Background and objectives: Almost every fifth stroke occurs during sleep. Data about characteristics and etiology of stroke during sleep are conflicting. We investigated the association of the activity at stroke onset (onset during night sleep vs. onset while awake) with stroke subtypes of different etiology.
Methods: A total of 1448 patients with first-ever stroke with known time of symptom presentation were prospectively evaluated. Statistical comparisons were performed between patients with stroke during sleep and stroke while awake in terms of demographic features, known risk factors, vascular comorbidities, and stroke subtypes. Multiple variable logistic regression analyses were performed to identify predictor variables (including stroke risk factors and stroke subtypes) for stroke during sleep.
Results: Stroke during sleep was documented in 264 cases (18.2%). In subjects with stroke during sleep, lacunar infarction was the most prevalent stroke subtype (39%), while in patients with stroke while awake, small-vessel disease was the underlying mechanism significantly (P < .001) less often (13.8%). In contrast, patients with stroke while awake suffered significantly (P < .001) more frequently from intracerebral hemorrhage (18.2%) and cardioembolic stroke (34.9%) when compared with subjects with stroke during sleep (6.4% and 18.9%, respectively). The multiple variable logistic regression model identified the following factors as independent predictors of stroke during sleep: atrial fibrillation (odds ratio: 0.346, 95% confidence interval: 0.237-0.505, P < .001) and intracerebral hemorrhage versus ischemic stroke (odds ratio: 0.238, 95% confidence interval: 0.138-0.410, P < .001). Lacunar infarction was the only ischemic stroke subgroup that was positively associated with stroke during sleep (odds ratio: 2.568, 95% confidence interval: 1.447-4.560, P < .001).
Conclusions: There are significant differences between stroke during sleep and stroke while awake concerning vascular risk profile and stroke etiologic subtypes.