Probability of Regaining Dexterity in the Flaccid
Upper Limb
Impact of Severity of Paresis and Time Since Onset in Acute Stroke
Gert Kwakkel, PhD; Boudewijn J. Kollen; Jeroen van der Grond, PhD; Arie J.H. Prevo, PhD
Downloaded from http://stroke.ahajournals.org/ by guest on July 28, 2017
Background and Purpose—To improve the accuracy of early postonset prediction of motor recovery in the flaccid
hemiplegic arm, the effects of change in motor function over time on the accuracy of prediction were evaluated, and a
prediction model for the probability of regaining dexterity at 6 months was developed.
Methods—In 102 stroke patients, dexterity and paresis were measured with the Action Research Arm Test, Motricity
Index, and Fugl-Meyer motor evaluation. For model development, 23 candidate determinants were selected. Logistic
regression analysis was used for prognostic factors and model development.
Results—At 6 months, some dexterity in the paretic arm was found in 38%, and complete functional recovery was
seen in 11.6% of the patients. Total anterior circulation infarcts, right hemisphere strokes, homonymous
hemianopia, visual gaze deficit, visual inattention, and paresis were statistically significant related to a poor arm
function. Motricity Index leg scores of at least 25 points in the first week and Fugl-Meyer arm scores of 11 points
in the second week increasing to 19 points in the fourth week raised the probability of developing some dexterity
(Action Research Arm Test ⱖ10 points) from 74% (positive predictive value [PPV], 0.74; 95% confidence interval
[CI], 0.63 to 0.86) to 94% (PPV, 0.83; 95% CI, 0.76 to 0.91) at 6 months. No change in probabilities of prediction
dexterity was found after 4 weeks.
Conclusions—Based on the Fugl-Meyer scores of the flaccid arm, optimal prediction of arm function outcome at 6 months
can be made within 4 weeks after onset. Lack of voluntary motor control of the leg in the first week with no emergence
of arm synergies at 4 weeks is associated with poor outcome at 6 months. (Stroke. 2003;34:2181-2186.)
Key Words: cerebrovascular disorders 䡲 prognosis 䡲 upper extremity
A
motion of the paretic arm within the first weeks after
stroke is considered to be a good prognostic sign.2,3,7 Most
likely, the length of time during which one finds a lack of
improvement reflects the intrinsic cerebral damage and
should be seen as an important predictor of poor outcome.7
Knowledge on outcome of the upper limb is of paramount
interest to clinicians to optimize their treatment goals and
to inform patients properly. In those cases in which some
return of dexterity is expected, training the paretic arm is
justified. However, if the prognosis is poor, teaching the
patient to deal with existing deficits may be more realistic,
thus allowing for the use of compensating strategies. The
first aim of the present study was to develop a prediction
model for estimating the probability of obtaining dexterity
of the paretic arm at 6 months in patients with a severe
middle cerebral artery (MCA) stroke. The second objective
was to investigate the effects of poststroke time on
probabilities for developing dexterity at 6 months after
stroke.
lthough prospective epidemiological studies are lacking, findings of a number of longitudinal studies
indicate that, in 30%1 to 66%2,3 of hemiplegic stroke
patients, the paretic arm remains without function when
measured 6 months after stroke, whereas only 5% to 20%
demonstrate complete functional recovery.1,4 To date, most
studies showed that type and localization of stroke and
initial severity of paresis of the upper limb are some of the
best predictors for outcome at 6 months.2–7 The findings
from all longitudinal studies with repeated measurements
in time indicate that recovery of neurological impairment
and disability shows a nonlinear pattern as a function of
time,1– 6 but only few patients show additional improvement after 3 months after stroke. In addition, most studies
found that the optimal prediction of outcome can be made
within 4 to 5 weeks after stroke onset.1–3,6 The absence of
a measurable grip function at ⬇1 month after stroke was
found to be indicative of a poor functional recovery of the
hemiplegic arm,1–3 whereas early return of voluntary
Received December 30, 2002; final revision received May 9, 2003; accepted May 23, 2003.
From the Rehabilitation Centre “de Hoogstraat” (G.K., A.J.H.P) and Department of Neuroradiology (J. vd G.), Rudolf Magnus Institute of
Neuroscience, University Medical Centre, Utrecht, and Department of Physiotherapy, Isala Clinics, Zwolle (B.J.K.), the Netherlands.
