Prestroke physical activity is associated with severity and long-term outcome from first-ever stroke

Epidemiologic studies have consistently suggested an association between physical activity and the risk of stroke.¹ Physical activity is recommended to reduce the risk of first ever and possibly the risk of recurrent stroke.² Stroke severity has marked implications for the patients and society. More severe strokes result in worse outcomes, including longer rehabilitation, absence from work, and higher mortality.³ A retrospective cross-sectional study investigation of pharmacologic and clinical factors in relation to stroke severity showed that physical activity prior to stroke was inversely associated with stroke severity and short-term outcome.⁴

In this study we investigate if prestroke level of physical activity influences stroke severity and long-term outcome in patients with first-ever stroke enrolled in the ExStroke Pilot Trial.⁵

METHODS

This is a retrospective analysis in a prospectively collected patient population. Patients enrolled in the present study had a first-time ever stroke and were included in the ExStroke Pilot Trial. The design of the ExStroke Pilot Trial has been described elsewhere.⁵ Briefly, the ExStroke Pilot Trial is a multicenter, multinational, randomized trial in which patients with ischemic stroke were randomized to an intervention consisting of repeated encouragement and verbal instruction in being physically active. The control group received standard care and information of the possible benefits of physical activity but no active encouragement. Patients aged 40 years or above were included if they fulfilled the following criteria: ischemic stroke diagnosis, CT scan/MRI excluding hemorrhage or other focal pathology not indicating stroke, ability to walk unassisted, and inclusion within 90 days after stroke. Patients were excluded if they were bedridden, unable to understand the trial, reluctant to provide informed consent, or had a modified Rankin Score ≥4 prior to index stroke. All patients were followed for 2 years with four follow-up visits the first year and two follow-up visits the second year. Patients were included from nine centers in four countries (China, Estonia, Poland, Denmark). At baseline a thorough medical history was obtained focusing on risk for stroke as well as other diseases that would compromise physical activities. Comorbidity was assessed using the Charlson comorbidity index¹¹ modified to use in patients with stroke.¹² The Charlson index is a comorbidity scoring system that includes weighting factors based on disease severity. The Charlson index is validated for the use in ischemic stroke outcome studies.¹²

RESULTS

A total of 314 patients participated in the ExStroke Pilot trial, of which 265 (84%) were included in the present analyses with a first ever stroke (135 from the intervention group and 130 from the control group). Mean (SD) age was 68.2 (12.2) years and 44.3% of the patients were female. The distribution of baseline characteristics is shown in table 1.

Univariable analyses showed that the level of prestroke physical activity (PASE score) was one of the main predictors of stroke severity and long-term outcome with increasing ORs for a less severe stroke and better outcome with increasing level of physical activity (table 2). The OR for a milder stroke severity was 1.53 (95% CI 1.25–1.86) and could be expressed as linear (χ² = 2.75, 2 df; p = 0.25). The OR for a higher modified Rankin score was 0.64 (95% CI 0.52–0.79), and could be expressed as linear (χ² = 1.64, 2 df; p = 0.44).

Concordantly, in multivariable analyses the OR for a higher SSS score increased with the level of physical activity (figure 1). The OR for the difference between each quartile was 1.43 (95% CI 1.16–1.78, p = 0.001) and could be expressed as linear (χ² = 2.29, 2 df; p = 0.32).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Cumulative OR for a higher Scandinavian Stroke Scale score according to level of prestroke physical activity

The association could be considered linear with increasing OR for a higher Scandinavian Stroke Scale score with increasing level of physical activity (OR 1.43, CI 1.16–1.78). The results are adjusted for age, gender, Charlson score, educational level, fasting glucose level, and history of diabetes mellitus.

