the alpha angle only can be measured in the superior and
anterior portion respectively, which may therefore be lower and prevalences
might be underestimated. Secondly, we were unable to include a healthy
control group, due to ethical considerations related to radiation in the pelvic
area. As an alternative, we selected a control group of comparable age from
the radiology databases of two hospitals. Since these radiographs were mainly
made in response to complaints related to the hip, the control group might not
completely represent a healthy adolescent population. However, because none
of the controls visited the hospital again with a minimum of two years after
the radiographs were obtained, we consider the control group as a relatively
normal adolescent population. Still, the prevalence of complaints related with
cam-type deformity might have been overestimated in our controls relative
152
Development of cam deformity
Chapter 9
153
Chapter 10
ABSTRACT
Background: A cam deformity is a major risk factor for hip osteoarthritis, and
its formation is thought to be influenced by high impact sporting activities during
growth.
Methods: Preprofessional soccer players (n=63; mean age, 14.43 years; range
12-19 years) participated both at baseline and follow-up (mean follow-up, 2.4
± 0.06 years). At both time points, standardised anteroposterior and frog-leg
lateral radiographs were obtained. For each hip, the alpha angle was measured,
and the anterosuperior head-neck junction was classified by a 3-point visual
system as normal, flattened, or having a prominence. Differences between
baseline and follow-up values for the alpha angle and the prevalence of each
visual hip classification were calculated. Additionally, the amount of internal
hip rotation, growth plate extension into the neck, and neck shaft angle were
determined.
156
Development of cam deformity: 2 years follow-up
INTRODUCTION
Osteoarthritis (OA) of the hip is a highly prevalent and disabling disease, especially
among former athletes.30,48 The role of femoral morphological characteristics,
in particular a cam deformity, has become increasingly recognized in the
pathophysiology of OA.75,101,203,221,313 A cam deformity is a nonspherical extension
of the femoral head anterosuperiorly that may lead to intra-articular damage
by a motion dependent inclusion of the cam deformity into the acetabulum,
known as cam impingement.55,313,314 It consistently shows an association with
hip pain, limited internal rotation, labral tears, chondral lesions, and ultimately
hip OA.57,63,203,208,216
The cause of a cam deformity is largely unknown, although several reports
suggest that participating in high impact sports during growth plays an important
role.23,226,227 Two recent studies reported a markedly higher prevalence of cam
deformity in asymptomatic adolescents participating in soccer and basketball,
respectively, as compared with their non athlete controls.226,227 These findings are
supported by other studies that reported a high prevalence of cam deformities in
both symptomatic and asymptomatic athletic adults.278,315-317
It was shown in a cross-sectional study that a cam deformity is radiographically
visible from the age of 13 years and may evolve in time.226 However, to the
best of our knowledge, no prospective data are available on how a cam
deformity evolves during skeletal maturation and whether it can evolve after
skeletal maturation. Furthermore, it is unknown whether the formation of a cam
deformity can be predicted from other radiological or clinical features before
a cam deformity appears radiographically. As high impact sport activities may
influence the formation of a cam deformity, a better insight into the time of
onset and development of a cam deformity might provide clues for prevention.
The primary aim of this study was to examine whether a cam deformity can
evolve over time in adolescents and whether such evolution continues after
Chapter 10
skeletal maturation. A secondary aim was to study whether other clinical or
radiographic features are associated with or are predictive for the formation of
a cam deformity.
METHODS
Participants
For this study, all 141 elite preprofessional soccer players aged 12 to 19 years
who practiced in the selection teams of the Feyenoord Soccer Club (Rotterdam,
the Netherlands), performing at the highest national level, were invited to
participate at baseline. Exclusion criteria were any hip disorders, but none of
the eligible participants were excluded because of these criteria. Of 141 eligible
157
Chapter 10
Radiographs
Most cam deformities are located in the anterosuperior head-neck junction.67
Therefore, both an anteroposterior (AP) pelvic view and a frog-leg lateral view
were obtained at both time points according to a standardized protocol, which
has been described in the baseline study.226
Quantification of morphology
Both the commonly used alpha angle and a semiquantitative visual classification
system were used to define the presence of a cam deformity.224,305 The latter
was used because the alpha angle might be less accurate for cam deformities
in skeletally immature individuals, as the contour of the femoral head appears
more oval when the growth plate is open, resulting in large alpha angles.226
Using Matlab (version 7.1.0, MathWorks Inc, Natick, Massachusetts, USA), the
alpha angle was automatically calculated in all AP and frog-leg lateral radiographs
from a point set that was manually positioned along the contour of the bone
(figure 1).
The alpha angle is measured by first fitting a circle to the femoral head. A line
is then drawn through the center of the neck and the center of the head, and
a second line is drawn from the center of the head to the point where the head
neck junction first departs from the circle. The angle between these two lines is
the alpha angle.224 We defined the presence of a cam deformity using a recently
validated alpha angle threshold of >60° and the presence of a pathological cam
deformity using an alpha angle >78°.318
All radiographs were also classified semiquantitatively by an experienced
orthopaedic surgeon and musculoskeletal radiologist based on consensus, using
a 3-point visual system (figure 2).226,305 The anterior head-neck junction was
graded as the following:
158
Development of cam deformity: 2 years follow-up
Figure 1. To measure the alpha angle, Matlab was used to calculate the best fitting circle
around the femoral head, the center of the head, and the middle of the neck, from a point
set along the contour of the bone. The alpha angle is the angle between the line through
the neck and the point where the head-neck junction deviates from the circle, with the
latter also calculated by Matlab. (A) A normal alpha angle of 42° is shown in a 16-year-old
soccer players at baseline. (B) An abnormal alpha angle of 76° is shown in a 19-year-old
soccer player at baseline. (From Agricola R. et al. The development of cam-type deformity
in adolescent and young male soccer players. Am J Sports Med. 2012;40(5):1099-1106.
Reprinted with permission)
Chapter 10
radiographs were presented in a randomized order to the observers, who were
blinded to any baseline data of the soccer player. This method showed an
intraclass correlation coefficient (ICC) of 0.82 for intra observer reliability in the
baseline study.226
Radiographic and clinical parameters that might be predictive for, or are
associated with, a cam deformity were assessed at baseline and included the
anterosuperior growth plate extension as described by Siebenrock et al.25, the
neck shaft angle (NSA) as measured on the AP radiographs (figure 3), and the
amount of internal hip rotation. The passive amount of internal hip rotation was
measured by a goniometer in a supine position in 90° of hip flexion and 90° of
knee flexion. To determine whether those parameters can predict the formation
of a cam deformity, we studied all hips without a cam deformity at baseline and
159
Chapter 10
compared the parameters between hips that had developed a cam deformity
with those that had not at follow-up.
Figure 2. The AP pelvic (left) and frog-leg lateral (right) radiographs showing typical
examples illustrating the 3 categories of the visual classification system used. Hips
were graded as normal (top), flattened (middle, black arrow), or having a prominence
(bottom, white arrow). (From Agricola R. et al. The development of cam-type deformity
in adolescent and young male soccer players. Am J Sports Med. 2012;40(5):1099-1106.
Reprinted with permission)
160
Development of cam deformity: 2 years follow-up
Figure 3. Neck-shaft angle (NSA) and growth plate extension. The NSA was calculated by
the line through the center of the neck and the center of the head, and the line parallel to
the femoral shaft, as determined by the direction of the shaft below the minor trochanter.
The amount of growth plate extension (E1) into the neck was calculated by first drawing
a line perpendicular to the line through the center of the neck and the center of the head
at the intersection with the medial femoral head. The distance from this line to the end
point of the growth plate was measured and divided by the diameter of the femoral head.
The reliability of the growth plate extension and NSA measurements has not
been described previously in this cohort. Therefore, intraobserver repeatability
Chapter 10
was assessed by a single investigator reading 10 randomly selected blinded
radiographs on 2 occasions. Interobserver reproducibility was assessed by
another observer reading the same 10 radiographs.
Statistical analysis
The reliability of the growth plate extension and NSA measurements were
assessed using ICC (2-way random, absolute agreement). A hip was classified
as having a cam deformity when an alpha angle >60º was measured in at
least one radiographic view, and it was classified as being flattened or having a
prominence when such was visually graded in at least 1 radiographic view. The
difference in prevalence numbers and difference in alpha angle between baseline
and follow-up were tested using Generalized Estimating Equations (GEEs). By
161
Chapter 10
using a GEE logistic regression model, we could model the correlation between
left and right hips. The prevalence at both time points is presented for the
entire group, by age category (12-13 years, 14-15 years, and >16 years at
baseline), and by growth plate status at baseline (open vs closed). The Pearson
correlation coefficient was used to describe the correlation between the amount
of growth plate extension and the alpha angle. The association of the amount of
growth plate extension, NSA, and internal rotation with the presence of a cam
deformity was tested using GEEs, all adjusted for age. Finally, in hips without a
cam deformity at baseline (either hips with an alpha angle <60° or hips graded
as normal), they were tested using GEEs on whether the above-mentioned
parameters could predict the formation of a cam deformity at follow-up. Each
parameter was tested separately and corrected for age. Differences in baseline
characteristics between included participants and drop-outs were tested using
an independent-samples t test.
RESULTS
Participant characteristics
Demographic data of the participants are summarized in Table 1. The mean
follow-up time was 2.4 years (range, 2.3-2.6). All included participants continued
to play soccer at a high level during follow-up.
162
Table 2. Prevalence of cam deformities at baseline and follow-up based on visual hip classification and alpha angle.
163
Development of cam deformity: 2 years follow-up
Chapter 10
Chapter 10
Figure 4. The prevalence of cam deformities at baseline and follow-up based on (A) hip
classification and (B) alpha angle is presented for all hips (left column), by age category
(middle columns), and by growth plate status at baseline (right columns). The number
of hips in each category is indicated on top of each column. AA, alpha angle; B, baseline;
FU, follow-up; GP, growth plate; y, years, *Statistically significant difference (p<0.05)
between baseline and follow-up.
Based on the visual grades, the prevalence of a prominence per hip in the
entire group increased from 7.1% at baseline to 22.2% at follow-up (p<0.001)
(Figure 6). Of the soccer players aged 12 or 13 at baseline, 84.1% had a normal
appearance of the head-neck junction, which decreased significantly to only
43.2% at follow-up (p<0.001). These soccer players predominantly acquired
a flattened head-neck junction (p=0.002) (Figure 5). Among players aged
≥14 years at baseline, 15 of the 47 hips (32%) with a flattened head-neck
junction evolved into a prominence, although the difference in prevalence was
not statistically significant in the subgroup of soccer players aged 14 to 15
years at baseline (6.2% to 27.1%; p=0.19), while it showed a strong trend
towards a significant difference for those aged >16 years at baseline (14.7%
to 35.3%; p=0.07) (Figure 4 and 6). Of the hips with an open growth plate
at baseline, the prevalence of normal hips decreased significantly (57.3% to
31.3%; p=0.001), and the prevalence of a prominence increased significantly
164
Development of cam deformity: 2 years follow-up
Chapter 10
Figure 5. Typical examples of hips with a normal, concave head-neck junction at baseline
(left, arrows) that developed into a flattened head-neck junction at follow-up (right, arrow
heads). This was typically seen in soccer players aged 12 to 13 years at baseline. The
corresponding AP view is shown in the upper left corner of each frog-leg lateral radiograph.
165
Chapter 10
Associated parameters
Considering all hips at baseline, a decreased NSA (mean 129.1° vs 133.6°;
p=0.001) and an increased epiphyseal extension (mean 1.54 vs 1.43; p=0.001)
were associated with the presence of a cam deformity based on an alpha angle
>60°. The correlation coefficient between amount of epiphyseal extension and
the alpha angle in the AP view was r=0.62 (p<0.001) for all hips at baseline,
and r=0.81 (p<0.001) for hips with a cam deformity. Limited internal rotation
was only associated with an alpha angle >60° from the age of 13 years (mean,
20.6° in hips with a cam deformity vs 26.1° in those without; p=0.009).
Regarding the visual grading, no association was found with the NSA, but the
epiphyseal extension was significantly associated (p=0.002) with mean values
of 1.39 ±0.16 for normal hips, 1.53 ±0.16 for hips with a flattened head-neck
junction, and 1.69 ±0.18 for hips with a prominence. In all hips, internal rotation
was associated with prevalence of the hip classifications (p=0.001), with mean
values of 29.5° ±7.0° for normal hips, 22.6° ±8.8° for hips with a flattened
head-neck junction, and 16.4° ±7.3° for hips with a prominence.
The ICC of the growth plate extension was 0.97 for intra-observer and 0.97 for
inter-observer reliability. For the NSA, the ICC score was 0.89 for intra-observer
and 0.92 for inter-observer reliability.
166
Development of cam deformity: 2 years follow-up
Chapter 10
Figure 6. Typical examples of hips with a flattened head-neck junction at baseline (left,
arrows) that evolved into a prominence at follow-up (right, arrow heads). This was
typically seen in soccer players aged 14 to 16 years at baseline. The corresponding AP
view is shown in the upper left corner of each frog-leg lateral radiograph.
167
Chapter 10
Figure 7. The morphology of hips with a closed growth plate at baseline (left) did not
change during follow-up (right). Hips with a normal head-neck junction (top), flattened
head-neck junction (middle), or prominence (bottom) appeared equal. The corresponding
AP view is shown in the upper left corner of each frog-leg lateral radiograph.
Predictive parameters
At baseline, 64 hips had a visually normal appearance of the anterosuperior
head-neck junction, of which 35 developed either a flattening or prominence at
168
Development of cam deformity: 2 years follow-up
follow-up. Regarding the visual grading system, none of the tested parameters
could significantly predict which hips were going to develop a flattening or
prominence at follow-up. Regarding the alpha angle, 80 hips did not have a cam
deformity (alpha angle <60°) at baseline, but 15 of those hips had an alpha
angle >60° at follow-up. Of those hips, both limited internal rotation (odds
ratio [OR], 0.90; 95% confidence interval [CI], 0.84-0.98; p=0.009) for every
increase in the degree of internal rotation and a small NSA (OR, 0.82; 95% CI,
0.70-0.96; p=0.015) for every increase in the degree of NSA were significantly
predictive for the formation of a cam deformity, also after correction for the
baseline alpha angle. The amount of epiphyseal extension was not predictive for
the formation of a cam deformity (p=0.73).
DISCUSSION
A cam deformity has been recognized as a cause of hip pain and limited function
in young adults and is a major risk factor for development of hip OA.75,203,216,319
Recent studies reported a high prevalence of cam deformities in young adults
participating in high-impact sporting activities.226,227,278,315,317 This is the first
prospective study on the development of a cam deformity in which we showed
that a cam deformity is gradually acquired and probably only during skeletal
maturation. The formation of a cam deformity may therefore be a result of
structural adaptation to high-impact sporting activities during growth, when the
skeleton is highly sensitive to mechanical loading.320
The formation of a cam deformity occurred during skeletal maturation, as was
shown by both the visual grading system and the alpha angle. Especially from
the age of 12 to 14 years, half of the normal hips acquired a flattened head-
neck junction (Figure 4 and 5). From the age of 14 years until growth plate
closure, the flattening continued to evolve into a prominence in a substantial
number of hips (figure 6). After closure of the proximal femoral growth plate,
Chapter 10
we did not observe a significant increase in the prevalence or severity of a cam
deformity, either based on the visual hip classifications or the alpha angle. The
slight nonsignificant increase in the prevalence of a prominence after skeletal
maturation was within the range of measurement error and caused by the
fact that follow-up radiographs were graded blinded, without information on
the baseline radiographs, as a comparison between the baseline and follow-
up radiographs revealed that these were exact copies of each other (figure 7).
These results are supported by 2 large cohort studies of adults >45 years that
did not show an increase in the prevalence of cam deformities at a follow-up of
5 and 19 years, which favor the theory that a cam deformity is a growth-related
acquired phenomenon.63,203
169
Chapter 10
The correlation coefficient between the amount of internal rotation and the
alpha angle in the frog-leg view of the 30 hips with a closed growth plate
at baseline (r= -0.58 at baseline and -0.52 at follow-up) was similar to the
correlation found by Siebenrock et al. in young basketball players with a
closed growth plate (32 hips) and the alpha angle as measured at the 1 o’clock
position (r= -0.55).227 These correlations were also in agreement with those
of other reports.216,272,275,321,322 Interestingly, limited internal rotation was also a
significant predictor for a cam deformity in those hips with an alpha angle <60°
at baseline. A possible explanation is that a cartilage-like structure not visible
on radiographs, which will ossify into a cam deformity later, might already be
present and limit internal rotation.
The prevalence per person of a visually detected flattening or prominence in the
group with a closed growth plate was 86.7% at baseline and 80% at follow-up.
Based on the alpha angle, we found a prevalence of 60% at both time points.
These high prevalence numbers of a cam deformity correspond with the reported
prevalence among (a)symptomatic athletes participating in high-impact sports
such as soccer (60%-68%)315,316, football (72%)278, basketball (89%)227, and ice
hockey (75%)317. However, the prevalence of cam deformity depends on the
definition used and threshold value of the alpha angle.
The correlation between a cam deformity and the extension of the growth plate
was first reported by Siebenrock et al.25 Moreover, in a 1-year follow-up study, it
was shown that there is a physiological increase in epiphyseal extension during
growth until closure of the physis, which was supported by our data.323 An
extension beyond a certain physiological value is probably associated with the
formation of a cam deformity. The correlation in our study between the epiphyseal
extension and the alpha angle (r=0.62) in all 126 hips was mainly explained by
the correlation coefficient (r=0.81) in hips with a cam deformity (alpha angle
>60°), implicating that the larger the epiphyseal extension toward the neck, the
greater the cam deformity. Siebenrock et al. showed in a cross-sectional study
that hips of young adolescent basketball players with an open growth plate
that had not yet developed a cam deformity already had a greater epiphyseal
extension than their non-athletic peers.324 Because of their cross-sectional
design, it was unknown if these hips with a greater epiphyseal extension toward
the femoral neck would develop a cam deformity, but it was suggested that
this alteration in growth plate precedes a cam deformity as basketball players
are more likely to develop a cam deformity. However, in our prospective study,
we did not find the epiphyseal extension to be predictive. Regarding its strong
association with the presence of a cam deformity and correlation with the alpha
angle, our data show that the epiphyseal extension does not precede the cam
deformity but rather suggest that the altered epiphyseal extension toward the
femoral neck runs analogous with the gradual formation of the cam deformity.
170
Development of cam deformity: 2 years follow-up
Therefore, the growth plate might play an important role in the development of
a cam deformity.
A smaller NSA was significantly predictive for the formation of a cam deformity.
Of all hips without a cam deformity (alpha angle <60°) at baseline, those that
developed a cam deformity (alpha angle >60°) at follow-up had a significantly
smaller NSA, independent of age and baseline alpha angle. Moreover, the NSA
was strongly associated with the presence of a cam deformity at baseline, but
only in the age categories from 12 to 15 years and in those with an open
growth plate. In contrast, NSA was similar between hips with and without a cam
deformity in the older soccer players and in those with a closed growth plate.
Because the NSA decreases with age during growth325, this suggests that hips
have a greater risk for developing a cam deformity when the NSA reaches adult
values (more varus position) already in early adolescence. However, as the NSA
was not predictive with regard to visual grading, this point needs further study.
