Although obstructive sleep apnea (OSA) appears to be a cardiovascular risk factor, its frequency in patients with transient ischemic attack (TIA) and stroke remains poorly known. We prospectively studied 128 patients (mean +/- SD age = 59 +/- 15 years) with stroke (n = 75) or TIA (n = 53). Assessment included body mass index (BMI); history of snoring and daytime sleepiness; cardiovascular risk factors and diseases; and severity of stroke (Scandinavian Stroke Scale = SSS). Polysomnography (PSG) was obtained in 80 subjects (group 1), a mean of 9 days (range, 1-71 days) after TIA or stroke. In 48 subjects (group 2), PSG was not available, refused, or inadequate. Groups 1 and 2 were similar with the exception of gender distribution. Clinical and PSG data were compared to those of 25 healthy controls matched for age, gender, and BMI. An apnea-hypopnea index (AHI) > 10 was found in 62.5% of subjects and 12.5% of controls. Between patients and controls there was a significant difference in AHI (mean [range]: 28 (0-140) vs 5 (0-24), p < 0.001), maximal apnea duration (mean + SD: 37 +/- 23 vs 23 +/- 13 seconds, p = 0.009), and minimal oxygen saturation (mean + SD: 82 +/- 10% vs 90 +/- 5%, p < 0.001). Conversely, frequency and severity of OSA were similar in stroke and TIA subjects. Multiple regression analysis identified age, BMI, diabetes, and SSS as independent predictors of AHI. Sleep apnea has a high frequency in patients with TIA and stroke, particularly in older patients with high BMI, diabetes, and severe stroke. These results may have implications for prevention, acute treatment, and rehabilitation of patients with acute cerebrovascular diseases.