Prior TIA, lipid-lowering drug use, and physical activity decrease ischemic stroke severity

Experimental studies have shown that the brain can be protected against the consequences of ischemia by procedures applied a few days before, including a brief period of ischemia^1,2 or pharmacologic agents such as bacterial lipopolysaccharide or 3-nitro-propionic acid.^2–4 These preconditioning procedures could lead to an endogenous neuroprotection through the induction of cytoprotective proteins such as antioxidant enzymes, endothelial nitric oxide synthase (eNOS), or heat shock proteins, as well as through the inhibition of inflammatory, oxidative, or apoptotic pathways.^1–4 A preventive neuroprotection may also be induced in animals by the use of drugs that are able to modulate pathways involved in endogenous neuroprotection.¹ In this way, lipid-lowering drugs (both statins and fibrates) given before cerebral ischemia could lead to a decrease in brain infarct volume^5–7 when beneficial effects of drugs acting on the angiotensin system are also discussed.^8,9

Whether such a preventive neuroprotection may be induced in humans remains unclear.¹⁰ Previous studies suggested a protective effect of previous TIA^11–15 and of statin treatment.^16–21 Nevertheless, other protective or deleterious conditions should also be considered. Preexisting dementia, blood pressure, body temperature, or blood glucose level at admission should be evaluated as potentially deleterious.^22–25 Conversely, previous treatments, namely fibrates and drugs acting on the angiotensin system,^7–9 and previous regular physical activity²⁶ should be considered as potentially protective.

Before designing clinical trials to evaluate preventive neuroprotection procedures, the aim of this cross-sectional hospital-based study was to determine previous clinical or pharmacologic conditions, including both TIA and lipid-lowering drug treatment, that were associated with a lower severity or a better short-term outcome of supratentorial cerebral ischemia.

Methods.

Results.

During the study period, 362 patients (median age 70 years, range 16 to 97 years; 195 women) were included (table 1). Univariate analyses discriminated a number of clinical and pharmacologic conditions potentially associated with a lower severity (NIHSS 0 to 5) at admission and a better short-term outcome (mRS 0 to 1 and BI 95 to 100) (table 2). Among all factors, previous leisure-time physical activity was a main factor potentially associated with a lower initial severity and a better short-term outcome with a trend toward an effect of the weekly duration of physical exercise (table 3).

The logistic regression analysis with NIHSS 0 to 5 as a dependent variable found not having a previous stroke, regular practice of physical activity, not previously receiving anticoagulation, previous TIA, and lipid-lowering therapy as independent factors associated with a lower severity of cerebral ischemia at admission (table 4). Therefore, previous TIA, lipid-lowering drug use, and physical activity were significantly associated with a lower NIHSS at admission when only previous stroke was associated with a worse clinical deficit (table 5).

The logistic regression analysis with mRS 0 to 1 as a dependent variable found not having a previous stroke, not having arterial hypertension, leisure-time physical activity, and previous lipid-lowering therapy as independent factors associated with a good outcome (table 6). When variables assessable during hospitalization were also included in the model, factors independently associated with mRS 0 to 1 were lower NIHSS at admission, higher MMSE, not having a previous stroke, and not having arterial hypertension (table 6).

When BI 95 to 100 was used as dependent variable, we found lower age, previous lipid-lowering therapy, not having a previous stroke, leisure-time physical activity, and not having diabetes mellitus as independent factors associated with a good short-term outcome (table 7). Moreover, taking into account variables assessable during hospitalization, lower NIHSS, higher MMSE, lower age, leisure-time physical activity, and no previous antipsychotic treatment were independently associated with a good short-term outcome (table 7).

Finally, using worse short-term outcome defined by NIHSS worsening ≥ 1 point as a dependent variable (other variables included in the model: age; medical history of previous stroke; physical activity; previous treatment with lipid-lowering drug, anticoagulant, Ca blockers, antiarrhythmic or antidepressant drugs; NIHSS, IQCODE, systolic blood pressure, blood glucose, and total, HDL, and LDL cholesterol at admission and final diagnosis), we found that the main independent deleterious factor was blood glucose level at admission (adjusted OR 3.81, 95% CI 1.48 to 9.79, by 1 mmol/L increase) and, to a lesser degree, increasing age (adjusted OR 1.04, 95% CI 1.00 to 1.08, by 1-year increase) and NIHSS at admission (adjusted OR 1.06, 95% CI 1.00 to 1.13, by 1-point increase).

Discussion.

Previous TIA, lipid-lowering drug use, and leisure-time physical activity were independently associated with a lower severity of cerebral ischemia at admission and a better short-term outcome. In addition, previous stroke, increasing age, arterial hypertension, diabetes mellitus, blood glucose level at admission, cognitive decline, previous anticoagulant, or antipsychotic treatments were independently associated with a higher severity of cerebral ischemia and a worse short-term outcome.

Previous regular and moderate leisure-time physical activity contributed to a decrease in the initial severity of cerebral ischemia and was associated with a better short-term outcome. Epidemiologic and case–control studies have suggested that physical activity could decrease stroke incidence.^27,34,35 Whether the type and duration of physical activity are determinants of the protection remains unclear. Epidemiologic studies suggested that the effect of physical activity on stroke prevention could have a dose-dependent relationship for both intensity and duration.^27,34,35 Our study suggested a trend toward an exercise duration-dependent neuroprotective effect of physical activity, which is significant for exercise duration of more than 2 hours each week. Conversely, our study showed uncertain results below 2 hours’ duration, as well as of heavy intense physical activity. Nevertheless, the small number of subjects in this last subgroup did not allow a definitive conclusion. The neuroprotective effect of previous physical activity may also contribute to the apparent decrease in stroke incidence,^34,35 by reducing cerebral infarct size to levels that are not symptomatic. Our findings supported the results of experimental studies in which physical activity before ischemia induced neuroprotection.^26,36–38 In rats, both voluntary running and forced exercise have a neuroprotective effect,³⁶ which could result from an improvement of cerebral blood flow through an increase in eNOS expression or angiogenesis, or from the modulation of neuroplasticity through nerve growth factor or brain-derived neurotrophic factor expression.^36–38