Correspondence to Gert Kwakkel, PhD; Universitair Medisch Centrum Utrecht, Department Revalidatie en Voeding, Divisie Hersenen, Heidelberglaan
100, PO Box 85500, 3508 GA Utrecht, The Netherlands. E-mail [email protected]
© 2003 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org
DOI: 10.1161/01.STR.0000087172.16305.CD
2181
2182
Stroke
September 2003
Subjects and Methods
Design
Downloaded from http://stroke.ahajournals.org/ by guest on July 28, 2017
Over a period of 32 months, 102 stroke patients were recruited from
7 hospitals to participate in this prospective cohort study. All patients
except 1 participated in a randomized controlled trial on the effects
of intensive rehabilitation in acute MCA stroke.8 The original study
incorporated 3 treatment arms, all of which received a basic
rehabilitation program formulated on evidence-based guidelines and
consisting of an eclectic approach to neurofacilitation techniques
with emphasis on task-specific goals for the arm and leg. Arm
rehabilitation included functional exercises that facilitated arm and
hand activity such as leaning, punching a ball, grasping, and moving
objects. The key elements in leg rehabilitation were sitting, standing,
and performing weight-bearing exercises during standing and walking, with an emphasis on achieving stability and improving gait
velocity. If treatment at the disability level was not possible,
strengthening exercises for arms and legs were promoted. Each of
the 3 arms received additional therapy, consisting of either extra arm
or leg exercises or air splint immobilization (ie, control group). All
therapy was based on 1-to-1 contact between therapist and patient. In
total, this contact amounted to 1 daily hour of physiotherapeutic and
occupational therapeutic intervention during 5 days a week over a
20-week period. After this period, decisions regarding type of
treatment were made by the relevant stroke management team. The
rehabilitation program can be characterized as intensive, fine-tuned
toward individual patient needs, and representative of today’s approach to care in many of the stroke service facilities around the
world.
Stroke diagnoses were based on the World Health Organization
definitions.9 The research protocol was approved by the ethics
committees of each participating facility.
Stroke Project [OCSP] classification);15 (5) days between stroke and
first assessment; (6) cognitive impairment (MMSE, 0 to 30 points);11
(7) consciousness during initial 24 hours after stroke onset (GCS, 0
to 15 points);12 (8) sitting balance (score 25 points on sitting balance
item of Trunk Control Test);8 (9) ADL score (Barthel Index [BI], 0
to 20 points);16 (10) urinary incontinence (scores 0 or 1 on the BI);16
(11) sensory deficit in the arm as determined by the Thumb-Finding
Test (0 to 3 points);17 (12) Orpington Prognostic Score (1.6 to 6.8
points); (13) homonymous hemianopia18 (no⫽0/yes⫽1); (14) inattention (1 if ⬎2 omissions on the letter-cancellation test);18 (15)
conjugate gaze19 (no⫽0/yes⫽1); (16) social support (no⫽0/yes⫽1);
(17) type of (additional) therapy (air splint, arm, or leg training; 0 to
2)8; and (18) severity and extent of paresis of upper and lower
extremity motor function in arm and leg as assessed by the Motricity
Index (MI)20 and motor parts of the Fugl-Meyer (FM) score.21 In the
MI, muscle strength was measured for upper extremity (MI arm, 0 to
100) in which 100 points represents normal strength and lower
extremity (MI leg, 0 to 100) separately, whereas FM motor scores
were subdivided into FM arm score (including wrist, 0 to 52), FM
hand score (0 to 14 points), and FM leg score (0 to 34 points) in
which the maximum score represents no synergism.
Procedure
The research protocol started within 14 days after stroke onset.
Clinical outcome parameters such as motor function tests (FM and
MI), Thumb-Finding Test, inattention, ARAT, and ADLs were
measured weekly during the first 10 weeks of the study. Final
outcome was defined at 6 months after stroke. With the exception of
the GCS and SAN tests, all measurements were performed by 1
independent investigator (G.K.).
Subjects
Data Analysis
Stroke patients met the following admission criteria. Subjects (1)
suffered a primary ischemic MCA stroke as revealed by CAT or MRI
scan; (2) were 30 to 80 years of age; (3) presented with an impaired
motor function of the upper and lower extremities; (4) lacked a
complicating medical history that may be restrictive to activities of
daily living (ADL); and (5) demonstrated no severe deficits in
communication, memory, or understanding. A speech therapist
assessed the ability to communicate and accepted a cutoff point of
the 50th percentile corrected for age on the Dutch Foundation
Aphasia test (SAN).10 The Mini-Mental State Examination (MMSE)
was applied to screen for cognitive impairment.11 Patients with an
MMSE score of ⱖ24 points were included in the trial.11 Within 24
hours after stroke onset, all patients were assessed by a neurologist.
