OBJECTIVE To determine the within-subject variability and to estimate the quantity of occult aspiration of nasopharyngeal secretions during sleep in normal humans. METHODS Prospective duplicate full-night sleep studies. METHODS Pulmonary sleep laboratory, university hospital. METHODS Ten normal male volunteers aged 22 to 55 years. METHODS Two full-night polysomnographic recordings with infusion of 2 mL/h radioactive 99mTc tracer into the nasopharynx through a small catheter during EEG-documented sleep. Standard lung scans were conducted immediately following final awakening. Aspiration was defined as the presence of radioactivity in the pulmonary parenchyma on two separate views. RESULTS A mean sleep efficiency of 85.7 +/- 2.6% was found with no difference between the two study nights. A total of 5 of the 10 subjects studied had tracer evident in the pulmonary parenchyma following final awakening. Three had the tracer apparent following the first-night study and four had tracer apparent following the second-night study. Thus, two subjects aspirated on both nights. Comparing the subjects who aspirated with those who did not, no significant difference could be found for age, time spent in bed, sleep efficiency, apnea-hypopnea index, arousal plus awakening index, or percent of sleep time spent in a supine position. The quantities of tracer aspirated were on the order of magnitude of 0.01 to 0.2 mL. CONCLUSIONS Aspiration measured by this technique occurs commonly in healthy young men during sleep, is unrelated to sleep quality, and is variable within subjects studied on more than one occasion. The quantity aspirated is of an order of magnitude likely to contain bacterial organisms in physiologically significant quantities.