To assess the effect of 3 oz of 80-proof alcohol on the frequency and severity of obstructive sleep apnea (OSA), we studied six OSA patients and six healthy subjects on 2 nights. During the 1st night, when no alcohol was given, five patients demonstrated mild and one severe OSA episodes associated with a decline in arterial oxygen saturation to at least 92% (hypoxic event). On the 2nd night after ingesting 3 oz of alcohol just prior to bedtime, all the patients demonstrated a significant increase in the number and/or severity of hypoxic events compared with the no-alcohol night. Furthermore, the most severe hypoxic events occurred within 80 to 160 min after sleep onset, a significantly shorter latency after sleep onset than on the no-alcohol night. In contrast, the healthy subjects had no incidents of hypoxic events or breathing abnormalities during sleep after ingesting 0.8 gm/kg of alcohol. Possible mechanisms for these results are discussed. An OSA provocation test using alcohol is proposed during a 2nd night of evaluation for patients with mild to moderate or intermittent OSA conditions, but not for patients demonstrating severe hypoxic events or with alcohol intolerance. The alcohol provocation test would serve to determine the influence of alcohol on the frequency and severity of hypoxic events, providing the patient with a measure of the adverse effects of social drinking on their condition.