After the diagnosis was confirmed and the clinical symptoms such as
level of consciousness (Glasgow Coma Scale [GCS])12 were recorded, the neurologist then referred the stroke patient complying
with the selection criteria to the observer (G.K.) for recruitment and
further assessment, which occurred within the initial 2 poststroke
weeks.
To investigate the possible association between return of dexterity on
ARAT (ie, ⬎9 points) and independent variables, univariate logistic
regression analysis was applied, and odds ratios and 95% confidence
intervals (CIs) were calculated. Although Van der Lee et al14 used 6
points as the cutoff, a more conservative approach was used in this
study to ensure the exclusion of possible false positives. With this in
mind, ordinal scaled determinants were dichotomized (0/1) on the
basis of clinical grounds; otherwise, the optimal cutoff point for each
determinant was determined by applying a receiver-operating characteristic. This area under the curve can be interpreted as the
probability of correctly predicting patients on ARAT. On the basis of
sensitivity/1⫺specificity and maximal area under the curve for each
cutoff score, the optimal dichotomization was estimated for each
candidate determinant separately. Finally, to control for change in
the optimal cutoff points as a result of functional recovery, the
receiver-operating characteristic curves for each candidate determinant were recalculated weekly until the 10th week after stroke.
Based on univariate logistic regression analysis, significant determinants were selected for the subsequent development of a multivariate logistic model necessary for the prediction of the return of
some (ie, ARAT ⱖ10 points) or no (ie, ARAT ⱕ9) dexterity 6
months after stroke. Because of the large number of variables with
respect to number of patients involved, the maximum likelihood
estimation of parameters in the multivariate model was conditional
on the basis of a forward, stepwise approach.22 Collinearity between
included determinants was removed if correlation coefficients of the
correlation matrix in the model were ⱖ0.7. Finally, probabilities of
developing dexterity at 6 months after stroke and their 95% CIs were
calculated from the derived multivariate models for each week using
the constants and regression coefficients of the predictor variables in
the following equation:
Dependent Variable
The outcome of hemiplegic arm function was assessed with the
Action Research Arm test (ARAT).13 This unidimensional hierarchical scale14 consists of 19 functional movement tasks that are divided
into 4 domains: grasp, grip, pinch, and gross movement.13 Patients
with scores of ⱕ9 points are not able to perform more than gross
movements such as grasp, grip, and pinch. Therefore, any ARAT
score of at least 10 points indicates some functional ability of the
paretic hand and scored a 1 in the prediction model, whereas patients
who failed to regain some dexterity (ie, ARAT score of ⱕ9 points)
scored a 0.
Independent Variables
Candidate determinants for development of a prediction model
included the following: (1) age; (2) sex; (3) localization of stroke; (4)
type of stroke (ie, according to the Bamford Oxford Community
P⫽1/共1⫹共exp[⫺共B0⫹B1X1⫹B2X2⫹B3X3 . . . ⫹BnXn)]))
Each hypothesis was tested 2 tailed with a level of significance of
0.05.
Kwakkel et al
Probability of Regaining Dexterity in Acute Stroke
TABLE 1. Patient Characteristics Measured Within 14 Days
After Stroke
Group
Total
n
102
Sex, F/M
41/61
Mean (SD) age, y*
65.4 (10.5)
Median MMSE (0–30)*
27 (24–27.5)
Hemisphere of stroke, L/R
41/61
Type of stroke (OCSP)
TACI
55
PACI
33
LACI
14
OPS (1.6–6.8)*
2183
points). From these patients, 11.6% reached a complete
functional recovery in dexterity of the paretic arm (ie, 57
points on ARAT) at 6 months.
Univariate Associations Between Dependent and
Independent Variables
Table 2 shows odds ratios and their 95% CIs as determined by
univariate logistic regression analysis for the second week
after stroke. Of 23 variables, 18 were significantly related to
the probability of the return of dexterity on ARAT obtained 6
months after stroke. The highest odds were found for the total
scores of FM motor scores of upper extremity (UE) (ie, hand
and arm scores together), their subscores (ie, FM hand and
arm score separately), and the MI arm and leg scores.
