1255
Ischemic Stroke: Relation of Age, Lesion Location, and Initial
Neurologic Deficit to Functional Outcome
Stephen N. Macciocchi, PhD, Paul T. Diamond, MD, Wayne M. Alves, PhD, Tracie Mertz, PhD
ABSTRACT. Macciocchi SN, Diamond PT, Alves WM,
Mertz T. Ischemic stroke: relation of age, lesion location, and
initial neurologic deficit to functional outcome. Arch Phys Med
Rehabil 1998;79:1255-7.
Objective: Establish the relation between age, gender, initial
neurologic deficit, stroke location, prior stroke, hemisphere of
stroke, and functional outcome in ischemic stroke.
Design: Single group, multivariate, repeated measures de-
sign with 327 persons having ischemic stroke recruited from 20
participating centers.
Setting: Twenty European stroke centers.
Patients: Consecutive admissions of men and women be-
tween the ages of 40 and 85yrs with a hemispheric stroke
caused by middle cerebral artery ischemia and a Unified
Neurological Stroke Scale score of 5 to 24.
Interventions: Inpatients enrolled in the trial received tradi-
tional rehabilitation therapies including physical therapy, occu-
pational therapy, and speech therapy when appropriate.
Main Outcome Measures: Barthel Index computed at 7 to
10 days and 3 months poststroke.
Results: Positive functional outcomes were significantly
related to the absence of prior strokes, a younger age, a less
severe initial neurologic deficit, stroke involving cortical struc-
tures, and dominant (left hemisphere) lesions.
Conclusions: Despite some inconsistencies in existing litera-
ture, standardized prospective examination of outcome after
stroke clearly demonstrated the effect of age, initial severity of
stroke, and lesion location as predictors of functional outcome.
© 1998 by the American Congress of Rehabilitation Medi-
cine and the American Academy of Physical Medicine and
Rehabilitation
TROKE IS ONE OF the most frequently occurring, dis-
Sincreased
abling, and costly diseases. The prevalence of stroke,
rate of survival, and expense of rehabilitation have
resulted in pressure to identify valid, reliable predictors of
outcome. In this study, several predictors of functional outcome
in stroke are discussed and empirically examined. Although
many predictors of outcome have been studied, sometimes with
conflicting results, we selected variables with conceptual and
empirical merit despite mixed findings in the literature.
Some authors have shown that increased age predicts poor
outcome after stroke, 1,2 whereas other researchers maintain that
From the Department of Physical Medicine and Rehabilitation (Drs. Macciocchi,
Diamond, Mertz), the Neuroclinical Trials Center, Virginia Neurological Institute
(Drs. Macciocchi, Alves), the Department of Neurological Surgery (Drs. Diamond,
Alves), and the Department of Psychiatric Medicine (Dr. Macciocchi), University of
Virginia Health Sciences Center, Charlottesville, VA.
Submitted for publication January 9, 1998. Accepted in revised form June 1, 1998.
No commercial party having a direct financial interest in the results of the research
supporting this article has or will confer a benefit upon the authors or upon any
organization with which the authors are associated.
Reprint requests to Steven N. Macciocchi, PhD, Department of Physical Medicine
and Rehabilitation, Box 522, HSC, Charlottesville, VA 22908
© 1998 by the American Congress of Rehabilitation Medicine and the American
Academy of Physical Medicine and Rehabilitation
0003-9993/98/7910-481853.00/0
age is not an independent predictor of functional outcome. 3
Nonetheless, Alexander I found initial severity of stroke and age
to be the most powerful predictors of functional recovery.
Similarly, Kotila's group 4 showed that age greater than 65 yrs
had a significant negative impact on discharge from the
hospital, adequate performance of activities of daily living
(ADL), and return to work. Other researchers have found
survival to be decreased in each successive age group)
Differences in association between age and outcome across
studies might be due to the correlation of age with comorbidi-
ties such as medical, psychosocial, and psychiatric disorders,
which may not emerge as independent predictors of outcome. 1.6
For example, history of prior stroke, as well as a history of
hypertension, diabetes, or cardiac disease, have been associated
with poorer outcome after stroke, 6,7 and the impact of these
comorbid disorders may be greater in older persons.