The exact cause of a cam deformity is still unknown, but the current study
provides some important indications toward a better understanding of its
development. Several hypotheses have been proposed, of which some are and
some are not supported by our results. First, it has been suggested that a cam
deformity is a result of secondary remodelling due to OA, which is unlikely for
the majority of cam deformities considering the age of onset. Second, it is
probably not caused by remodelling after a subclinical slipped capital femoral
epiphysis (SCFE) because none of the subjects who developed a cam deformity
during follow-up showed any clinical or radiographic signs of a (subclinical)
SCFE at baseline.23,294 Another proposed hypothesis on the formation of a cam
deformity is a delayed separation of the initially common physis of the greater
trochanter and the femoral head.309 We could not confirm this hypothesis, as all
of the 12-year-old boys in this study had a wide distance between the growth
plate of the femoral head and greater trochanter. We instead observed that
after separation, the distance between the physis of the greater trochanter and
Chapter 10
the femoral head decreased again during adolescence (figure 8). It is possible
that subtle stresses in the proximal femoral growth plate and its connection via
the neck isthmus with the growth plate of the greater trochanter might play an
important role. As the femoral head classically slips toward a posteroinferior
direction with an SCFE, the anterosuperiorly located cam deformity might very
well be a preventive mechanism for a slip, like a ‘clamp’. Another explanation is
that the remnant of the formerly common epiphysis, the femoral neck isthmus,
might continue to have a potential for bone apposition, leading to a broader
neck and aspherical femoral head. This is supported by a study that reports a
bone-bridge formation between the greater trochanter and the femoral head
as a variation in ossification pattern during adolescence.326 The formation of a
cam deformity probably results from an interplay between the proximal femoral
growth plate, the femoral neck isthmus, and the growth plate of the greater
171
Chapter 10
Figure 8. The distance between the lateral part of the proximal femoral growth plate and
the growth plate of the major trochanter is shown (left). During adolescence, this distance
becomes smaller and the head and trochanter might eventually almost connect (right).
172
Development of cam deformity: 2 years follow-up
Chapter 10
173
Chapter 11
ABSTRACT
Objective A cam-type deformity drastically increases the risk of hip
osteoarthritis. Since this type of skeletal anomaly is more prevalent among
young active adults, it is hypothesized that the loading conditions experienced
during certain types of vigorous physical activities stimulates formation of cam-
type deformity. We further hypothesize that the growth plate shape modulates
the influence of mechanical factors on the development of cam-type deformity.
Design We used finite element (FE) models of the proximal femur with an
open growth plate to study whether mechanical factors could explain the
development of cam-type deformity in adolescents. Four different loading
conditions (representing gait, internal rotation, external rotation, and flexion)
and three different levels of growth plate extension towards the femoral neck
were considered. Mechanical stimuli at the tissue level were calculated by
means of the osteogenic index for all loading conditions and growth plate shape
variations.
Results It was found that loading conditions and growth plate shape influence
the distribution of osteogenic index in hips with an open growth plate, thereby
driving the development of cam-type deformity. In particular, external rotation
and flexion and a larger growth plate extension towards the femoral neck
increases the chance of cam-type deformity.
176
Effect of loading on cam deformity
INTRODUCTION
In cam femoroacetabular impingement (FAI), an osseous bump at the femoral
head neck junction results in an abnormal contact between the acetabulum
and femoral head during hip movements. The impingement causes higher
compression and shear forces in the joint that might damage the articular
cartilage and labrum.55 Therefore, cam impingement is thought to be an
important factor in the development of osteoarthritis (OA) of the hip especially
in young and active adults.203 A recent prospective cohort study showed that
the presence of a cam deformity at baseline in hips without OA confers a 10-
fold increased risk for development of end-stage OA within 5 years.203 Previous
studies have focused on improving the diagnosis options,67,263,328 characterizing
the related symptoms,308 and identifying the best treatments possibilities for
cam FAI.329-332 Nevertheless, the etiology of cam impingement is not yet well
understood.
The formation of the osseous bump that causes the impingement might be due
to childhood hip diseases such as slipped capital femoral epiphysis (SCFE)333-335
or Legg-Calve-Perthes disease.336 Additionally, incorrect healing after a femoral
neck fracture can result in a bump at the anterosuperior region.337,338 However,
cam-type deformity is in most cases observed without signs of any pediatric
disease or a former fracture. As cams are generally diagnosed in young and
active adults,23,226,227,285,339 it has been suggested that excessive femoral loading
due to a high level of physical activity during skeletal development might trigger
the abnormal morphology.340 Agricola et al.226 studied adolescent soccer players
and compared them with a control group of the same age (12 to 18 years).
They observed that cam-type deformities are more prevalent within the athletes
group. Furthermore, a cam-type deformity was already present from the age
of 13, before complete closure of the growth plate. After 2,5 years follow-up, it
was observed that a cam-type deformity in the soccer players only developed
during skeletal maturation, when the growth plate is open. Similar trends were
observed in adolescent basketball players as studied by Siebenrock et al.227 who
Chapter 11
177
Chapter 11
METHODS
Geometry and mesh generation
The geometry of the proximal femur was extracted from a CT dataset of a young
and healthy individual (male, 12 years old, left leg) using the image processing
software Mimics 14.0 (Materialise, Belgium). The CT-scans were obtained for
other unrelated medical reasons and came from a database of the Erasmus
Medical Centre Rotterdam, The Netherlands. The images were acquired using
Siemens SOMATOM Emotion 6 with an in-plane isotropic voxel size of 0.545 mm
and a slice thickness of 1.3 mm. That is thin enough to account for the complex
innominate structure.347
A convergence study was conducted using both 10-node quadratic tetrahedral
elements (C3D10) and 4-node linear elements (C3D4). The elements of a selected
region of the femoral head were chosen for the convergence study. It was found
that when 500,000 degrees-of-freedom are used, the calculated field variables
such as strain energy and displacement of the elements converge within 1%
both for linear and quadratic elements. Therefore, the mesh generated using
178
Effect of loading on cam deformity
the 3D femoral geometry was considered to represent the convex growth plate
shapes (CGP1-CGP3). The parameters of the growth plate shapes are presented
in Table 1.
179
Chapter 11
A C A
r1
B r2
B
D
Table 1
CGP1 3 33 30 1:6
CGP2 3 32 29 1:6
CGP3 3 32 29 1:6
t: thickness of the growth plate, 2C-ratio: height to length ration of the growth plate
1
Material properties
The material properties of the model are based on the x-ray attenuation values
measured in Hounsfield Unit’s (HU) derived from CT images. The average
HU value for each element was computed and 200 bone density values were
assigned according to the following linear relation:
ρ(g/cm3) = α + βHU (1)
Since there was no calibration phantom included in the scans, the α and β values
were determined based on the average value of the cortical bone density and air
density. The air density was set to 1.225×10-3 g/cm³. For the cortical bone, an
average density of 0.9 g/cm³ was chosen.348 Using the HU values corresponding
to average cortical bone density and to air, the parameters of Equation (1) were
determined as α = 0.45, β = 0.44×10-3.
The relationship between apparent bone density and elastic modulus was based
on the experiment by Öhman et al. 348
on the cortical bone tissue of children:
Ε(MPa) = 12900ρ 2
(2)
180
Effect of loading on cam deformity
The Poisson’s ratio was fixed at 0.3 similar to other studies that have used
Öhman’s relationship.349 The growth plate was modeled with a constant Young’s
modulus of 6 MPa and a Poisson’s ratio of υ=0.49 to account for almost
incompressible material.350 FE models were solved using an implicit nonlinear FE
solver, namely Abaqus Standard (Dassault Systems Simulia).
Loading Conditions
Various loading conditions were considered to study the effects of loading
conditions on the formation of cam-type deformity. The literature was consulted
to find the appropriate range of load magnitudes and inclination angles. The
first ‘control’ loading scenario was based on the resultant hip contact force of
the normal walking cycle of the study of study by Bergmann et al.351. Although
the measurements were carried out for arthroplasty patients and may not
represent the loading conditions of healthy individuals, they were considered
to be accurate enough for the purpose of the current study. The peak force was
1200 N, i.e. 250% body weight (BW) with inclination angles of 15º and 35º in
the coronal and transverse planes, respectively. Since the angle in the sagittal
plane was not specified, a value of 20º was used based on the normal gait in the
study by Carriero et al.352
It is hypothesized that specific types of movements trigger the formation of
cam-type deformity. For example, 90 degrees of flexion together with internal
rotation causes the impingement of the cam-type deformity into the acetabulum.
Therefore, it has been suggested that these motions might also trigger the
formation of the cam-type deformity itself. Carriero et al. studied the gait
cycle of children with cerebral palsy who have muscle spasticity that results in
excessive internal hip rotation. The peak inclination values of this study were
used in establishing loading case 2 that represents internal rotation. In loading
case 3 that represented external rotation, the load was placed close to the
position where the cam-type deformity normally develops, i.e. more laterally
in the lateral anterosuperior region. For loading case 4, different motions were
Chapter 11
181
Chapter 11
Table 2
Case 1 15 35 20 1200 - -
Internal rotation
External rotation
Hip flexion
Case 4 15 80 70 1200 - -
The inclination angles are positive in lateral or posterior direction.
35º
8º
Figure 2. Four loading cases depicted in the coronal, sagittal, and transverse views
Fig. 7 Four loading cases presented in the coronal, sagittal and transverse view. Loading scenarios 2 and 3 are extended with the implementation of a muscle
force of 364N. The hip contact force magnitude is 1200N.
20
182
Effect of loading on cam deformity
Analysis of growth
According to Hueter-Volkmann’s law, pressure on the epiphyseal plate retards
bone growth, while tension accelerates it.354,355 This principle can be simulated
using mechanobiological models such as those proposed by Stokes356 and
Carter346. In this study, we used Carter’s model to predict bone growth. In
Carter’s model, the rate of bone growth is dependent on both biological and
mechanical components.345 The biological factors include intrinsic genetic
factors, hormonal regulation, and nutrients. The biological component is usually
set as a constant when studying the effects of mechanical loads on growth. The
mechanical component can be described by the osteogenic index (OI):346
OI = ασS + bσH (3)
where σS is the peak octahedral shear stress and σH is the peak hydrostatic
stress. It is assumed that cyclic hydrostatic stress inhibits bone growth while
cyclic octahedral shear stress stimulates growth . The constants a and b
343
regulate the contributions of both components on the growth rate. Since bone
grows between 40 and 80% of the normal size without mechanical loading, the
ratio of b to a should be set between 0.3 and 1 to ensure the contribution of
mechanical loading is about 50% of the contribution of biological factors. In this
study, the biological component was not included and b was set to 0.5.357,358
RESULTS
As the epiphyseal extension, Q1, increases from 1.2 (CGP1) to 1.6 (CGP3),
the values of the osteogenic index (OI, a measure of the relevant mechanical
loading of the tissue) calculated for the proximal and distal sides of the growth
plate deviate from each other (Figure 3). The difference between proximal and
distal sides is the largest close to the region where cam-type deformity normally
develops (Figure 3). The same trend holds for all loading cases (Figure 3).
However, there is a major difference between the various loading conditions in
terms of the side that experiences the highest values of OI. For loading cases
Chapter 11
1 (normal gait) and 2 (internal rotation), the level of osteogenic index is higher
on the proximal side of the growth plate, whereas the reverse holds for loading
cases 3 (external rotation) and 4 (flexion).
For CGP1 (Q1= 1.2), loading case 3 results in a compression peak at the lateral-
anterior region that is reflected in the low OI values in that same region (Figure
4). In loading case 2, the peak is located at the medial-anterior region (Figure
4). Higher growth rates appear when the load is directed more parallel to the
growth plate axis on the opposite site of the contact force, as in the extreme
183
Chapter 11
loading cases 3 and 4. In CGP3 (Q1= 1.6) under loading case 3, a lateral edge
occurs under tension resulting in high OI values laterally (Figure 4). A clear
increase in OI can be observed when comparing loading cases 3 and 4 with 1
and 2 (Figure 4).
CGP1
CGP2
CGP3
Figure 3. The distribution of OI for different loading cases and convex growth plate shapes.
Osteogenic�Index�
� Case�1�� � � Case�2�� � � Case�3�� � � Case�4�
� Loading case 1 Loading case 2 Loading case 3 Loading case 4
Osteogenic
� index, MPa
� � � GP1�
� � � � �
�
�
�
�
�
�
�
� CGP1 Cam�Region�
�
�
� � � GP3�
CGP3
Figure 4.andThe
Fig. 15 GP1 distribution
GP3 under of OI in the growth plate for different loading cases and two
four loading conditions.
different growth plate shapes.
From the clinical follow-up study,236 a femur of an adolescent soccer player with
an epiphyseal extension, Q1, of 1.6 at baseline who appeared to have developed
30
184
Effect of loading on cam deformity
a cam-type deformity at two years follow-up (Figure 5a-b) was selected for
comparison with simulation results. Formation of cam-type deformity was found
anterosuperiorly while bone resorption was observed inferiorly (Figure 5b).
The formation of cam-type deformity was found for most femurs that had an
open growth plate at baseline. After epiphyseal closure, the morphology did not
change substantially (see e.g. Figure 5a-b). The FE model could explain these
clinical findings where the distribution of OI of the proximal femur with CGP3
(corresponding to Q1=1.6) under loading case 3 shows high OI values in the
epiphysis distributed along the top of the growth plate towards the lateral side
(Figure 5c). Low OI values were found in the area under the medial side of the
growth plate (Figure 5c). The scaled plot and contour plot of the growth plate
indicate high growth rates at the lateral edge of the growth plate (Figure 5d).
Bone�Formation�
Bone�Resorption�
Chapter 11
185
31
Chapter 11
Another representative femur with a low extension at the baseline (Q1=1.2) was
also selected from the prospective study (Figure 6). The femoral head was found
to deflect minimally at the anterosuperior side (Figure 6a). Under the superior
side of the growth plate, low OI values were observed (Figure 6c-d). The high
values were distributed along the growth plate (Figure 6c-d).
Posterior�
Medial�
Lateral�
Anterior�
DISCUSSION
A larger epiphyseal extension results in large differences between the OI values
on the proximal and distal sides of the growth plate particularly for loading cases
3 and 4 (external rotation and flexion, Figure 3). For the largest 32epiphyseal
extension (Q1=1.6), some of the highest OI values are found on the lateral side
186
Effect of loading on cam deformity
of the femur in the area where cam-type deformity normally develops (Figure 3).
On the other hand, OI values are relatively low on the medial side of the femur
(Figure 3). As a result of this imbalance and concentration of high OI values
close to the site of cam-type deformity, the anomalistic bone growth occurs.
These observations match the results of our prospective follow-up study236
where cam-type deformity is found to be associated with a large epiphyseal
extension. Detailed analysis of a representative radiograph from the cam-type
deformity group (Figure 5) shows that both bone apposition and bone resorption
occur during the development of cam-type deformity (Figure 5b). The location
of the bone apposition is coincident with the location where high OI values are
found in the FE analysis (Figure 5b-c) whereas bone resorption occurs on the
medial side of the femur where low OI values are found (Figure 5b-c).
The results of this study clearly show that the development of cam-type deformity
is directly related to the physical activities undertaken by the individual before
growth plate closure. Both the level of OI values and the distribution of OI
within the epiphysis are significantly different between gait (loading case 1) and
external rotation and flexion (loading cases 3 and 4) (Figure 3). As for the level
of OI, higher OI values are calculated for loading cases 3 and 4 (Figure 3) as
compared to loading case 1. In addition, the areas with high osteogenic index
values are located closer to the cam-type deformity in loading cases 3 and 4 as
compared to loading case 1 (Figure 3). It is therefore expected that people who
are exposed to loading cases 3 and 4 (such as soccer players and basketball
players) are more likely to develop cam-type deformity as compared to the
control group whose loading is predominantly described by patterns similar
to loading case 1. These phenomena combined with the fact that the loading
conditions are more dynamic in sports (impact-like forces) increase the chance
of cam-type deformity in young individuals practicing in (pre-professional)
sports requiring external rotation and flexion of the hip while weight-bearing
such as young soccer players. These observations are in agreement with the
results of clinical studies that have shown young active adults participating in
Chapter 11
187
Chapter 11
Stress, MPa
Figure 7. Hydrostatic stress contour plots for loading case 3 with CGP1 (see methods for
description).
Those relatively large compressive stresses tend to inhibit bone growth while
tensile stresses can be found on the medial side of the growth plate that
stimulate bone growth. As a result of this mismatch between the growth of
the lateral and medial sides, the growth plate may start to deflect and the
epiphyseal extension may increase. It could therefore be postulated that the
loading patterns experienced by soccer players may be partially responsible
for the large epiphyseal extension seen in the cam-type deformity groups.236
Once the epiphyseal extension increases due to excessive external rotation, the
distribution of the OI caused by large epiphyseal extension further increases the
chance of cam-type deformity, as described above.
In summary, the results presented here support both hypotheses of the study
and show that both growth plate shape and loading conditions contribute
to increased mechanical stimuli that initiate the development of cam-type
deformity. Moreover, there is a modulation between the shape of growth plate
and the loading conditions experienced during vigorous sport activities. As
far the shape of the growth plate is concerned, a large epiphyseal extension
increases the chance of formation of cam-type deformity. The presented results
are in agreement with the observations reported in a our recent prospective
follow-up study of young soccer players236 and reveal the role of mechanical
factors in the etiology of FAI. These findings might be useful when considering
implementation of preventive strategies on the formation of cam-type deformity,
which might have a considerable impact on the incidence of osteoarthritis.
188
Chapter 12
ABSTRACT
Femoroacetabular impingement (FAI) is characterized by abnormal contact
between the proximal femur and the acetabulum. Two subtypes have been
described: pincer impingement, caused by an overcovered acetabulum; and
cam impingement, which occurs as a result of an aspherical femoral head (cam
abnormality). A strong correlation exists between cam impingement and the
subsequent development of hip osteoarthritis (OA). Major cam deformities
confer a high risk of OA. However, the association between cam abnormalities
and the pathology of OA has been difficult to compare between studies, as
different methods have been used to define the abnormality. Cam abnormalities
are acquired during skeletal growth and could be influenced by high impact
sporting activities. Preventative treatments aiming to reduce the incidence of
cam abnormalities and subsequent OA could, therefore, be developed. In this
Perspective, we discuss the current understanding of FAI, focusing on cam
abnormalities and their association with OA.
190
Cam impingement: an opinion
INTRODUCTION
Historically, most hip osteoarthritis (OA) has been defined as idiopathic,
as the cause is often unknown. Over the past decade, subtle morphological
abnormalities have been increasingly recognised as potential risk factors for OA.
The observation that a nonspherical femoral head might lead to OA of the hip
was first made at the beginning of the last century, but was only sporadically
reported thereafter.24,50,359
Morphological abnormalities in the hip joint can cause femoroacetabular
impingement (FAI), which can, in turn, cause motion-dependent soft-tissue
damage.55 Two types of FAI exist, one caused by an abnormality in acetabular
shape or orientation (pincer type), the other caused by a shape abnormality in
the proximal femur (cam type). In pincer impingement, the abnormal contact
between the femoral neck and the overcovered acetabular rim is thought to
cause an impaction type of impingement which compresses the labrum, and
may subsequently lead to circumferential cartilage damage. The cam type of
FAI is caused by extra bone formation at the anterolateral head–neck junction,
which results in a nonspherical, cam-shaped abnormality (figure 1). The cam
is forced into the acetabulum during flexion and internal rotation of the hip, a
process referred to as cam impingement (Figure 1). Cam impingement can cause
an inclusion type of damage to the acetabular labrum and articular cartilage,
which is typically accompanied by detachment of the acetabular cartilage from
the subchondral bone. This biomechanically based hypothesis of how damage
is caused by cam impingement is supported by intra-operative findings in
symptomatic patients with cam abnormalities—acetabular cartilage delamination
has been found in the anterosuperior quadrant of the joint, corresponding to
the site where the cam abnormality is forced into the acetabulum.57 These
observations suggest a relationship between cam impingement and OA, but the
epidemiological evidence for this link is scarce.