4.4 (3.6–5.0)
TFT (0–3)*
1 (0–2)
GCS (0–15)*
15 (15–15)
Cognitive disturbances, n/y
Downloaded from http://stroke.ahajournals.org/ by guest on July 28, 2017
Aphasia
27/75
Inattention
51/51
Impairments of vision, n/y
Hemianopia
69/33
Visual gaze deficit
77/25
Urinary incontinence, n/y
52/50
Time between stroke onset and first
assessment (mean), d
MI-arm (0–100)*
7.3 (2.8)
0 (0–30.5)
MI-leg (0–100)*
23 (0–47)
FM arm (0–52)*
7 (4.5–16)
FM hand (0–14)*
0 (0–0.5)
FM leg (0–34)*
11 (6–20)
Sitting balance, n/y
28/74
BI (0–20), %*
6 (3–7.2)
FAC (0–5)*
0 (0–1)
ARAT (0–57)*
0 (0–1)
OCSP indicates Oxford Community Stroke Project; TACI, total anterior
circulation infarcts; PACI, partial anterior circulation infarcts; LACI, lacunar
anterior circulation infarcts; OPS, Orpington Prognostic Score (range, 1.6 to
6.8); TFT, Thumb-Finding test (0 –3); and FAC, functional ambulation categories
(range, 0 to 5; ⬍5 indicates reduced ambulatory capabilities). For the MMSE,
the range is 0 to 30; ⬍23 indicates impaired cognition. The range for the GCS
is 0 to 15; ⬍15 indicates reduced vigilance. The range for the MI is 0 to 200,
and ⬍200 indicates reduced muscle strength. The range for the BI is 0 to 20,
with 19 to 20 indicating independence. For ARAT, the range is 0 to 57; ⬍10
indicates no dexterity.
*Median values (interquartile ranges).
Results
Table 1 presents the main patient characteristics. On the first
assessment, all patients showed a BI of ⱕ45% (ⱕ9 points),
which corresponds to (very) severely disabled.22 None of the
patients demonstrated some dexterity or were able to walk
independently at the end of the first week (ie, week 1). Two
patients died within 6 months after stroke as a result of a
recurrent stroke or oncological comorbidity. At 6 months
after stroke, 33% of the patients were classified as independent (BI, 19 or 20 points). Approximately 38% showed some
recovery in dexterity of the hemiplegic arm (ARAT ⱖ10
Multivariate Modeling
Table 3 shows the variables included in the prediction model
at their optimal cutoff points and the probabilities of achieving dexterity 6 months after stroke. Depending on the
poststroke week of assessment, FM UE alone and MI leg
were included in the multivariate model. The correlation
between the determinants MI leg and FM UE varied from
0.33 for week 2 to 0.37 for week 3, indicating no problems
with near collinearity.
As shown in Table 3, in the first week after stroke onset,
only MI leg was included into the model. On the basis of
regression coefficients and constants, the probability of
achieving some dexterity was estimated at 0.74 if the patient
reached an MI leg score of ⱖ25 points at stroke onset and
0.14 if the MI leg score was ⬍25 points.
After the FM UE scores were included, the model performance in weeks 2 and 3 showed an increase in probabilities
for achieving dexterity of 0.89 (95% CI, 0.71 to 0.97) and
0.90 (95% CI, 0.74 to 0.96), respectively.
From week 4 on, the model was based on FM UE scores
alone, without MI leg scores. A maximal probability of 0.94
(95% CI, 0.74 to 0.99) was achieved in week 4 in those
patients who reached an FM UE score of ⱖ19 points),
whereas a probability of 0.09 (95% CI, 0.04 to 0.19) was
found for those patients who did not reach this level at week
4 after stroke.
In addition, it was found that the optimal cutoff point for FM
UE, which changed from 11 points at week 2 to 19 points at
week 4, was highly associated with motor control of the paretic
hand and the ability to move in postsynergic patterns. Further
analysis showed that the FM UE scores for each week were
strongly associated with the FM hand scores. Correlation coefficients between FM arm and FM hand scores increased from
0.58 (P⬍0.001) in week 1 to 0.93 (P⬍0.001) in week 4 on.
Recalculating the probabilities from 5 weeks on showed no
significant change in time.
