To assess etiology of ischemic stroke in patients with obstructive sleep apnea (OSA) compared with controls. This information may aid in determining how OSA increases stroke risk and facilitate recurrent stroke prevention in patients with OSA.

Retrospective, case-control study.

Academic tertiary referral center.

Consecutive patients who underwent polysomnography and had an ischemic stroke within 1 year were identified. Stroke subtype was determined using two validated algorithms. Polysomnographic results were used to separate patients into OSA cases and controls. Information regarding cardiovascular risks, neuroimaging, and echocardiographic data were collected.

N/A.

In 53 subjects, cardioembolic (CE) strokes were more common among OSA cases than controls (72% versus 33%, P = 0.01). The majority of CE strokes occurred in those with moderate to severe OSA. Atrial fibrillation (AF) was more frequent in OSA cases (59% versus 24%, P = 0.01). The association between OSA and CE stroke remained significant after controlling for AF (P = 0.03, odds ratio 4.5).

There appears to be a strong association between obstructive sleep apnea (OSA) and cardioembolic (CE) stroke. In patients with OSA presenting with cryptogenic stroke, high clinical suspicion for CE is warranted. This may lead to consideration of diagnostic studies to identify CE risk factors such as paroxysmal atrial fibrillation (AF). CE strokes are more common in patients with OSA even after adjusting for AF. This finding may reflect a high rate of occult paroxysmal AF in this population; alternatively, OSA may lead to CE strokes through mechanisms independent of AF.