The presence of a cam abnormality can be found radiographically in 15–
25% of males and 0–15% in females in the general population.60,275 Treating
the cam abnormality surgically has become increasingly common, but the
underlying etiology is unknown. Cam abnormalities might be acquired during
skeletal maturation and influenced by certain loading patterns of the hip.226,227
In this Perspectives article, we discuss the current knowledge within the field
of FAI, focusing on cam impingement, and especially on the aetiology of
cam abnormalities, the definition of abnormal in this context, and how these
Chapter 12
191
Chapter 12
192
Cam impingement: an opinion
the alpha angle, and total hip arthroplasty was found with an adjusted OR of
1.05 for every one degree increase in alpha angle (P = 0.001).63
193
Table 1. Studies presenting descriptive data of the alpha angle
194
Sympt/ Age mean Hips Men Imaging modality Alpha angle
Study Alpha angle mean (sd, range)
Asympt (sd, range) (n) (%) view/location threshold
Chapter 12
Gosvig Asympt 62 (men) (3) 2110 37.6% X-ray: AP 69°, 83° (1) Right hip 51.7° (13.5, 34.0-100.0)
et al. (NA, 23-93) Left hip 53.1° (13.9, 30.0-94.0)
65 (women) 3496 X-ray: AP 51°, 57° (1) Right hip 44.4° (5.5, 34.0-90.0)
(NA, 22-92) Left hip 45.5° (5.1, 32.0-108.0)
Pollard Asympt 48 (men) 78 47.0% X-ray: cross-table 62° (2) Men: 48° (8, 95% CI 45-50)
et al. (12, 28-69)
44 (women) 88 X-ray: cross-table 62° (2) Women: 47° (8, 95% CI 45-49)
(11, 22-67)
Fraitzl Asympt 47 (men) 170 50.1% X-ray: AP 70° (2) Men: 49.4° (10.5, 34.5-88.3)
et al. (17, NA) Frog leg 70° (2) Women: 49.1° (10.6, 27.0-87.9)
55 (women) 169 X-ray: AP 61° (2) Men: 45.0° (8.0, 29.5-77.6)
(17, NA) Frog leg 66° (2) Women: 46.1° (9.9, 23.0 - 82.5)
Notzli Sympt 35 39 66.7% MRA NA All sympt:74° (5.4, 55-95)
et al. (9, 18-57) obl. ax/ant
Asympt 30 35 48.6% MRI NA All asympt: 42° (2.2, 33-48)
(5, 20-46) obl. ax/ant
Kassarjian Sympt 37 42 55.0% MRA 55° All: 69.7° (NA, 41 - 91)
et al. (NA, 17-67) obl. sag/ant
Beaule Sympt 41 36 56.7% CT 50.5° All sympt: 66.4° (17.2, 39-94)
et al. (NA, 25-54) sag. ax/ant Men 73.3° (NA, NA) women 58.7° (NA, NA)
Asympt 37 20 58.3% CT 50.5° All asympt: 43.8° (4.46, 39-48)
(NA, 18-70) sag. ax/ant Men 43.7° (NA, NA) women 43.8° (NA, NA)
Pfirrmann Sympt 29 50 60.0% MRA NA Cam hips (4): 68° (19, NA), 81° (15, NA), 53° (20, NA)
et al. (NA, 19-48)
rad/ant, ant-sup, sup Pincer hips (4): 54° (11, NA), 66° (19, NA), 42° (5, NA)
Clohisy Sympt 32 61 74.0% X-ray: AP NA All sympt: 71.5° (sd 19.4, 38-132)
et al. (NA, 14-53) cross-table, frog leg 58.8° (sd 15.7, 31-101), 65.2° (15.0, 38-114)
Asympt 35 24 54.0% X-ray: AP NA All asympt: 51.2° (sd 15.7, 36-94)
(NA, 18-49) cross-table, frog leg 47.2° (15.4, 30-92), 43.7° (12.1, 31-76)
Toogood Asympt 44 400 50.0% Photographed NA All: 53.5° (12.68, 31-112), 45.6° (10.5, 17-79)
et al. (NA, 18-89) anterior, lateral
Table 1. Studies presenting descriptive data of the alpha angle (Continued)
195
Cam impingement: an opinion
Chapter 12
Chapter 12
Data from the CHECK cohort,203 a large nationwide Dutch prospective cohort
study, confirmed the association between a cam abnormality and an elevated
risk of OA in 723 subjects who had no radiographic evidence of OA upon entry
into the cohort but developed OA after 5 years. The presence of a moderate
(alpha angle >60°) or severe (alpha angle >83°) cam abnormality at baseline
was significantly associated with the development of end-stage OA, defined
by Kellgren & Lawrence grade 3 or 4, or total hip arthroplasty within 5 years
(OR = 3.7, 95% CI 1.7–8.0 and OR = 9.7, 95% CI 4.7–19.8 for moderate and
severe cam abnormalities, respectively).203 Notably, if a severe cam abnormality
was accompanied by limited internal hip rotation (≤20°), the OR was further
increased. Moreover, larger alpha angles, which often imply more severe cam
abnormalities, are associated with a substantially increased risk of cartilage
damage and OA.63,203,262
The prevalence of cam abnormalities in these epidemiological studies might be
underestimated, as only anterioposterior radiographs were used for assessment.
Ideally, a potential cam abnormality is assessed using 3-dimensional radial images,
which are better than standard radiography for visualizing cam abnormalities
located anterolaterally.67 In addition, almost all published longitudinal cohorts
consisted of only middle-aged individuals, despite the relationship between cam
impingement and OA being poorly defined in individuals 20–45 years of age.
Although cam impingement is strongly correlated with the development of OA in
patients ≥45 years of age, there are indications that cartilage damage caused
by a cam abnormality might already exist from the age of 20 years. In a study
of symptomatic patients with a mean age of 28 years, cam abnormalities were
often found in hips with the types of intra-articular damage, such as cartilage
delamination, that are associated with OA of the hip.59 In only one study264
has the relationship between cam abnormality and intra-articular hip damage
been examined in asymptomatic younger individuals. In this study, MRI was
used to assess hip morphology in this cross-sectional study of 244 young males
(mean age 19.9 years). A decreased combined femoral and acetabular cartilage
thickness was found in individuals with a visually scored definite or severe cam
abnormality.264 The OR for labral lesions was 2.77 (95% CI 4.6–10.4) when
comparing those with and without cam abnormalities. The association between
labral tears and cam abnormalities, however, should be interpreted with caution,
as labral tears are highly prevalent in the asymptomatic population.264
196
Cam impingement: an opinion
Developmental causes
The prevalence of cam abnormalities has geographical variability. The prevalence
in Chinese people is substantially lower than in white people.38 Although cultural
differences in activities that involve hip loading and subsequent adaptation during
growth could exist, genetic differences between Chinese and white people could
also contribute to the difference in the prevalence of cam abnormalities. Siblings
of patients with cam abnormalities had a relative risk of 2.8 of having a cam
abnormality compared with the patients’ spouses.360 However, siblings might
participate in similar sporting activities during growth, and so far only indirect
associations between genetics and cam abnormalities have been documented.
Although some genetic involvement is likely, current evidence is scarce.
Environmental causes
Indications exist to support the hypothesis that a substantial proportion of cam
abnormalities are acquired during skeletal maturation. First, a cam abnormality
is not present, or at least not radiographically visible, in boys <13 years of
age.226 Second, the formation of a cam abnormality in individuals 13–14 years of
Chapter 12
197
Chapter 12
Most cam abnormalities are probably acquired during growth and are heavily
influenced by high impact and high intensity athletic activities. Interestingly,
this theory implies that the formation of a cam abnormality could be prevented
198
Cam impingement: an opinion
199
Chapter 12
Advanced imaging modalities are also being developed to predict the presence
of dynamic impingement. Acetabular chondrolabral damage at the typical
anterosuperior location in an individual with a cam abnormality could support
the diagnosis of cam impingement. Quantitative analysis of biochemical cartilage
composition using techniques such as delayed gadolinium-enhanced MRI of
cartilage, T1 rho, and T2 mapping, could provide such evidence in the very early
stages of cartilage damage.262 Motion simulations of the hip, based on MRI or
CT, provide further insight into the range of motion, and the location of bone
impingement.361,362 Logically, other causes that might lead to groin pain should
be ruled out.
CONCLUSIONS
Cam abnormalities are prevalent in the general population and probably result
from high impact activities during bone maturation. The diagnosis of cam
impingement should be based on both radiographic and clinical findings. Cam
impingement is strongly associated with the development of OA and future
studies should therefore focus on the pathophysiology of cam abnormalities.
Altering activity levels during adolescence could then be used as strategies to
prevent cam abnormalities from developing, thereby eliminating a major risk
factor for hip OA.
200
Chapter 13
Summary
Summary
203
Chapter 13
204
Summary
view of the hip was included and it was demonstrated that the risk of hip OA
significantly increased when acetabular dysplasia was present both laterally and
anteriorly in one hip (odds ratios between 4.9 and 6.5). Current evidence for
the relationship between a pincer deformity and OA is conflicting and limited by
retrospective or cross-sectional study designs. We found a protective effect of
a pincer deformity when present both anteriorly and laterally in one hip, with a
corresponding odds ratio of 0.34 (95% CI 0.1-0.9) for incident hip OA (K&L ≥2).
This finding was supported by the fact that none of these hips developed end-
stage OA (K&L 3,4, or THR) within 5 years’ follow-up. The currently available
literature suggests that pincer impingement is a mechanism which might lead
to pain and hip damage as shown by studies of symptomatic patients. However,
our data suggests that pincer impingement is not associated with hip OA on an
epidemiological scale but might even be protective. In chapter 6, we looked to
see whether cam impingement at baseline was associated with the development
of hip OA over a period of 5 years. We showed that a cam deformity was strongly
associated with the development of end-stage OA. A moderate cam deformity
(alpha angle >60°) and a severe cam deformity (alpha angle >83°) resulted
in respective odds ratios of 3.7 (95% CI 1.7-8.0) and 9.7 (95% CI 4.7-19.8)
for end-stage OA. The combination of a severe cam deformity and decreased
internal rotation –a clinical sign of cam impingement– at baseline resulted in an
even more pronounced odds ratio of 25.2 (95% CI 7.9-80.6) and in a positive
predictive value of 52.6%. The data suggest that the risk of OA is higher when
the cam deformity is more pronounced.
to repetitive collisions that damage the soft-tissue structures within the joint
205
Chapter 13
itself’. This definition contains five essential elements: (1) abnormal morphology
of the femur (cam) and/or acetabulum (pincer), (2) abnormal contact between
these two structures, (3) especially vigorous supraphysiological motion that
results in such abnormal contact and collision, (4) repetitive motion resulting
in the continuous insult, and (5) the presence of soft tissue damage. The most
investigated subcategory of FAI is cam impingement, in which a cam-shaped
femoral head causes impingement. However, there is no consensus on how to
define and quantify the presence of a cam deformity. It is usually quantified by
the alpha angle, which measures the extent to which the femoral head deviates
from being spherical, but there is no consensus on which alpha angle threshold
to use to define the presence of a cam deformity. In Chapter 8, we proposed
alpha angle threshold values to define the presence of a cam deformity and a
pathological cam deformity. By combining alpha angles of two large prospective
cohort studies (CHECK cohort and Chingford cohort), comprising almost 3000
hips, an alpha angle threshold for defining the presence of a cam deformity
was determined based on the finding of a bimodal distribution of the alpha
angle. Interestingly, a bimodal distribution indicates two different populations,
one without cam deformity and one with cam deformity. The optimal threshold
that distinguishes between both distributions was assessed and resulted in an
alpha angle threshold of 60°. As greater alpha angles substantially increase
the risk for OA, we also determined a pathological alpha angle threshold based
on developing end-stage OA at follow-up. A pathological threshold of 78° was
proposed, as this value resulted in the maximum area under the ROC curve
(highest sum of sensitivity and specificity), which was 0.69 (95% CI 0.62-0.75)
for end-stage OA.
206
Summary
the amount of growth plate extension towards the femoral neck is strongly
207
Chapter 13
208
Chapter 14
General discussion
General discussion and future perspectives
GENERAL DISCUSSION
In this thesis, the rise (skeletal development) and fall (OA) of the hip was
studied. In the first part, the role of hip morphology in the development of
hip OA was investigated. In the second part, we defined FAI and the presence
of a cam deformity. In the third part, we studied if and how a cam deformity
developed during skeletal maturation and how this was influenced by athletic
activities.
211
Chapter 14
shape during skeletal maturation as has been shown in chapter 9 and 10 of this
thesis. Finally, there is a relationship between heavy physical workload and OA;
farmers have an up to 10-fold increased risk to develop OA.45 Like high impact
athletic activities, heavy physical workload during skeletal growth might also
lead to an abnormal hip shape, although no studies are available on this topic.
Besides the fact that these factors are important in establishing the shape of
the hip during skeletal growth, they are also important during mid adulthood
because they might determine whether or not an abnormal hip shape will lead
to development of OA. For example, OA susceptible SNPs might affect cartilage
structure or metabolism such that it becomes more vulnerable to biomechanical
stresses of an abnormal hip shape. Also, certain athletic activities and
occupations reported to be at risk of OA might modify the association between
a cam deformity and hip OA as they require impingement movements (flexion
+ internal rotation). These movements might cause damage in the presence of
a cam deformity, while somebody with a cam deformity not experiencing these
movements repetitively will probably not develop OA. Thus, the main risk factor
for hip OA is probably an abnormal shape of the hip joint, while this relationship
is modified by genetics and environmental factors. The different types of hip
shape which are considered ‘abnormal’ will be discussed below.
Figure 1. A non-optimal hip shape might be the main trigger for hip OA. This pathway
might be influenced by genetics and environmental factors twice in life: during skeletal
maturation, since these factors co-determine the final hip shape and during adulthood, when
these factors might modify the association between an abnormal hip shape and hip OA.
Acetabular dysplasia
Acetabular dysplasia is characterised by acetabular undercoverage of the femoral
head, leading to a decreased contact area which might result in higher static loads
on the anterosuperior acetabular labrum and cartilage. Mild acetabular dysplasia
has previously been associated with hip OA in prospective studies,53,222 but the
results from cross-sectional and retrospective studies are conflicting.241,242 Also,
the prospective studies included primarily older people (mean ages between 65
and 70 years) and evidence for the relationship between acetabular dysplasia
and OA at the age of 50-60 years is limited. It has been suggested that the
212
General discussion and future perspectives
213
Chapter 14
Pincer impingement
The presence of a pincer deformity is thought to result in labral damage by
the linear impaction of the femoral neck against the acetabular rim. During
abutment, the femoral neck might create a leverage effect, so that the femoral
head is forced in the posteroinferior portion of the acetabulum, leading to a
so-called contre-coup lesion (figure 2 E-G). Although this pattern of labral
and cartilage damage damage has been demonstrated intra-operatively in
symptomatic hips with a pincer deformity,57 scarce epidemiological data on the
risk of development of hip OA is available. In chapter 5, we could not identify
such an association between a pincer deformity at baseline and development of
hip OA five years on. In contrast to our expectations, a pincer deformity might
even have a protective effect on the risk of developing OA when acetabular
overcoverage is present both anteriorly and laterally in one hip. Regarding the
results of the properly designed studies, the question arises whether such as
pincer impingement truly exists. Based on clinical observations and case series,
a pincer deformity is likely to be a cause of hip pain and subsequent OA in
some patients, but on an epidemiological scale a pincer deformity is probably
not associated with hip OA. One explanation might be that there is a mismatch
between individuals with a radiographic pincer deformity and individuals who
214
General discussion and future perspectives
Cam impingement
A cam deformity is an extra bone formation in the anterolateral head-neck
junction, resulting in a nonspherical femoral head which can be forced into the
acetabulum during flexion and internal rotation of the hip (figure 2 B-D). Using
the definition for the presence of a cam deformity from chapter 8, we showed
in chapter 6 a strong association between cam impingement and development
of especially fast progression to end-stage OA. The presence of a radiographic
cam deformity resulted in a grossly fourfold increased risk for development of
hip OA, and a severe cam deformity (alpha angle >83°) resulted in a tenfold
increased risk. These results are in agreement with other studies showing an
association between a cam deformity and hip OA and with studies showing
a correlation between the magnitude of the alpha angle and the severity of
acetabular cartilage damage at the side where the cam deformity is forced into
the acetabulum.57,59,60,63-65,208,244,283,377
Despite its strong association with hip OA, there are still more people with a cam
deformity that will not develop OA. As a cam deformity is a potential modifiable
risk factor, it is important to recognise which individuals will eventually develop
OA. Distinguishing these people who might benefit from surgery should be done
before too much intra-articular damage has occurred, as post operative results
become worse with more advanced cartilage damage.166 Our proposed clinical
definition of FAI which comprises five essential elements can help to distinguish
between people that will or will not develop OA. In most cohort studies only
the first element, the presence of abnormal morphology (the radiographic
presence of a cam deformity), is taken into account. Therefore, we attempted to
include the second element, abnormal contact between the cam deformity and
the acetabulum, as well. We defined abnormal contact by a decreased internal
hip rotation (<20°) in 90° of flexion, which is a typical clinical sign of cam
impingement.216 By combining the presence of a radiographic cam deformity
Chapter 14
215
Chapter 14
CHECK cohort
CHECK is a nationwide prospective cohort study of 1002 individuals between 45
and 65 years who visited their general practitioner (GP) for the first time with
pain and/or stiffness in their hip and/or knee. These inclusion criteria were set
to study the factors related to development and progression of hip and knee OA.
Of the included individuals, 80% were women. Using these criteria, it happened
216
General discussion and future perspectives
that none of the individuals had radiographic signs of definite hip OA at baseline
with a K&L score of 0 in 76% of the hips and a K&L score of 1 in 24%.
These inclusion criteria should be kept in mind by interpreting the results of
chapter 3, 4, 5 and 6. The results found in our studies might be different in open
population cohorts, where a relatively smaller proportion of people will have
hip pain. Still, the results can be generalised to a relevant group of individuals,
namely those who visit their GP for the first time. We showed that besides
clinical examination, a simple and inexpensive pelvic radiograph can provide
considerably more information on the prediction of future hip OA. Furthermore,
the minimum age to enter the cohort study was 45 years, which might have
influenced the strength of association between hip shape and development of
OA. Individuals with a cam deformity are at high risk of fast progression to
end-stage OA and preliminary results indicate that the strength of association
becomes weaker when the follow-up time is longer. When somebody with a cam
deformity has not developed OA at a certain age, he or she will be less likely
to develop OA as a result of the cam deformity. Therefore, in cohorts including
older people, the strength of association between a cam deformity and hip OA
is probably lower than in the CHECK cohort, because people will develop OA for
other reasons than a cam deformity. On the other hand, a large number of people
that developed OA as a result of a cam deformity might have been excluded
from the CHECK cohort because they probably had hip complaints before the
age of 45 years. Conversely, the predictive value of a cam deformity might be
higher in younger people than the individuals included in the CHECK cohort, but
no such studies are available yet. This is an interesting point considering the fact
that a cam deformity is present immediately after skeletal growth. It is unknown
from which age someone with a cam deformity will show the first signs of hip OA
and this should be subject of future studies.
In the CHECK cohort, anteroposterior (AP) pelvic and False Profile lateral (FP)
radiographs were obtained both at baseline and 5 years follow-up according to a
standardized protocol. There are no other large cohorts available that obtained
both radiographic views and the addition of lateral radiographs considerably
increased the predictive value of future hip OA, especially by a more precise
quantification of acetabular coverage. Moreover, at follow-up we were able
to quantify the joint space at the locations where impingement occurs. In
addition, an FP view is advocated to be more sensitive in the quantification
of joint space narrowing than an AP view.246 Despite the advantages of two
radiographic views, three-dimensional imaging techniques like MRI or CT might
capture more morphological aspects of the hip joint.67 As a result, we might
have slightly underestimated the prevalence of FAI morphology. The effect of
this underestimation on the prediction of hip OA is unknown.
Chapter 14
217
Chapter 14
Finally, the prospective nature of this cohort study together with the fact that
none of the hips showed definite signs of radiographic OA at baseline were
of considerable importance. Especially in a more advanced stage of hip OA,
the femoral head might become non-spherical as a consequence of OA. The
baseline shape aspects that we found to be associated with hip OA at follow-up
were most likely the cause of hip OA. Also, we could exclude a cam deformity
from being a consequence of OA. These shape aspects might therefore be used
in a predictive model that can assess the chances of developing hip OA when
somebody visits their GP for their first time with hip complaints.
218
General discussion and future perspectives
219
Chapter 14
220
General discussion and future perspectives
This hypothesis can explain the high prevalence of cam deformity in athletes
practising high impact sport. It also explains the strong association between
a cam deformity and amount of growth plate extension towards the neck,
which runs analogous with the formation of a cam deformity, while it does not
precede a cam deformity. This is in line with the findings of chapter 10. Finally,
it may also explain the higher prevalence of cam deformities in males than
in females. In general, fewer females than males practise high impact sports
during adolescence. In addition, as females are skeletally mature at a younger
age, the intensity they practise sport at the time that a cam deformity develops
is probably less.