Discussion
It is important to note that this study was restricted to a
homogeneous group of patients participating in a 3-arm randomized intervention trial8 with first-ever severe stroke in the
territory of the MCA resulting in complete hemiplegia. In most
cases (76%), this hemiplegia incorporated a paralysis of the
upper limb at onset. For this group of patients in particular, early,
2184
Stroke
September 2003
TABLE 2. Acute (Within 2 Weeks) Impairments and Disabilities Associated With
Outcome of Dexterity at 6 Months After Stroke as Determined by Logistic
Regression Coefficients
n⫽102
Determinant
Odds Ratio
95% CI
P
Downloaded from http://stroke.ahajournals.org/ by guest on July 28, 2017
Sex (M/F)
0.59
0.34–1.02
0.060
Age (0⬍70; 1ⱖ70)
0.87
0.38–2.01
0.750
Hemisphere of stroke (L/R)
0.51
0.30–0.89
0.018
Type of stroke (OCSP)
0.76
0.65–0.90
0.001
Level of consciousness at onset (GCS)
0.97
0.94–0.99
0.019
Days between stroke onset and first assessment
0.94
0.89–0.98
0.009
MMSE, no/yes
0.78
0.47–1.31
0.363
Visual gaze deficit
0.26
0.10–0.71
0.008
Homonymous hemianopia
0.29
0.13–0.68
0.004
Visual inattention
0.26
0.13–0.53
⬍0.001
Thumb-Finding test score
0.53
0.37–0.75
⬍0.001
OPS
0.85
0.77–0.94
0.001
Urinary incontinence (0⬍2; 1⫽2)
0.36
0.19–0.74
0.003
Sitting balance (0⬍25; 1⫽25)
0.87
0.55–1.38*
MI arm score (0⬍11; 1ⱖ11)
28.33
9.18–87.46*
⬍0.001
0.553
MI leg score (0⬍25; 1ⱖ25)
24.80
7.96–77.30*
⬍0.001
FM arm (hand included) score (0⬍11; 1ⱖ11)
36.40
10.44–126.97*
⬍0.001
FM arm (hand excluded) score (0⬍10; 1ⱖ10)
22.71
7.63–67.63*
⬍0.001
FM hand score (ⱖ1⫽1)
19.77
3.25–175.8*
0.002
FM leg score (ⱖ12⫽1)
7.00
2.70–18.72
⬍0.001
BI (at admission)
1.39
1.18–1.64
⬍0.001
Social support, no/yes
0.71
0.42–1.19
0.191
Treatment type†
1.96
1.16–3.31
0.012
OCSP indicates Bamford Classification according to the Oxford Community Stroke Project; GCS, Glasgow
Coma Scale (1–15); TFT, Thumb-Finding Test (0–3); OPS, Orpington Prognostic Score (1.6– 6.8); MMSE,
Mini-Mental State Examination (0–30); MI, Motricity Index (total score 0–200); FM, Fugl-Meyer score (total
score 0 –114; motor part of the FM score used only); BI, Barthel Index (0–100).
*Skewness caused by the few individuals (ie, minimally 5) in the class that represented the false
negatives.
†Treatment type: air splint⫽0; leg training⫽1; arm training⫽2 (see Reference 8 for details).
accurate prediction of dexterity is often perceived as difficult and
unreliable. The present study shows that 62% failed to achieve
some dexterity at 6 months, indicating that the prognosis for
functional outcome in MCA stroke is poor.
The present study demonstrates some important clinical findings for the rehabilitation management of stroke patients with a
virtual paralysis of the upper limb at onset. First, prediction in
the first week after an MCA stroke leading to a flaccid arm can
best be based on muscle strength of the hemiplegic leg. This
finding suggests that patients who developed some voluntary
movement over hip, knee, and/or foot (ie, ⱖ25 points on MI leg)
in the first week after stroke had about a 74% chance of
regaining some dexterity, whereas absence of voluntary leg
movements reduced this probability to 14%. This finding is in
agreement with the results of a prospective study from Wade and
colleagues2 and indicates that severity and extent of infarction in
the territory of the MCA in the acute phase correspond not only
to the severity of paresis of the arm but also to the extent in
which the lower limb is involved. Moreover, Shelton and
Reding23 recently showed that stroke patients with an additional
lower limb plegia within 2 weeks after MCA stroke had a highly
predictable poor outcome for the return of isolated arm or hand
movements after 6 months on the basis of the FM motor scores.
