1629

REVIEW ARTICLE

Motor Recovery After Stroke: A Systematic Review of the

Literature
Henk T. Hendricks, MD, Jacques van Limbeek, MD, PhD, Alexander C. Geurts, MD, PhD,
Machiel J. Zwarts, MD, PhD
ABSTRACT. Hendricks HT, van Limbeek J, Geurts AC, idly improves during the first few days after stroke. Initial
Zwarts MJ. Motor recovery after stroke: a systematic review of paralysis implies the worst prognosis for subsequent motor
the literature. Arch Phys Med Rehabil 2002;83:1629-37. recovery. Remarkably, the prognostic accuracy of MEPs ap-
pears much higher than that of clinical examination for differ-
Objective: To collect and integrate existing data concerning ent subgroups of patients.
the occurrence, extent, time course, and prognostic determi- Key Words: Evoked potentials, motor; Rehabilitation; Re-
nants of motor recovery after stroke using a systematic meth- view literature [publication type]; Stroke.
odologic approach. © 2002 by the American Congress of Rehabilitation Medi-
Data Sources: A computer-aided search in bibliographic cine and the American Academy of Physical Medicine and
databases was done of longitudinal cohort studies, original Rehabilitation
prognostic studies, and randomized controlled trials published
in the period 1966 to November 2001, which was expanded by
ARLY PREDICTION OF FUNCTIONAL outcome re-
references from retrieved articles and narrative reviews.
Study Selection: After a preliminary screening, internal, E mains an important topic in stroke management and related
research. Functional recovery is influenced by a variety of
external, and statistical validity was assessed by a priori meth-
odologic criteria, with special emphasis on the internal validity. biologic and environmental factors,1 and recovery profiles are
Data Extraction: The studies finally selected were dis- characterized by a high interindividual variability. A recent
cussed, based on the quantitative analysis of the outcome critical review2 of the literature on this matter indicated that
measures and prognostic determinants. Meta-analysis was pur- several clinical and demographic variables may be valid pre-
sued, but was not possible because of substantial heterogeneity. dictors of general functional recovery, including neurologic
Data Synthesis: The search resulted in 174 potentially rel- factors such as consciousness at onset, disorientation in time
evant studies, of which 80 passed the preliminary screening and place, sitting balance, and severity of motor deficits. In
and were subjected to further methodologic assessment; 14 other previously published reviews,3,4 severe hemiparesis has
studies were finally selected. Approximately 65% of the hos- been identified as a negative predictor for functional outcome.
pitalized stroke survivors with initial motor deficits of the Our systematic review is specifically focused on the resto-
lower extremity showed some degree of motor recovery. In the ration of motor deficits after stroke. Motor recovery seems to
case of paralysis, complete motor recovery occurred in less occur predominantly in the first few months after stroke, al-
than 15% of the patients, both for the upper and lower extrem- though some patients may show considerable recovery in later
ities. Hospitalized patients with small lacunar strokes showed phases. The initial grade of paresis is generally regarded as the
relatively good motor recovery. The recovery period in patients most important predictor for motor recovery; however, it is not
with severe stroke was twice as long as in patients with mild yet possible to predict accurately the occurrence and extent of
stroke. The initial grade of paresis was the most important motor recovery in individual patients during the (sub)acute
predictor for motor recovery (odds ratios [OR], ⬎4). Objective phase of their stroke. Some patients may show complete re-
analysis of the motor pathways by motor-evoked potentials covery, whereas in others the degree of paresis may not change
(MEPs) showed even higher ORs (ORs, ⬎20). at all. It is also difficult to give a precise time window for motor
Conclusions: Our knowledge of motor recovery after stroke recovery in individual patients. Despite the vast amount of
in more accurate, quantitative, and qualitive terms is still lim- literature on this subject, no attempt has been made to summa-
ited. Nevertheless, our data synthesis and quantitative analysis rize and integrate the findings of the most valid studies and to
comprises data from many methodologically robust studies, provide a quantitative summary estimate of motor recovery
which may support the clinician in the management of stroke after stroke.
patients. With respect to early prognosis of motor recovery, our Although fully dependent on the literature for its method-
review confirms clinical experience that the initial grade of ologic quality and accuracy of data presentation, we tried to
paresis (as measured on admission in the hospital) is the most answer the following a priori questions: (1) What proportion of
important predictor, although the accuracy of prediction rap- the acute stroke patients exhibiting motor deficits on hospital
admission shows motor recovery, and to what extent and over
what time period poststroke? and (2) Which prognostic factors
for motor recovery can be identified and what is their strength
From the Departments of Rehabilitation Medicine (Hendricks), and of Clinical in terms of recovery probability?
Neurophysiology (Zwarts), and SMK-Research, St. Maartenskliniek (Hendricks, van
Limbeek, Geurts), University Medical Center St. Radboud, Nijmegen, The Nether- METHODS
lands.
No commercial party having a direct financial interest in the results of the research Data Sources
supporting this article has or will confer a benefit upon the author(s) or upon any
organization with which the author(s) is/are associated. Relevant studies were primarily identified by consulting the
Reprint requests to Henk T. Hendricks, MD, Dept of Rehabilitation Medicine, following bibliographic databases: MEDLINE (1966 –Novem-
University Medical Center St. Radboud, Geert Grooteplein 10, 6500 HB Nijmegen,
The Netherlands, e-mail: [email protected].
ber 2001); PsychLit (1967–November 2001); Current Contents
0003-9993/02/8311-7180$35.00/0 (to December 2001); PubMed (to December 2001); EMBASE
doi:10.1053/apmr.2002.35473 (to December 2001); and the Cochrane database for clinical

