The Aphasia Rapid Test: An NIHSS-like Aphasia Test
The Aphasia Rapid Test: An NIHSS-like Aphasia Test
DOI 10.1007/s00415-013-6943-x
ORIGINAL COMMUNICATION
S. Chomel-Guillaume • Y. Samson
Received: 1 February 2013 / Revised: 21 April 2013 / Accepted: 26 April 2013 / Published online: 15 May 2013
Ó The Author(s) 2013. This article is published with open access at Springerlink.com
Abstract The Aphasia Rapid Test (ART) is a 26-point scale showed that D8 ART was the only significant predictor of
developed as a bedside assessment to rate aphasia severity in good (AHS 0–2) or poor (AHS 4–5) outcome. The ROC
acute stroke patients in \3 min. We tested its inter-rater curves analyzes showed areas under the curve above 0.9 for
reproducibility, its sensitivity to detect changes from Day 1 to good and poor outcome and revealed D8 ART best cut-off
Day 8, and the predictive value of D8 ART scores on the values of\12 for good and[21 for poor outcome, with more
3-month aphasia outcome assessed with the Aphasia Handicap than 90 % sensitivity and 80 % specificity. The ART is a
Score (AHS), a 0–5 ‘‘Rankin-like’’ score for aphasic disability. simple, rapid and reproducible language task, useful in mon-
The reproducibility was tested in 91 aphasic patients within itoring early aphasic changes in acute stroke patients and
one week of stroke onset. The inter-rater concordance coef- highly predictive of the 3-month verbal communication out-
ficient was 0.99 and the weighted Kappa value (jw) was 0.93. come. It should be easy to adapt to other languages.
The sensitivity was tested in 70 aphasic patients by measuring
changes in ART values between D1 and D8. Improvement Keywords Acute stroke Outcomes Functional
occurred in 46 patients (66 %) and aggravation in three recovery Language outcome Rating scale
patients (4 %). In these patients, a logistic regression analysis
Introduction
Electronic supplementary material The online version of this
article (doi:10.1007/s00415-013-6943-x) contains supplementary
material, which is available to authorized users. The severity of neurological deficit is now routinely quan-
tified in acute stroke patients using the NIH stroke scale
C. Azuar A. Leger Y. Samson (NIHSS), a standard in stroke care and research. This bed-
APHP Service des Urgences Cérébro-Vasculaires, Groupe
side neurological examination is simple, rapid, and repro-
Hospitalier Pitié-Salpêtrière, COGIMAGE CR ICM, ICM
UPMC Paris VI Inserm U 975, CNRS UMR 7225, ducible. It is sensitive enough to detect early changes in
75013 Paris, France neurological status, and has been shown to be highly pre-
dictive of stroke outcome (http://www.nihstrokescale.org).
C. Azuar C. Arbizu F. Henry-Amar
We reasoned that an NIHSS-like scale of language would be
APHP-Fédération de Neurologie, Institut de la Mémoire et de la
Maladie d’Alzheimer, Groupe Hospitalier Pitié-Salpêtrière, useful in grading the severity of aphasia in acute stroke
75013 Paris, France patients. Most aphasia rating scales are too long to be used
in acute stroke patients [6]. They are designed for trained
C. Azuar (&)
speech therapists and require specific material. Some bed-
Institut de la Mémoire et de la Maladie d’Alzheimer,
Groupe Hospitalier Pitié-Salpêtrière, 47, Boulevard de l’Hôpital, side assessments of aphasia have recently been developed
75013 Paris, France [2–4, 9], but their sensitivity in detecting rapid changes in
e-mail: [email protected] the severity of aphasia and in predicting language recovery
is limited or has not been evaluated [3, 4, 9].
