Background: In non-stroke patients, the severity of sleep apnea (SA) is known to be frequently related to the sleeping position, a condition called positional SA. In the present study, we investigated whether in acute stroke the occurrence of apneas was related to the positioning of patients, and whether a similar finding could be observed after rehabilitation. With the purpose of identifying patients potentially being in need of a SA treatment beyond rehabilitation, we furthermore looked for epidemiologic and clinical parameters being related to persistent SA 6 months after stroke.Patients and methods: Fifty-five acute stroke patients underwent cardiorespiratory polygraphy within 72 hours after onset of neurological symptoms and after 6 months. Apart from the total AHI (AHITOT), the AHI with the patient in supine position and the AHI with the patient in other positions were determined. In all patients, demographic data, NIH-stroke scale score and cumulative vascular risk factors were assessed.Results: In the initial sleep study, 78% of patients had an AHI≥10/h, of whom 65% fulfilled the criteria of positional SA. On follow-up, the incidence of SA declined to 49% with positional SA being present in 33%. Multivariate logistic regression analysis identified AHITOT on admission [OR=1.07 (1.002–1.13)] and cumulative vascular risk factors [OR=3.48 (1.34–9.05)] as independent predictors of persistent SA 6 months after stroke.Conclusion: According to our results, positional SA is a predominant feature in acute stroke and its incidence decreases significantly during the following months. These findings may have implications for SA treatment in patients with acute stroke.