The objective was to compare the prevalence of sleep apnea syndrome (SAS) in elderly individuals with a history of stroke to the one in individuals, matched by age and sex with a history of insomnia. To determine if previous subjective sleep complaints in the elderly are associated with stroke. A cross-sectional study with subsequent 3-year follow-up was designed in an acute geriatric inpatient unit. 19 subjects with documented stroke and 21 subjects with insomnia were included. All participants were assessed with a sleep questionnaire, an overnight polysomnographic examination including a recording of respiratory movements, and pulse oximetry. SAS was diagnosed in 68.4 p. 100 in the stroke group compared with 28.6 p. 100 in the insomnia group (p = 0.01). The median apnea/hypopnea index was significantly higher in the stroke group: 25 events per hour versus 2 in the insomnia group (p = 0.01). The median lowest oxygen saturation was not significantly different in the two groups (p = 0.3). Snoring and previous daytime sleepiness were both correlated with stroke (p = 0.05, p = 0.003). Among sleep complaints and cardiovascular risk factors, only a history of diabetes and previous daytime sleepiness were found to be significantly associated with stroke (p = 0.01, p = 0.002). Mortality was higher in SAS subjects (58.8 p. 100) than in non SAS subjects (33 p. 100). The difference was not statistically significant (p = 0.14), but a tendency could be noticed. SAS is a common finding in elderly individuals after a stroke. Physicians must be aware of the risks of prescribing sedatives and anxiolytics to these patients.