S. Chomel-Guillaume
Service de Neurologie, Centre Hospitalier de Versailles, We designed the Aphasia Rapid Test (ART) as an NI-
Le Chesnay, 78157 Paris, France HSS-like aphasia scale, based on the scoring of items that
123
J Neurol (2013) 260:2110–2117 2111
are commonly used in the neurological examination of autonomy of verbal communication), 4 = severe lan-
language in acute stroke patients. It has been designed to be guage-related disability (lack of effective verbal commu-
easy to translate into any language, and to be as little nication), 5 = mutism or total loss of verbal expression
language-specific as possible. It can be administered by any and comprehension. The scoring system is patient-oriented.
health care professional after brief training, without If oral communication is preserved, the examiner asks the
requiring any specific test material. The ART should not be patient or his/her proxy, if necessary, to self-rate his/her
used as a diagnostic tool since it does not discriminate language abilities using a semi-structured interview, as
between aphasia, apraxia of speech and dysarthria. How- described in the Online Resource 2. In case of ambiguity
ever, we reasoned that the ART may be useful to monitor between moderate (score 3) and severe (score 4) disability,
changes in aphasia severity during the acute stage and to the rule of thumb is to assign a score of 3 if it is possible to
predict aphasia prognosis. Here we first describe the ART, score activities of daily living using a scale such as the
its scoring system, and its reliability across two different Barthel index by oral communication with the patient
examiners. Next, we detail the sensitivity of the ART in alone, and to assign a score of 4 if the help of a proxy is
detecting change in language skills during the first week necessary. Since the AHS has not been published in a peer-
post-stroke. Finally, we present the value of the ART in reviewed journal, we retrospectively compared these
predicting language ability at three months using a differ- results with a conventional and well-established language
ent measure, the Aphasia Handicap Scale (AHS). testing battery. The search for patients of the follow-up
study, who received a formal assessment of aphasia by a
speech therapist three months post-stroke using the French
Methods version of the Boston Diagnostic Aphasia Examination
(BDAE) [8], yielded 37 patients. It should be stressed
Scoring systems however that functional scales (such as the AHS) and
impairment scales (such as the BDAE) are of course cor-
The ART was designed by two neurologists (YS and AL) related (see Online Resource 2) but are clearly different.
with extensive experience with aphasic patients, in an
attempt to quantify the severity of aphasia during the acute Patients and studies
phase of a stroke. Earlier, more complex versions were
discarded at preliminary stages, because they were found to All patients were recruited through our stroke unit. All
be unsuitable for bedside examination in acute stroke were right-handed, with French as their first language. The
patients or were found to have low reproducibility in pre- study was approved by the local ethics committee, and in
liminary investigations. The ART score ranges from 0 to 26, agreement with French legislation, informed consent was
with higher values indicating more severe impairment. The waived since assessing the severity of aphasia is part of
patient is successively asked to follow two simple orders standard care in stroke patients.
(maximum 2 points), one more complex order (3 points),
repeat three single words (6 points), repeat one sentence Inter-rater reliability of the ART
(2 points) and name three common objects (6 points). This is
followed by a 1-min verbal semantic fluency task (4 points). We included 91 patients with acute stroke confirmed by
The examiner additionally scores dysarthria (3 points) using MRI, and considered as aphasic by the neurologists and
the same scoring system as in the NIHSS. Since the ART has speech therapists of the stroke unit. Patients with impaired
been designed as a bedside clinical tool, there is no explicit consciousness were excluded. The patients were tested at a
time limit for patient responses. Table 1 shows an English median post-stroke delay of eight days (inter-quartile
version of the ART and explains the scoring system (the range, IQR: 7–10) by two independent examiners, who
Online Resource 1 shows the French version). administered the ART on the same day at a maximum
The outcome of aphasia was assessed three months interval of 12 h. The examiners were stroke neurologists or
post-stroke using the Aphasia Handicap Scale (AHS), a speech therapists who were not involved in the develop-
modified Rankin-score-like five-point scoring system for ment of the ART, and were blind to each other’s ratings.
handicap in verbal communication, designed by two of the The inter-rater reproducibility of the ART was assessed by
authors (YS and SCG) [1]. The scores are as follows: computing the coefficient of concordance, the weighted
0 = normal language, 1 = minor difficulties of language Kappa values (jw), and constructing a Bland–Altman plot of
without disability (no impact on normal life), 2 = the total ART scores of both examiners. In addition, weighted
mild language-related disability (without restrictions in Kappa values (jw) were calculated for the scores of each item.
the autonomy of verbal communication in daily life), All statistical analyses were carried out using MedCalc for
3 = moderate language-related disability (restricted Windows (version 11.6.1.0; http://www.medcalc.be).