The finding that a cam deformity does not develop or further increases in size
after closure of the growth plate was supported by data from the CHECK and
Chingford cohorts, where we did not find an increase in alpha angle between
baseline and follow-up. However, it has been shown that the bone mineral
density (BMD) of subchondral bone underlying a cam deformity is higher than
Chapter 14
221
Chapter 14
the BMD at the same location in hips without a cam deformity.380 The BMD of
the cam deformity under the cartilage layer was similar to or even somewhat
higher than normal compact bone in the femoral neck. However, it is unknown if
a higher BMD is already present when the cam deformity is formed. When this is
not the case, an explanation for the higher BMD might be that the stimulus for
bone apposition continuous, but that the newly formed bone is not capable of
changing the femoral morphology as it is now surrounded by a cortical envelope
that cannot expand. The higher BMD is most likely not explained by the abutment
of the cam deformity against the acetabulum, as asymptomatic individuals with
a cam deformity showed the same BMD as symptomatic individuals with a cam
deformity.380 This would suggest that before growth plate closure the bone can
adapt its morphology (modelling) to mechanical load and after growth plate
closure only changes in bone architecture (remodelling) occur.
Combining the results of our studies and other data from literature, I hypothesise
that a cam deformity is a bone adaptation and results from high impact athletic
activities during skeletal maturation. Loading patterns of especially flexion and
external rotation while weightbearing influence bone formation and growth plate
orientation, which together have a synergistic effect on the formation of extra
bone in the anterolateral head-neck junction.
222
General discussion and future perspectives
CONCLUSIONS
We showed that hip shape plays an important biomechanical role in the
development of hip OA. By using statistical shape modelling we were able
to identify all hip shape variants of participants of the Dutch CHECK cohort.
Based on specific shape patterns of the hip in those without OA at baseline, we
could make a good prediction of those who will develop future hip OA. Some of
these shape patterns could even predict future hip OA in a completely different
population from that of the United Kingdom. A broad and short femoral neck are
two examples of such shape patterns that are associated with hip OA. We also
identified that a non-spherical femoral head could predict the development of
hip OA, which has been linked to causing hip OA by a process referred to as FAI.
The clinical and radiographic definitions for this pathological entity have been
provided in the current thesis. We specifically studied the relationship between
cam and pincer impingement and the subsequent risk of developing OA. Pincer
impingement did not increase the risk of hip OA, but cam impingement was
strongly associated with especially fast progressing hip OA. The presence of a
cam deformity can confer up to a 10-fold increased risk for the development
of end-stage hip OA. Besides the fact that a cam deformity is probably the
most important risk factor of hip OA, it is also a potential modifiable risk factor.
We showed that a cam deformity only develops during skeletal maturation
most likely as a result of high impact athletic activities during adolescence.
More specifically, when the growth plate is open, loading patterns of flexion
and external rotation while weight bearing might stimulate the formation of
a cam deformity. Regarding the strong association between a cam deformity
and development of hip OA and the potential modifiable character of a cam
deformity, preventing the development of a cam deformity might substantially
decrease the future prevalence of hip OA.
Chapter 14
223
Chapter 14
FUTURE PERSPECTIVES
Based on the outcomes of this thesis, three main topics will be discussed in this
chapter: 1. The relationship between FAI and OA: what do we not know? 2. The
formation of a cam deformity: which future studies are needed to better unravel
the etiology of a cam deformity? And 3. Can osteoarthritis be prevented by
impeding the development of a cam deformity?
224
General discussion and future perspectives
225
Chapter 14
226
General discussion and future perspectives
Chapter 14
227
Appendices
Nederlandstalige samenvatting
References
PhD portfolio
List of publications
Curriculum Vitae
Dankwoord
Nederlandstalige samenvatting
231
Nederlandstalige samenvatting
Figuur 1. Schematische weergave van FAI. A. Dankzij een sferische heupkop met een
congruent acetabulum heeft de heup een grote bewegingsvrijheid. B,C,D. Een cam
deformiteit (B) kan vooral tijdens flexie en endorotatie van de heup impingement in het
acetabulum veroorzaken (C), wat kan leiden tot schade van het kraakbeen en labrum,
met name antero-superieur (D). E,F,G. Een pincer deformiteit (E) kan vooral tijdens flexie
van de heup leiden tot impingement tegen de femurhals (F), wat tot gegeneraliseerde
kraakbeen en labrumschade zou kunnen leiden in het hele acetabulum (G).
232
Nederlandstalige samenvatting
preventief kunnen zijn. Tijdens deze behandeling wordt het additionele bot van
de asferische heupkop verwijderd zodat er weer een sferische heupkop ontstaat.
Echter, de lange termijn resultaten van deze ingreep zijn nog niet bekend en tot
op heden zijn er nog geen gerandomiseerde studies uitgevoerd. In hoofdstuk
3 hebben we de associatie van de algemene heupvorm op baseline en het
risico op het ontwikkelen van artrose na 5 jaar bestudeerd. Hiervoor hebben
we gebruik gemaakt van het CHECK cohort, een Nederlands cohort van 1002
mannen en vrouwen tussen de 45 en 65 jaar die werden geïncludeerd wanneer
ze voor het eerst bij de huisarts kwamen met klachten van pijn en/of stijfheid
in de knie en/of heup. De algemene vorm van de heup werd op anteroposterior
(AP) röntgenfoto’s gekwantificeerd met Statistical Shape Moddelling (SSM).
Deze software ‘herkent’ op basis van statistische principes alle vormvariaties
die in een populatie aanwezig zijn. Hierdoor is het mogelijk om zonder een
vooraf gedefinieerde hypothese over welke vormaspecten relevant zijn, van alle
vormvariaties op baseline de kans op het krijgen van heupartrose te berekenen.
Tijdens de 5-jaars follow-up werd artrose gedefinieerd volgens de American
College of Rheumatology (ACR) criteria en tevens als het hebben van een
heupprothese op follow-up. Wij lieten zien dat bij mensen die op baseline nog
geen artrose hebben, specifieke vormaspecten kunnen voorspellen of iemand
een heupprothese heeft na 5 jaar. Vooral een brede en korte femurhals, een
geretroverteerd acetabulum en een cam vorm van de kop-hals overgang konden
voorspellen of iemand een heupprothese na 5 jaar zou hebben. Alhoewel de
heupvorm niet kon voorspellen of iemand tijdens follow-up aan de ACR criteria
zou voldoen, kon het wel twee ACR criteria afzonderlijk voorspellen, namelijk het
hebben van een beperkte endorotatie en heuppijn tijdens follow-up. Gebaseerd
op deze resultaten werd in hoofdstuk 4 de generaliseerbaarheid van deze
vormaspecten bestudeerd. Door een identiek shape model in het Chingford
cohort (een open populatie cohort van vrouwen tussen 45 en 65 jaar uit
Londen) te gebruiken, konden we vergelijken hoe consistent de voorspellende
vormaspecten van het CHECK cohort zijn. In overeenstemming met de literatuur
vonden we zowel in het CHECK cohort als in het Chingford cohort voorspellende
vormaspecten van de heup voor het krijgen van een heupprothese. Eén van deze
vormaspecten was significant voorspellend voor het krijgen van een heupprothese
in zowel het CHECK cohort als het Chingford cohort. Daarnaast vertoonden
twee vormaspecten die significant geassocieerd waren met een heupprothese
in het CHECK cohort ook een sterke trend in dezelfde richting in het Chingford
cohort. Dit wil zeggen dat bepaalde vormaspecten bij vrouwen generaliseerbaar
zijn in de voorspelling van het krijgen van een heupprothese, ongeacht de
karakteristieken van de populatie, het röntgenprotocol of de follow-up tijd (5
jaar in het CHECK cohort en 19 jaar in het Chingford cohort). Daarentegen
zijn andere vormaspecten die geassocieerd zijn met een heupprothese juist wel
233
Nederlandstalige samenvatting
234
Nederlandstalige samenvatting
235
Nederlandstalige samenvatting
236
Nederlandstalige samenvatting
237
Nederlandstalige samenvatting
Conclusie
Wij hebben aangetoond dat de vorm van de heup een belangrijke etiologische
rol speelt in het krijgen van heupartrose. Met SSM konden wij alle heupvorm
variaties in het CHECK cohort kwantificeren en met een aantal specifieke
vormvariaties een goede voorspelling maken van welke heupen na 5 jaar artrose
zouden ontwikkelen. Een aantal van deze vormvariaties zijn generaliseerbaar
naar andere populaties en konden ook voorspellen welke heupen in het Chingford
cohort na 19 jaar artrose zouden ontwikkelen. Een brede en korte femurhals
zijn twee voorbeelden van heupvormen die geassocieerd zijn met artrose.
Naast het bestuderen van algemene vormvariaties hebben we aangetoond dat
heupdysplasie geassocieerd is met het krijgen van artrose. Tenslotte hebben
we FAI gedefinieerd en onderzocht of pincer impingement (overhelling van
het acetabulum door een pincer deformiteit) en cam impingement (asferisch
femurkop door een cam deformiteit) geassocieerd waren met artrose. Pincer
impingement was niet geassocieerd met het ontwikkelen van heupartrose (het
beschermt juist), terwijl cam impingement wel sterk geassocieerd was met
heupartrose. Personen met een cam deformiteit hadden grofweg een 10-voudig
verhoogde kans op het krijgen van artrose. Bovendien is een cam deformiteit
een mogelijk vermijdbare risicofactor. Wij hebben namelijk aangetoond dat een
cam deformiteit vanaf 13 jarige leeftijd ontstaat en veel vaker voorkomt bij
jeugdvoetballers dan bij ‘niet sportende controles’. Tevens ontwikkelt een cam
deformiteit alleen maar tijdens skeletale groei en verandert de heupvorm niet
meer nadat de proximale femorale groeischijf gesloten is. In de periode van
botrijping leidt een te hoge belasting van de heup, waarschijnlijk vooral door
bewegingen van flexie en exorotatie, tot adaptatie van het bot waardoor een cam
deformiteit ontstaat. De sterke associatie van een cam deformiteit met artrose
in combinatie met het feit dat het onstaan van een cam deformiteit mogelijk
vermijdbaar is, suggereert een potentiële preventieve target om heupartrose in
de toekomst te voorkomen.
238
Appendices
Nederlandstalige samenvatting
References
PhD portfolio
List of publications
Curriculum Vitae
Dankwoord
References
REFERENCES
1. Hogervorst T, Bouma HW, de Vos J. Evolution of the hip and pelvis. Acta Orthop Suppl 2009; 80:
1-39.
2. Cadet ER, Chan AK, Vorys GC, Gardner T, Yin B. Investigation of the preservation of the fluid seal
effect in the repaired, partially resected, and reconstructed acetabular labrum in a cadaveric hip
model. Am J Sports Med 2012; 40: 2218-2223.
3. Tan V, Seldes RM, Katz MA, Freedhand AM, Klimkiewicz JJ, Fitzgerald RH, Jr. Contribution of
acetabular labrum to articulating surface area and femoral head coverage in adult hip joints: an
anatomic study in cadavera. Am J Orthop (Belle Mead NJ) 2001; 30: 809-812.
4. Jay GD, Torres JR, Rhee DK, Helminen HJ, Hytinnen MM, Cha CJ, et al. Association between
friction and wear in diarthrodial joints lacking lubricin. Arthritis Rheum 2007; 56: 3662-3669.
5. Buckwalter JA. Skeletal growth and development: clinical issues and basic science advances1998.
6. Siffert RS. Patterns of deformity of the developing hip. Clin Orthop Relat Res 1981: 14-29.
7. Wolff J. Das gesetz der transformation der knochen. Berlin: Hirschwald1892.
8. Roux W. Der Kampf der Theile im Organismus. Leipzig: W. Engelmann1881.
9. Huiskes R, Ruimerman R, van Lenthe GH, Janssen JD. Effects of mechanical forces on maintenance
and adaptation of form in trabecular bone. Nature 2000; 405: 704-706.
10. Mullender MG, Huiskes R. Proposal for the regulatory mechanism of Wolff’s law. J Orthop Res
1995; 13: 503-512.
11. Bass SL, Saxon L, Daly RM, Turner CH, Robling AG, Seeman E, et al. The effect of mechanical
loading on the size and shape of bone in pre-, peri-, and postpubertal girls: a study in tennis
players. J Bone Miner Res 2002; 17: 2274-2280.
12. Huddleston AL, Rockwell D, Kulund DN, Harrison RB. Bone mass in lifetime tennis athletes. JAMA
1980; 244: 1107-1109.
13. Calbet JA, Moysi JS, Dorado C, Rodriguez LP. Bone mineral content and density in professional
tennis players. Calcif Tissue Int 1998; 62: 491-496.
14. Maimoun L, Coste O, Mura T, Philibert P, Galtier F, Mariano-Goulart D, et al. Specific bone mass
acquisition in elite female athletes. J Clin Endocrinol Metab 2013; 98: 2844-2853.
15. Maimoun L, Coste O, Philibert P, Briot K, Mura T, Galtier F, et al. Peripubertal female athletes in
high-impact sports show improved bone mass acquisition and bone geometry. Metabolism 2013;
62: 1088-1098.
16. Vicente-Rodriguez G, Ara I, Perez-Gomez J, Serrano-Sanchez JA, Dorado C, Calbet JA. High
femoral bone mineral density accretion in prepubertal soccer players. Med Sci Sports Exerc
2004; 36: 1789-1795.
17. Tenforde AS, Fredericson M. Influence of sports participation on bone health in the young athlete:
a review of the literature. PM R 2011; 3: 861-867.
18. Vico L, Collet P, Guignandon A, Lafage-Proust MH, Thomas T, Rehaillia M, et al. Effects of long-
term microgravity exposure on cancellous and cortical weight-bearing bones of cosmonauts.
Lancet 2000; 355: 1607-1611.
19. Inoue M, Tanaka H, Moriwake T, Oka M, Sekiguchi C, Seino Y. Altered biochemical markers of
bone turnover in humans during 120 days of bed rest. Bone 2000; 26: 281-286.
20. Winkelmann WW. Type-B-IIIa hip rotationplasty: an alternative operation for the treatment of
malignant tumors of the femur in early childhood. J Bone Joint Surg Am 2000; 82: 814-828.
21. Wyland DJ, Pill SG, Shanley E, Clark JC, Hawkins RJ, Noonan TJ, et al. Bony adaptation of the
proximal humerus and glenoid correlate within the throwing shoulder of professional baseball
pitchers. Am J Sports Med 2012; 40: 1858-1862.
22. Thijs Y, Bellemans J, Rombaut L, Witvrouw E. Is high-impact sports participation associated with
bowlegs in adolescent boys? Med Sci Sports Exerc 2012; 44: 993-998.
23. Murray RO, Duncan C. Athletic activity in adolescence as an etiological factor in degenerative hip
disease. J Bone Joint Surg Br 1971; 53: 406-419.
24. Murray RO. The aetiology of primary osteoarthritis of the hip. Br J Radiol 1965; 38: 810-824.
241
References
25. Siebenrock KA, Wahab KH, Werlen S, Kalhor M, Leunig M, Ganz R. Abnormal extension of the
femoral head epiphysis as a cause of cam impingement. Clin Orthop Relat Res 2004: 54-60.
26. Buckwalter JA, Saltzman C, Brown T. The impact of osteoarthritis: implications for research. Clin
Orthop Relat Res 2004: S6-15.
27. Reginster JY. The prevalence and burden of arthritis. Rheumatology (Oxford) 2002; 41 Supp 1:
3-6.
28. Nho SJ, Kymes SM, Callaghan JJ, Felson DT. The burden of hip osteoarthritis in the United States:
epidemiologic and economic considerations. J Am Acad Orthop Surg 2013; 21 Suppl 1: S1-6.
29. Bitton R. The economic burden of osteoarthritis. Am J Manag Care 2009; 15: S230-235.
30. Arden N, Nevitt MC. Osteoarthritis: epidemiology. Best Pract Res Clin Rheumatol 2006; 20: 3-25.
31. Hermans J, Koopmanschap MA, Bierma-Zeinstra SM, van Linge JH, Verhaar JA, Reijman M, et al.
Productivity costs and medical costs among working patients with knee osteoarthritis. Arthritis
Care Res (Hoboken) 2012; 64: 853-861.
32. Altman R, Alarcon G, Appelrouth D, Bloch D, Borenstein D, Brandt K, et al. The American College
of Rheumatology criteria for the classification and reporting of osteoarthritis of the hip. Arthritis
Rheum 1991; 34: 505-514.
33. Kellgren JH, Lawrence JS. Radiological assessment of rheumatoid arthritis. Ann Rheum Dis 1957;
16: 485-493.
34. van Saase JL, van Romunde LK, Cats A, Vandenbroucke JP, Valkenburg HA. Epidemiology of
osteoarthritis: Zoetermeer survey. Comparison of radiological osteoarthritis in a Dutch population
with that in 10 other populations. Ann Rheum Dis 1989; 48: 271-280.
35. Nevitt MC, Lane NE, Scott JC, Hochberg MC, Pressman AR, Genant HK, et al. Radiographic
osteoarthritis of the hip and bone mineral density. The Study of Osteoporotic Fractures Research
Group. Arthritis Rheum 1995; 38: 907-916.
36. Jordan JM, Helmick CG, Renner JB, Luta G, Dragomir AD, Woodard J, et al. Prevalence of
hip symptoms and radiographic and symptomatic hip osteoarthritis in African Americans and
Caucasians: the Johnston County Osteoarthritis Project. J Rheumatol 2009; 36: 809-815.
37. http://issuu.com/bladerboek/docs/lroi-rapportage?e=3144185/1050596.
38. Dudda M, Kim YJ, Zhang Y, Nevitt MC, Xu L, Niu J, et al. Morphologic differences between the
hips of Chinese women and white women: could they account for the ethnic difference in the
prevalence of hip osteoarthritis? Arthritis Rheum 2011; 63: 2992-2999.
39. Nevitt MC, Xu L, Zhang Y, Lui LY, Yu W, Lane NE, et al. Very low prevalence of hip osteoarthritis
among Chinese elderly in Beijing, China, compared with whites in the United States: the Beijing
osteoarthritis study. Arthritis Rheum 2002; 46: 1773-1779.
40. Reynard LN, Loughlin J. Insights from human genetic studies into the pathways involved in
osteoarthritis. Nat Rev Rheumatol 2013; 9: 573-583.
41. Evangelou E, Kerkhof HJ, Styrkarsdottir U, Ntzani EE, Bos SD, Esko T, et al. A meta-analysis of
genome-wide association studies identifies novel variants associated with osteoarthritis of the
hip. Ann Rheum Dis 2013.
42. Reijman M, Pols HA, Bergink AP, Hazes JM, Belo JN, Lievense AM, et al. Body mass index
associated with onset and progression of osteoarthritis of the knee but not of the hip: the
Rotterdam Study. Ann Rheum Dis 2007; 66: 158-162.
43. Lievense AM, Bierma-Zeinstra SM, Verhagen AP, van Baar ME, Verhaar JA, Koes BW. Influence
of obesity on the development of osteoarthritis of the hip: a systematic review. Rheumatology
(Oxford) 2002; 41: 1155-1162.
44. Franklin J, Ingvarsson T, Englund M, Lohmander LS. Sex differences in the association between
body mass index and total hip or knee joint replacement resulting from osteoarthritis. Ann
Rheum Dis 2009; 68: 536-540.
45. Bierma-Zeinstra SM, Koes BW. Risk factors and prognostic factors of hip and knee osteoarthritis.
Nat Clin Pract Rheumatol 2007; 3: 78-85.
46. Lievense A, Bierma-Zeinstra S, Verhagen A, Verhaar J, Koes B. Influence of work on the
development of osteoarthritis of the hip: a systematic review. J Rheumatol 2001; 28: 2520-
2528.
242
References
47. Lievense AM, Bierma-Zeinstra SM, Verhagen AP, Bernsen RM, Verhaar JA, Koes BW. Influence of
sporting activities on the development of osteoarthritis of the hip: a systematic review. Arthritis
Rheum 2003; 49: 228-236.
48. Tveit M, Rosengren BE, Nilsson JA, Karlsson MK. Former male elite athletes have a higher
prevalence of osteoarthritis and arthroplasty in the hip and knee than expected. Am J Sports
Med 2012; 40: 527-533.
