In treating obstructive sleep apnea, positive pressure applied through the nose (CPAP) might cause a reflex increase in upper airway muscle activity or might enlarge the airway passively. We studied the effect of CPAP applied by a nasal mask on the electromyographic (EMG) activation of the alae nasi and genioglossal muscles in 8 patients with obstructive apneas during sleep, and correlated EMG activity with concentrations of oxygenation by ear oximeter, and with the end-expiratory position of the rib cage and abdomen by DC-coupled inductance plethysmography. One to 3 cm H2O of CPAP did not eliminate the cyclic occurrence of obstructive apneas. The greatest tonic and phasic EMG activity occurred at apnea termination; the least occurred at apnea onset. With 13 to 15 cm H2O CPAP, apneas were eliminated; mean oxygen saturation rose from 84 +/- 6% (mean +/- SD) to 92 +/- 2%, and EMG activity was reduced or eliminated. With abrupt lowering of CPAP, end-expiratory positions fell, and an obstructive apnea ensued; however, EMG activity did not immediately return. We conclude that the elimination of apneas with CPAP is not attributed to increased EMG activity in the upper airway. The reduction in EMG activity observed with nasal CPAP was closely related to the improvement in hemoglobin oxygen saturation. Therefore, CPAP may act as a pneumatic splint and passively open the upper airway to prevent obstructive apnea.