Similarly, the association between PASE quartiles and long-term outcome (mRS) showed that with increasing levels of physical activity the OR for a higher mRS decreased (figure 2). The OR between each PASE quartile was 0.78 (95% CI 0.62–0.99, p = 0.04) and could be expressed as linear (χ² = 0.32, 2 df; p = 0.85). Including and excluding mean arterial blood pressure and fasting glucose levels from the analysis did not alter the results. Excluding the SSS score from the analysis strengthened the association between prestroke level of physical activity and mRS (OR 0.75, CI: 0.59–0.94, p_trend = 0.01). Of note, using the PASE score as a continuous variable did not alter the results.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Cumulative OR for a higher modified Rankin score according to level of prestroke physical activity

The association could be considered linear with decreasing OR for a higher modified Rankin score with increasing level of physical activity (OR 0.79 CI: 0.62–0.99). The results are adjusted for fasting glucose level, age, educational level, history of atrial fibrillation, diabetes mellitus, depression, Charlson comorbidity index, availability of caretaker, stroke diagnosis (TOAST), randomization group, and Scandinavian Stroke Scale score.

A total of 21 strokes, 2 myocardial infarctions, and 14 deaths occurred during follow-up. There was no significant association between prestroke level of physical activity and recurrent stroke (hazard ratio 0.88, 95% CI 0.56–1.37) or recurrent stroke, myocardial infarction, or death (hazard ratio 0.86, 95% CI 0.60–1.24).

DISCUSSION

In the present study we observed that a higher prestroke level of physical activity was associated with a less severe stroke, and that higher activity levels were reported in patients with the most favorable outcomes. The results were consistent in uni- and multivariable analyses, and the association could be expressed linearly, suggesting a dose-response association.

It has previously been reported that increased level of physical activity is associated with milder stroke severity.⁴ Our results confirm these findings, establishing a firm association between reported level of physical activity and stroke severity. Our results also demonstrate a significant association between reported level of physical activity and long-term outcome of first-ever stroke after adjusting for possible confounding factors.

The prestroke level of physical activity may affect the prognosis after stroke in many ways. Physical activity modifies traditional risk factors for cardiovascular diseases, which could lead to a reduction in the risk of recurrent stroke. In the present study we could not observe an effect of prestroke physical activity on long-term outcome mediated through glucose and blood pressure levels. Animal studies have reported that physical activity may have a neuroprotective effect by enhancing endothelial nitric oxide synthase expression, which results in reduced cerebral infarct size and less severe stroke.¹⁷ We observed that excluding the SSS score from the results strengthened the association between prestroke level of physical activity and long-term outcome, suggesting that some of the effect of prestroke physical activity on long-term outcome is mediated though the SSS score.

Guidelines for the primary prevention of ischemic stroke reported that physical inactivity was a modifiable risk factor, which requires greater emphasis in stroke prevention campaigns.¹⁸ The results from the present study suggest that physically active persons may also have milder strokes and a better outcome in the long run.

This study is strengthened by the use of ordinal logistic regression statistics for all analyses. Ordinal scales have often been dichotomized prior to analyses, which leads to loss of statistical power and to controversies as to where the appropriate cutoff point should be. In addition, physical activity was assessed using the PASE, which is a validated questionnaire specifically developed to assess physical activity in the elderly. This will reduce a potential age bias, as age-neutral questionnaires tend to underestimate physical activity in the elderly.¹⁹ The study was conducted as a multicenter study, which adds to the generalizability of the results.

The main limitation of this study is that prestroke level of physical activity was assessed retrospectively. We cannot exclude that there are confounding factors, e.g., participating in rehabilitation programs and cognitive status, which we have not taken into consideration and which might explain the association between physical activity and long-term outcome reported in this study. The results in this study are based on patients with mild ischemic strokes and of Caucasian and Asian ethnicity. Testing the hypothesis in a broader selection of patients with stroke and in other ethnic groups could add to further clarify the association between physical activity and long-term outcome.

ACKNOWLEDGMENT

The authors thank the ExStroke Pilot Trial Group for collaboration. They also thank Anders Pedersen, RN, Marianne Lindahl, PT, MPH, Lotte Hansen, PT, Tang Ning, MD, and Sun Xiangrong, MD.