What Determines Good Recovery in Patients With the Most Severe Strokes?
The Copenhagen Stroke Study
Background and Purpose—Even patients with the most severe strokes sometimes experience a remarkably good recovery. We evaluated possible predictors of a good outcome to search for new therapeutic strategies.
Methods—We included the 223 patients (19%) with the most severe strokes (Scandinavian Stroke Scale score <15 points) from the 1197 unselected patients in the Copenhagen Stroke Study. Of these, 139 (62%) died in the hospital and were excluded. The 26 survivors (31%) with a good functional outcome (Barthel Index ≥50 points) were compared with the 58 survivors (69%) with a poor functional outcome (Barthel Index <50 points). The predictive value of the following factors was examined in a multivariate logistic regression model: age; sex; a spouse; work; home care before stroke; initial stroke severity; blood pressure, blood glucose, and body temperature on admission; stroke subtype; neurological impairment 1 week after onset; diabetes; hypertension; atrial fibrillation; ischemic heart disease; previous stroke; and other disabling disease.
Results—Decreasing age (odds ratio [OR], 0.50 per 10-year decrease; 95% CI, 0.25 to 0.99; P=0.04), a spouse (OR, 3.1; 95% CI, 1.1 to 8.8; P=0.03), decreasing body temperature on admission (OR, 1.8 per 1°C decrease; 95% CI, 1.1 to 3.1; P=0.01), and neurological recovery after 1 week (OR, 3.2 per 10-point increase in Scandinavian Stroke Scale score; 95% CI, 1.1 to 7.8; P=0.01) were all independent predictors of good functional outcome.
Conclusions—Patients with the most severe strokes who achieve a good functional outcome are generally characterized by younger age, the presence of a spouse at home, and early neurological recovery. Body temperature was a strong predictor of good functional outcome and the only potentially modifiable factor. We suggest that a randomized controlled trial be undertaken to evaluate whether active reduction of body temperature can improve the generally poor prognosis of patients with the most severe strokes.
The patients with the most severe strokes account for 20% of the total stroke population,1 and they represent a major therapeutic challenge: the majority of these patients die from their stroke, and more than half of the survivors remain severely disabled.2 However, after completed rehabilitation on a dedicated stroke unit, a third of those who survive can be discharged back to their own homes to a somewhat independent life with no or only mild or moderate disability.3 Knowledge of factors of importance for good functional recovery is essential to patient evaluation, treatment planning, guidance of patient and relatives, and eventually also in the search for new therapeutic strategies.
This study evaluates the prognostic importance of various stroke characteristics and social, demographic, and medical factors in patients with the most severe strokes in the unselected population of the Copenhagen Stroke Study.
Subjects and Methods
The Copenhagen Stroke Study is a truly observational, prospective, and community-based study of 1197 unselected patients with acute stroke consecutively admitted to the 60-bed stroke unit of the neurological department at Bispebjerg Hospital, Copenhagen.3 This stroke unit serves all stroke patients admitted to the hospital from a well-defined area of Copenhagen with 239 886 inhabitants. They are admitted without regard to age, severity of the stroke, or condition before the stroke. Hospital care is free, and the stroke admission rate in the Copenhagen area is high, at 88%.4 Diagnostic procedures, treatment, and all stages of rehabilitation take place within the stroke unit. Patients are not discharged until the rehabilitation team judges that further in-hospital improvement is unlikely; referral to other departments or hospitals for further rehabilitation is therefore not required.
In this part of the Copenhagen Stroke Study we included patients with only the most severe strokes1 : those who had a Scandinavian Stroke Scale (SSS) score on acute admission <15 points and who survived. Nine hundred seventy-four patients (81%) were excluded because initial stroke severity was not very severe (SSS score ≥15 points on admission). One hundred thirty-nine (62%) of the patients with a very severe stroke (SSS score <15 points on admission) were excluded because they died in the hospital. The cutoff point of 15 points for the SSS score is arbitrary and was chosen because we have previously shown that patients with an initial stroke severity <15 points generally have a poor prognosis and represent the most severe strokes.1 Thus, a total of 84 patients (7%) who survived a very severe stroke were included in the analyses. These patients were stratified into 2 groups: (1) those who eventually had a good functional outcome, ie, a Barthel Index (BI) score of ≥50 points after completed rehabilitation, and (2) those who had a poor functional outcome, ie, a BI score <50 points after completed rehabilitation. A cutoff of 50 points for the BI score was chosen arbitrarily to dichotomize the variable. This seems justified since only 8% of the patients with a score ≥50 points eventually had to be discharged to a nursing home facility compared with 91% of the patients with a score of <50 points.