49. Smith-Petersen MN. Treatment of malum coxae senilis, old slipped upper femoral epiphysis,
intrapelvic protrusion of the acetabulum, and coxa plana by means of acetabuloplasty. J Bone
Joint Surg Am 1936; 18: 869-880.
50. Solomon L. Patterns of osteoarthritis of the hip. J Bone Joint Surg Br 1976; 58: 176-183.
51. Stulberg SD, Cooperman DR, Wallensten R. The natural history of Legg-Calve-Perthes disease. J
Bone Joint Surg Am 1981; 63: 1095-1108.
52. Okano K, Enomoto H, Osaki M, Shindo H. Outcome of rotational acetabular osteotomy for early
hip osteoarthritis secondary to dysplasia related to femoral head shape: 49 hips followed for 10-
17 years. Acta Orthop 2008; 79: 12-17.
53. Reijman M, Hazes JM, Pols HA, Koes BW, Bierma-Zeinstra SM. Acetabular dysplasia predicts
incident osteoarthritis of the hip: the Rotterdam study. Arthritis Rheum 2005; 52: 787-793.
54. Stulberg SDC, L.D. Harris, W.H. Ramsey, P.L. MacEwan, G.D. Unrecognized childhood hip
disease: a major cause of idiopathic osteoarthritis of the hip. The hip Proceedings of the third
open scientific meetings of the hip society 1975.
55. Ganz R, Parvizi J, Beck M, Leunig M, Notzli H, Siebenrock KA. Femoroacetabular impingement: a
cause for osteoarthritis of the hip. Clin Orthop Relat Res 2003: 112-120.
56. Ganz R, Gill TJ, Gautier E, Ganz K, Krugel N, Berlemann U. Surgical dislocation of the adult hip
a technique with full access to the femoral head and acetabulum without the risk of avascular
necrosis. J Bone Joint Surg Br 2001; 83: 1119-1124.
57. Beck M, Kalhor M, Leunig M, Ganz R. Hip morphology influences the pattern of damage to the
acetabular cartilage: femoroacetabular impingement as a cause of early osteoarthritis of the hip.
J Bone Joint Surg Br 2005; 87: 1012-1018.
58. http://www.youtube.com/watch?v=qkVxHnCyuxc.
59. Anderson LA, Peters CL, Park BB, Stoddard GJ, Erickson JA, Crim JR. Acetabular cartilage
delamination in femoroacetabular impingement. Risk factors and magnetic resonance imaging
diagnosis. J Bone Joint Surg Am 2009; 91: 305-313.
60. Gosvig KK, Jacobsen S, Sonne-Holm S, Palm H, Troelsen A. Prevalence of malformations of the
hip joint and their relationship to sex, groin pain, and risk of osteoarthritis: a population-based
survey. J Bone Joint Surg Am 2010; 92: 1162-1169.
61. Giori NJ, Trousdale RT. Acetabular retroversion is associated with osteoarthritis of the hip. Clin
Orthop Relat Res 2003: 263-269.
62. Seessle J, Gohdes A, Gotthardt DN, Pfeiffenberger J, Eckert N, Stremmel W, et al. Alterations of
lipid metabolism in Wilson disease. Lipids Health Dis 2011; 10: 83.
63. Nicholls AS, Kiran A, Pollard TC, Hart DJ, Arden CP, Spector T, et al. The association between hip
morphology parameters and nineteen-year risk of end-stage osteoarthritis of the hip: a nested
case-control study. Arthritis Rheum 2011; 63: 3392-3400.
64. Bardakos NV, Villar RN. Predictors of progression of osteoarthritis in femoroacetabular
impingement: a radiological study with a minimum of ten years follow-up. J Bone Joint Surg Br
2009; 91: 162-169.
65. Clohisy JC, Dobson MA, Robison JF, Warth LC, Zheng J, Liu SS, et al. Radiographic structural
abnormalities associated with premature, natural hip-joint failure. J Bone Joint Surg Am 2011;
93 Suppl 2: 3-9.
66. Chung CY, Park MS, Lee KM, Lee SH, Kim TK, Kim KW, et al. Hip osteoarthritis and risk factors in
elderly Korean population. Osteoarthritis Cartilage 2010; 18: 312-316.
67. Ito K, Minka MA, 2nd, Leunig M, Werlen S, Ganz R. Femoroacetabular impingement and the cam-
effect. A MRI-based quantitative anatomical study of the femoral head-neck offset. J Bone Joint
Surg Br 2001; 83: 171-176.
243
References
68. Woolf AD, Pfleger B. Burden of major musculoskeletal conditions. Bull World Health Organ 2003;
81: 646-656.
69. Hiligsmann M, Cooper C, Arden N, Boers M, Branco JC, Luisa Brandi M, et al. Health economics
in the field of osteoarthritis: an expert’s consensus paper from the European Society for Clinical
and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO). Semin Arthritis Rheum 2013;
43: 303-313.
70. Carr AJ, Robertsson O, Graves S, Price AJ, Arden NK, Judge A, et al. Knee replacement. Lancet
2012; 379: 1331-1340.
71. Pivec R, Johnson AJ, Mears SC, Mont MA. Hip arthroplasty. Lancet 2012; 380: 1768-1777.
72. Bijlsma JW, Berenbaum F, Lafeber FP. Osteoarthritis: an update with relevance for clinical
practice. Lancet 2011; 377: 2115-2126.
73. Vincent TL. Targeting mechanotransduction pathways in osteoarthritis: a focus on the pericellular
matrix. Curr Opin Pharmacol 2013; 13: 449-454.
74. Agricola R, Heijboer MP, Roze RH, Reijman M, Bierma-Zeinstra SM, Verhaar JA, et al. Pincer
deformity does not lead to osteoarthritis of the hip whereas acetabular dysplasia does: acetabular
coverage and development of osteoarthritis in a nationwide prospective cohort study (CHECK).
Osteoarthritis Cartilage 2013; 21: 1514-1521.
75. Agricola R, Waarsing JH, Arden NK, Carr AJ, Bierma-Zeinstra SM, Thomas GE, et al. Cam
impingement of the hip--a risk factor for hip osteoarthritis. Nat Rev Rheumatol 2013; 9: 630-
634.
76. Neogi T, Bowes MA, Niu J, De Souza KM, Vincent GR, Goggins J, et al. Magnetic resonance
imaging-based three-dimensional bone shape of the knee predicts onset of knee osteoarthritis:
data from the osteoarthritis initiative. Arthritis Rheum 2013; 65: 2048-2058.
77. Sharma L, Chmiel JS, Almagor O, Felson D, Guermazi A, Roemer F, et al. The role of varus and
valgus alignment in the initial development of knee cartilage damage by MRI: the MOST study.
Ann Rheum Dis 2013; 72: 235-240.
78. Felson DT, Niu J, Gross KD, Englund M, Sharma L, Cooke TD, et al. Valgus malalignment is a risk
factor for lateral knee osteoarthritis incidence and progression: findings from the Multicenter
Osteoarthritis Study and the Osteoarthritis Initiative. Arthritis Rheum 2013; 65: 355-362.
79. Harvey WF, Yang M, Cooke TD, Segal NA, Lane N, Lewis CE, et al. Association of leg-length
inequality with knee osteoarthritis: a cohort study. Ann Intern Med 2010; 152: 287-295.
80. Wang Y, Wluka AE, Berry PA, Siew T, Teichtahl AJ, Urquhart DM, et al. Increase in vastus medialis
cross-sectional area is associated with reduced pain, cartilage loss, and joint replacement risk in
knee osteoarthritis. Arthritis Rheum 2012; 64: 3917-3925.
81. Siebenrock KA, Kaschka I, Frauchiger L, Werlen S, Schwab JM. Prevalence of cam-type deformity
and hip pain in elite ice hockey players before and after the end of growth. Am J Sports Med
2013; 41: 2308-2313.
82. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi H, Deyo RA, et al. Estimates of the
prevalence of arthritis and other rheumatic conditions in the United States. Part II. Arthritis
Rheum 2008; 58: 26-35.
83. Nevitt MC, Zhang Y, Javaid MK, Neogi T, Curtis JR, Niu J, et al. High systemic bone mineral
density increases the risk of incident knee OA and joint space narrowing, but not radiographic
progression of existing knee OA: the MOST study. Ann Rheum Dis 2010; 69: 163-168.
84. Muthuri SG, McWilliams DF, Doherty M, Zhang W. History of knee injuries and knee osteoarthritis:
a meta-analysis of observational studies. Osteoarthritis Cartilage 2011; 19: 1286-1293.
85. Conde J, Scotece M, Gomez R, Lopez V, Gomez-Reino JJ, Gualillo O. Adipokines and osteoarthritis:
novel molecules involved in the pathogenesis and progression of disease. Arthritis 2011; 2011:
203901.
86. Blagojevic M, Jinks C, Jeffery A, Jordan KP. Risk factors for onset of osteoarthritis of the knee in
older adults: a systematic review and meta-analysis. Osteoarthritis Cartilage 2010; 18: 24-33.
87. Zhang Y, Niu J, Felson DT, Choi HK, Nevitt M, Neogi T. Methodologic challenges in studying risk
factors for progression of knee osteoarthritis. Arthritis Care Res (Hoboken) 2010; 62: 1527-
1532.
244
References
88. Mork PJ, Holtermann A, Nilsen TI. Effect of body mass index and physical exercise on risk of
knee and hip osteoarthritis: longitudinal data from the Norwegian HUNT Study. J Epidemiol
Community Health 2012; 66: 678-683.
89. Zeggini E, Panoutsopoulou K, Southam L, Rayner NW, Day-Williams AG, Lopes MC, et al.
Identification of new susceptibility loci for osteoarthritis (arcOGEN): a genome-wide association
study. Lancet 2012; 380: 815-823.
90. Hochberg MC, Yerges-Armstrong L, Yau M, Mitchell BD. Genetic epidemiology of osteoarthritis:
recent developments and future directions. Curr Opin Rheumatol 2013; 25: 192-197.
91. Buckwalter JA, Mankin HJ, Grodzinsky AJ. Articular cartilage and osteoarthritis. Instr Course Lect
2005; 54: 465-480.
92. Burleigh A, Chanalaris A, Gardiner MD, Driscoll C, Boruc O, Saklatvala J, et al. Joint immobilization
prevents murine osteoarthritis and reveals the highly mechanosensitive nature of protease
expression in vivo. Arthritis Rheum 2012; 64: 2278-2288.
93. Kim HA, Cho ML, Choi HY, Yoon CS, Jhun JY, Oh HJ, et al. The catabolic pathway mediated by
Toll-like receptors in human osteoarthritic chondrocytes. Arthritis Rheum 2006; 54: 2152-2163.
94. Liu-Bryan R, Terkeltaub R. Chondrocyte innate immune myeloid differentiation factor
88-dependent signaling drives procatabolic effects of the endogenous Toll-like receptor 2/Toll-like
receptor 4 ligands low molecular weight hyaluronan and high mobility group box chromosomal
protein 1 in mice. Arthritis Rheum 2010; 62: 2004-2012.
95. Liu-Bryan R, Pritzker K, Firestein GS, Terkeltaub R. TLR2 signaling in chondrocytes drives
calcium pyrophosphate dihydrate and monosodium urate crystal-induced nitric oxide generation.
J Immunol 2005; 174: 5016-5023.
96. Wang Q, Rozelle AL, Lepus CM, Scanzello CR, Song JJ, Larsen DM, et al. Identification of a central
role for complement in osteoarthritis. Nat Med 2011; 17: 1674-1679.
97. Happonen KE, Saxne T, Aspberg A, Morgelin M, Heinegard D, Blom AM. Regulation of complement
by cartilage oligomeric matrix protein allows for a novel molecular diagnostic principle in
rheumatoid arthritis. Arthritis Rheum 2010; 62: 3574-3583.
98. Sjoberg AP, Manderson GA, Morgelin M, Day AJ, Heinegard D, Blom AM. Short leucine-rich
glycoproteins of the extracellular matrix display diverse patterns of complement interaction and
activation. Mol Immunol 2009; 46: 830-839.
99. Loeser RF, Yammani RR, Carlson CS, Chen H, Cole A, Im HJ, et al. Articular chondrocytes express
the receptor for advanced glycation end products: Potential role in osteoarthritis. Arthritis Rheum
2005; 52: 2376-2385.
100. Rasheed Z, Akhtar N, Haqqi TM. Advanced glycation end products induce the expression of
interleukin-6 and interleukin-8 by receptor for advanced glycation end product-mediated
activation of mitogen-activated protein kinases and nuclear factor-kappaB in human osteoarthritis
chondrocytes. Rheumatology (Oxford) 2011; 50: 838-851.
101. Weinans H, Siebelt M, Agricola R, Botter SM, Piscaer TM, Waarsing JH. Pathophysiology of peri-
articular bone changes in osteoarthritis. Bone 2012; 51: 190-196.
102. Cox LG, van Donkelaar CC, van Rietbergen B, Emans PJ, Ito K. Decreased bone tissue
mineralization can partly explain subchondral sclerosis observed in osteoarthritis. Bone 2012;
50: 1152-1161.
103. Taljanovic MS, Graham AR, Benjamin JB, Gmitro AF, Krupinski EA, Schwartz SA, et al. Bone
marrow edema pattern in advanced hip osteoarthritis: quantitative assessment with magnetic
resonance imaging and correlation with clinical examination, radiographic findings, and
histopathology. Skeletal Radiol 2008; 37: 423-431.
104. Buckland-Wright C. Subchondral bone changes in hand and knee osteoarthritis detected by
radiography. Osteoarthritis Cartilage 2004; 12 Suppl A: S10-19.
105. Dieppe P, Cushnaghan J, Young P, Kirwan J. Prediction of the progression of joint space narrowing
in osteoarthritis of the knee by bone scintigraphy. Ann Rheum Dis 1993; 52: 557-563.
106. Guermazi A, Roemer FW, Felson DT, Brandt KD. Motion for debate: osteoarthritis clinical trials
have not identified efficacious therapies because traditional imaging outcome measures are
inadequate. Arthritis Rheum 2013; 65: 2748-2758.
245
References
107. Sanchez C, Pesesse L, Gabay O, Delcour JP, Msika P, Baudouin C, et al. Regulation of subchondral
bone osteoblast metabolism by cyclic compression. Arthritis Rheum 2012; 64: 1193-1203.
108. Scanzello CR, Goldring SR. The role of synovitis in osteoarthritis pathogenesis. Bone 2012; 51:
249-257.
109. Bastow ER, Byers S, Golub SB, Clarkin CE, Pitsillides AA, Fosang AJ. Hyaluronan synthesis and
degradation in cartilage and bone. Cell Mol Life Sci 2008; 65: 395-413.
110. Marcelino J, Carpten JD, Suwairi WM, Gutierrez OM, Schwartz S, Robbins C, et al. CACP, encoding
a secreted proteoglycan, is mutated in camptodactyly-arthropathy-coxa vara-pericarditis
syndrome. Nat Genet 1999; 23: 319-322.
111. Ludwig TE, McAllister JR, Lun V, Wiley JP, Schmidt TA. Diminished cartilage-lubricating ability of
human osteoarthritic synovial fluid deficient in proteoglycan 4: Restoration through proteoglycan
4 supplementation. Arthritis Rheum 2012; 64: 3963-3971.
112. Baker K, Grainger A, Niu J, Clancy M, Guermazi A, Crema M, et al. Relation of synovitis to knee
pain using contrast-enhanced MRIs. Ann Rheum Dis 2010; 69: 1779-1783.
113. Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD, et al. Presence of MRI-detected
joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at
30-month follow-up: the MOST study. Ann Rheum Dis 2011; 70: 1804-1809.
114. Jin X, Beguerie JR, Zhang W, Blizzard L, Otahal P, Jones G, et al. Circulating C reactive protein in
osteoarthritis: a systematic review and meta-analysis. Ann Rheum Dis 2013.
115. Haugen IK, Ramachandran VS, Misra D, Neogi T, Niu J, Yang T, et al. Hand osteoarthritis in
relation to mortality and incidence of cardiovascular disease: data from the Framingham Heart
Study. Ann Rheum Dis 2013.
116. Yusuf E, Nelissen RG, Ioan-Facsinay A, Stojanovic-Susulic V, DeGroot J, van Osch G, et al.
Association between weight or body mass index and hand osteoarthritis: a systematic review.
Ann Rheum Dis 2010; 69: 761-765.
117. Yusuf E. Metabolic factors in osteoarthritis: obese people do not walk on their hands. Arthritis
Res Ther 2012; 14: 123.
118. Malfait AM, Schnitzer TJ. Towards a mechanism-based approach to pain management in
osteoarthritis. Nat Rev Rheumatol 2013; 9: 654-664.
119. Dieppe PA, Lohmander LS. Pathogenesis and management of pain in osteoarthritis. Lancet 2005;
365: 965-973.
120. Palmer AJ, Brown CP, McNally EG, Price AJ, Tracey I, Jezzard P, et al. Non-invasive imaging of
cartilage in early osteoarthritis. Bone Joint J 2013; 95-B: 738-746.
121. Altman RD, Gold GE. Atlas of individual radiographic features in osteoarthritis, revised.
Osteoarthritis Cartilage 2007; 15 Suppl A: A1-56.
122. Gossec L, Jordan JM, Lam MA, Fang F, Renner JB, Davis A, et al. Comparative evaluation of three
semi-quantitative radiographic grading techniques for hip osteoarthritis in terms of validity and
reproducibility in 1404 radiographs: report of the OARSI-OMERACT Task Force. Osteoarthritis
Cartilage 2009; 17: 182-187.
123. Gossec L, Jordan JM, Mazzuca SA, Lam MA, Suarez-Almazor ME, Renner JB, et al. Comparative
evaluation of three semi-quantitative radiographic grading techniques for knee osteoarthritis in
terms of validity and reproducibility in 1759 X-rays: report of the OARSI-OMERACT task force.
Osteoarthritis Cartilage 2008; 16: 742-748.
124. Guermazi A, Hunter DJ, Roemer FW. Plain radiography and magnetic resonance imaging
diagnostics in osteoarthritis: validated staging and scoring. J Bone Joint Surg Am 2009; 91 Suppl
1: 54-62.
125. Javaid MK, Lynch JA, Tolstykh I, Guermazi A, Roemer F, Aliabadi P, et al. Pre-radiographic MRI
findings are associated with onset of knee symptoms: the most study. Osteoarthritis Cartilage
2010; 18: 323-328.
126. Guermazi A, Niu J, Hayashi D, Roemer FW, Englund M, Neogi T, et al. Prevalence of abnormalities
in knees detected by MRI in adults without knee osteoarthritis: population based observational
study (Framingham Osteoarthritis Study). BMJ 2012; 345: e5339.
127. Hunter DJ, Zhang YQ, Tu X, Lavalley M, Niu JB, Amin S, et al. Change in joint space width:
246
References
hyaline articular cartilage loss or alteration in meniscus? Arthritis Rheum 2006; 54: 2488-2495.
128. Kinds MB, Vincken KL, Hoppinga TN, Bleys RL, Viergever MA, Marijnissen AC, et al. Influence
of variation in semiflexed knee positioning during image acquisition on separate quantitative
radiographic parameters of osteoarthritis, measured by Knee Images Digital Analysis.
Osteoarthritis Cartilage 2012; 20: 997-1003.
129. Reichmann WM, Maillefert JF, Hunter DJ, Katz JN, Conaghan PG, Losina E. Responsiveness to
change and reliability of measurement of radiographic joint space width in osteoarthritis of the
knee: a systematic review. Osteoarthritis Cartilage 2011; 19: 550-556.
130. Wirth W, Buck R, Nevitt M, Le Graverand MP, Benichou O, Dreher D, et al. MRI-based extended
ordered values more efficiently differentiate cartilage loss in knees with and without joint space
narrowing than region-specific approaches using MRI or radiography--data from the OA initiative.
Osteoarthritis Cartilage 2011; 19: 689-699.
131. Conaghan PG, Hunter DJ, Maillefert JF, Reichmann WM, Losina E. Summary and recommendations
of the OARSI FDA osteoarthritis Assessment of Structural Change Working Group. Osteoarthritis
Cartilage 2011; 19: 606-610.