After the first week, however, the strongest clinical factor that
predicts outcome of dexterity at 6 months is severity of paresis
of the arm. In addition, it was found that the optimal prediction
of outcome of dexterity can be made within the first month after
stroke by measuring motor recovery of the upper limb. At the
end of week 4, a probability of ⬇94% was found in those
patients who had an FM UE score of ⱖ19 points. In contrast, the
chance to achieve some dexterity at 6 months dropped to only
9% in those patients who failed to achieve this level of motor
performance within 4 weeks. Remarkably, no further improvement in accuracy of prediction was found after 4 weeks,
suggesting that long-term outcome of dexterity can already be
optimally predicted within this time frame. In agreement with
previous reports,1–3,6 this latter finding suggests that the time
window for predicting the return of dexterity is limited to only 1
month after onset.
Kwakkel et al
TABLE 3.
Probability of Regaining Dexterity in Acute Stroke
Probabilities of Achieving Dexterity at 6 Months After Stroke
ARAT ⱖ 10 at 6 months
Determinants
Model week 1
Optimal cutoff
FM
UE
MI
leg
True Negatives,
n
False Negatives, False Positives, True Positives,
n
n
n
1/(1⫹(EXP(⫺(⫺1.856⫹2.886 䡠 MI leg)))
Not included
MI leg ⱖ25
⫹
32
5
5
14
⫺
Model week 2
Optimal cutoff
Model week 3
Downloaded from http://stroke.ahajournals.org/ by guest on July 28, 2017
Optimal cutoff
Model week 4
Optimal cutoff
1/(1⫹(EXP(⫺(⫺2.470⫹2.512 䡠 FM UE⫹2.070 䡠 MI leg)))
FM UE ⱖ11
MI leg ⱖ25
⫹
⫹
⫹
⫺
56
4
10
26
⫺
⫹
0.40 (0.14–0.73)
⫺
0.08 (0.03–0.18)
1/(1⫹(EXP(⫺(⫺2.303⫹3.258 䡠 FM UE⫹1.204 䡠 MI leg)))
FM UE ⱖ13
MI leg ⱖ33
⫹
⫹
⫹
⫺
52
7
8
28
⫺
⫹
0.25 (0.09–0.53)
⫺
⫺
0.09 (0.03–0.20)
1/(1⫹(EXP(⫺(⫺2.269⫹4.976 䡠 FM UE)))
FM UE ⱖ19
Not included
58
2
6
30
1/(1⫹(EXP(⫺(⫺2.470⫹5.447 䡠 FM UE)))
FM UE ⱖ22
Not included
58
2
4
32
1/(1⫹(EXP(⫺(⫺2.909⫹5.681 䡠 FM UE)))
FM UE ⱖ23
Not included
⫹
Optimal cutoff
55
2
3
32
1/(1⫹(EXP(⫺(⫺2.461⫹5.156 䡠 FM UE)))
FM UE ⱖ24
Not included
⫹
Optimal cutoff
56
2
5
31
1/(1⫹(EXP(⫺(⫺2.434⫹5.141 䡠 FM UE)))
FM UE ⱖ26
Not included
⫹
Optimal cutoff
57
2
5
30
1/(1⫹(EXP(⫺(⫺2.416⫹5.850 䡠 FM UE)))
FM UE ⱖ28
Not included
⫹
Optimal cutoff
0.94 (0.74–0.99)
0.08 (0.03–0.18)
56
1
5
31
⫺
Model week 10
0.94 (0.74–0.99)
0.08 (0.04–0.18)
⫺
Model week 9
0.94 (0.72–0.99)
0.05 (0.02–0.14)
⫺
Model week 8
0.94 (0.74–0.99)
0.06 (0.02–0.14)
⫺
Model week 7
0.94 (0.74–0.99)
0.09 (0.04–0.19)
⫹
Optimal cutoff
0.90 (0.74–0.96)
0.79 (0.54–0.88)
⫺
Model week 6
0.89 (0.71–0.97)
0.51 (0.21–0.81)
⫺
⫹
Optimal cutoff
0.74 (0.41–0.92)
0.14 (0.02–0.25)
⫺
Model week 5
Probability
(95% CI)
0.97 (0.78–1.00)
0.08 (0.03–0.18)
1/(1⫹(EXP(⫺(⫺2.872⫹6.306 䡠 FM UE)))
FM UE ⱖ28
⫹
⫺
Not included
53
1
3
31
0.97 (0.76–1.00)
0.05 (0.02–0.15)
ARAT indicates Action Research Arm test; MI leg, Motricity Index leg (0 –100); Not included, not included into the prediction model; FM UE,
Fugl-Meyer score upper extremity (0 – 66).