Arch Phys Med Rehabil Vol 83, November 2002

1630 MOTOR RECOVERY AFTER STROKE, Hendricks

trials (to December 2001). The key words used were: stroke, scored yes/no, which resulted in a maximum sum score of 18
cerebral hemorrhage, cerebral infarction, motor recovery, mo- points. The selected studies were independently analyzed by 2
tor function, impairments, motor control, spontaneous recov- authors (HTH, JvL). In cases of disagreement, consensus was
ery, and rehabilitation. The references used in the retrieved pursued.
articles, as well as in narrative reviews, were also reviewed. Assessment of internal validity was comprised of the fol-
lowing items. The study population had to be homogeneous
Study Selection with respect to diagnosis and disease stage. Therefore, a con-
Preliminary screening. A preliminary screening was con- firmatory diagnosis, by computed tomography (CT) or mag-
ducted to select longitudinal cohort studies in which at least netic resonance imaging (MRI), was required in at least 90% of
some standardized assessment of motor deficits was used at the cases, and patients had to be included in the study within 1
stroke onset and at some point during follow-up. Assessment week after stroke onset. Both prognostic determinants and
by global stroke scales that include evaluation of motor func- outcome measures had to have been assessed by using stan-
tions was accepted. Prognostic studies that evaluated specific dardized tests. Outcome measurements had to be repeated after
diagnostic procedures (particularly evoked potentials) with re- a period of at least 3 months. In the case of heterogeneity of
spect to the prognosis of motor recovery were also included. In initial motor impairments or other possible prognostic variables
addition, the control groups in randomized clinical trials such as stroke type, first or recurrent stroke, subgroup analysis
(RCTs; receiving placebo treatment) were eligible. Only stud- was required. The percentage of patients lost to follow-up was
ies published in English, French, German, and Dutch were not to exceed 20%. The reasons for loss to follow-up had to be
included. Case studies, letters, abstracts, comments, and pre- given to judge selective loss. Furthermore, these cases must
liminary reports were excluded. Each study had to include have been managed adequately in the analysis. Death or stroke
more than 20 patients. recurrence during the study was regarded as loss to follow-up
Assessment of methodologic quality. All studies emerg- if these cases were not yet handled as such. External validity
ing from the preliminary screening were subjected to a system- was assessed as follows. To prevent selection bias, the study
atic review using a checklist with a priori defined methodologic population had to be extracted from a community base or a
criteria. This checklist was constructed according to a system general hospital base. Inclusion and exclusion criteria had to be
originally developed for evaluating RCTs.5,6 The character of described accurately, and the relevant characteristics should
our review, with its special emphasis on prognostic factors, have been given, including age, gender, and comorbidity. Fi-
demanded specific adaptations.7,8 The constructed checklist nally, statistical analysis was assessed. It was first determined
(table 1) assessed internal validity (11 items), external validity whether the statistical analysis was described clearly. The
(3 items), and statistical validity (4 items). All items were requested sample size was calculated in relation to the statis-
tical analysis performed. The ratio of the number of patients
and the number of prognostic determinants had to equal or
exceed 10. There had to be sufficient control for known con-
Table 1: Methodologic Checklist founders in the research design or in the analysis, if applicable.
The statistical analysis also had to be appropriate for the design
Internal validity
used and for the research objective.
1. Diagnosis confirmed by CT or MRI in at least 90% of the
The internal validity was judged as the most critical aspect of
cases
the selected studies, in particular the study entry (2) and the
2. Study entry within 1-wk poststroke onset
homogeneity (7) criteria (table 1). To be included for quanti-
3. Standardized assessment of possible prognostic
tative analysis and final discussion, the study entry and homo-
determinants
geneity criteria had to be fulfilled, and the total minimal inter-
4. Standardized outcome measures
nal validity score had to be at least 9 (maximum 11). The
5. Repeated measurements during the observation period
external and statistical validity criteria were considered of
6. Homogeneity of study sample or subgroups analysis done
secondary importance, and their minimally required sum score
with respect to stroke type, subarachnoidal hemorrhage, and
had to be at least 5 (maximum 7). Studies fulfilling our inclu-
recurrent stroke
sion criteria were used as primary evidence with respect to the
7. Homogeneity of study sample or subgroup analysis done
research questions. Other studies that did not meet the critieria
with respect to initial impairments and severity of stroke
could have been used as secondary evidence.
8. Minimal observation period of 3mo
9. Loss to follow-up ⬍20% Data Extraction
10. Description of relevant characteristics related to loss to
follow-up
In studies in which the results could be expressed as pro-
11. Adequate management of loss to follow-up
portions, odds ratios (ORs) for the occurrence of motor recov-
External validity
ery and their confidence intervals (CIs) were calculated. In the
12. Hospital- or community-based
case of continuous outcome variables, standardized effect sizes
13. Description of inclusion and exclusion criteria
(z scores) were calculated in order to compare (group) differ-
14. Demographic characteristics are given, including age,
ences as a function of the pooled standard deviations: z
gender, and comorbidity
score⫽(xa⫺xb)/PSD, where xa and xb are means of samples a
Statistical validity
and b, and PSD is the pooled standard deviation.9 By conven-
15. Statistical analysis described
tion, the cutoff criterion for considering a particular effect
16. Adequate sample size
clinically relevant was a (group) difference of at least 1 stan-
17. Statistical control for confounding, if applicable
dard score (z⫽1.0). Originally, a meta-analysis was pursued to
18. Appropriate statistical analysis done in relation to design
generate summary estimates.
used Data Synthesis
Abbreviations: CT, computed tomography; MRI, magnetic reso- The numbers of studies resulting from the primary search,
nance imaging. the preliminary screening, and from the systematic method-