Our study also confirmed previous studies of the reduced severity of cerebral ischemia in patients with a previous TIA.^11–15 In contrast, one study demonstrated no beneficial effect of previous TIA.³⁵ However, this study mainly evaluated the protective effect of TIAs regarding the timing (less or more than 1 week before stroke onset) and duration, without any comparisons to stroke patients without previous TIA.³⁹ By taking into account control patients without previous TIA, two studies suggested that the neuroprotection afforded by TIA was more prominent when one or more TIAs of up to 20 minutes’ duration occurred 1 to 7 days before stroke.^12,15 In our study, we did not take into account duration of symptoms or territory in which TIA occurred. However, we found no clear differences with the timing criterion used (<3 months, 3 months to 1 year, >1 year). Whether only ipsilateral TIA induces neuroprotection remains also unsettled,^12,14,15 but some experimental data showed that neuroprotection of both hemispheres could be induced by a previous unilateral transient focal ischemia.⁴⁰ Mechanisms involved in TIA-induced neuroprotection in humans remain uncertain. Whether it is the consequence of an ischemic preconditioning rather than that of secondary prevention measures remains unclear. Taking into account a large number of previous treatments including antiplatelet and anticoagulant, we found that previous TIA remains independently associated with a lower severity of cerebral ischemia. Such a result could argue in favor of underlying ischemic preconditioning mechanisms.

We found that a previous lipid-lowering therapy (statin or fibrate) was also independently associated with a lower severity of cerebral ischemia and a better short-term outcome. Such a neuroprotective effect was already suggested with statins^16–21 but not with fibrates. Statins and fibrates sharing similar pleiotropic effects potentially leading to neuroprotection,^5–7 we combined both therapies in our analysis. Experimental models showed that neuroprotection is developed independently of the metabolic effects of these drugs.^5–7 In this way, conversely to some previous studies,^16–20 we also analyzed the impact on the short-term outcome of cholesterol and triglycerides levels at admission. We did not find any independent relationship despite a trend toward an effect of HDL or LDL cholesterol levels. It is possible that the impact of lipid was underestimated in our study because a recent study showed that lipid levels rapidly decreased after stroke onset.⁴¹ Conversely, in another previous study, it was shown that a lower cholesterol level could be associated with a poor clinical outcome.⁴² This finding remains to be discussed because HDL cholesterol level and lipid-lowering therapy were not taken into account in the analyses, and because patients with a lower cholesterol level were also more likely to have cardioembolic stroke leading to more severe prognosis.⁴² The mechanisms by which statins and fibrates induce a neuroprotection in cerebral ischemia remain speculative: statins could interfere with platelet aggregation,⁶ whereas both statins and fibrates may have anti-inflammatory, antioxidative, and antiapoptotic properties.^5–7 These pathways are also involved in the development of endogenous neuroprotection.¹

Our study suggested a deleterious effect of increasing age, arterial hypertension, diabetes mellitus, blood glucose level at admission, and cognitive decline. Some of these factors were already described as potent determinants of worse outcome.^22,25,43 Previous studies showed that increasing age and preexisting dementia could be associated with a higher rate of mortality after stroke.^22,43 In our study, preexisting dementia was not an independent predictor of initial severity or of short-term outcome. Previous lipid-lowering drugs in stroke patients could have masked the detrimental effect of preexisting dementia because we have shown in a previous study that both statins and fibrates could slow the progression of dementia in patients with Alzheimer disease.⁴⁴ In contrast, lower MMSE score was associated with worse short-term outcome. According to the time from stroke onset, it could be that MMSE mainly reflected stroke severity rather than was a marker of cognitive decline. In accordance with recent data, our study also suggested a deleterious effect of antipsychotic drugs.⁴⁵ This hypothesis is currently under debate, but according to the few number of patients receiving antipsychotic drugs in our study, results should be analyzed with caution.

Our study has several strengths. The design of data collection in this cross-sectional hospital-based study allowed us to evaluate the effects of a large number of variables on initial clinical severity and short-term outcome in a well-defined and homogeneous population of supratentorial ischemic stroke patients. To limit recall bias and missing data, our data collection process also included the interview of the patients, their family, and their general practitioner to more precisely assess many variables, particularly previous treatments and leisure-time physical activity. Moreover, the statistical analysis process was planned to select only the most relevant variables to include in the logistic regression analyses. It allowed us to avoid the inclusion of too large a number, or of colinear variables that could lead to bias the results.

Our study also has limitations. Foremost, it was a nonrandomized hospital-based study focusing on supratentorial ischemic stroke, and thus these data should only be considered hypothesis generating. Results could probably not be generalized to all patients with ischemic stroke. Because of the observational study design, we cannot rule out that the association between TIA, previous lipid-lowering drug use, physical activity, and outcome was confounded by other factors that were not taken into account in our analysis. However, these findings remained similar after adjustment for factors that potentially influence stroke severity. These results also remained similar irrespective of the criterion used (NIHSS 0 to 5, mRS 0 to 1, BI 95 to 100). It is possible that different cutoff values may have led to different results,³⁰ but those cutoff values were prespecified. According to the assessment of a large number of variables, it could be argued that not decreasing the p value could lead to misinterpretation. Nevertheless, the logistic regression analysis including the calculation of adjusted OR partially solves such a methodologic issue.