132. Eckstein F, Mc Culloch CE, Lynch JA, Nevitt M, Kwoh CK, Maschek S, et al. How do short-term
rates of femorotibial cartilage change compare to long-term changes? Four year follow-up data
from the osteoarthritis initiative. Osteoarthritis Cartilage 2012; 20: 1250-1257.
133. Eckstein F, Hudelmaier M, Cahue S, Marshall M, Sharma L. Medial-to-lateral ratio of tibiofemoral
subchondral bone area is adapted to alignment and mechanical load. Calcif Tissue Int 2009; 84:
186-194.
134. Cotofana S, Buck R, Wirth W, Roemer F, Duryea J, Nevitt M, et al. Cartilage thickening in early
radiographic knee osteoarthritis: a within-person, between-knee comparison. Arthritis Care Res
(Hoboken) 2012; 64: 1681-1690.
135. Zilkens C, Miese F, Herten M, Kurzidem S, Jager M, Konig D, et al. Validity of gradient-echo three-
dimensional delayed gadolinium-enhanced magnetic resonance imaging of hip joint cartilage: a
histologically controlled study. Eur J Radiol 2013; 82: e81-86.
136. Van Ginckel A, Baelde N, Almqvist KF, Roosen P, McNair P, Witvrouw E. Functional adaptation
of knee cartilage in asymptomatic female novice runners compared to sedentary controls.
A longitudinal analysis using delayed Gadolinium Enhanced Magnetic Resonance Imaging of
Cartilage (dGEMRIC). Osteoarthritis Cartilage 2010; 18: 1564-1569.
137. Zilkens C, Miese F, Kim YJ, Jager M, Mamisch TC, Hosalkar H, et al. Direct comparison of intra-
articular versus intravenous delayed gadolinium-enhanced MRI of hip joint cartilage. J Magn
Reson Imaging 2014; 39: 94-102.
138. Baum T, Joseph GB, Karampinos DC, Jungmann PM, Link TM, Bauer JS. Cartilage and meniscal
T2 relaxation time as non-invasive biomarker for knee osteoarthritis and cartilage repair
procedures. Osteoarthritis Cartilage 2013; 21: 1474-1484.
139. Nishioka H, Hirose J, Nakamura E, Oniki Y, Takada K, Yamashita Y, et al. T1rho and T2 mapping
reveal the in vivo extracellular matrix of articular cartilage. J Magn Reson Imaging 2012; 35:
147-155.
140. Kijowski R, Blankenbaker DG, Munoz Del Rio A, Baer GS, Graf BK. Evaluation of the articular
cartilage of the knee joint: value of adding a T2 mapping sequence to a routine MR imaging
protocol. Radiology 2013; 267: 503-513.
141. Prasad AP, Nardo L, Schooler J, Joseph GB, Link TM. T(1)rho and T(2) relaxation times predict
progression of knee osteoarthritis. Osteoarthritis Cartilage 2013; 21: 69-76.
142. Takayama Y, Hatakenaka M, Tsushima H, Okazaki K, Yoshiura T, Yonezawa M, et al. T1rho is
superior to T2 mapping for the evaluation of articular cartilage denaturalization with osteoarthritis:
radiological-pathological correlation after total knee arthroplasty. Eur J Radiol 2013; 82: e192-
198.
143. Bittersohl B, Miese FR, Hosalkar HS, Mamisch TC, Antoch G, Krauspe R, et al. T2* mapping of
acetabular and femoral hip joint cartilage at 3 T: a prospective controlled study. Invest Radiol
2012; 47: 392-397.
144. Raya JG, Horng A, Dietrich O, Krasnokutsky S, Beltran LS, Storey P, et al. Articular cartilage: in
vivo diffusion-tensor imaging. Radiology 2012; 262: 550-559.
247
References
145. Roemer FW, Felson DT, Wang K, Crema MD, Neogi T, Zhang Y, et al. Co-localisation of non-
cartilaginous articular pathology increases risk of cartilage loss in the tibiofemoral joint--the
MOST study. Ann Rheum Dis 2013; 72: 942-948.
146. Guermazi A, Roemer FW, Haugen IK, Crema MD, Hayashi D. MRI-based semiquantitative scoring
of joint pathology in osteoarthritis. Nat Rev Rheumatol 2013; 9: 236-251.
147. Keen HI, Conaghan PG. Ultrasonography in osteoarthritis. Radiol Clin North Am 2009; 47: 581-
594.
148. Nicolaou S, Liang T, Murphy DT, Korzan JR, Ouellette H, Munk P. Dual-energy CT: a promising
new technique for assessment of the musculoskeletal system. AJR Am J Roentgenol 2012; 199:
S78-86.
149. van Spil WE, DeGroot J, Lems WF, Oostveen JC, Lafeber FP. Serum and urinary biochemical
markers for knee and hip-osteoarthritis: a systematic review applying the consensus BIPED
criteria. Osteoarthritis Cartilage 2010; 18: 605-612.
150. Garnero P, Piperno M, Gineyts E, Christgau S, Delmas PD, Vignon E. Cross sectional evaluation
of biochemical markers of bone, cartilage, and synovial tissue metabolism in patients with knee
osteoarthritis: relations with disease activity and joint damage. Ann Rheum Dis 2001; 60: 619-
626.
151. Hoch JM, Mattacola CG, Medina McKeon JM, Howard JS, Lattermann C. Serum cartilage oligomeric
matrix protein (sCOMP) is elevated in patients with knee osteoarthritis: a systematic review and
meta-analysis. Osteoarthritis Cartilage 2011; 19: 1396-1404.
152. Dam EB, Loog M, Christiansen C, Byrjalsen I, Folkesson J, Nielsen M, et al. Identification of
progressors in osteoarthritis by combining biochemical and MRI-based markers. Arthritis Res
Ther 2009; 11: R115.
153. Catterall JB, Hsueh MF, Stabler TV, McCudden CR, Bolognesi M, Zura R, et al. Protein modification
by deamidation indicates variations in joint extracellular matrix turnover. J Biol Chem 2012; 287:
4640-4651.
154. Reijman M, Hazes JM, Bierma-Zeinstra SM, Koes BW, Christgau S, Christiansen C, et al. A new
marker for osteoarthritis: cross-sectional and longitudinal approach. Arthritis Rheum 2004; 50:
2471-2478.
155. Blumenfeld O, Williams FM, Hart DJ, Spector TD, Arden N, Livshits G. Association between
cartilage and bone biomarkers and incidence of radiographic knee osteoarthritis (RKOA) in UK
females: a prospective study. Osteoarthritis Cartilage 2013; 21: 923-929.
156. Golightly YM, Marshall SW, Kraus VB, Renner JB, Villaveces A, Casteel C, et al. Biomarkers of
incident radiographic knee osteoarthritis: do they vary by chronic knee symptoms? Arthritis
Rheum 2011; 63: 2276-2283.
157. van Spil WE, Drossaers-Bakker KW, Lafeber FP. Associations of CTX-II with biochemical markers
of bone turnover raise questions on its tissue origin: data from CHECK, a cohort study of early
osteoarthritis. Ann Rheum Dis 2013; 72: 29-36.
158. Rousseau J, Garnero P. Biological markers in osteoarthritis. Bone 2012; 51: 265-277.
159. Onnerfjord P, Khabut A, Reinholt FP, Svensson O, Heinegard D. Quantitative proteomic analysis
of eight cartilaginous tissues reveals characteristic differences as well as similarities between
subgroups. J Biol Chem 2012; 287: 18913-18924.
160. Felson DT, Zhang Y, Anthony JM, Naimark A, Anderson JJ. Weight loss reduces the risk for
symptomatic knee osteoarthritis in women. The Framingham Study. Ann Intern Med 1992; 116:
535-539.
161. Gudbergsen H, Boesen M, Lohmander LS, Christensen R, Henriksen M, Bartels EM, et al. Weight
loss is effective for symptomatic relief in obese subjects with knee osteoarthritis independently
of joint damage severity assessed by high-field MRI and radiography. Osteoarthritis Cartilage
2012; 20: 495-502.
162. Anandacoomarasamy A, Leibman S, Smith G, Caterson I, Giuffre B, Fransen M, et al. Weight
loss in obese people has structure-modifying effects on medial but not on lateral knee articular
cartilage. Ann Rheum Dis 2012; 71: 26-32.
248
References
163. Richette P, Poitou C, Garnero P, Vicaut E, Bouillot JL, Lacorte JM, et al. Benefits of massive weight
loss on symptoms, systemic inflammation and cartilage turnover in obese patients with knee
osteoarthritis. Ann Rheum Dis 2011; 70: 139-144.
164. Uthman OA, van der Windt DA, Jordan JL, Dziedzic KS, Healey EL, Peat GM, et al. Exercise for
lower limb osteoarthritis: systematic review incorporating trial sequential analysis and network
meta-analysis. BMJ 2013; 347: f5555.
165. Hartig-Andreasen C, Troelsen A, Thillemann TM, Soballe K. What factors predict failure 4 to 12
years after periacetabular osteotomy? Clin Orthop Relat Res 2012; 470: 2978-2987.
166. Clohisy JC, St John LC, Schutz AL. Surgical treatment of femoroacetabular impingement: a
systematic review of the literature. Clin Orthop Relat Res 2010; 468: 555-564.
167. Wiegant K, van Roermund PM, Intema F, Cotofana S, Eckstein F, Mastbergen SC, et al. Sustained
clinical and structural benefit after joint distraction in the treatment of severe knee osteoarthritis.
Osteoarthritis Cartilage 2013; 21: 1660-1667.
168. Spahn G, Hofmann GO, von Engelhardt LV, Li M, Neubauer H, Klinger HM. The impact of a
high tibial valgus osteotomy and unicondylar medial arthroplasty on the treatment for knee
osteoarthritis: a meta-analysis. Knee Surg Sports Traumatol Arthrosc 2013; 21: 96-112.
169. Palmer AJ, Thomas GE, Pollard TC, Rombach I, Taylor A, Arden N, et al. The feasibility of
performing a randomised controlled trial for femoroacetabular impingement surgery. Bone Joint
Res 2013; 2: 33-40.
170. Mollon B, Kandel R, Chahal J, Theodoropoulos J. The clinical status of cartilage tissue regeneration
in humans. Osteoarthritis Cartilage 2013; 21: 1824-1833.
171. Harris JD, Siston RA, Pan X, Flanigan DC. Autologous chondrocyte implantation: a systematic
review. J Bone Joint Surg Am 2010; 92: 2220-2233.
172. National Institute for Health and Care Excellence. Osteoarthritis: care and management in
adults.
173. Hochberg MC, Altman RD, April KT, Benkhalti M, Guyatt G, McGowan J, et al. American College
of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic
therapies in osteoarthritis of the hand, hip, and knee. Arthritis Care Res (Hoboken) 2012; 64:
465-474.
174. Henrotin Y, Lambert C. Chondroitin and glucosamine in the management of osteoarthritis: an
update. Curr Rheumatol Rep 2013; 15: 361.
175. Block JA, Oegema TR, Sandy JD, Plaas A. The effects of oral glucosamine on joint health: is a
change in research approach needed? Osteoarthritis Cartilage 2010; 18: 5-11.
176. McAlindon TE, Bannuru RR, Sullivan MC, Arden NK, Berenbaum F, Bierma-Zeinstra SM, et
al. OARSI guidelines for the non-surgical management of knee osteoarthritis. Osteoarthritis
Cartilage 2014; 22: 363-388.
177. Jevsevar DS. Treatment of osteoarthritis of the knee: evidence-based guideline, 2nd edition. J
Am Acad Orthop Surg 2013; 21: 571-576.
178. Rutjes AW, Juni P, da Costa BR, Trelle S, Nuesch E, Reichenbach S. Viscosupplementation for
osteoarthritis of the knee: a systematic review and meta-analysis. Ann Intern Med 2012; 157:
180-191.
179. Schmidt TA, Gastelum NS, Nguyen QT, Schumacher BL, Sah RL. Boundary lubrication of articular
cartilage: role of synovial fluid constituents. Arthritis Rheum 2007; 56: 882-891.
180. Ruan MZ, Erez A, Guse K, Dawson B, Bertin T, Chen Y, et al. Proteoglycan 4 expression protects
against the development of osteoarthritis. Sci Transl Med 2013; 5: 176ra134.
181. da Costa BR, Nuesch E, Reichenbach S, Juni P, Rutjes AW. Doxycycline for osteoarthritis of the
knee or hip. Cochrane Database Syst Rev 2012; 11: CD007323.
182. Krzeski P, Buckland-Wright C, Balint G, Cline GA, Stoner K, Lyon R, et al. Development of
musculoskeletal toxicity without clear benefit after administration of PG-116800, a matrix
metalloproteinase inhibitor, to patients with knee osteoarthritis: a randomized, 12-month,
double-blind, placebo-controlled study. Arthritis Res Ther 2007; 9: R109.
249
References
183. Hellio le Graverand MP, Clemmer RS, Redifer P, Brunell RM, Hayes CW, Brandt KD, et al. A 2-year
randomised, double-blind, placebo-controlled, multicentre study of oral selective iNOS inhibitor,
cindunistat (SD-6010), in patients with symptomatic osteoarthritis of the knee. Ann Rheum Dis
2013; 72: 187-195.
184. Spector TD, Conaghan PG, Buckland-Wright JC, Garnero P, Cline GA, Beary JF, et al. Effect
of risedronate on joint structure and symptoms of knee osteoarthritis: results of the BRISK
randomized, controlled trial [ISRCTN01928173]. Arthritis Res Ther 2005; 7: R625-633.
185. Bingham CO, 3rd, Buckland-Wright JC, Garnero P, Cohen SB, Dougados M, Adami S, et al.
Risedronate decreases biochemical markers of cartilage degradation but does not decrease
symptoms or slow radiographic progression in patients with medial compartment osteoarthritis
of the knee: results of the two-year multinational knee osteoarthritis structural arthritis study.
Arthritis Rheum 2006; 54: 3494-3507.
186. Laslett LL, Dore DA, Quinn SJ, Boon P, Ryan E, Winzenberg TM, et al. Zoledronic acid reduces
knee pain and bone marrow lesions over 1 year: a randomised controlled trial. Ann Rheum Dis
2012; 71: 1322-1328.
187. Henrotin Y, Labasse A, Zheng SX, Galais P, Tsouderos Y, Crielaard JM, et al. Strontium ranelate
increases cartilage matrix formation. J Bone Miner Res 2001; 16: 299-308.
188. Reginster JY, Badurski J, Bellamy N, Bensen W, Chapurlat R, Chevalier X, et al. Efficacy and
safety of strontium ranelate in the treatment of knee osteoarthritis: results of a double-blind,
randomised placebo-controlled trial. Ann Rheum Dis 2013; 72: 179-186.
189. Raynauld JP, Buckland-Wright C, Ward R, Choquette D, Haraoui B, Martel-Pelletier J, et al.
Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a
randomized, double-blind, placebo-controlled trial. Arthritis Rheum 2003; 48: 370-377.
190. Wenham CY, Grainger AJ, Hensor EM, Caperon AR, Ash ZR, Conaghan PG. Methotrexate for pain
relief in knee osteoarthritis: an open-label study. Rheumatology (Oxford) 2013; 52: 888-892.
191. Chevalier X, Eymard F, Richette P. Biologic agents in osteoarthritis: hopes and disappointments.
Nat Rev Rheumatol 2013; 9: 400-410.
192. Chevalier X, Goupille P, Beaulieu AD, Burch FX, Bensen WG, Conrozier T, et al. Intraarticular
injection of anakinra in osteoarthritis of the knee: a multicenter, randomized, double-blind,
placebo-controlled study. Arthritis Rheum 2009; 61: 344-352.
193. Cohen SB, Proudman S, Kivitz AJ, Burch FX, Donohue JP, Burstein D, et al. A randomized,
double-blind study of AMG 108 (a fully human monoclonal antibody to IL-1R1) in patients with
osteoarthritis of the knee. Arthritis Res Ther 2011; 13: R125.
194. Kapoor M, Martel-Pelletier J, Lajeunesse D, Pelletier JP, Fahmi H. Role of proinflammatory
cytokines in the pathophysiology of osteoarthritis. Nat Rev Rheumatol 2011; 7: 33-42.
195. Verbruggen G, Wittoek R, Vander Cruyssen B, Elewaut D. Tumour necrosis factor blockade
for the treatment of erosive osteoarthritis of the interphalangeal finger joints: a double blind,
randomised trial on structure modification. Ann Rheum Dis 2012; 71: 891-898.
196. Maksymowych WP, Russell AS, Chiu P, Yan A, Jones N, Clare T, et al. Targeting tumour necrosis
factor alleviates signs and symptoms of inflammatory osteoarthritis of the knee. Arthritis Res
Ther 2012; 14: R206.
197. Sandker MJ, Petit A, Redout EM, Siebelt M, Muller B, Bruin P, et al. In situ forming acyl-capped
PCLA-PEG-PCLA triblock copolymer based hydrogels. Biomaterials 2013; 34: 8002-8011.
198. Johnson K, Zhu S, Tremblay MS, Payette JN, Wang J, Bouchez LC, et al. A stem cell-based
approach to cartilage repair. Science 2012; 336: 717-721.
199. Mazzuca SA, Brandt KD, Chakr R, Lane KA. Varus malalignment negates the structure-modifying
benefits of doxycycline in obese women with knee osteoarthritis. Osteoarthritis Cartilage 2010;
18: 1008-1011.
200. Seidel MF, Wise BL, Lane NE. Nerve growth factor: an update on the science and therapy.
Osteoarthritis Cartilage 2013; 21: 1223-1228.
201. Conaghan PG. Osteoarthritis in 2012: Parallel evolution of OA phenotypes and therapies. Nat Rev
Rheumatol 2013; 9: 68-70.
250
References
202. Cushnaghan J, Dieppe P. Study of 500 patients with limb joint osteoarthritis. I. Analysis by age,
sex, and distribution of symptomatic joint sites. Ann Rheum Dis 1991; 50: 8-13.
203. Agricola R, Heijboer MP, Bierma-Zeinstra SM, Verhaar JA, Weinans H, Waarsing JH. Cam
impingement causes osteoarthritis of the hip: a nationwide prospective cohort study (CHECK).
Ann Rheum Dis 2013; 72: 918-923.
204. Baker-LePain JC, Lane NE. Relationship between joint shape and the development of osteoarthritis.
Curr Opin Rheumatol 2010; 22: 538-543.
205. Cootes TF, Taylor CJ, Cooper DH, Graham J. Active Shape Models - Their Training and Application.
Computer Vision and Image Understanding 1995; 61: 38-59.
206. Gregory JS, Waarsing JH, Day J, Pols HA, Reijman M, Weinans H, et al. Early identification of
radiographic osteoarthritis of the hip using an active shape model to quantify changes in bone
morphometric features: can hip shape tell us anything about the progression of osteoarthritis?
Arthritis Rheum 2007; 56: 3634-3643.
207. Kinds MB, Welsing PM, Vignon EP, Bijlsma JW, Viergever MA, Marijnissen AC, et al. A systematic
review of the association between radiographic and clinical osteoarthritis of hip and knee.
Osteoarthritis Cartilage 2011; 19: 768-778.
208. Doherty M, Courtney P, Doherty S, Jenkins W, Maciewicz RA, Muir K, et al. Nonspherical femoral
head shape (pistol grip deformity), neck shaft angle, and risk of hip osteoarthritis: a case-control
study. Arthritis Rheum 2008; 58: 3172-3182.
209. Waarsing JH, Rozendaal RM, Verhaar JA, Bierma-Zeinstra SM, Weinans H. A statistical model of
shape and density of the proximal femur in relation to radiological and clinical OA of the hip.
Osteoarthritis Cartilage 2010; 18: 787-794.
210. Wesseling J, Dekker J, van den Berg WB, Bierma-Zeinstra SM, Boers M, Cats HA, et al. CHECK
(Cohort Hip and Cohort Knee): similarities and differences with the Osteoarthritis Initiative. Ann
Rheum Dis 2009; 68: 1413-1419.
211. Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957; 16:
494-502.