Arch Phys Med Rehabil Vol 83, November 2002

 MOTOR RECOVERY AFTER STROKE, Hendricks 1631

Many of the finally selected studies still had serious meth-

odologic limitations (table 2). Several studies failed on the
homogeneity criterion with respect to recurrent stroke15,42,43,57
or stroke type.29 Another frequent limitation was the minimally
requested follow-up period of 3 months.13,24,70 Biller et al13
studied the rate of neurologic recovery only in the first hours
after stroke, and Rapisarda et al70 studied motor recovery after
hand muscle paralysis only in the first 2 weeks after stroke
onset. Conclusions on late motor recovery could not be drawn
from these studies. Two studies15,43 failed on the diagnosis
criterion (confirmation by CT and MRI). Duncan et al25 did not
adequately cope with patients lost to follow-up. Biller13 per-
formed no formal control for confounding, even though their
patient sample was relatively small. Bonita and Beaglehole15
inappropriately calculated ORs. Dominkus et al24 failed to
describe clear inclusion and exclusion criteria.
In table 3, study design, base population, sample size, study
aim, relevant data for our review, type of assessment, fol-
low-up period, sample homogeneity or stratification, and main
outcome are given for the studies that were finally included.
Calculated ORs for the proportions and z scores for the degree
or time course of recovery are given in table 4. Because of
considerable methodologic diversity, such as varying stratifi-
cation procedures, motor assessment, and follow-up periods,
meta-analysis was not possible.
Synthesis of Main Outcomes
With respect to the research questions, a synthesis of the
main outcomes of the finally included studies was as follows.
What proportion of acute stroke patients exhibiting motor
deficits on hospital admission showed motor recovery, and if
so, to what extent and in what time period? Bonita and
Beaglehole15 studied the natural history of hemiparesis after
acute stroke. On admission, 89% of the patients had a hemi-
paresis. The proportion of survivors showing complete or par-
tial motor recovery is given in table 3. There was no differen-
tiation between arm and leg recovery. More detailed data
concerning motor recovery of the upper and lower extremity
can be derived from other community-based studies43,57 (table
3). Jorgensen et al43 studied motor recovery of the lower
extremity in acute stroke patients, stratified according to the
severity of initial motor deficits, as measured on admission in
Fig 1. Flow diagram for final selection of studies. the hospital. The subgroups were as follows: paralysis, severe
paresis, moderate paresis, mild paresis, and no paresis. For
example, motor recovery in patients with moderate leg paresis
ologic assessment for final selection and discussion are sum- was as follows: complete motor recovery at the end of reha-
marized in a flow diagram (fig 1). The primary search resulted bilitation was seen in 44% of the patients, 29% showed partial
in 174 potentially relevant studies: 107 references from the recovery, whereas 20% experienced no changes and 7% dete-
databases and 67 references from retrieved articles. Eighty riorated (table 3). Approximately 65% of all survivors with
studies10-89 passed the preliminary screening and were sub- motor deficits of the leg at admission showed motor recovery.
jected to a systematic, methodologic assessment. Several stud- It was remarkable that a considerable number of the patients
ies26,60,63 were identified in which previously published deteriorated, a fact on which the researchers did not comment.
follow-up data were presented in an alternative way. Such du- The deterioration might be explained by stroke recurrence, too
plicative sources were excluded. Eight studies12,36,38,44,46,50,73,87 early initial assessment in the case of progressive stroke, or
had a minimal internal validity score of 9 but failed criteria 2 initial motor assessment influenced by apraxia or neglect. For
or 7 (table 1). Another study48 had a maximal internal validity the upper extremity, Nakayama et al57 showed in a homoge-
score but failed the minimally requested statistical criteria. neous patient sample with severe arm paresis (little or no active
Ultimately, 14 studies13-15,24,25,27,29,40,42,43,57,61,70,72 met all a pri- movement) on admission that 14% of the patients experienced
ori methodologic criteria and were further reviewed. The re- complete motor recovery and 30% partial recovery (table 3).
sults of the methodologic assessments of these studies are listed As far as a comparison between community-based studies
in table 2. Four community-based studies15,42,43,57 were in- and smaller, more selected samples was possible, there seemed
cluded that comprise the vast majority of the patients. One to exist no great difference with respect to proportions of
RCT29 was selected, from which the data of the control group patients who showed motor recovery (tables 3, 4). One remark
were used. Five studies24,27,40,61,70 were selected that assessed about the generalizability should be made. All selected studies
specifically the prognostic value of evoked potentials in motor concerned patients who were admitted to a hospital. However,
recovery. in general, not all stroke patients will be admitted. For instance,

Arch Phys Med Rehabil Vol 83, November 2002

1632 MOTOR RECOVERY AFTER STROKE, Hendricks

Table 2: Methodologic Assessment of Selected Studies

Internal Validity Score External Validity Score Statistical Validity Score

Reference Total Score (Insufficient Items) (Insufficient Items) (Insufficient Items)

Biller et al13 15 10 (8) 3 2 (16, 17)

Bonita and Beaglehole15 15 9 (1, 6) 3 3 (18)
Dominkus et al24 16 10 (8) 2 (13) 4
Duncan et al25 16 9 (10, 11) 3 4
Escudero et al27 18 11 3 4
Feys et al29 17 10 (6) 3 4
Hendricks et al40 18 11 3 4
Jorgensen et al42 17 10 (6) 3 4
Jorgensen et al43 16 9 (1, 6) 3 4
Nakayama et al57 17 10 (6) 3 4
Samuelsson et al72 18 11 3 4
Rapisarda et al70 17 10 (8) 3 4
Palliyath61 18 11 3 4
Binkofski et al14 18 11 3 4

NOTE: See table 1 for checklist items.

patients with mild motor deficits and no self-care problems will (table 4). Rapisarda et al70 studied motor recovery in 26 acute
often be treated as outpatients. stroke patients with hand paralysis (table 3); 7 patients (27%)
As to the extent of motor recovery, the data from the studies showed partial hand motor recovery at 2 weeks poststroke
appear to be rather vague, particularly with respect to partial (Medical Research Council [MRC] Scale91 score range, 1– 4).
motor recovery. Only few (rather small) selected stud- Which prognostic factors of motor recovery can be identi-
ies13,24,40,70 supplied more detailed information on this issue, fied and what is their magnitude in terms of ORs? Based on
showing a broad range from little to nearly complete motor the studies discussed previously, the initial grade of paresis
recovery. appears the most important clinical predictor for motor recov-
One study was specifically aimed at the temporal aspects of ery, as could be expressed by different ORs for different grades
recovery. Jorgensen et al42 studied the time course of neuro- of initial motor deficits (table 4). For example, Bonita and
logic recovery in acute stroke patients stratified according to Beaglehole15 found a significant association between motor
initial stroke severity, as measured on admission in the hospi- recovery and initial grade of hemiparesis, as measured at
tal. There was a substantial difference in time course between admission. They calculated an OR of 10.8 for recovery after
the strata (tables 3, 4) showing a recovery period approxi- mild versus severe initial paresis, including nonsurvivors. Cal-
mately twice as long for patients with severe paresis (mean, culating motor recovery in survivors only, the OR was still 8.7
15wk) compared with those with only mild paresis (mean, (CI, 4.43–17.06) (table 4). In the study of Jorgensen et al,43 the
6.5wk), resulting in a z score of 12.18. This finding is consis- calculated ORs indicated that a patient with initial mild leg
tent with the studies of Bonita and Beaglehole15 and Duncan et paresis was 4 times as likely to show motor recovery as a
al,25 who both found that most of the overall improvement in patient with initial leg paralysis. Comparable ORs could be
motor functions occurred within the first month after stroke, calculated for the upper extremity. Dominkus et al24 assessed
although some degree of motor recovery continued in some motor recovery in the upper extremity (shoulder flexion, elbow
patients for up to 6 months, especially in the initially severe flexion, handgrip) by means of the Motricity Index.21 A patient
subgroups. with initial paresis was 4.58 as likely to show motor recovery
Valid data concerning late motor recovery appear to be as a patient with initial paralysis. Quantitative analysis of the
rather scarce in the literature considered in the present system- data of Escudero et al27 resulted in a much higher OR of 24.0
atic review. In an RCT29 concerning a therapeutic intervention (CI, 4.50 –127.96) for motor recovery after initial moderate
for improving motor and functional recovery of the upper paresis (MRC score range, 2–3) versus severe paresis or pa-
extremity, mean changes of the Fugl-Meyer Motor Assess- ralysis (MRC score range, 0 –1) (table 4), indicating an extreme
ment90 (FMA) of 12.1 points in the first half year and 4.3 points difference in recovery potential for the muscle group studied
in the second half year poststroke were found in the control (abductor pollicis brevis). It should be argued here that valid
group (50 patients). Because of the limited data presented, z assessment of motor recovery by the MRC is not possible in
scores could not be calculated to quantify differences in the this small muscle group.
degree of recovery during the first versus second half year Five studies24,27,40,61,70 were specifically aimed at determin-
poststroke. Yet substantial secondary evidence concerning late ing the predictive value of evoked potentials for motor recov-
recovery is available.16,35,44,45,56,58,59,68,75,78 These studies sug- ery after stroke. Motor-evoked potentials (MEPs) recorded in
gest that, in some patients, late motor recovery may occur even the early phase (first days) poststroke appeared highly predic-
several months after stroke. Because most of these studies were tive for the occurrence of motor recovery, as could be illus-
conducted in rehabilitation centers, these observations were trated by very high ORs for present versus absent evoked
made in selected patient populations. potentials (table 4). Despite the differences in timing of the
Two studies13,70 specifically addressed early recovery. Biller assessments (on day 1 clinical examination vs several days
et al13 studied the degree of spontaneous neurologic improve- poststroke), evoked potentials seemed to be considerably more
ment (predominantly motor functions) during the first hours predictive than clinical examination. MEPs assess objectively
after admission (table 3). A z score of .26 indicated no sub- and quantitatively the integrity of the motor pathways and may
stantial neurologic recovery in the first hours after admission generate valid prognostic information because postlesional re-