123
2112 J Neurol (2013) 260:2110–2117
Table 1 Instructions and scoring system for the Aphasia Rapid Test
Instructions Score
Sensitivity of the ART during the first week post stroke D0 and D8 in the whole group of patients and the pro-
portion of patients who had a significant change between
We included another population of 70 consecutive patients D0 and D8 ART values. We also investigated whether
who met the following criteria: admission to the stroke unit patients with good, intermediate or poor 3-month outcome
within 12 h of the onset of a first-ever stroke; left MCA differed in D0 and D8 ART values by running a two-way
infarct confirmed by MRI; acute aphasia noted by the repeated measure ANOVA.
neurologist on duty; lack of consciousness disorders; ART
performed within 24 h of stroke onset (D0) and at eight - Prediction of 3-month aphasia outcome
days (D8); and an AHS score obtained at three months
during patient follow-up by a stroke neurologist blind to This was done in the same population of 70 patients. We
the D0 and D8 ART scores. The ART data were not used to conducted two stepwise logistic regression models, the first
plan speech therapy strategies and all patients received predicting good language outcome (AHS 0–2) and the
speech therapy as usual in the stroke unit and during second predicting poor language outcome (AHS score
rehabilitation. We determined the changes in ART between 4–5). In both models, the independent variables were
123
J Neurol (2013) 260:2110–2117 2113
gender, age, recombinant tissue plasminogen activator on D8, DART) was above 2 points (i.e., revealing an
(rt-PA) treatment, and D0 and D8 ART scores. The variables improvement) in 46 patients (66 %) and below -2 points
were retained in the final model at p \ 0.01. In addition, (i.e., revealing an aggravation of aphasia) in three patients
Receiver Operating Characteristic (ROC) curves were (4 %), who all suffered from an enlargement of their infarct
generated to compare the predictive value of the ART score during the first few days of stroke. Figure 2 shows the ART
at D0 and D8 for good and poor outcomes. Since AHS values at D0 and D8 in the subgroups of patients with good
has not been published in a peer review journal, we per- (AHS 0–2, 33 patients, 47 %), intermediate (AHS 3, 22
formed similar analyses in the subgroup of 37 patients, patients, 31 %), and poor (AHS 4–5, 15 patients, 21 %)
which had a formal assessment of aphasia by a speech language 3-month outcome groups. The two-way ANOVA
therapist three months post-stroke using the French version for group and time showed significant group
of the Boston Diagnostic Aphasia Examination (BDAE) [8]. (F(2,134) = 39.5, p \ 0.0001) and time (F(2,134) = 52.9,
p \ 0.0001) effects. The Fig. 2 also shows that D0 ART
was lower in the good recovery group and that the changes
Results between D0 and D8 in ART scores differed across the three
groups, as confirmed by a significant group X time inter-
Inter-rater reliability of the ART action [F(2,134) = 4.8, p \ 0.01]. This is also shown by
an ANOVA for DART (F: 12.5, p \ 0.0001). Post hoc tests
The mean age (±SD) of the 91 patients (52 men and 39 showed significant differences (p \ 0.05) between good
women) was 63.96 ± 19.3 years. They had ischemic (DART = 11.2 ± 7.9), intermediate (DART = 6.2 ± 8.3)
(n = 80) or hemorrhagic (n = 11) stroke. The median and poor (DART = -0.7 ± 6.1) recovery groups. In
ART value was 11 (IQR: 4.25–24) for Rater 1 and 12 summary, ART appears to be highly sensitive to change in
(4.25–24) for Rater 2, and the mean (±SD) ART value was the first week post-stroke.
13.4 (±9.51) for Rater 1 and 13.49 (±9.52) for Rater 2.