212. Bierma-Zeinstra SM, Bohnen AM, Ramlal R, Ridderikhoff J, Verhaar JA, Prins A. Comparison
between two devices for measuring hip joint motions. Clin Rehabil 1998; 12: 497-505.
213. Barr RJ, Gregory JS, Reid DM, Aspden RM, Yoshida K, Hosie G, et al. Predicting OA progression
to total hip replacement: can we do better than risk factors alone using active shape modelling
as an imaging biomarker? Rheumatology (Oxford) 2012; 51: 562-570.
214. Verkleij SP, Hoekstra T, Rozendaal RM, Waarsing JH, Koes BW, Luijsterburg PA, et al. Defining
discriminative pain trajectories in hip osteoarthritis over a 2-year time period. Ann Rheum Dis
2012.
215. Reichenbach S, Juni P, Nuesch E, Frey F, Ganz R, Leunig M. An examination chair to measure
internal rotation of the hip in routine settings: a validation study. Osteoarthritis Cartilage 2010;
18: 365-371.
216. Wyss TF, Clark JM, Weishaupt D, Notzli HP. Correlation between internal rotation and bony
anatomy in the hip. Clin Orthop Relat Res 2007; 460: 152-158.
217. Waarsing JH, Kloppenburg M, Slagboom PE, Kroon HM, Houwing-Duistermaat JJ, Weinans H,
et al. Osteoarthritis susceptibility genes influence the association between hip morphology and
osteoarthritis. Arthritis Rheum 2011; 63: 1349-1354.
218. Kim WY, Hutchinson CE, Andrew JG, Allen PD. The relationship between acetabular retroversion
and osteoarthritis of the hip. J Bone Joint Surg Br 2006; 88: 727-729.
219. Ezoe M, Naito M, Inoue T. The prevalence of acetabular retroversion among various disorders of
the hip. J Bone Joint Surg Am 2006; 88: 372-379.
220. Dougados M, Gueguen A, Nguyen M, Berdah L, Lequesne M, Mazieres B, et al. Requirement for
total hip arthroplasty: an outcome measure of hip osteoarthritis? J Rheumatol 1999; 26: 855-
861.
221. Agricola R, Reijman M, Bierma-Zeinstra SM, Verhaar JA, Weinans H, Waarsing JH. Total hip
replacement but not clinical osteoarthritis can be predicted by the shape of the hip: a prospective
cohort study (CHECK). Osteoarthritis Cartilage 2013; 21: 559-564.
251
References
222. Lane NE, Lin P, Christiansen L, Gore LR, Williams EN, Hochberg MC, et al. Association of mild
acetabular dysplasia with an increased risk of incident hip osteoarthritis in elderly white women:
the study of osteoporotic fractures. Arthritis Rheum 2000; 43: 400-404.
223. Agricola R, Waarsing J.H., Thomas G.E., Carr A.J., Reijman M., Bierma-Zeinstra S.M.A., Glyn-
Jones S., Weinans H., Arden N.K. . Cam impingement: defining the presence of a cam deformity
by the alpha angle; data from the CHECK cohort and Chingford cohort. Osteoarthritis Cartilage
[in press] 2013.
224. Notzli HP, Wyss TF, Stoecklin CH, Schmid MR, Treiber K, Hodler J. The contour of the femoral
head-neck junction as a predictor for the risk of anterior impingement. J Bone Joint Surg Br
2002; 84: 556-560.
225. Wiberg. Studies on dysplastic acetabula and congenital subluxation of the hip joint. With special
reference to the complications of osteoarthritis. Acta Chirurgica Scand 1939; 58 (suppl): 5-135.
226. Agricola R, Bessems JH, Ginai AZ, Heijboer MP, van der Heijden RA, Verhaar JA, et al. The
development of Cam-type deformity in adolescent and young male soccer players. Am J Sports
Med 2012; 40: 1099-1106.
227. Siebenrock KA, Ferner F, Noble PC, Santore RF, Werlen S, Mamisch TC. The cam-type deformity
of the proximal femur arises in childhood in response to vigorous sporting activity. Clin Orthop
Relat Res 2011; 469: 3229-3240.
228. Lynch JA, Parimi N, Chaganti RK, Nevitt MC, Lane NE. The association of proximal femoral shape
and incident radiographic hip OA in elderly women. Osteoarthritis Cartilage 2009; 17: 1313-
1318.
229. Nelson AE, Liu F, Lynch JA, Renner JB, Schwartz TA, Lane NE, et al. Association of incident
symptomatic hip osteoarthritis with differences in hip shape by active shape modeling: the
johnston county osteoarthritis project. Arthritis Care Res (Hoboken) 2014; 66: 74-81.
230. Baker-Lepain JC, Lynch JA, Parimi N, McCulloch CE, Nevitt MC, Corr M, et al. Variant alleles of
the Wnt antagonist FRZB are determinants of hip shape and modify the relationship between hip
shape and osteoarthritis. Arthritis Rheum 2012; 64: 1457-1465.
231. Sarkalkan N, Weinans H, Zadpoor AA. Statistical shape and appearance models of bones. Bone
2013; 60C: 129-140.
232. Schiphof D, Boers M, Bierma-Zeinstra SM. Differences in descriptions of Kellgren and Lawrence
grades of knee osteoarthritis. Ann Rheum Dis 2008; 67: 1034-1036.
233. Cootes TF TC, Cooper DH, Graham J. Active shape models - their training and application.
Comput Vis Image und 1995; 61: 38-59.
234. Sarkalkan N, Weinans H, Zadpoor AA. Statistical shape and appearance models of bones. Bone
2014; 60C: 129-140.
235. Castano-Betancourt MC, Van Meurs JB, Bierma-Zeinstra S, Rivadeneira F, Hofman A, Weinans
H, et al. The contribution of hip geometry to the prediction of hip osteoarthritis. Osteoarthritis
Cartilage 2013; 21: 1530-1536.
236. Agricola R, Heijboer MP, Ginai AZ, Roels P, Zadpoor AA, Verhaar JA, et al. A Cam Deformity Is
Gradually Acquired During Skeletal Maturation in Adolescent and Young Male Soccer Players: A
Prospective Study With Minimum 2-Year Follow-up. Am J Sports Med 2014; 42: 798-806.
237. Lindner C, Thiagarajah S, Wilkinson JM, Wallis GA, Cootes TF. Development of a fully automatic
shape model matching (FASMM) system to derive statistical shape models from radiographs:
application to the accurate capture and global representation of proximal femur shape.
Osteoarthritis Cartilage 2013; 21: 1537-1544.
238. Felson DT, Lawrence RC, Dieppe PA, Hirsch R, Helmick CG, Jordan JM, et al. Osteoarthritis: new
insights. Part 1: the disease and its risk factors. Ann Intern Med 2000; 133: 635-646.
239. Agricola R, Reijman M, Bierma-Zeinstra SM, Verhaar JA, Weinans H, Waarsing JH. Total Hip
Replacement but not Clinical Osteoarthritis can be Predicted by the Shape of the Hip: A
Prospective Cohort Study (Check). Osteoarthritis Cartilage 2013.
240. Ganz R, Leunig M, Leunig-Ganz K, Harris WH. The etiology of osteoarthritis of the hip: an
integrated mechanical concept. Clin Orthop Relat Res 2008; 466: 264-272.
252
References
253
References
262. Pollard TC, McNally EG, Wilson DC, Wilson DR, Madler B, Watson M, et al. Localized cartilage
assessment with three-dimensional dGEMRIC in asymptomatic hips with normal morphology and
cam deformity. J Bone Joint Surg Am 2010; 92: 2557-2569.
263. Meyer DC, Beck M, Ellis T, Ganz R, Leunig M. Comparison of six radiographic projections to
assess femoral head/neck asphericity. Clin Orthop Relat Res 2006; 445: 181-185.
264. Reichenbach S, Leunig M, Werlen S, Nuesch E, Pfirrmann CW, Bonel H, et al. Association between
cam-type deformities and magnetic resonance imaging-detected structural hip damage: a cross-
sectional study in young men. Arthritis Rheum 2011; 63: 4023-4030.
265. Altman RD, Bloch DA, Dougados M, Hochberg M, Lohmander S, Pavelka K, et al. Measurement of
structural progression in osteoarthritis of the hip: the Barcelona consensus group. Osteoarthritis
Cartilage 2004; 12: 515-524.
266. Lavigne M, Parvizi J, Beck M, Siebenrock KA, Ganz R, Leunig M. Anterior femoroacetabular
impingement: part I. Techniques of joint preserving surgery. Clin Orthop Relat Res 2004: 61-66.
267. Philippon MJ, Briggs KK, Yen YM, Kuppersmith DA. Outcomes following hip arthroscopy for
femoroacetabular impingement with associated chondrolabral dysfunction: minimum two-year
follow-up. J Bone Joint Surg Br 2009; 91: 16-23.
268. Sink EL, Gralla J, Ryba A, Dayton M. Clinical presentation of femoroacetabular impingement in
adolescents. J Pediatr Orthop 2008; 28: 806-811.
269. Parvizi J, Leunig M, Ganz R. Femoroacetabular impingement. J Am Acad Orthop Surg 2007; 15:
561-570.
270. Reynolds D, Lucas J, Klaue K. Retroversion of the acetabulum. A cause of hip pain. J Bone Joint
Surg Br 1999; 81: 281-288.
271. Resnick DN, G. Degenerative disease of extraspinal locations, in Resnick D, Niwayama G, eds:.
Diagnosis of Bone and Joint Disorders, Philadelphia, PA, WB Saunders 1981: 1322-1335.
272. Reichenbach S, Juni P, Werlen S, Nuesch E, Pfirrmann CW, Trelle S, et al. Prevalence of cam-type
deformity on hip magnetic resonance imaging in young males: a cross-sectional study. Arthritis
Care Res (Hoboken) 2010; 62: 1319-1327.
273. Kang AC, Gooding AJ, Coates MH, Goh TD, Armour P, Rietveld J. Computed tomography assessment
of hip joints in asymptomatic individuals in relation to femoroacetabular impingement. Am J
Sports Med 2010; 38: 1160-1165.
274. Pollard TC, Villar RN, Norton MR, Fern ED, Williams MR, Simpson DJ, et al. Femoroacetabular
impingement and classification of the cam deformity: the reference interval in normal hips. Acta
Orthop 2010; 81: 134-141.
275. Hack K, Di Primio G, Rakhra K, Beaule PE. Prevalence of cam-type femoroacetabular impingement
morphology in asymptomatic volunteers. J Bone Joint Surg Am 2010; 92: 2436-2444.
276. Laborie LB, Lehmann TG, Engesaeter IO, Eastwood DM, Engesaeter LB, Rosendahl K. Prevalence
of radiographic findings thought to be associated with femoroacetabular impingement in a
population-based cohort of 2081 healthy young adults. Radiology 2011; 260: 494-502.
277. Jung KA, Restrepo C, Hellman M, AbdelSalam H, Morrison W, Parvizi J. The prevalence of cam-
type femoroacetabular deformity in asymptomatic adults. J Bone Joint Surg Br 2011; 93: 1303-
1307.
278. Kapron AL, Anderson AE, Aoki SK, Phillips LG, Petron DJ, Toth R, et al. Radiographic prevalence
of femoroacetabular impingement in collegiate football players: AAOS Exhibit Selection. J Bone
Joint Surg Am 2011; 93: e111(111-110).
279. Hartofilakidis G, Bardakos NV, Babis GC, Georgiades G. An examination of the association
between different morphotypes of femoroacetabular impingement in asymptomatic subjects and
the development of osteoarthritis of the hip. J Bone Joint Surg Br 2011; 93: 580-586.
280. Dudda M, Albers C, Mamisch TC, Werlen S, Beck M. Do normal radiographs exclude asphericity
of the femoral head-neck junction? Clin Orthop Relat Res 2009; 467: 651-659.
281. Wassilew GI, Heller MO, Diederichs G, Janz V, Wenzl M, Perka C. Standardized AP radiographs do
not provide reliable diagnostic measures for the assessment of acetabular retroversion. J Orthop
Res 2012; 30: 1369-1376.
254
References
282. Beaule PE, Hynes K, Parker G, Kemp KA. Can the alpha angle assessment of cam impingement
predict acetabular cartilage delamination? Clin Orthop Relat Res; 470: 3361-3367.
283. Johnston TL, Schenker ML, Briggs KK, Philippon MJ. Relationship between offset angle alpha and
hip chondral injury in femoroacetabular impingement. Arthroscopy 2008; 24: 669-675.
284. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of
clinical measurement. Lancet 1986; 1: 307-310.
285. Fraitzl CR, Kappe T, Pennekamp F, Reichel H, Billich C. Femoral head-neck offset measurements
in 339 subjects: distribution and implications for femoroacetabular impingement. Knee Surg
Sports Traumatol Arthrosc; 21: 1212-1217.
286. Sutter R, Dietrich TJ, Zingg PO, Pfirrmann CW. How useful is the alpha angle for discriminating
between symptomatic patients with cam-type femoroacetabular impingement and asymptomatic
volunteers? Radiology 2012; 264: 514-521.
287. Carlisle JC, Zebala LP, Shia DS, Hunt D, Morgan PM, Prather H, et al. Reliability of various
observers in determining common radiographic parameters of adult hip structural anatomy. Iowa
Orthop J 2011; 31: 52-58.
288. Lohan DG, Seeger LL, Motamedi K, Hame S, Sayre J. Cam-type femoral-acetabular impingement:
is the alpha angle the best MR arthrography has to offer? Skeletal Radiol 2009; 38: 855-862.
289. Barton C, Salineros MJ, Rakhra KS, Beaule PE. Validity of the alpha angle measurement on plain
radiographs in the evaluation of cam-type femoroacetabular impingement. Clin Orthop Relat Res
2011; 469: 464-469.
290. Vignon E, Valat JP, Rossignol M, Avouac B, Rozenberg S, Thoumie P, et al. Osteoarthritis of the
knee and hip and activity: a systematic international review and synthesis (OASIS). Joint Bone
Spine 2006; 73: 442-455.
291. Lindberg H, Roos H, Gardsell P. Prevalence of coxarthrosis in former soccer players. 286 players
compared with matched controls. Acta Orthop Scand 1993; 64: 165-167.
292. Harris WH. Etiology of osteoarthritis of the hip. Clin Orthop Relat Res 1986: 20-33.
293. Solomon L, Schnitzler CM. Pathogenetic types of coxarthrosis and implications for treatment.
Arch Orthop Trauma Surg 1983; 101: 259-261.
294. Goodman DA, Feighan JE, Smith AD, Latimer B, Buly RL, Cooperman DR. Subclinical slipped
capital femoral epiphysis. Relationship to osteoarthrosis of the hip. J Bone Joint Surg Am 1997;
79: 1489-1497.
295. Tanzer M, Noiseux N. Osseous abnormalities and early osteoarthritis: the role of hip impingement.
Clin Orthop Relat Res 2004: 170-177.
296. Nepple JJ, Carlisle JC, Nunley RM, Clohisy JC. Clinical and Radiographic Predictors of Intra-
articular Hip Disease in Arthroscopy. Am J Sports Med.
297. Kassarjian A, Yoon LS, Belzile E, Connolly SA, Millis MB, Palmer WE. Triad of MR arthrographic
findings in patients with cam-type femoroacetabular impingement. Radiology 2005; 236: 588-
592.
298. Hogstrom M, Nordstrom A, Alfredson H, Lorentzon R, Thorsen K, Nordstrom P. Current physical
activity is related to bone mineral density in males but not in females. Int J Sports Med 2007;
28: 431-436.
299. Karlsson KM, Karlsson C, Ahlborg HG, Valdimarsson O, Ljunghall S. The duration of exercise as
a regulator of bone turnover. Calcif Tissue Int 2003; 73: 350-355.
300. Karlsson KM, Karlsson C, Ahlborg HG, Valdimarsson O, Ljunghall S, Obrant KJ. Bone turnover
responses to changed physical activity. Calcif Tissue Int 2003; 72: 675-680.
301. Zouch M, Jaffre C, Thomas T, Frere D, Courteix D, Vico L, et al. Long-term soccer practice
increases bone mineral content gain in prepubescent boys. Joint Bone Spine 2008; 75: 41-49.
302. Fraitzl CR, Kafer W, Nelitz M, Reichel H. Radiological evidence of femoroacetabular impingement
in mild slipped capital femoral epiphysis: a mean follow-up of 14.4 years after pinning in situ. J
Bone Joint Surg Br 2007; 89: 1592-1596.
303. Leunig M, Beaule PE, Ganz R. The concept of femoroacetabular impingement: current status and
future perspectives. Clin Orthop Relat Res 2009; 467: 616-622.
255
References
304. Clohisy JC, Nunley RM, Otto RJ, Schoenecker PL. The frog-leg lateral radiograph accurately
visualized hip cam impingement abnormalities. Clin Orthop Relat Res 2007; 462: 115-121.
305. Clohisy JC, Carlisle JC, Beaule PE, Kim YJ, Trousdale RT, Sierra RJ, et al. A systematic approach
to the plain radiographic evaluation of the young adult hip. J Bone Joint Surg Am 2008; 90 Suppl
4: 47-66.
306. Acheson RM. The Oxford method of assessing skeletal maturity. Clin Orthop 1957; 10: 19-39.
307. Samora JB, Ng VY, Ellis TJ. Femoroacetabular impingement: a common cause of hip pain in
young adults. Clin J Sport Med; 21: 51-56.
308. Burnett RS, Della Rocca GJ, Prather H, Curry M, Maloney WJ, Clohisy JC. Clinical presentation of
patients with tears of the acetabular labrum. J Bone Joint Surg Am 2006; 88: 1448-1457.
309. Serrat MA, Reno PL, McCollum MA, Meindl RS, Lovejoy CO. Variation in mammalian proximal
femoral development: comparative analysis of two distinct ossification patterns. J Anat 2007;
210: 249-258.
310. Tachdjian. Tachdjian’s pediatric orthopaedics. Volume 4th, Saunders Elsevier 2007.
311. Domayer SE, Ziebarth K, Chan J, Bixby S, Mamisch TC, Kim YJ. Femoroacetabular cam-type
impingement: Diagnostic sensitivity and specificity of radiographic views compared to radial
MRI. Eur J Radiol 2010.
312. Ochoa LM, Dawson L, Patzkowski JC, Hsu JR. Radiographic prevalence of femoroacetabular
impingement in a young population with hip complaints is high. Clin Orthop Relat Res; 468:
2710-2714.
313. Wudbhav N. Sankar MN, Javad Parvizi, David T. Felson, Rintje Agricola, Michael Leunig.
Femoroacetabular impingement: Defining the condition and its role in the pathofysiology of
osteoarthritis. J Am Acad Orthop Surg 2013; July 2013: S7-S15.
314. Sankar WN, Nevitt M, Parvizi J, Felson DT, Agricola R, Leunig M. Femoroacetabular impingement:
defining the condition and its role in the pathophysiology of osteoarthritis. J Am Acad Orthop
Surg 2013; 21 Suppl 1: S7-S15.
315. Gerhardt MB, Romero AA, Silvers HJ, Harris DJ, Watanabe D, Mandelbaum BR. The prevalence
of radiographic hip abnormalities in elite soccer players. Am J Sports Med 2012; 40: 584-588.
316. Johnson AC, Shaman MA, Ryan TG. Femoroacetabular impingement in former high-level youth
soccer players. Am J Sports Med 2012; 40: 1342-1346.
317. Philippon MJ, Ho CP, Briggs KK, Stull J, Laprade RF. Prevalence of Increased Alpha Angles as a
Measure of Cam-Type Femoroacetabular Impingement in Youth Ice Hockey Players. Am J Sports
Med 2013.
318. Agricola R, Waarsing JH, Thomas GE, Carr AJ, Reijman M, Bierma-Zeinstra SM, et al. Cam
impingement: defining the presence of a cam deformity by the alpha angle: Data from the
CHECK cohort and Chingford cohort. Osteoarthritis Cartilage 2013.
319. Larson CM, Sikka RS, Sardelli MC, Byrd JW, Kelly BT, Jain RK, et al. Increasing Alpha Angle
is Predictive of Athletic-Related “Hip” and “Groin” Pain in Collegiate National Football League
Prospects. Arthroscopy 2013; 29: 405-410.