Arch Phys Med Rehabil Vol 83, November 2002

 Table 3: Overview of the Finally Included Studies With Respect to Aim, Follow-Up, and Main Study Outcome

Motor/ Proportion of Motor and

Study Sample Subtracted Neurologic Follow-Up Sample Homogeneity Neurologic Recovery, Degree, or
Reference Design Size Aim of Study Data Assessment Period and Stratification Time Course of Recovery

Biller et al13 Ps, Hb 29 Neurologic recovery in Neurologic recovery in NIH 6h after admission Moderate to severe neurologic 52% recovery, 41% no change, 7%
the first hours the first hours deficit deterioration
Binkofski et al14 Ps, Hb 52 Prognostic value of Motor recovery arm Own scale¶ 6mo Severe paresis (33%) 45% complete or partial recovery
lesion size
Bonita and Ps, Cb 680 Recovery from Recovery from SSS motor 6mo Severe (33.8%) 7% complete, 31% partial
Beaglehole15 hemiparesis hemiparesis part Moderate (25.3%) 22% complete, 31% partial
Mild (30%) 46% complete
Dominkus et al24 Ps, Hb 33 Prognostic value of Motor recovery arm Motricity 2mo Paralysis 36% partial recovery
MEPs Index Paresis 50% partial recovery
Duncan et al25 Ps, Hb 146 Recovery from Recovery from FMA 6mo* Severe motor deficit (31%), Most recovery in the first month,
hemiparesis hemiparesis moderately severe (12%), regardless initial severity, no
moderate (21%), mild (36%) significant recovery after 3mo
Escudero et al27 Ps, Hb 54 Prognostic value of Motor recovery hand MRC 6mo MRC 0–1 (40%) 33% recovery MRC 4
MEPs muscle MRC 2–3 (32%) 92% recovery MRC 4
MRC 4 (28%) —
Feys et al29 RCT, Hb 108 Effect of additional Motor recovery arm FMA 12mo Moderate to severe paresis FMA improvement 12.1 and 4.3 in
sensorimotor first and second half-year,
treatment respectively
Hendricks et al40 Ps, Hb 29 Prognostic value of Motor recovery arm FMA 1–4y Paralysis 10% complete, 13% partial
SEPs and MEPs
Jorgensen et al42 Ps, Cb 1197 Time course neurologic Time course neurologic SSS Until death or end of Very severe (9%) 13 (11.6–14.4) wk†
recovery recovery rehabilitation*; Severe (12%) 15 (13–17) wk†
41⫾46d Moderate (29%) 10.5 (9.5–11.5) wk†
Mild (50%) 6.5 (5.4–7.6) wk†
Jorgensen et al43 Ps, Cb 804 Recovery of walking Recovery from leg SSS motor Until death or end of Paralysis (19%) 14% complete, 31% partial
abilities paralysis and paresis part rehabilitation*; Severe paresis (10%) 34% complete, 41% partial, 4% no
35⫾41d change, 21% deterioration
Moderate paresis (11%) 44% complete, 29% partial, 20%
no change, 7% deterioration
Mild paresis (25%) 76% complete, 19% no change, 5%
deterioration
No paresis (35%) 24% deterioration
Nakayama et al57 Ps, Cb 214 Recovery of upper- Recovery from severe SSS motor Until death or end of Homogeneous, severe arm 14% Complete, 30% partial
extremity function arm paresis part rehabilitation*; paresis (34%)
71⫾53d
MOTOR RECOVERY AFTER STROKE, Hendricks

Palliyath61 Ps, Hb 38 Prognostic value of Motor recovery arm MRC 3mo Paralysis (24%) —
MEPs Paresis (76%) 62% partial recovery
Moderate paresis (67%) 100% Complete or partial recovery
Rapisarda et al70 Ps, Hb 26 Prognostic value of Motor recovery hand MRC 2wk Hand paralysis 27% partial, at 2wk poststroke
MEPs modified㛳
Samuelsson et al72 Ps, Hb 81 Functional outcome in Motor recovery arm Own scale‡ 36mo Severe (17%) At 6mo, 83% of patients with
lacunar infarction and leg Moderate (14%) initial severe deficits showed
Mild (69%) some degree of motor recovery

NOTE. For the Scandinavian Neurological Stroke Scale,92 total score is 0 –58; very severe is 0 –14, severe is 15–29, moderate is 30 – 44, and mild is 45–58. For the scale’s motor part, severe equals little or no active movement; moderate
equals movement against gravity or resistance but limited in range of motion and not in an uncontrolled fashion; and mild equals functionally insignificant impairment of fine movements.
Abbreviations: Ps, prognostic cohort study; Hb, hospital-based; NIH, National Institutes of Health Stroke Scale93 modified (predominantly motor functions); Cb, community-based; SEPs, somatosensory evoked potentials; FMA,
Fugl-Meyer Motor Assessment90; SSS, Scandinavian Neurological Stroke Scale.92
* Weekly assessment.
†
Best neurologic recovery as measured in 95% of the patients, expressed as weeks poststroke (with CIs).
‡
Authors used a motor scale that categorizes motor deficits as severe, moderate, and mild, where severe is severe hemiparesis, cannot elevate the arm and leg against gravity, the hand cannot be used functionally, and walking with
aid is not possible; moderate is moderate hemiparesis, can elevate the arm and leg against gravity, and skilled movement of the hand and walking are clearly affected but the patient can walk with or without aid; and mild is mild
hemiparesis defined as motor function that permits a full range of movement but with reduced strength, and the patient is able to perform all ordinary motor activities without aid.
㛳
Modified MRC Scale range: 0, no movement; 1, movement, only if gravity is removed; 2, weakness against gravity; 3, weakness against slight resistance; 4, weakness against strong resistance; and 5, normal.
¶
Standardized multifactorial scale, assessing force of the arm, grip force, dexterity, coordination, apraxia, motor neglect, and subjective self-assessment. The scale ranges from 0 (normal) to 32 (complete loss of motor functions).
1633