Stroke was defined according to the World Health Organization criteria as rapidly developed clinical signs of focal disturbance, lasting >24 hours or leading to death, with no apparent cause other than vascular origin. Subarachnoid hemorrhage was not included.5
The SSS was used to assess neurological impairment6 (Table 1⇓). The SSS evaluates level of consciousness; eye movement; motor power in arm, hand, and leg; orientation; aphasia; facial paresis; and gait on a total score from 0 to 58 (maximum). It was recorded on acute admission, the day after admission, weekly during the hospital stay, and at discharge.
The BI was used to assess functional disability.7 It evaluates basic activities of daily living, such as feeding, grooming, transfer, dressing, toileting, bathing, walking, incontinence of bowel and bladder, and stair walking, on a total score from 0 to 100 (independent functional level) points. The BI was assessed by the nursing and training staff during the first week, weekly throughout hospital stay, and at discharge after completed rehabilitation.
The following factors were considered in the analysis of predictors of good outcome: age; sex; marital status; home help (ie, a professional who helps with cleaning, shopping, etc); stroke severity on admission (SSS); stroke subtype (hemorrhage/infarct); blood pressure, blood glucose, and spontaneous body temperature on admission; diabetes; atrial fibrillation; hypertension; ischemic heart disease; previous stroke; previous transient ischemic attack; daily alcohol consumption; and other disabling comorbidity. The classification of specific factors was as follows: (1) diabetes—patients with known diabetes before stroke and patients with diabetes diagnosed after stroke onset either during the hospital stay or because admission plasma glucose was >11 mmol/L, in accordance with the World Health Organization diagnostic criteria for diabetes8 ; (2) atrial fibrillation—if present on admission ECG; (3) hypertension—on antihypertensive treatment at the time of admission or hypertension diagnosed during the hospital stay by repeated detection of blood pressure ≥160/95 mm Hg; (4) ischemic heart disease—a history of ischemic heart disease or ischemic heart disease diagnosed during the hospital stay; (5) daily alcohol consumption—1 drink per day (12 g alcohol) or more on average; (6) comorbidity—information concerning comorbidity believed to interfere with the patient’s ability to perform basic activities of daily living was obtained on admission and included disabling diseases other than previous stroke (eg, amputation, multiple sclerosis, severe dementia, heart failure, latent or persistent respiratory insufficiency); various comorbidities were not registered separately, and the influence of specific comorbidities was therefore not evaluated; (7) temperature—body temperature on acute admission was recorded with a Diatek model 9000 infrared aural thermometer (Diatek); this device registers tympanic membrane temperature, which correlates well with core body temperature9 ; and (8) stroke subtype—hemorrhage or ischemic infarction as seen on CT scan.10
A priori we believed that young age, female sex, a spouse, no home help before stroke, less severe stroke, infarction as opposed to hemorrhage, high blood pressure, no elevated blood glucose, low spontaneous body temperature, no diabetes, no atrial fibrillation, hypertension, no ischemic heart disease, no previous stroke, previous transient ischemic attack, no daily alcohol consumption, and no other disabling comorbidity were all candidates to be potential predictors of good outcome.
Statistical analysis was done with the SPSS package for Windows.11 Student’s t test was used in univariate analysis for continuous data and the χ2 for noncontinuous data. To evaluate the prognostic importance of multiple factors independently of other factors, a multiple logistic regression model for poor versus good outcome was created. Poor functional outcome was coded as 0, and good functional outcome was given the value of 1. All variables of interest were tested backward to fit the full model with all the potential explanatory variables. Unimportant variables were then removed one at a time until all those remaining in the model contributed with probability value <0.2. The 2 methods of multiple regression analysis, backward and forward regression, often yield the same model, but differences are not uncommon. To select the “best model,” backward logistic regression including only variables with a P value <0.2 was followed by a forward logistic regression including the same variables. The 2 methods yielded similar models. The explanatory power of the equation was determined by the value of correct classification (the percentage of patients correctly classified as achieving a poor or a good functional outcome from the information given by the independent variables included in the final model). The required 2-tailed significance level for all tests was set at 0.05.
Stroke was very severe in 223 (19%) of the patients (SSS score <15 points on admission). One hundred thirty-nine (62%) of these patients died in the hospital. In survivors, good functional outcome (a BI score of ≥50 points after completed rehabilitation) was achieved by 26 (31%), and 58 patients (69%) had a poor functional outcome (a final BI score <50 points).
