We evaluated self-administered questionnaires and short sleep studies in screening for sleep-disordered breathing (SDB) in 40 hypertensive men ages 36-66 unselected for symptoms. Each subject completed a questionnaire including questions on sleep-related symptoms and underwent overnight polysomnography in which we evaluated the apnea-hypopnea index (AHI) and the percentage of time during which arterial O2 saturation was less than 90% (T90). The first 90 min of overnight study was evaluated separately, and 10 subjects with an AHI greater than or equal to 10 also underwent late afternoon nap study. By overnight polysomnography, 48% of the cohort had an AHI greater than or equal to 10, and 35% had a T90 greater than or equal to 10%. Using linear regression, we found no features of the symptom questionnaire that strongly predicted AHI. Only self-reported snoring and baseline arterial Po2 significantly predicted T90. The AHI and T90 were not significantly correlated. Considering an AHI greater than or equal to 10 in the overnight study as "abnormal" and an AHI greater than or equal to 10 on the short study as a "positive" test, the specificity of the AHI in the first 90 min was 100% (21/21), and the sensitivity was 42% (8/19). The sensitivity of the nap study was 60% (6/10). We conclude that in a cohort unselected for symptoms, the ability of self-administered questionnaires to predict SDB was low; short studies were only moderately sensitive for detecting an AHI greater than or equal to 10, and the AHI was not a major determinant of nocturnal desaturation.