Arch Phys Med Rehabil Vol 83, November 2002

1634 MOTOR RECOVERY AFTER STROKE, Hendricks

Table 4: ORs and z Scores for Relevant Outcome Measures and Predictive Factors

Reference Outcome Measures OR (95% CI) Score

Biller et al13 Early motor recovery: admission neurologic scores vs neurologic scores at 6h — .26
Bonita &
Beaglehole15 Complete motor recovery in initial mild vs initial severe motor deficit 8.70 (4.43–17.06) —
Dominkus et al24 Any motor recovery in initial paresis vs paralysis 4.58 (0.73–28.64) —
Any motor recovery in small (⬍ 0.5cm) vs large (⬎2.5cm) infarction 5.14 (0.47–55.64) —
Any motor recovery in subcortical vs cortical infarction 2.22 (0.36–13.53) —
Any motor recovery in present vs absent MEP 28.00 (2.65–295.73) —
Escudero et al27 Any motor recovery in moderate paresis (MRC 2–4) vs severe paresis (MRC 0–1) 24.00 (4.50–127.96) —
Any motor recovery in the total group: present vs absent MEP 56.00 (9.20–340.52) —
Hendricks et al40 Any motor recovery of upper-extremity paralysis in present vs absent MEP 46 (6.75–313.3) —
Any motor recovery of upper-extremity paralysis in present vs absent SEP 6.66 (1.1–39.2) —
Jorgensen et al42 Time course of neurologic recovery in initial very severe deficit vs mild deficit — 11.3
Time course of neurologic recovery in initial severe deficit vs mild deficit 12.18
Time course of neurologic recovery in initial moderate vs mild deficit 7.52
Jorgensen et al43 Any motor recovery of lower extremity in initial mild paresis vs initial paralysis 4.00 (2.21–7.26) —
Any motor recovery of lower extremity in initial mild paresis vs severe paresis 1.07 (0.53–2.15) —
Any motor recovery of lower extremity in initial severe paresis vs initial paralysis 3.72 (1.71–8.10) —
Nakayama et al57 Any motor recovery of upper extremity in infarction vs hemorrhage 1.28 (0.35–4.70) —
Rapisarda et al70 Any recovery of hand muscle strength in present initial MEP vs absent initial MEP 21.80 (2.54–257.57) —
Palliyath61 Any motor recovery of upper extremity in present vs absent MEP 108.00 (8.78–1327.83) —
Binkofski et al14 Any motor recovery of upper extremity in severe vs moderate paresis 21.17 (2.55–76.53) —

covery appears strongly influenced by a critical residual spared (table 3). The supplied data were insufficient for further sub-
function.14 Especially in noncooperative or severely cogni- group analysis. Nakayama et al57 found no significant differ-
tively impaired patients (eg, global aphasia, apraxia, neglect), ences between patients with and without motor recovery of the
the clinical evaluation is often questionable and thus inconclu- upper extremity for gender, age, and stroke type (infarction or
sive with respect to prognosis. hemorrhage).
The timing of the prognostic assessment appears of great None of the selected studies discriminated hemispheric from
importance. Duncan et al25 prospectively measured the FMA90 brainstem stroke. Secondary evidence concerning the prognos-
up to 6 months in a cohort of 104 acute hospitalized stroke tic value of lesion site can be derived from the prognostic study
patients, stratified according to the severity of initial motor of Turney et al82 (internal validity, 7; validity sum score, 14).
deficits. Regression analysis revealed that, on day 1, the initial This study was aimed at recovery profiles in first-ever hemi-
FMA motor score accounted for only half of the variance in spheric versus brainstem infarctions; 87% of the hospitalized
motor functions at 6 months, whereas the 5-day motor and patients with hemispheric infarctions (n⫽505) showed motor
sensory scores explained 74% of the variance and the 30-day deficits at onset versus 78% of the patients with brainstem
motor score explained 86% of the variance. These results infarctions (n⫽188). At 1-year follow-up, the proportions of
suggest that very early prediction of motor recovery based on patients with residual motor deficits had declined to 57% and
clinical examination alone may be precarious and that the 59%, respectively. An OR of .93 (CI, 0.62–1.37) indicated no
accuracy of prediction may rapidly increase within a few days significant difference between motor recovery in hemispheric
after the stroke. and brainstem infarctions.
The selected studies were also assessed for other prognostic Several other studies provide secondary evidence concerning
factors that might be associated with motor recovery. Bonita our second research question. Katrak et al44,45 studied early
and Beaglehole15 performed subgroup analysis for different voluntary shoulder movements (shoulder shrug) as a prognostic
age and gender; they reported no significant association be- indicator for recovery of hand paralysis in patients who were
tween these patient characteristics and motor recovery. admitted to a rehabilitation center. Initial shoulder shrug in 71
Dominkus et al24 supplied sufficient data to perform subgroup patients (examined 11d poststroke; range, 0 –23d) predicted
analysis for different lesion size and site (table 4). Motor good hand movement and hand function.45 Other clinical de-
recovery seemed better for small than for large lesions terminants, such as prolonged muscular flaccidity31 and lack of
(OR⫽5.14; CI, .47–55.64) and for subcortical compared with early grip strength,79 have also been suggested as negative
cortical lesion (OR⫽2.22; CI, .36 –13.53). Binkofski et al,14 on predictors for motor recovery. However, there is insufficient
the other hand, found no significant correlation between the evidence for the prognostic value of these more specific clinical
initial lesion size (MRI, 1–3d after stroke onset) and recovery signs.
of motor functions of the upper extremity. There was also no
correlation between the changes of the lesion size (as measured CONCLUSIONS
in proton density MRI) and motor deficits. Samuelsson et al72 This review attempted to establish quantitative estimates of
studied motor recovery in a homogeneous sample of patients different aspects of motor recovery after stroke, including
with lacunar stroke. Although the method of motor assessment prognostic factors, by using only primary evidence from meth-
was self-designed by the authors and rather global (table 3), odologically well-conducted studies. From a total of 174 stud-
some conclusions with respect to motor recovery can be drawn. ies, only 14 studies were included for further review and
Even in the subgroup of patients with severe motor deficits, quantitative analysis. The main problem in the final analysis
some degree of motor recovery occurred in 83% of the patients was the pluriformity in the applied assessment procedures for