The inter-rater agreement was good, with a coefficient Prediction of 3-month aphasia outcome
of concordance of 0.990 (95 % confidence interval, CI:
0.985–0.993; p \ 0.0001, Fig. 1a) and a jw of 0.934 (95 % Good (AHS 0–2) and poor (AHS 4–5) outcome
CI: 0.909–0.958). The jw of each item is shown in Table 2,
and ranged from 0.967 for the denomination of the watch In the logistic regression analysis, the ART score at D8
to 0.854 for the scoring of dysarthria. The Bland–Altman remained the only significant predictor of good (odds ratio,
plot (Fig. 1b) showed that there was no test–retest effect OR: 0.75, 95 % CI: 0.66–0.85, p \ 0.0001, accuracy:
and that ART reproducibility was stable across all degrees 88.6 %), or poor outcome (OR: 1.60, 95 % CI: 1.15–2.23,
of aphasia severity. A difference of more than two points p = 0.005, accuracy: 88.6 %), and age, gender, rt-PA
indicated a significant change in aphasia severity. The treatment, and D0 ART scores were not retained in the final
mean duration of ART administration was calculated for logistic regression models. The ROC analysis of the D8
the first 58 patients and was found to be 177 s, including ART score showed that the area under the curve (AUC)
the 1-min fluency task. Thus, the ART is quick to admin- was very high for good (0.926, 95 % CI: 0.838–0.975,
ister, with no test–retest effects, and good rater reliability. p \ 0.0001) and poor (0.955, 95 % CI: 0.876–0.990,
p \ 0.0001) outcomes. A comparison of the AUC gener-
Sensitivity of the ART during the first week post stroke ated with ART scores at D8 and D0 confirmed that ART
was a better predictor at D8 than at D0 of both good
The mean age (±SD) of the 70 patients (41 men, 29 (p = 0.02) and poor (p = 0.007) outcomes (Fig. 3). The
women) was 61.2 ± 15.6 years, and they had a median best prediction of good recovery (AHS 0–2) was yielded by
initial NIHSS score of 16 (IQR: 8–22). Thirty-two patients a D8 ART score of \12, which was associated with a
(46 %) were treated with intravenous rt-PA. At D0, the sensitivity of 93.9 % (95 % CI: 79.7–99.1), a specificity of
mean (±SD) ART score was 19.6 (±7.8) and the median 83.8 % (95 % CI: 68.0–93.8), a positive predictive value of
score 24 (IQR: 13–25, range 1–26). The ART score cor- 83.8 % and a negative predictive value of 93.9 %, whereas
related with the initial NIHSS score at D0 (r: 0.635, the best prediction of poor recovery (AHS 4–5) was
p \ 0.0001) and at D8 (r: 0.525, p \ 0.0001) but not with observed with a D8 ART score of [21, associated with a
age, gender or rt-PA treatment (stepwise multiple regres- sensitivity of 93.3 % (95 % CI: 68.0–98.9), a specificity of
sion). At D8, the ART score had significantly decreased 89.1 % (95 % CI: 77.7–95.9), a positive predictive value of
(p \ 0.0001) with a mean (±SD) value of 12.5 (±9.5) and 70.0 %, and a negative predictive value of 98.0 %. Fig-
a median of 10 (IQR: 4–23). The difference in the ART ure 4 shows the distribution of the AHS at three months as
score between D0 and D8 (i.e., the score on D0—the score a function of these D8 ART thresholds. Note that most of
123
2114 J Neurol (2013) 260:2110–2117
Fig. 1 High inter-rater reproducibility of the ART tested in 91 Altman plot showing that ART reproducibility is stable across all
aphasic patients. a ART scores rated on the same day by two degrees of aphasia severity, with no test–retest effect. Note that a
independent examiners (Raters 1 and 2) showing a coefficient of difference of [2 points indicates a significant change in aphasia
concordance of 0.990 and a weighted kappa value of 0.934. b Bland– severity
Weighted kappa 0.874 0.863 0.874 0.921 0.833 0.842 0.967 0.889 0.885 0.854 0.925
123
J Neurol (2013) 260:2110–2117 2115
Fig. 3 ROC curves for good (AHS 0–2) and poor (AHS 4–5) 93.9 % sensitivity and 83.8 % specificity. b ROC analysis of D0 (blue
3-month language outcomes. a ROC analysis of D0 (blue curve) and curve) and D8 ART scores (red curve) for poor outcome (AHS 4–5).