Table 2⇓ depicts comparisons of stroke severity, length of hospital stay, and discharge placement between the 2 groups. Initial stroke severity (as measured by the SSS score on admission) was quite similar between groups; in addition, the frequencies of patients with decreased consciousness on admission (evaluated by the SSS subscore for consciousness) were comparable. Only 8% of patients with a good functional outcome were discharged to a nursing home after completed rehabilitation compared with 91% of the patients with a poor functional outcome (Table 2⇓).
Table 3⇓ shows a comparison of potential predictors of good functional outcome between the 2 groups. Patients with a good functional outcome were younger, more often married, working, and had a higher frequency of daily alcohol consumption and a lower frequency of other disabling comorbidity in the univariate analyses. However, most of these differences were explained by the difference in age between groups. The multivariate logistic regression analysis of factors obtainable on acute admission revealed 3 independent factors as predictors of good outcome: age (odds ratio [OR], 0.50 per 10-year decrease; 95% CI, 0.25 to 0.99; P=0.04), a spouse at home (OR, 3.1; 95% CI, 1.1 to 8.8; P=0.03), and spontaneous body temperature on admission (OR, 1.8 per 1°C decrease; 95% CI, 1.1 to 3.1; P=0.01) (Figure 1⇓). In this model, correct classification was obtained in 75% of the cases.
Figure 2⇓ shows neurological impairment in the 2 groups as a function of time from stroke onset. Neurological impairment at the time of acute admission was identical in the 2 groups, but the day after admission a diversion in scores was evident, reaching a significant level after 1 week (SSS score 1 week after admission was 27.4 [SD 19.8] points in patients who eventually had a good functional outcome compared with 13.3 [8.2] points in patients who eventually had a poor functional outcome; P<0.001). Adding neurological impairment at week 1 to the multivariate logistic regression model of predictors of good outcome increased the rate of correct classification to 85%. Neurological impairment at week 1 was an independent predictor of good outcome (OR, 3.2 per 10-point increase in SSS score; 95% CI, 1.3 to 7.8; P=0.01) (Figure 1⇑). The difference in neurological impairment after 1 week was maintained throughout the hospital stay.
Figure 3⇓ depicts the level of disability in the 2 groups as a function of time from stroke onset. There was a marked difference in BI score 1 week after admission (28.6 points [SD 3.2] versus 3.2 points [SD 5.8]; P<0.001), and this difference increased steadily throughout hospital stay to 79.2 points (17.3) versus 10.2 (13.5) after completed rehabilitation (P<0.001).
The initial severity of the stroke is the single most important prognostic factor in stroke.12 However, even patients with the most severe strokes sometime experience a remarkably good recovery. In this study, 32% of the survivors with the most severe strokes (SSS score <15 points on admission) achieved a good functional outcome.
Prognostic factors found exclusively in patients with the most severe strokes have not been examined before. We found 4 factors related to a good functional outcome in this group: decreasing age, a spouse at home, spontaneous body temperature on admission, and early neurological recovery. A correct classification was achieved in 85% of the cases in the logistic regression prediction model with these 4 factors: age, presence of a spouse at home, body temperature, and SSS score after 1 week. A 10-year decrease in age corresponded to a 50% reduction in the relative risk of a poor outcome. That functional outcome decreases with increasing age is in agreement with previous findings in the general stroke population.13 14 This relation is most likely due to an age-linked reduced ability for functional compensation15 and is not to any greater extent due to a relation between increasing age and comorbidity since these parameters were not highly correlated (Spearman correlation coefficient, 0.11). The existence of a spouse at home increased the relative chance of a good outcome 3-fold, independent of age, sex, stroke severity, and other factors. This is probably a reflection of the importance of a good social network to outcome in general.16 17
We have previously reported a relation between spontaneous body temperature and mortality/neurological impairment in survivors.12 18 That a good functional outcome is so strongly related to body temperature, as reported here, is nevertheless surprising: a 1°C decrease in body temperature corresponded to an almost doubled relative chance of a good functional outcome, independent of other factors. This finding may suggest that lowering body temperature could point to a treatment effect, provided that the relation is causal.