In addition to age, initial stroke severity appears strongly
related to functional outcomes. 1,6,8-1°Initial stroke severity and
side of lesion have also been shown to interact. Persons with
severe functional impairment on admission following right
hemisphere lesions appear to demonstrate less improvement
than those with left hemisphere lesions. 1 Although these
findings are not unique, 11 some researchers have not found a
difference in outcome related to lesion location, 2 and recent
literature reviews suggest that hemisphere of stroke does not
predict outcome. 6
Despite the need for valid outcome data, differences in
methodology complicate the interpretation of results across
studies. Problems frequently identified in stroke studies include
heterogeneity of the stroke samples, including variability in
diagnosis (cerebral hemorrhage vs ischemia), mixing of single
and recurrent stroke, failure to control for prior stroke, and
combining strokes of different acuity, for example, acute
hospitalization and rehabilitation. 7,a2 In addition to these con-
cerns, studying only the most severely affected stroke patients
is problematic when improved scores on functional measures
could be attributed to statistical regression to the mean rather
than to genuine functional improvement. Moreover, the instru-
ments used to measure outcomes are not necessarily quantita-
tively or qualitatively interchangeable, thereby making study
comparisons difficult. Finally, many stroke studies are retrospec-
tive, quasiexperimental designs which enhance potential threats
to internal validity (selection bias, for example). Retrospective
studies that do not take into account patient selection and
attrition due to mortality run the risk of biased results and
misrepresentation of the larger stroke population.
Because findings on the effects of age and lesion location on
outcome are in conflict, the present study was undertaken to
prospectively identify predictors of stroke outcome. Examined
variables included patient's age, lesion location (right vs left;
cortical vs subcortical), initial neurologic deficit, history of
previous stroke, and comorbid medical disorders. To control for
methodologic variability, standardized protocols were used to
study participants from stroke onset to 3 months poststroke.
Despite some differences of opinion in the literature, we
hypothesized that age, history of prior stroke, lesion location,
Arch Phys Med Rehabil Vol 79, October 1998
1256 OUTCOME AFTER ISCHEMIC STROKE, Macciocchi
and initial neurologic deficit would be strong predictors of
functional outcome at 3 months.
METHOD
Subjects
All patients admitted at 20 participating European sites with
a diagnosis of ischemic stroke were considered for inclusion.
Inpatients between the ages of 40 and 85 years with a
hemispheric stroke due to focal brain ischemia in the middle
cerebral artery territory and a deficit between 5 and 24 on the
Unified Neurological Stroke Scale ( U N N S ) 13 w e r e considered
for inclusion. Patients with a history of coma, uncontrolled
chronic hypertension (systolic >210 mmHg or diastolic
>120mmHg), intracranial hemorrhage on computed tomogra-
phy (CT) scan (ie, subarachnoid hemorrhage or gross intracere-
bral hemorrhage), or with one or more epileptic seizures not
attributable to the acute ischemic stroke were excluded. The
sample of eligible patients were Caucasian (n = 328), with the
men (n = 179) having a mean age of 68.5 yrs (SD = 10.3) and
the women (n = 149) a mean age of 72.2 yrs (SD = 10.0).
Ninety percent of all patients with moderate or severe
disability, defined by the Glasgow Outcome Scale, were
involved in a formal rehabilitation program. All patients with
moderate and severe disability received physical therapy, 72%
received occupational therapy, and 47% received speech therapy.
Dose and duration of treatment varied depending on functional
skill deficits.
Procedure
Patients who met the inclusion criteria completed a standard-
ized protocol involving medical and neurological examinations,
various scaling procedures including the UNSS, 13 the Barthel
Index (BI), 14 and radiologic procedures (CT). The UNSS is a
clinical rating scale with a range of 0 to 32 used by clinicians to
rate level of consciousness, orientation, language, and motor
dysfunction including ann, hand, and leg movements and gait;
motor items account for 68% of its total score) 3 The B114 is a
well-established measure of functional skills with a range of 0
to 100 that has been widely applied in outcome studies and
indexes 10 functional skill areas, such as feeding, transfers,
bathing, walking and other ADL.
Medical-neurological examinations and scaling procedures
were completed 7 to 10 days poststroke (T1) and 3 months
poststroke (T2) by investigators at individual sites. Procedures
included recording relevant variables such as age and comorbid
medical disorders (hypertension, diabetes, cardiac disease), as
well as neurologic status (UNSS) and functional status (BI).
Radiologic procedures were analyzed by radiologists and coded
by investigative staff. Hemisphere of stroke and involvement of
cortical and/or subcortical structures were recorded.
Statistical analysis. Analysis of variance was used to
examine the effect of lesion side on functional outcome. Lesion
side served as the independent variable. The distribution of
frequency of lesion side and location in the sample was
analyzed using X2 analyses. Finally, logistic regression was
used to delineate a model of functional outcome and odds ratios
(ORs) were computed for each variable. For example, the OR
was computed comparing the odds of a good outcome given the
patient had a left side lesion compared to the odds of a good
outcome if the patient had a right side lesion.