D8 ART scores (red curve) for good outcome (AHS 0–2). The area The AUC was significantly larger at D8 (0.955, 95 % CI:
under the curve (AUC) was significantly larger at D8 (0.926, 95 % 0.876–0.990) than at D0 (0.766, 95 % CI: 0.650–0.859, p = 0.007).
CI: 0.838–0.975) than at D0 (0.811, 95 % CI: 0.700–0.895, A D8 ART value of [21 predicted poor outcome with 93.3 %
p = 0.02). A D8 ART value of \12 predicted good outcome with sensitivity and 89.1 % specificity
123
2116 J Neurol (2013) 260:2110–2117
improvement was found in 36–57 % of the patients [5, 7, based on the AHS, an unpublished verbal communication
10]. handicap score adapted from the modified Rankin score.
The severity of initial aphasia is considered the best However, more than half of the patients had a BDAE
clinical predictor of the outcome of language function [11, 3 months post-stroke, and the ART appeared to be in good
12, 14]. The prognosis value of the D8 ART score is agreement with the BDAE aphasia severity rating scale.
consistent with other studies, where initial aphasia is often Third, since the ART was developed in French, although it
assessed around one week after stroke [6, 11–13]. How- was designed to be easy to translate, it needs to be tested
ever, several new findings emerge from our study. First, it and validated in other languages by independent studies.
is not only the ART score per se but also the recovery of Finally, the ART has not been directly compared with
basic language functions during the first week post-stroke comprehensive aphasia rating scales, and does not allow us
(i.e., the difference in the score of some items tested by the to classify patients in classic aphasic syndromes. Consid-
ART between D0 and D8) that appears to be an important ering that our subjects are consecutive acute stroke
predictor of later language outcome. A comparison of ART patients, such an analysis may not be feasible, and in any
values in good, intermediate, and poor outcome groups case, the use of ART scoring in an acute stroke unit cannot
shows not only significant group and time effects but also a and should not replace comprehensive language assessment
significant group-time interaction, indicating that early in stabilized patients in speech therapy departments.
recovery differs across groups. This difference is confirmed In summary, the ART appears to be a simple, rapid, and
by a comparison of DART across the three groups. This reproducible language-focused stroke scale to quantify the
also explains why the D8 ART score, which integrates D0 severity of initial aphasia and to monitor early changes in
ART and DART, is a better predictor of language outcome acute stroke patients. It is an accurate predictor of verbal
than the D0 ART score, as shown by logistic regression and communication outcome at three months. This may be of
ROC curve analyses. Second, the ROC curve-based pre- importance for patient stratification in future trials testing
dictions concerning good (AHS 0–2) and poor (AHS 4–5) the effect of early therapeutic intervention after stroke on
outcomes were surprisingly accurate, with an AUC greater aphasia recovery. In addition, since the only language-
than 0.9 and accuracy greater than 85 % in each case. specific items on the test are three words and one sentence
Indeed the cut-off values (D8 ART \12 for good and [21 to be repeated, it should be easy to adapt to other languages.
for poor outcome) yielded a sensitivity of [90 % and a
specificity of [80 %, better than results obtained using Conflicts of interest On behalf of all authors, the corresponding
author states that there is no conflicts of interest.
more complex language tests [6], and similar to those
recently reported by a sophisticated model combining Open Access This article is distributed under the terms of the
language tests and functional MRI results [13]. In addition, Creative Commons Attribution License which permits any use, dis-
as shown in Fig. 4, most patients with intermediate D8 tribution, and reproduction in any medium, provided the original
author(s) and the source are credited.