Degree of neurological impairment 1 week after stroke onset added substantial information regarding whether or not a good functional outcome could eventually be achieved. Initial neurological impairment was identical in patients with good and poor functional outcome, but neurological recovery within the first week markedly differed between groups, and this difference explained most of the difference in both neurological and functional outcome after completed rehabilitation. This suggests that early neurological recovery is the basis for good functional outcome in patients with the most severe strokes. Early spontaneous reperfusion of the ischemic penumbra surrounding the core area of infarction could in part be the pathophysiological background for early neurological recovery.19 20
Other comorbidity such as diabetes, hypertension, ischemic heart disease, and other disabling disease did not predict functional outcome, nor did the type of stroke. This confirms previous findings in the general stroke population.3 10 21 22 23 24 It should be emphasized that function was defined by the ability to perform basic activities of daily living as measured by the BI. An influence on the ability to perform instrumental activities of daily living or higher cognitive functions can therefore not be excluded. We have previously reported low acute systolic blood pressure to predict poor outcome in the subgroup of patients with early stroke in progression.25 In the present study of unselected patients with the most severe strokes, blood pressure did not predict functional outcome. Blood pressure seems important to the development of early stroke in progression, but not to the prognosis of the stroke population as a whole.
Body temperature is a strong predictor of functional outcome in patients who survive the most severe strokes. It was also the only potentially modifiable outcome predictor. We suggest that a randomized controlled trial be performed to evaluate whether active reduction of body temperature can improve the generally poor prognosis of the patients with the most severe strokes.
This study was supported by grants from the Danish Health Foundation and the Danish Heart Foundation.
Reprint requests to Henrik Stig Jørgensen, MD, DMSci, Department of Neurology, Bispebjerg Hospital, Bispebjerg bakke 23, DK-2400 Copenhagen NV, Denmark.
- Received March 26, 1999.
- Revision received May 17, 1999.
- Accepted June 4, 1999.
- Copyright © 1999 by American Heart Association
Jørgensen HS, Nakayama H, Raaschou HO, Olsen TS. Acute stroke: prognosis and a prediction of the effect of medical treatment on outcome and health care utilization: the Copenhagen Stroke Study. Neurology. 1997;49:1335–1342.
Jørgensen HS. The Copenhagen Stroke Study experience. J Stroke Cerebrovasc Dis. 1996;6:5–16.
Jørgensen HS, Plesner AM, Larsen K, Hübbe P. Marked increase in stroke incidence in men between 1972 and 1990 in Frederiksberg, Denmark. Stroke. 1992;23:1701–1704.
Report of the WHO Task Force on Stroke and Other Cerebrovascular Disorders. Stroke–1989: recommendations on stroke preventions, diagnosis, and therapy. Stroke. 1989;20:1407–1431.
Scandinavian Stroke Study Group. Multicenter trial of hemodilution in ischemic stroke: background and study protocol. Stroke. 1985;16:885–890.
WHO Study Group on Diabetes Mellitus. WHO Technical Report Series 727. Geneva, Switzerland: World Health Organization; 1985.
Chamberlain JM, Grandner J, Rubinoff JL, Klein BL, Waisman Y, Huey M. Comparison of a tympanic thermometer to rectal and oral thermometers in a pediatric emergency department. Clin Pediatr. 1991;30(suppl 4):24–29.
SPSS. Statistical Package for Social Sciences [computer program]. Windows version, release 8.0. Chicago, Ill: SPSS; 1998.
Nakayama H, Jørgensen HS, Raaschou HO, Olsen TS. The influence of age on stroke outcome: the Copenhagen Stroke Study. Stroke. 1994;25:808–813.
Kalra L. Does age affect benefits of stroke unit rehabilitation? Stroke.. 1994;25:346–351.
Jørgensen HS, Nakayama H, Reith J, Raaschou HO, Olsen TS. Factors delaying hospital admission in acute stroke: the Copenhagen Stroke Study. Neurology. 1996;47:383–387.
Jørgensen HS, Nakayama H, Raaschou HO, Olsen TS. Acute stroke care and rehabilitation: an analysis of the direct cost and its clinical and social determinants: the Copenhagen Stroke Study. Stroke. 1997;28:1138–1141.
Jørgensen HS, Nakayama H, Reith J, Raaschou HO, Olsen TS. Acute stroke with atrial fibrillation: the Copenhagen Stroke Study. Stroke. 1996;27:1765–1769.
Jørgensen HS, Nakayama H, Raaschou HO, Olsen TS. Stroke in patients with diabetes: the Copenhagen Stroke Study. Stroke. 1994;25:1977–1984.
Jørgensen HS, Nakayama H, Reith J, Raaschou HO, Olsen TS. Stroke recurrence: predictors, severity and prognosis: the Copenhagen Stroke Study. Neurology. 1997;48:891–895.
Censori B, Camerlingo M, Casto L. Prognostic factors in first-ever stroke in the carotid artery territory seen within 6 hours after stroke onset. Stroke.. 1993;24:532–535.