RESULTS
The X 2 analyses revealed no differences in the frequency of
cerebral lesions in the right and left hemispheres (49.7% and
Arch Phys Med Rehabil Vol 79, October 1998
50.3%, respectively). Temporal lesions were more commonly
associated with dominant stroke (p < .003), but there was no
significant difference in the distribution of frontal and parietal
lesions in left and right hemisphere strokes. Finally, cortical and
subcortical lesions were equally common in both right and left
stroke groups (p < .884). Comparison of lesion location (hemi-
sphere) at T1 did not reveal significant differences on the UNSS
or BI.
At T2 analysis of variance revealed no effect for lesion side
on the UNSS or BI. Analysis of change in UNSS and BI from T1
to T2 revealed a similar nonsignificant effect for lesion side. In
contrast, overall change in BI from T1 to T2 (mean = 21.7) was
significant (t = 14.8, p < .001), while change in UNSS from T1
to T2 (mean = 3.2) was not significant.
Finally, a stepwise logistic regression was used to identify
variables significantly related to functional outcome and the OR
or risk of a poor outcome was calculated. The BI was
dichotomized into good (>60) and poor (<60) outcomes, with
criteria for good outcome established by center participants.
The dichotomizing of outcome was established empirically
with clear demarcation of patient by clusters above and below
BI = 60. The mean BI scores are listed in table 1. Variables
significantly related to outcome included age, history of prior
stroke, initial neurologic deficit, and lesion location (table 2).
Risk of a poor outcome increased with age, and patients with a
prior stroke also were at greater risk for a poor outcome
(OR = .28). Left cerebral hemisphere lesions were associated
with good outcomes (OR = 5.46) as were cortical lesions
(OR = 3.23).
DISCUSSION
When ischemic stroke was systematically studied over a
period of 3 months, variables including age, history of prior
stroke, initial neurologic deficit, and lesion location were found
to be significantly related to functional outcome. Comorbid
medical disorders (hypertension, diabetes, cardiac disease)
were not found to be significantly related to functional outcome.
Unfortunately, the relationship between neurorehabilitative
therapies and outcome was obscured by variability in dose and
duration of the therapeutic interventions. Because of design
limitations this study did not reach conclusions regarding the
impact of specific rehabilitation interventions on functional
outcome.
In this study, multiple sites were used to recruit stroke
patients, which reduces the bias introduced by a single center or
enrollment site. Standardized inclusion/exclusion criteria pro-
vided a homogeneous stroke population with generally equiva-
lent frequencies of dominant and nondominant ischemic lesions
of comparable size. For example, the frequency of right and left
hemisphere lesions were essentially equivalent and lesion
locations were equally distributed in both cerebral hemispheres
were equally distributed. Most importantly, patients were
Table 1: UNSS and BI by Lesion Side and Gender
T1 and T2 Poststroke
UNSS BI
Side of
Lesion/Gender n T1 T2 T1 T2
Left
Male 84 19.6 (9.6) 23.7 (8,6) 46,6 (40.0) 69.5 (35.8)
Female 78 20,2 (9.1) 22.9 (9.3) 47.5 (36.7) 66.9 (37.2)
Right
Male 94 22.3 (7.8) 25.2 (7.2) 52.7 (39.6) 74.1 (31.4)
Female 71 20.0 (8.5) 23.9 (8.0) 40,0 (37,1) 62.8 (36.2)
Values reported as mean (SD),
 OUTCOME AFTER ISCHEMIC STROKE, Macciocchi
Table 2: Relation of Predictor Variables to BI Score
at 3 Months Poststroke
P severe, initial neurological deficits have the least chance of
Estimate SE Value Odds Ratio
Intercept 4.72 2,080 .023
UNSS at day 7-10 .40 .058 <.001 3.16
Age -.17 .037 <.001 70yrsvs60yrs .19
80yrs vs 60yrs .40
Side lesion 1.70 .527 . 0 0 1 Left vs right 5.46
Cortical lesion 1.17 .551 .034 Cortical vs other 3.23
Stroke experience -1.28 .631 .042 Prior stroke vs none .28
Abbreviation: SE, standard error of the estimate.
followed prospectively from onset of stroke to 3 months after
their initial hospitalization, a period of time when much, but not
all, restitution in function is observed.