ART scores had an intermediate outcome (AHS 3). The
good predictive value of D8 ART was also observed when
good and poor outcome groups were defined using the
BDAE aphasia severity rating scale. Age and gender were
not predictors of outcome in this study, in agreement with a
References
recent review that concluded that gender and age did not
significantly impact recovery patterns in post-stroke apha- 1. Chomel-Guillaume S, Laloup G, Bernard I (2010) Les aphasies
sia [12]. It should be noted that the lack of an impact of evaluation et reeducation. Elsevier Masson, Issy les moulineaux
rt-PA treatment on recovery cannot be interpreted as a lack 2. Cloutman L, Newhart M, Davis C, Heidler-Gary J, Hillis AE
of an effect of rt-PA on aphasia outcome, since all patients (2009) Acute recovery of oral word production following stroke:
patterns of performance as predictors of recovery. Behav Neurol
eligible for rt-PA were thus treated, instead of being ran- 21:145–153
domized into treatment and non-treatment groups. In 3. Enderby P, Crow E (1996) Frenchay aphasia screening test:
addition, dramatic recovery of aphasia may have occurred validity and comparability. Disabil Rehabil 18:238–240
in some rt-PA treated patients before the D0 ART scoring, 4. Flamand-Roze C, Falissard B, Roze E, Maintigneux L, Beziz J,
Chacon A, Join-Lambert C, Adams D, Denier C (2011) Valida-
which was done after thrombolysis. tion of a new language screening tool for patients with acute
The study also has certain obvious limitations. First, as stroke: the language screening test (LAST). Stroke 42:1224–1229
already stated, the ART should not be used as a diagnostic 5. Inatomi Y, Yonehara T, Omiya S, Hashimoto Y, Hirano T,
test for aphasia, since some of the scored items may also be Uchino M (2008) Aphasia during the acute phase in ischemic
stroke. Cerebrovasc Dis 25:316–323
affected by dysarthria, speech apraxia, buccofacial apraxia, 6. Laska AC, Bartfai A, Hellblom A, Murray V, Kahan T (2007)
ideomotor apraxia, executive dysfunction, or attentional Clinical and prognostic properties of standardized and functional
fluctuations. Second, the 3-month language outcome was aphasia assessments. J Rehabil Med 39:387–392
123
J Neurol (2013) 260:2110–2117 2117
7. Maas MB, Lev MH, Ay H, Singhal AB, Greer DM, Smith WS, analysis of admission CT perfusion scans. AJNR Am J Neuro-
Harris GJ, Halpern EF, Koroshetz WJ, Furie KL (2012) The radiol 31:1661–1668
prognosis for aphasia in stroke. J Stroke Cerebrovasc Dis 11. Pedersen PM, Vinter K, Olsen TS (2004) Aphasia after stroke:
21:350–357 type, severity and prognosis. The Copenhagen aphasia study.
8. Mazaux JM, Orgogozo JM (1982) Echelle d’évaluation de Cerebrovasc Dis 17:35–43
l’aphasie. Editions Scientifiques et Psychologiques, Issy les 12. Plowman E, Hentz B, Ellis C (2012) Post-stroke aphasia prog-
Moulineaux nosis: a review of patient-related and stroke-related factors.
9. Nakase-Thompson R, Manning E, Sherer M, Yablon SA, Gont- J Eval Clin Pract 18:689–694
kovsky SL, Vickery C (2005) Brief assessment of severe lan- 13. Saur D, Ronneberger O, Kummerer D, Mader I, Weiller C,
guage impairments: initial validation of the Mississippi aphasia Kloppel S (2010) Early functional magnetic resonance imaging
screening test. Brain Inj 19:685–691 activations predict language outcome after stroke. Brain
10. Payabvash S, Kamalian S, Fung S, Wang Y, Passanese J, Ka- 133:1252–1264
malian S, Souza LC, Kemmling A, Harris GJ, Halpern EF, 14. Wade DT, Hewer RL, David RM, Enderby PM (1986) Aphasia
Gonzalez RG, Furie KL, Lev MH (2010) Predicting language after stroke: natural history and associated deficits. J Neurol
improvement in acute stroke patients presenting with aphasia: a Neurosurg Psychiatry 49:11–16
multivariate logistic model using location-weighted atlas-based
123