2185
2186
Stroke
September 2003
Downloaded from http://stroke.ahajournals.org/ by guest on July 28, 2017
It is important to note that the optimal cutoff point for FM UE
and MI leg scores is not fixed in time but is dependent on the
poststroke time of assessment. Apparently, a gradual improvement in motor function during the first 2 months after stroke
onset is required to develop some dexterity at 6 months. This
finding underscores that severity of hemiparesis and lack of
substantial improvement of motor function in the first weeks are
factors associated with poor outcome. Therefore, the amount of
change in motor function, particularly within the first weeks,
may be regarded as an independent factor for the probability of
developing a long-term functional upper limb, reflecting the
initial severity of intrinsic cerebral damage and the process of
neurological recovery. Conversely, the above findings suggest
that, in patients without some dexterity at onset in whom motor
recovery is lacking in the first 4 weeks, the outcome of the upper
limb is likely to be poor at 6 months.
Future studies should focus on understanding the mechanisms
that define the critical time window of functional recovery after
stroke. Better understanding of the different time frames for
mechanisms that contribute to functional recovery such as
plasticity, a gradual reversal of diaschisis,24 and behavioral
mechanisms that allow compensation strategies25 may have a
significant impact on rehabilitation management of patients.
The present study predicted dexterity established at 6 months
after stroke by analyzing repeated measurements of the variables
of interest at fixed times after stroke. This approach allows us to
determine the impact of early change in determinants such as
strength on the accuracy of outcome. In our opinion, this
approach realistically reflects the clinical practice of repeated
observations in time. Therefore, we recommend that in future
studies repeated measurement designs be applied to such issues
as the way in which recovery of clinically relevant determinants
may independently affect the final recovery of outcome.
The present study had some limitations. Because of the
homogeneous patient population used, external validity may
have suffered. In addition, all subjects recruited to participate in
this cohort also participated in a 3-group randomized controlled
trial.8 The applied treatment is representative of today’s approach to care in many of the stroke service facilities around the
world. Small differences were found in favor of a 20-week upper
extremity training program.8 This finding is reflected in the
significant odds found for the treatment arm in the univariate
logistic regression analysis. However, the term “treatment type”
as a candidate determinant was not included in the multivariate
model for predicting functional outcome of the upper limb at 6
months, indicating that prediction results were not affected by
the type of treatment to which patients were randomized.
Obviously, the limited time window for the prediction of
regaining dexterity compared with the effects of upper limb
treatment beyond 1 month suggests that the differences in effects
are restricted to those patients with some return of dexterity. This
finding is in agreement with the results of several other randomized controlled trials.26,27 However, both limitations (ie, homogeneity and the intervention) may have resulted in an underestimation of the probabilities of predictions.
Acknowledgments
This study was a part of a research project supported by a grant from
the Netherlands Heart Foundation (reference No. 93.134). We
gratefully acknowledge J.W.R. Twisk, PhD (Institute for Research in
Extramural Medicine, VU University Medical Centre), for providing
statistical advice and Els van Deventer (Department of Neurology)
for collecting and copying relevant references.
References
1. Heller A, Wade DT, Wood VA, Sunderland A, Langton Hewer RL. Arm
function after stroke: measurement and recovery over the first three
months. J Neurol Neurosurg Psychiatry. 1987;50:714 –719.
2. Wade DT, Langton Hewer RL, Wood VA, Skilbeck CE, Ismail HM. The
hemiplegic arm and recovery. J Neurol Neurosurg Psychiatry. 1983;46:521–524.
3. Sunderland A, Tinson DJ, Bradley L, Langton Hewer RL. Arm function after
stroke: an evaluation of grip strength as a measure of recovery and prognostic
indicator. J Neurol Neurosurg Psychiatry. 1989;52:1267–1272.
4. Nakayama H, Jørgensen HS, Raaschou HO, Olsen TS. Recovery of upper
extremity function in stroke patients: the Copenhagen Stroke Study. Arch
Phys Med Rehabil. 1995;76:27–32.
5. Loewen SC, Anderson BA. Predictors of stroke outcome using objective
measurement scales. Stroke. 1990;21:78 – 81.
6. Duncan PW, Goldstein LB, Matchar D, Divine GW, Feussner J. Measurement of motor recovery after stroke: outcome assessment and sample
size requirements. Stroke. 1992;23:1084 –1089.