Arch Phys Med Rehabil Vol 83, November 2002

 MOTOR RECOVERY AFTER STROKE, Hendricks 1635

motor recovery, follow-up periods, and stratification proce- 9. Greene S, Buchbinder R, Glazier R, Forbes R. Systematic review
dures, suggesting that meta-analysis was not possible. How- of randomised controlled trials of interventions for painful shoul-
ever, our data synthesis and quantitative analysis comprised der: selection criteria, outcome assessment and efficacy. BMJ
data from many methodologically robust studies that may 1998;316:354-60.
support the clinician in the management of stroke patients. 10. Andrews K, Brocklehurst J, Richards B, Laycock P. The rate of
recovery from stroke—and its measurement. Int Rehabil Med
Some observations are as follows. Approximately 65% of
1981;3:155-61.
the hospitalized stroke survivors with initial motor deficits of 11. Arac N, Sagduyu A, Binai S, Ertekin C. Prognostic value of
the lower extremity show some degree of motor recovery. For transcranial magnetic stimulation. Stroke 1994;25:2183-6.
the upper extremity, data were insufficient to give an overall 12. Argentino C, Sacchetti M, Toni D, et al. GM1 ganglioside therapy
estimate. In the case of initial paralysis, complete motor recov- in acute ischemic stroke. Stroke 1989;20:1143-9.
ery occurs in less than 15% of the patients, both for the upper 13. Biller J, Love B, Marsh E, et al. Spontaneous improvement after
and lower extremities. Very little valid information is available acute ischemic stroke. A pilot study. Stroke 1990;21:1008-12.
about the extent of motor recovery. As for the time course of 14. Binkofski F, Seitz R, Hackländer T, Pawelec D, Mau J, Freund
recovery, the recovery period in patients with severe stroke was H-J. Recovery of motor functions following hemiparetic stroke: a
twice as long as in patients with mild stroke. With respect to clinical and magnetic resonance morphometric study. Cerebrovasc
early prognosis of motor recovery, our review confirmed clin- Dis 2001;11:273-81.
ical experience that the initial grade of paresis is the most 15. Bonita R, Beaglehole R. Recovery of motor function after stroke.
important predictor, although the accuracy of prediction rap- Stroke 1988;19:1497-500.
idly improves during the first few days after stroke. Initial 16. Broeks J, Lankhorst G, Rumping K, Prevo A. The long-term
paralysis implies the worst prognosis for subsequent motor outcome of arm function after stroke: results of a follow-up study.
recovery. Remarkably, the prognostic accuracy of evoked po- Disabil Rehabil 1999;21;357-64.
tentials appears much higher than that of clinical examination 17. Chae J, Bethoux F, Bohinc T, Dobos L, Davis T, Friedl A.
for different subgroups of patients, although the ORs showed Neuromuscular stimulation for upper extremity motor and func-
tional recovery in acute hemiplegia. Stroke 1998;29:975-9.
wide CIs because of generally small sample sizes. Patients with 18. Chen Q, Ling R. A 1-4 year follow-up study of 306 cases of
small lacunar strokes seemed to show relatively good recovery stroke. Stroke 1985;16:323-7.
profiles compared with larger hemispheric strokes. Perhaps 19. Chen C-L, Tang F-T, Chen H-C, Chung C-Y, Wong M-K. Brain
somewhat counterintuitive for clinical reasoning, there was no lesion size and location: effects on motor recovery and functional
epidemiologic evidence of a systematic difference in recovery outcome in stroke patients. Arch Phys Med Rehabil 2000;81:447-
potential between hemorrhages and infarctions or between 52.
brainstem and hemispheric infarctions. 20. Crow J, Lincoln N, Nouri F, De Weerdt W. The effectiveness of
In conclusion, this review showed that our knowledge of EMG biofeedback in the treatment of arm function after stroke. Int
motor recovery after stroke in more accurate, quantitative, and Disabil Stud 1989;11:155-60.
21. Demeurisse G, Demol O, Robaye E. Motor evaluation in vascular
qualitive terms is still much more limited than it is perceived by hemiplegia. Eur Neurol 1980:19;382-9.
many. The lack of precise and valid epidemiologic data is in 22. Denes G, Semenza C, Stoppa E, Lis A. Unilateral spatial neglect
contrast with the general idea of global predictability of post- and recovery from hemiplegia. A follow-up study. Brain 1982;
stroke motor recovery based on the initial severity of motor 105:543-52.
deficits. The initial motor assessment may be invalid because 23. Dickstein R, Hocherman S, Pillar T, Shaham R. Stroke rehabili-
of apraxia, neglect, or aphasia. For clinical and policy pur- tation. Three exercise therapy approaches. Phys Ther 1986;66:
poses, a much more precise prognosis is required in individual 1233-8.
patients. Hence, further research should focus on a more pre- 24. Dominkus M, Grisold W, Jelinek V. Transcranial electrical motor
cise prediction of the degree of motor recovery after stroke evoked potentials as a prognostic indicator for motor recovery in
stroke patients. J Neurol Neurosurg Psychiatry 1990;53:745-8.
based on physical characteristics, type, size, and site of the 25. Duncan P, Goldstein L, Matchar D, Divine G, Feussner J. Mea-
lesion. Special attention should be paid to the clinical prognos- surement of motor recovery after stroke. Outcome assessment and
tic value of MEPs. sample size requirements. Stroke 1992;23:1084-9.
26. Duncan P, Goldstein L, Horner R, Landsman P, Samsa G,
References Matchar D. Similar motor recovery of upper and lower extremities
1. Bach-y-Rita P, Bach-y-Rita E. Biological and psychosocial factors after stroke. Stroke 1994;25:1181-8.
in recovery from brain damage in humans. Can J Psychol 1990; 27. Escudero J, Sancho J, Bautista D, Escudero M, Lopez-Trigo J.
44:148-65. Prognostic value of motor evoked potential obtained by transcra-
2. Kwakkel G, Wagenaar R, Kollen B, Lankhorst G. Predicting nial magnetic brain stimulation in motor function recovery in
disability in stroke—a critical review of the literature. Age Ageing patients with acute ischemic stroke. Stroke 1998;29:1854-9.
1996;25:479-89. 28. Faghri P, Rodgers M, Glaser R, Bors J, Ho C, Akuthota P. The
3. Hier D, Edelstein G. Deriving clinical prediction rules from stroke effects of functional electrical stimulation on shoulder subluxa-
outcome research. Stroke 1991;22:1431-6. tion, arm function recovery, and shoulder pain in hemiplegic
4. Dombovy M, Sandok B, Basford J. Rehabilitation for stroke: a stroke patients. Arch Phys Med Rehabil 1994;75:23-9.
review. Stroke 1986;17:363-9. 29. Feys H, De Weerdt W, Selz B, et al. Effect of a therapeutic
5. Chalmers T, Smith H, Blackburn B, et al. A method for assessing intervention for the hemiplegic upper limp in the acute phase after
the quality of a randomized control trial. Control Clin Trials stroke. Stroke 1998;29:785-92.
1981;2:31-49. 30. Feys H, Hetebrij J, Wilms G, Dom R, De Weerdt W. Predicting
6. Cook D, Sackett D, Spitzer W. Methodologic guidelines for arm recovery following stroke: value of site of lesion. Acta Neurol
systematic reviews of randomized control trials in health care Scand 2000;102:371-7.
from the Potsdam consultation on meta-analysis. J Clin Epidemiol 31. Formisano R, Barbanti P, Catarci T, De Vuono G, Calisse P,
1995;48:167-71. Razzano C. Prolonged muscular flaccidity: frequency and associ-
7. Laupacis A, Wells G, Richardson W, Tugwell P. Users’ guide to ation with an unilateral spatial neglect after stroke. Acta Neurol
the medical literature. V. How to use an article about prognosis. Scand 1993;88:313-5.
JAMA 1994;272:234-47. 32. Goldstein LB. Common drugs may influence motor recovery after
8. Jenicek M. Meta-analysis in medicine. Where we are and where stroke. The Sygen In Acute Stroke Study Investigators. Neurology
we want to go. J Clin Epidemiol 1989;42:35-44. 1995;45:865-71.

