Article under review: Andersen KK, Olsen TS, Dehlendorff C, Kammersgaard LP. Hemorrhagic and ischemic strokes compared: stroke severity, mortality, and risk factors. Stroke 2009;40(6):2068-72.
Purpose: To compare the stroke severity, mortality, and cardiovascular risk factors in a large population-based cohort of incident hospitalized cases.
Methods: This study utilized data collected in the Danish National Indicator Project, which began in 2001 and includes information on all stroke admissions in Denmark. Approximately 39,500 patients with incident stroke were identified and followed until death or censoring in 2007. Admission stroke severity was measured by the Scandinavian Stroke Scale, which incorporates assessments of level of consciousness, orientation, aphasia, eye movements, facial paresis, gait, and power in the arm, hand, and leg. Cardiovascular risk factor assessments included information on alcohol intake (high = >14 drinks/week in women or > 21 drinks/week in men), smoking (current, former, or never), diabetes, atrial fibrillation, hypertension, current or prior myocardial infarction, prior stroke, and intermittent claudication. Survival of patients in the database was monitored through the Danish Registry of Persons. A multivariate logistic regression model was created to determine the independent association of cardiovascular risk factors on stroke type. A multivariate Poisson regression survival model was created to identify independent predictors of death in a subset of more than 25,000 individuals with a complete data set.
Results: Overall, 35,491 (90%) had ischemic stroke and 3993 (10%) had hemorrhagic stroke. Similarly, for those with a complete data set, 23,150 (92%) had ischemic stroke and 1973 (8%) had hemorrhagic stroke.
Hemorrhagic strokes tended to be more severe than ischemic strokes, so that as impairments increased (as measured by the Scandinavian Stroke Scale), the proportion of cases that were hemorrhagic increased as well from 2% for patients with the mildest strokes to 30% for those with the most severe strokes.
Cardiovascular risk factors independently favoring ischemic stroke as opposed to hemorrhagic stroke were diabetes, atrial fibrillation, current or prior myocardial infarction, previous stroke, and intermittent claudication. High alcohol intake and smoking favored hemorrhagic stroke. Gender and hypertension did not favor either stroke type.
Patients with hemorrhagic stroke had a higher risk of all-cause death compared to patients with ischemic stroke (hazard ratio 1.6), even after adjustment for age, gender, initial stroke severity, and cardiovascular risk factors. About half (49%) of those with hemorrhagic stroke died by follow-up compared to only about a quarter (26%) of those with ischemic stroke. Other factors that contributed independently to death included smoking, atrial fibrillation, prior or current myocardial infarction, prior stroke, older age, and degree of impairment. The hazard ratio of mortality for hemorrhagic stroke versus ischemic stroke varied over time: initially patients with hemorrhagic stroke had a 4-fold excess mortality, which decreased to 2.5-fold by 1 week, then 1.5-fold by 3 weeks, and then gradually further, so that by about 3 months the risk of death from that point was similar in the 2 types of stroke (see the graph of case fatality by days after stroke below; note the approximately parallel lines after 30 days post-stroke).
Conclusions and commentary: Strokes are generally more severe in patients with hemorrhagic strokes and such strokes are associated with significantly greater early mortality in the first 3 months after stroke onset.
As the authors note, "knowledge on the relative role of risk factors in hemorrhagic versus ischemic strokes is still inconsistent." In other words, for many risk factors, well-designed studies that tried to address whether individual cardiovascular risk factors were more predictive of 1 stroke type or the other came to differing conclusions. In this study, cardiovascular risk factors independently favoring ischemic stroke as opposed to hemorrhagic stroke were diabetes, atrial fibrillation, current or prior myocardial infarction, previous stroke, and intermittent claudication. Diabetes has been consistently identified as more predictive of ischemic than hemorrhagic stroke (Jamrozik et al 1994; Jørgensen et al 1995; Hajat et al 2001), as generally have prior stroke or TIA (Hajat et al 2001; Liu et al 2005) and ischemic heart disease and/or myocardial infarction (Jørgensen et al 1995; Liu et al 2005). Also in the present study, high alcohol intake and smoking favored hemorrhagic stroke, whereas in prior studies smoking has not generally favored 1 stroke type or the other (Jamrozik et al 1994; Jørgensen et al 1995; Hajat et al 2001) or has instead been associated more with ischemic strokes (Liu et al 2005). Several epidemiologic studies addressing multiple factors have not identified a higher risk of hemorrhagic stroke than ischemic stroke with high alcohol intake (Jørgensen et al 1995; Hajat et al 2001), but high alcohol intake is a recognized risk factor for hemorrhagic stroke (Qureshi et al 2001; Ariesen et al 2003; Ferro 2006; Hänggi and Steiger 2008).