320. Mirtz TA, Chandler JP, Eyers CM. The effects of physical activity on the epiphyseal growth plates:
a review of the literature on normal physiology and clinical implications. J Clin Med Res 2011; 3:
1-7.
321. Kapron AL, Anderson AE, Peters CL, Phillips LG, Stoddard GJ, Petron DJ, et al. Hip internal
rotation is correlated to radiographic findings of cam femoroacetabular impingement in collegiate
football players. Arthroscopy 2012; 28: 1661-1670.
322. Audenaert EA, Peeters I, Vigneron L, Baelde N, Pattyn C. Hip morphological characteristics and
range of internal rotation in femoroacetabular impingement. Am J Sports Med 2012; 40: 1329-
1336.
323. Kienle KP, Keck J, Werlen S, Kim YJ, Siebenrock KA, Mamisch TC. Femoral morphology and
epiphyseal growth plate changes of the hip during maturation: MR assessments in a 1-year
follow-up on a cross-sectional asymptomatic cohort in the age range of 9-17 years. Skeletal
Radiol 2012; 41: 1381-1390.
256
References
324. Siebenrock KA, Behning A, Mamisch TC, Schwab JM. Growth plate alteration precedes cam-type
deformity in elite basketball players. Clin Orthop Relat Res 2013; 471: 1084-1091.
325. Tonnis D. Normal values of the hip joint for the evaluation of X-rays in children and adults. Clin
Orthop Relat Res 1976: 39-47.
326. Lütken P. Bone-bridge formation between the greater trochanter and the femoral head: A normal
variation of the pattern of the ossification in the upper end of the femur in adolescence. Acta
Orthop Scand. 1961; 31: 209-215.
327. Kolo FC, Charbonnier C, Pfirrmann CW, Duc SR, Lubbeke A, Duthon VB, et al. Extreme hip
motion in professional ballet dancers: dynamic and morphological evaluation based on magnetic
resonance imaging. Skeletal Radiol 2013; 42: 689-698.
328. Clohisy JC, Knaus ER, Hunt DM, Lesher JM, Harris-Hayes M, Prather H. Clinical presentation of
patients with symptomatic anterior hip impingement. Clin Orthop Relat Res 2009; 467: 638-644.
329. Beck M, Leunig M, Parvizi J, Boutier V, Wyss D, Ganz R. Anterior femoroacetabular impingement:
part II. Midterm results of surgical treatment. Clinical Orthopaedics and Related Research 2004;
418: 67-73.
330. Larson CM, Giveans MR. Arthroscopic debridement versus refixation of the acetabular labrum
associated with femoroacetabular impingement. Arthroscopy 2009; 25: 369-376.
331. Lincoln M, Johnston K, Muldoon M, Santore R. Combined arthroscopic and modified open
approach for cam femoroacetabular impingement: a preliminary experience. Arthroscopy 2009;
25: 392-399.
332. Naal FD, Miozzari HH, Wyss TF, Notzli HP. Surgical hip dislocation for the treatment of
femoroacetabular impingement in high-level athletes. American Journal of Sports Medicine
2011; 39: 544-550.
333. Leunig M, Casillas MM, Hamlet M, Hersche O, Notzli H, Slongo T, et al. Slipped capital femoral
epiphysis: early mechanical damage to the acetabular cartilage by a prominent femoral
metaphysis. Acta Orthopaedica 2000; 71: 370-375.
334. Rab GT. The geometry of slipped capital femoral epiphysis: implications for movement,
impingement, and corrective osteotomy. Journal of Pediatric Orthopaedics 1999; 19: 419-424.
335. Hosalkar HS, Pandya NK, Bomar JD, Wenger DR. Hip impingement in slipped capital femoral
epiphysis: a changing perspective. Journal of Children’s Orthopaedics 2012: 1-12.
336. Snow SW, Keret D, Scarangella S, Bowen JR. Anterior Impingement of the Femoral-Head - a Late
Phenomenon of Legg-Calve-Perthes Disease. Journal of Pediatric Orthopaedics 1993; 13: 286-
289.
337. Eijer H, Myers SR, Ganz R. Anterior femoroacetabular impingement after femoral neck fractures.
Journal of Orthopaedic Trauma 2001; 15: 475-481.
338. Strehl A, Ganz R. [Anterior femoroacetabular impingement after healed femoral neck fractures].
Unfallchirurg 2005; 108: 263-273.
339. Keogh MJ, Batt ME. A review of femoroacetabular impingement in athletes. Sports Med 2008;
38: 863-878.
340. Ng VY, Ellis TJ. More than just a bump: cam-type femoroacetabular impingement and the
evolution of the femoral neck. Hip International 2011; 21: 1-8.
341. Carter DR, Van Der Meulen MC, Beaupre GS. Mechanical factors in bone growth and development.
Bone 1996; 18: S5-S10.
342. Timothy A. Mirtza c, Judy P. Chandlerb, Christina M. Eyersb. The Effects of Physical Activity on
the Epiphyseal Growth Plates: A Review of the Literature on Normal Physiology and Clinical
Implications. Journal of Clinical Medicine Research 2010; 3(1): 1-7.
343. Carter DR, Beaupré GS. Skeletal function and form: mechanobiology of skeletal development,
aging, and regeneration, Cambridge University Press 2007.
344. Carter DR, Orr TE, Fyhrie DP. Relationships between loading history and femoral cancellous bone
architecture. Journal of Biomechanics 1989; 22: 231-244.
345. Carter DR, Orr TE, Fyhrie DP, Schurman DJ. Influences of mechanical stress on prenatal and
postnatal skeletal development. Clinical Orthopaedics and Related Research 1987; 219: 237-
250.
257
References
346. Carter DR, Wong M. Mechanical stresses and endochondral ossification in the chondroepiphysis.
Journal of Orthopaedic Research 1988; 6: 148-154.
347. Murphy MJ. The importance of computed tomography slice thickness in radiographic patient
positioning for radiosurgery. Medical Physics 1999; 26: 171-175.
348. Ohman C, Baleani M, Pani C, Taddei F, Alberghini M, Viceconti M, et al. Compressive behaviour
of child and adult cortical bone. Bone 2011; 49: 769-776.
349. Fishkin Z, Armstrong DG, Shah H, Patra A, Mihalko WM. Proximal femoral physis shear in slipped
capital femoral epiphysis-a finite element study. Journal of Pediatric Orthopaedics 2006; 26:
291-294.
350. Carter DR, Wong M. The role of mechanical loading histories in the development of diarthrodial
joints. Journal of Orthopaedic Research 2005; 6: 804-816.
351. Bergmann G, Deuretzbacher G, Heller M, Graichen F, Rohlmann A, Strauss J, et al. Hip contact
forces and gait patterns from routine activities. Journal of Biomechanics 2001; 34: 859-871.
352. Carriero A, Zavatsky A, Stebbins J, Theologis T, Lenaerts G, Jonkers I, et al. Influence of altered
gait patterns on the hip joint contact forces. Computer Methods in Biomechanics and Biomedical
Engineering 2012; in press.
353. Bergmann GeCU. OrthoLoad, Loading of Orthopaedic Implants. 2008 medizin Berlin.
354. Lin H, Aubin CE, Parent S, Villemure I. Mechanobiological bone growth: comparative analysis
of two biomechanical modeling approaches. Medical and Biological Engineering and Computing
2009; 47: 357-366.
355. Villemure I, Stokes IA. Growth plate mechanics and mechanobiology. A survey of present
understanding. Journal of Biomechanics 2009; 42: 1793-1803.
356. Stokes IA, Laible JP. Three-dimensional osseo-ligamentous model of the thorax representing
initiation of scoliosis by asymmetric growth. Journal of Biomechanics 1990; 23: 589-595.
357. Carriero A, Jonkers I, Shefelbine SJ. Mechanobiological prediction of proximal femoral deformities
in children with cerebral palsy. Computer Methods in Biomechanics and Biomedical Engineering
2011; 14: 253-262.
358. Shefelbine SJ, Carter DR. Mechanobiological predictions of femoral anteversion in cerebral palsy.
Annals of Biomedical Engineering 2004; 32: 297-305.
359. Smith-Petersen MN. Treatment of malum coxae senilis, old slipped upper femoral epiphysis,
intrapelvic protrusion of the acetabulum, and coxa plana by means of acetabuloplasty. J Bone
Joint Surg Am 1936; 18: 869-880.
360. Pollard TC, Villar RN, Norton MR, Fern ED, Williams MR, Murray DW, et al. Genetic influences in
the aetiology of femoroacetabular impingement: a sibling study. J Bone Joint Surg Br 2010; 92:
209-216.
361. Krekel PR, Vochteloo AJ, Bloem RM, Nelissen RG. Femoroacetabular impingement and its
implications on range of motion: a case report. J Med Case Rep 2011; 5: 143.
362. Tannast M, Kubiak-Langer M, Langlotz F, Puls M, Murphy SB, Siebenrock KA. Noninvasive three-
dimensional assessment of femoroacetabular impingement. J Orthop Res 2007; 25: 122-131.
363. Nemes S, Gordon M, Rogmark C, Rolfson O. Projections of total hip replacement in Sweden from
2013 to 2030. Acta Orthop 2014; 85: 238-243.
364. Haugen IK, Cotofana S, Englund M, Kvien TK, Dreher D, Nevitt M, et al. Hand joint space
narrowing and osteophytes are associated with magnetic resonance imaging-defined knee
cartilage thickness and radiographic knee osteoarthritis: data from the Osteoarthritis Initiative.
J Rheumatol 2012; 39: 161-166.
365. Bijsterbosch J, Meulenbelt I, Watt I, Rosendaal FR, Huizinga TW, Kloppenburg M. Clustering of
hand osteoarthritis progression and its relationship to progression of osteoarthritis at the knee.
Ann Rheum Dis 2013.
366. Srikanth VK, Fryer JL, Zhai G, Winzenberg TM, Hosmer D, Jones G. A meta-analysis of sex
differences prevalence, incidence and severity of osteoarthritis. Osteoarthritis Cartilage 2005;
13: 769-781.
367. Miyamoto Y, Mabuchi A, Shi D, Kubo T, Takatori Y, Saito S, et al. A functional polymorphism in the
5’ UTR of GDF5 is associated with susceptibility to osteoarthritis. Nat Genet 2007; 39: 529-533.
258
References
259
Appendices
Nederlandstalige samenvatting
References
PhD portfolio
List of publications
Curriculum Vitae
Dankwoord
PhD portfolio
1. PhD training
Workload
Year (Hours/ECTS)
In-depth courses
- Statistics: Cohort studies (ESP 39, NIHES) 2012 0.7
- Statistics: Logistic regression (ESP 66, NIHES) 2012 1.4
263
PhD portfolio
Workload
Year (Hours/ECTS)
264
PhD portfolio
Workload
Year (Hours/ECTS)
2. Teaching activities
Workload
Year (Hours/ECTS)
265
Appendices
Nederlandstalige samenvatting
References
PhD portfolio
List of publications
Curriculum Vitae
Dankwoord
List of publications
LIST OF PUBLICATIONS
Glyn-Jones S, Palmer AJR, Agricola R, Price A.J, Vincent TL, Weinans H, Carr
AJ.Osteoarthritis: Is it possible to diagnose and treat early disease? Lancet.
Accepted for publication
Siebelt M, Agricola R, Weinans H, Kim YJ. The role of imaging in early hip
osteoarthritis. Osteoarthritis and Cartilage. 2014 Oct;22(10):1470-80
Agricola R, Heijboer MP, Ginai AZ, Roels P, Zadpoor AA, Verhaar JAN, Weinans
H,Waarsing JH. A cam deformity is gradually acquired during skeletal
maturation in adolescent and young male soccer players: a prospective study
with a minimum of 2 years follow-up. American Journal of Sports Medicine.
2014 Apr;42(4):798-806
Agricola R, Waarsing JH, Arden NK, Carr AJ, Bierma-Zeinstra SMA, Thomas GE,
Weinans H, Glyn-Jones S. Cam impingement of the hip - a risk factor for hip
osteoarthritis. Nature Reviews Rheumatology 2013 Oct;9(10):630-4
269
List of publications
Agricola R, Bessems JH, Ginai AZ, Heijboer MP, van der Heijden RA, Verhaar
JAN, Weinans H, Waarsing JH. The development of cam-type deformity
in adolescent and young male soccer players. American Journal of Sports
Medicine. 2012 May;40(5):1099-106
Agricola R, Leyland KM, Bierma-Zeinstra SMA, Thomas GE, Emans PE, Spector
TD, Weinans H, Waarsing JH, Arden NK. Validation of statistical shape
modeling to predict hip osteoarthritis in females: data from two prospective
cohort studies (CHECK and Chingford) Rheumatology. Revision submitted.
Roze RH, Bierma-Zeinstra SMA, Agricola R, Oei EHG, Waarsing JH. Sex
related differences in MRI features of between two different osteoarthritis
subpopulations: data from the Osteoarthritis Initiative. Submitted
270
Appendices
Nederlandstalige samenvatting
References
PhD portfolio
List of publications
Curriculum Vitae
Dankwoord
Curriculum Vitae
CURRICULUM VITAE
Rintje Agricola werd geboren op 25 november 1986 te
Heerenveen waar hij tevens in 2005 zijn VWO diploma
aan het Bornego College behaalde. In datzelfde jaar ging
hij bewegingswetenschappen studeren aan de
Rijksuniversiteit Groningen. Na het behalen van zijn
propedeuse startte hij in 2006 de studie geneeskunde aan
de Erasmus Universiteit Rotterdam. Tijdens zijn
afstudeeronderzoek naar de ontwikkeling van cam-
deformiteiten bij de afdeling orthopaedie van het Erasmus
MC, werd zijn belangstelling gewekt voor wetenschappelijk
onderzoek. In 2011 heeft hij zijn doctoraal behaald. Toen
vervolgens de mogelijkheid zich voordeed om zijn afstudeeronderzoek als
promovendus voort te zetten, greep hij die kans vol enthousiasme. Onder
begeleiding van prof.dr.ir. H. Weinans, prof.dr. J.A.N. Verhaar en dr.ir. J.H.
Waarsing heeft hij 2.5 jaar onderzoek gedaan naar morfologische risicofactoren
van heupartrose en hoe deze ontstaan in de jeugd. In 2013 is hij gestart met
zijn co-schappen in diverse ziekenhuizen in de regio Rotterdam. Zijn co-schappen
werden 2.5 maand onderbroken om een fellowship te volgen in het Nuffield
Orthopaedic Centre en Botnar Research Centre aan de University of Oxford.
Onder begeleiding van prof. N.K. Arden heeft hij zijn kennis op het gebied van
morfologische kwantificatie van de heup verder kunnen uitbreiden. In februari
2015 zal hij zijn artsdiploma in ontvangst mogen nemen en aansluitend een
fellowship in Doha, Qatar volgen. In juli 2015 zal hij met zijn vooropleiding
algemene heelkunde starten in het IJsselland ziekenhuis te Capelle aan den
IJssel (opleider dr. I. Dawson). Zijn opleiding tot orthopaedisch chirurg zal in
2017 aanvangen en hij zal gaan werken in het Erasmus MC te Rotterdam
(opleider prof. dr. J.A.N. Verhaar) en in de Reinier de Graaf Groep te Delft
(opleider dr. R.M. Bloem).
273
Appendices
Nederlandstalige samenvatting
References
PhD portfolio
List of publications
Curriculum Vitae
Dankwoord
Dankwoord
DANKWOORD
Tot slot mijn dank aan alle mensen die een bijdrage hebben geleverd aan de
totstandkoming van dit proefschrift. Een aantal mensen wil ik in het bijzonder
bedanken.
Promotor prof. dr. ir. H. Weinans, beste Harrie, jouw enthousiasme werkt
aanstekelijk. Bedankt voor de prachtige kans die jij mij geboden hebt om mezelf
te ontplooien als wetenschapper. Jouw begeleiding gedurende de afgelopen
jaren heb ik enorm gewaardeerd en ook het feit dat je mij vrij liet om binnen dit
onderwerp mijn eigen koers te bepalen. Ik heb de afgelopen jaren ontzettend
veel van je geleerd en hoop in de toekomst nog intensief met je te kunnen
samenwerken.
Promotor prof. dr. J.A.N. Verhaar, beste professor, uw klinische blik heeft het
wetenschappelijk werk van dit proefschrift naar een hoger niveau getild. Bedankt
dat u mij heeft begeleid met een scherpe en kritische blik. Als mijn toekomstig
opleider hoop ik ook op klinisch gebied veel van u te kunnen leren. Bedankt voor
de mogelijkheden die u mij geboden heeft.
Co-promotor dr. ir. J.H. Waarsing, beste Erwin, als dagelijkse begeleider heb
ik natuurlijk heel veel aan jou te danken; met name ook in het begin bij de
sturing van mijn onderzoek: daar waar ik tien horden tegelijk wilde nemen,
adviseerde jij om de horden weloverwogen en één voor één te nemen. Daarnaast
sta ik er nog altijd van versteld hoe snel jij data kunt interpreteren en in een
simpele boodschap weet over te dragen. Doordat jij mij voor een groot deel
wetenschappelijk hebt opgevoed, zal er in mijn onderzoek altijd een vleugje
‘Erwin’ te bespeuren zijn. Ik hoop dat we onze samenwerking nog lange tijd
kunnen voortzetten.
277
Dankwoord
Drs. Heijboer, u was de eerste bij wie ik op de deur klopte met de vraag: ‘kan
ik onderzoek bij Orthopaedie doen’? Heel veel dank dat u mij het vertrouwen
schonk en die deur voor mij hebt geopend.
Beste Gert, het is aanstekelijk om te zien hoe gepassioneerd jij met je vak bezig
bent. Ik denk dat we nog heel veel mooie heupprojecten kunnen opzetten.
Beste collegae en vrienden van de 16e en HS, beste Johan, Marjan, Marianne,
Anna, Maarten, Roberto, Wu, Yvonne, Mieke, Gerjo, Holger, Gerben, Wendy,
Sandra, Nicole, Esther, Ruud, Sander, Stefan, Marloes, Nienke, Max, Job, Belle,
Eline, Desiree, Vincent, Guus, Tijs en Robert-Jan: bedankt voor jullie steun en
gezelligheid.
Dear colleagues from Oxford, dear professor Arden, professor Carr, professor
Glyn-Jones, Kirsten, Antony, Geraint, thank you for the wonderful and interesting
time I had at the NOC and Botnar Institute: it was very fruitful.
Dear Anne-Marie, thank you for revising the discussion of my thesis even until
02:00 in the morning.
Hartelijk dank aan de deelnemers en organisatie van het CHECK cohort, zonder
jullie inzet zou dit fantastische initiatief en daarmee ook het eerste deel van mijn
proefschrift niet bestaan.
278
Dankwoord
Beste familie Rulkens, Sylvia, gelukkig kon ik met al mijn vragen en advies over
promoveren bij jou als expert terecht. Marc, ook jij wist me van de nodige raad
te voorzien omtrent mijn vervolgopleiding en sollicitatie; bedankt.
Lieve Herman en Klaske, lieve heit en mem, promoveren gaat met vallen
en opstaan, net zoals in het echte leven. Bedankt voor jullie interesse en
onvoorwaardelijke steun. Mem, wat had ik je het eindresultaat toch graag laten
zien. Heit, je doorzettingsvermogen is een voorbeeld voor mij.
Lieve Alex, Nida en Celine, jullie zijn mijn tweede thuis. Het is niet alleen heerlijk
om elke keer weer ondergedompeld te worden in la dolce vita, maar bovenal om
een warme plek te hebben. Bedankt dat ik altijd bij jullie terecht kon, of het nou
ging om de technische aspecten van dit proefschrift of als uitlaatklep wanneer
het even tegen zat.
Lieve Cathelijne, jou liefde geeft mij een ongelimiteerde dosis energie. Je bent
mijn steun, mijn maatje en mijn allermooiste. Elke dag naast jou wakker worden
geeft een ultiem gevoel van geluk; we gaan nog heel wat mooie avonturen
beleven. Bedankt voor het meedenken, je begrip en kritische blik. Ik zal beloven
dat ik vanaf nu iets minder onderzoek in de avonduurtjes ga doen..
279