Arch Phys Med Rehabil Vol 83, November 2002

1636 MOTOR RECOVERY AFTER STROKE, Hendricks

33. Gott P, Karnaze D, Fisher M. Assessment of median nerve so- 58. Newman M. The process of recovery after hemiplegia. Stroke
matosensory evoked potentials in cerebral ischemia. Stroke 1990; 1972;3:702-10.
21:1167-71. 59. Olsen T. Improvement of function and motor impairment after
34. Gowland C. Recovery of motor function following stroke: profile stroke. J Neurol Rehabil 1989;3:187-92.
and predictors. Physiother Can 1982;34:77-84. 60. Olsen T. Arm and leg paresis as outcome parameters in stroke
35. Gowland C. Predicting sensorimotor recovery following stroke rehabilitation. Stroke 1990;21:247-51.
rehabilitation. Physiother Can 1984;36:313-20. 61. Palliyath S. Role of central conduction time and motor evoked
36. Gray C, French J, Bates D, Cartlidge N, James O, Venables G. response amplitude in predicting stroke outcome. Electromyogr
Motor recovery following acute stroke. Age Ageing 1990;19:179- Clin Neurophysiol 2000;40:315-20.
84. 62. Pantano P, Formisano R, Ricci M, et al. Prolonged muscular
37. Gray C, French J, Venables G, Cartlidge N, James O, Bates D. A flaccidity after stroke. Morphological and functional brain alter-
randomized double-blind controlled trial of naftidrofuryl in acute ations. Brain 1995;118:1329-38.
stroke. Age Ageing 1990;19;356-66. 63. Pantano P, Formisano R, Ricci M, et al. Motor recovery after
38. Heald A, Bates D, Cartlidge N, French J, Miller S. Longitudinal stroke. Morphological and functional brain alterations. Brain
study of central motor conduction time following stroke. Brian 1996;119:1849-57.
1993;116:1371-85. 64. Parker V, Wade D, Hewer RL. Loss of arm function after stroke:
39. Heller A, Wade D, Wood V, Sunderland A, Hewer RL, Ward E. measurement, frequency, and recovery. Int Rehabil Med 1986;8:
Arm function after stroke: measurement and recovery over the 69-73.
first three months. J Neurol Neurosurg Psychiatry 1987;50:714-9. 65. Parry R, Lincoln N, Vass C. Effect of severity of arm impairment
40. Hendricks H, Hageman G, van Limbeek J. Prediction of recovery on response to additional physiotherapy early after stroke. Clin
from upper extremity paralysis after stroke by measuring motor Rehabil 1999;13:187-98.
evoked potentials. Scand J Rehabil Med 1997;29:155-9. 66. Partridge C, Edwards S, Johnston M. Recovery from physical
41. Horgan N, Finn A. Motor recovery following stroke: a basis for disability after stroke. Lancet 1987;14:373-5.
evaluation. Disabil Rehabil 1997;19:64-70. 67. Powell J, Pandyan D, Granat M, Cameron M, Stott D. Electrical
42. Jorgensen HS, Nakayama H, Raaschou H, Vive-Larsen J, Stoier stimulation of wrist extensors in poststroke hemiplegia. Stroke
M, Olsen T. Outcome and time course of recovery. Part II: Time 1999;30:1384-9.
course of recovery. The Copenhagen Stroke Study. Arch Phys 68. Prevo A, Dijkman M, Le Fevre F. [Impairment and disability in
Med Rehabil 1995;76:406-12. patients with a severe ischemic cerebral infarction at admission to
43. Jorgensen H, Nakayama H, Raaschou H, Olsen T. Recovery of the rehabilitation center and six months after stroke] [Dutch]. Ned
walking function in stroke patients: the Copenhagen Stroke Study. Tijdschr Geneeskd 1998;142:637-40.
Arch Phys Med Rehabil 1995;76:27-32. 69. Rand D, Weiss P, Gottlieb D. Does proprioceptive loss influence
44. Katrak P. Shoulder shrug—a prognostic sign for recovery of hand recovery of the upper extremity after stroke? Neurorehabil Neural
movement after stroke. Med J Aust 1990;152:297-301.
Repair 1999;13:15-21.
45. Katrak P, Bowring G, Conroy P, Chilvers M, Poulos R, McNeil D.
Predicting upper limp recovery after stroke: the place of early 70. Rapisarda G, Bastings E, de Noordhout AM, Pennisi G, Delwaide
shoulder and hand movement. Arch Phys Med Rehabil 1998;79: P. Can motor recovery in stroke patients be predicted by early
758-61. transcranial magnetic stimulation? Stroke 1996;27:2191-6.
46. Knopmann D, Rubens A. The validity of computed tomographic 71. Roth E, Heinemann A, Lovell L, Harvey R, McGuire J, Diaz S.
scan findings for the localization of cerebral functions. The rela- Impairment and disability: their relation during stroke rehabilita-
tionship between computed tomography and hemiparesis. Arch tion. Arch Phys Med Rehabil 1998:79:329-35.
Neurol 1986;43:328-32. 72. Samuelsson M, Söderfeldt B, Olsson GB. Functional outcome in
47. Kotila M, Waltimo O, Niemi M, Laaksonen R, Lempinen M. The patients with lacunar infarction. Stroke 1996;27:842-6.
profile of recovery from stroke and factors influencing outcome. 73. Ganglioside GM1 in acute ischemic stroke. The SASS Trial.
Stroke 1984;15:1039-44. Stroke 1994;25:1141-8.
48. Kusoffsky A, Wadell I, Nilsson B. The relationship between 74. Scheidtmann K, Fries W, Müller F, Koenig E. Effect of levodopa