7. Twitchell TE. The restoration of motor function following hemiplegia in
man. Brain. 1951;75:443– 480.
8. Kwakkel G, Wagenaar RC, Twisk JWR, Lankhorst GJ, Koetsier JC. Effects
of intensity of lower and upper extremity training after a primary middle
cerebral artery stroke: a randomised clinical trial. Lancet. 1999;354:191–196.
9. WHO Special Report. Stroke–1989: recommendations on stroke patients:
prevention, diagnosis, and therapy. Stroke. 1989;20:1407–1431.
10. Deelman BG, Liebrand WBG, Koning-Haanstra, van de Burg W. SAN test:
an aphasia test for language comprehension and linguistic usage. In: Construction and Standards. Lisse, Netherlands: Swets and Zeitlinger BV; 1981.
11. Folstein MF, Folstein SE, McHugh PR. Mini-Mental State: a practical
method for grading the cognitive state of patients for clinician. J Psychiatr Res. 1975;12:189 –198.
12. Teasdale G, Murray G, Parker L, Jennett B. Adding up the Glasgow
Coma Scale. Acta Neurochir. 1979;28(suppl 28):S13–S16.
13. Lyle RCA. Performance test for assessment of upper limb function in physical
rehabilitated treatment and research. Int J Rehabil. 1981;4:483–492.
14. van der Lee JH, Roorda LD, Beckerman H, Lankhorst GJ, Bouter LM.
Improving the Action Research Arm test: an unidimensional hierarchical
scale. Clin Rehabil. 2002;16:646 – 653.
15. Bamford J, Sandercock P, Dennis MS, Burn J, Warlow CP. Classification
and natural history of clinically identifiable subtypes of cerebral
infarction. Lancet. 1991;337:1521–1526.
16. Collin C, Wade DT, Davies S, Horne V. The Barthel ADL Index: a
reliability study. Int Disabil Stud. 1988;10:61– 63.
17. Presscott RJ, Garrway WM, Akthar AJ. Predicting functional outcome
following acute stroke using a standard clinical examination. Stroke.
1982;13:641– 647.
18. Lezak MD. Neurophysiological Assessment. 3rd ed. New York: Oxford
University Press; 1995.
19. Kelley RE, Kovacs AG. Horizontal gaze paresis in hemispheric stroke.
Stroke. 1986;17:1030 –1032.
20. Collin C, Wade D. Assessing motor impairment after stroke: a pilot
reliability study. J Neurol Neurosurg Psychiatry. 1990;53:576 –570.
21. Sanford J, Moreland J, Swanson LR, Stratford PW, Gowland C. Reliability of the Fugl-Meyer assessment for testing motor performance in
patients following stroke. Phys Ther. 1993;73:447– 455.
22. Kleinbaum DG, Kupper, LL, Muller KE. Applied Regression Analysis and Other
Multivariable Methods. Boston, Mass: PWS-KENT Publishing; 1988.
23. Shelton F, Reding MJ. Effect of lesion location on upper limb motor
recovery after stroke. Stroke. 2001;32:107–112.
24. Nudo RJ, Plautz EJ, Frost SB. The role of adaptive plasticity in recovery of
function after damage to motor cortex. Muscle Nerve. 2001;24:1000–1019.
25. Cirstea MC, Levin MF. Compensatory strategies for reaching in stroke.
Brain. 2000;123:940 –953.
26. Sunderland A, Tinson DJ, Bradley EL, Fletcher D, Langton-Hewer R,
Wade DT. Enhanced physical therapy improves recovery of arm function
after stroke: a randomized controlled trial. J Neurol Neurosurg Psychiatry. 1992;55:530 –535.
27. Lincoln NB, Parry RH, Vass CD. Randomized, controlled trial to evaluate
increased intensity of physiotherapy treatment of arm function after
stroke. Stroke. 1999;30:573–579.
Probability of Regaining Dexterity in the Flaccid Upper Limb: Impact of Severity of
Paresis and Time Since Onset in Acute Stroke
Gert Kwakkel, Boudewijn J. Kollen, Jeroen van der Grond and Arie J.H. Prevo
Downloaded from http://stroke.ahajournals.org/ by guest on July 28, 2017
Stroke. 2003;34:2181-2186; originally published online August 7, 2003;
doi: 10.1161/01.STR.0000087172.16305.CD
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2003 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/34/9/2181
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/