Association Between Ischemic Stroke Subtype and Stroke Severity: The Atherosclerosis Risk in Communities Study.

BACKGROUND AND OBJECTIVES

Disability after stroke occurs across ischemic stroke subtypes, with a suggestion that embolic strokes are more devastating. Whether this difference is as a result of differences in comorbidities or differences in severity at the time of the stroke event is not known. The primary hypothesis was that participants with embolic stroke would have more severe stroke at the time of admission and a higher risk of mortality, compared with thrombotic stroke participants even with consideration of confounders over time, with a secondary hypothesis that this association would differ by race and sex.

METHODS

Atherosclerosis Risk in Communities (ARIC) study participants with incident adjudicated ischemic stroke, stroke severity and mortality data, and complete covariates were included. Multinomial logistic regression models determined the association between stroke subtype (embolic vs thrombotic) and admission NIH Stroke Scale (NIHSS) category (minor [≤5], mild [6-10], moderate [11-15], severe [16-20], and very severe [>20]) adjusted for covariates from visits most proximal to the stroke. Separate ordinal logistic models evaluated for interaction by race and sex. Adjusted Cox proportional hazard models estimated the association between stroke subtype and all-cause mortality (through December 31, 2019).

RESULTS

Participants (N = 940) were mean age 71 years (SD = 9) at incident stroke, 51% female, and 38% Black. Using adjusted multinomial logistic regression, the risk of having a more severe stroke (reference NIHSS ≤5) was higher among embolic stroke vs thrombotic stroke patients, with a step-wise increase for embolic stroke patients when moving from mild (odds ratio [OR] 1.95, 95% CI 1.14-3.35) to very severe strokes (OR 4.95, 95% CI 2.34-10.48). After adjusting for atrial fibrillation, there was still a higher risk of having a worse NIHSS among embolic vs thrombotic strokes but with attenuation of effect (very severe stroke OR 3.91, 95% CI 1.76-8.67). Sex modified the association between stroke subtype and severity (embolic vs thrombotic stroke, interaction = 0.03, per severity category, females OR 2.38, 95% CI 1.55-3.66; males OR 1.75, 95% CI 1.09-2.82). The risk of death (median follow-up 5 years, interquartile range 1-12) was also increased for embolic vs thrombotic stroke patients (hazard ratio 1.66, 95% CI 1.41-1.97).

DISCUSSION

Embolic stroke was associated with greater stroke severity at the time of the event and a higher risk of death vs thrombotic stroke, even after careful adjustment for patient-level differences.