sensory impairment and motor recovery in patients with hemiple- in combination with physiotherapy on functional motor recovery
gia. Scand J Rehabil Med 1982;14:27-32. after stroke: a prospective, randomised, double-blind study. Lan-
49. La Joie W, Reddy N, Melvin J. Somatosensory evoked potentials: cet 2001;358:787-90.
their predictive value in right hemiplegia. Arch Phys Med Rehabil 75. Shah S, Harasymiw S, Stahl P. Stroke rehabilitation: outcome
1982;63:223-6. based on Brunnstrom recovery stages. Occup Ther J Res 1986;6:
50. Lampl Y, Gilad R, Eshel Y, Sarova-Pinhas I. Neurological and 365-76.
functional outcome in patients with supratentorial hemorrhages. A 76. Sivenius J, Pyorala K, Heinonen O, Salonen J, Riekkinen P. The
prospective study. Stroke 1995;26:2249-53. significance of intensity of rehabilitation of stroke—a controlled
51. Loewen S, Anderson B. Predictors of stroke outcome using ob- trial. Stroke 1985;16:928-31.
jective measurement scales. Stroke 1990;21:78-81. 77. Smith J, Brotheridge S, Young J. Patterns of hemiparesis recovery
52. Logigian M, Samuels M, Falconer J, Zagar R. Clinical exercise in lacunar and partial anterior circulation infarct stroke syn-
trial for stroke patients. Arch Phys Med Rehabil 1983;64:364-7. dromes. Clin Rehabil 2001;15:59-66.
53. McDowell F, Louis S. Improvement in motor performance in 78. Sonoda S, Chino N, Domen K, Saitoh E. Changes in impairment
paretic and paralyzed extremities following nonembolic cerebral and disability from the third to the sixth month after stroke and its
infarction. Stroke 1971;2:395-9. relationship evaluated by an artificial network. Am J Phys Med
54. Miyai I, Blau A, Reding M, Volpe B. Patients with stroke confined Rehabil 1997;76:395-400.
to basal ganglia have diminished response to rehabilitation efforts. 79. Sunderland A, Tinson D, Bradley E, Hewer RL. Arm function
Neurology 1997;48:95-101. after stroke. An evaluation of grip strength as a measure of
55. Miyai I, Suzuki T, Kang J, Kubota K, Volpe B. Middle cerebral recovery and a prognostic indicator. J Neurol Neurosurg Psychi-
artery stroke that includes the premotor cortex reduces mobility atry 1989;52:1267-72.
outcome. Stroke 1999;30:1380-3. 80. Sunderland A, Tinson D, Bradley E, Fletcher D, Hewer RL, Wade
56. Nakamura R, Moriyama S, Yamada Y, Seki K. Recovery of D. Enhanced physical therapy improves recovery of arm function
impaired motor function of the upper extremity after stroke. after stroke. A randomised controlled trial. J Neurol Neurosurg
Tohoku J Exp Med 1992;168:11-20. Psychiatry 1992;55:530-5.
57. Nakayama H, Jorgensen H, Raaschou H, Olsen T. Compensation 81. Trompetto C, Assini A, Buccolieri A, Marchese R, Abbruzzese G.
in recovery of upper extremity function after stroke: the Copen- Motor recovery following stroke: a transcranial magnetic stimu-
hagen Stroke Study. Arch Phys Med Rehabil 1994;75:852-7. lation study. Clin Neurophysiol 2000;111:1860-7.

Arch Phys Med Rehabil Vol 83, November 2002

 MOTOR RECOVERY AFTER STROKE, Hendricks 1637

82. Turney T, Garraway M, Whisnant J. The natural history of hemi- 88. Young J, Forster A. The Bradford community stroke trial: results
spheric and brainstem infarction in Rochester, Minnesota. Stroke at six months. BMJ 1992;304:1085-9.
1984;15:790-4. 89. Yoshimoto T, Ogawa A, Seki H, Kogure T, Suzuki J. Clinical
83. Vang C, Dunbabin D, Kilpatrick D. Correlation between func- course of acute middle cerebral artery occlusion. J Neurosurg
tional and electrophysiological recovery in acute ischemic stroke. 1986;65:326-30.
Stroke 1999;30:2126-30. 90. Fugl-Meyer A, Jaasko L, Leyman I, Olsson S, Steglind S. The
84. Visintin M, Barbeau H, Korner-Bitensky N, Mayo N. A new post-stroke hemiplegic patient. 1. A method for evaluation of
treatment approach to retrain gait in stroke patients through body physical performance. Scand J Rehabil Med 1975;7:13-31.
weight support and treadmill stimulation. Stroke 1998;29:1122-8. 91. Medical Research Council. Aids to examination of the peripheral
85. Walker M, Lincoln N. Factors influencing dressing performance nervous system. London: HMSO; 1976.
after stroke. J Neurol Neurosurg Psychiatry 1991;54:699-701. 92. Lindenstrom E, Boysen G, Christiansen LW, Hansen BR, Nielsen
86. Williams B, Galea M, Winter A. What is the functional out- PW. Reliability of Scandinavian neurological stroke scale. Cere-
come of the upper limp after stroke? Aust J Physiother 2001; brovasc Dis 1991;1:103-7.
47:19-27. 93. Brott T, Holinger AP Jr, Olinger CP, et al. Measurement of acute
87. Wood-Dauphinee S, Shapiro S, Bass E, et al. A randomized trial cerebral infarction: a clinical examination scale. Stroke 1989;20:
of team care following stroke. Stroke 1984;15:864-72. 864-70.

Arch Phys Med Rehabil Vol 83, November 2002