Our results support several previous findings. First, age did
have a significant effect on functional outcome. Moreover, this
effect appeared generally linear. Somewhat surprisingly, comor-
bid medical disorders did not appear to significantly contribute
to functional status poststroke. This finding is encouraging and
suggests that patients with comorbidities are likely to demon-
strate functional recovery despite their medical disorders.
Previous research on side of lesion has yielded contradictory
results. Despite reviews that found no differences in outcome
between dominant and nondominant cerebral lesions, 6 our
study supports findings obtained by other researchers who
found poorer functional outcomes in patients with nondominant
(right hemisphere) strokes. TM The difference in outcome does
not appear to be related to differential lesion size and location in
right and left hemisphere stroke, since no significant differences
in lesion distribution was apparent across groups. Although
speculative, patients with nondominant strokes are known to
evidence neurobehavioral problems, such as perceptual deficits,
anosognosia, and abulia that may play a role in determining
functional skill acquisition poststroke. Further research will be
needed to examine this hypothesis more closely.
Another interesting finding was the change in the BI over
time relative to the UNSS, The lack of change in the UNSS may
be due to the different focus of the instruments, with the BI
primarily focused on functional skills and the UNSS on motor
function including power, tone, and movement. Persons with
strokes may be able to demonstrate considerable functional
gains despite continued motor impairment. Our findings sug-
gest investigators should be sensitive to the measurement
properties of instruments used to document recovery post-
stroke.
The strengths of this study included standardized inclusion
criteria, systematic follow-up, and standardized measurement
techniques that allowed for an adequate test of previous
1257
findings. The model that emerged makes theoretical as well as
clinical sense. Our data show that older persons with more
significant functional improvement. More importantly, a history
of prior stroke was strongly related to risk of impaired
functional skills at 3 months after a second stroke. Patients with
left hemisphere and cortical lesions showed more favorable
odds of a positive outcome.
Future studies can examine the relation between the variables
validated in the present study and other variables such as degree
of motor impairment, cognitive dysfunction, neurobehavioral
impairment, and dose and frequency of rehabilitation interven-
tions. Prospective designs with reliable measurement standards
appear to offer the most value when studying these complex
problems.
References
1. Alexander MP. Stroke rehabilitation outcome: a potential use of
predictive variables to establish levels of care. Stroke 1994;25:
128-34.
2. Kotila M. Four year prognosis of patients under the age of 65
surviving their first ischemic brain infarction. Ann Cfin Res
1986;18:76-9.
3. Samuelsson M, Soderfeldt B, Olsson GB. Functional outcome in
patients with lacunar infarction. Stroke 1996;27:842-6.
4. Kotila M, Waltimo O, Niemi MJ, Laaksonen R, Lempinen M. The
profile of recovery from stroke and factors influencing outcome.
Stroke 1984; 15:1039-44.
5. Chen Q, Ling R. A 1-4 year follow-up study of 306 cases of stroke.
Stroke 1985;16:323-7.
6. Cifu DX, Lorish TR. Stroke rehabilitation. 5. Stroke outcome.
Arch Phys Med Rehabil 1994;75 Suppl:56-60.
7. Anderson TP. Studies up to 1980 on stroke rehabilitation out-
comes. Stroke 1990;21 Suppl II:43-5.
8. Chua KSG, Kong KH. Functional outcome in brain stem stroke
patients after rehabilitation. Arch Phys Med Rehabil 1996;77:
194-7.
9. Wade DT, Hewer RL. Functional abilities after stroke: measure-
ment, natural history and prognosis. J Neurol Neurosurg Psychia-
try 1987;50:177-82.
10. Granger CV, Hamilton BB, Gresham GE. The stroke rehabilitation
outcome study--Part 1: general description. Arch Plays Med
Rehabil 1988;69:506-9.
ll. Lehmann JF, DeLateur BJ, Fowler RS, Warren CG, Arnold R,
Schertzer G, et al. Stroke rehabilitation: outcome and prediction.
Arch Phys Med Rehabil 1975;56:375-82.
12. Jongbloed L. Problems of methodological heterogeneity in studies
predicting disability after stroke. Stroke 1990;21 Suppl II:32-4.
13. Orgogozo JM, Asplund K, Boyseu G. A unified form for neurologi-
cal scoring of hemispheric stroke with motor impairment. Stroke
1992;23:1678-9.
14. Mahoney FL, Barthel DW. Functional evaluation: Barthel Index.
Md Med J 1965;14:61-5.
Arch Phys Med Rehabi! Vol 79, October 1998