The ventilatory response to hypercapnic progressive hypoxia and the breathing pattern during steady-state hypercapnic hypoxia were compared before and after intravenous infusion of 3 mg of naloxone in a relatively large number of healthy adults (n = 21). In addition, the withdrawal response from hypercapnic hypoxia (modified transient O2 test) was measured to investigate the possible role of endogenous opioids in the peripheral chemoreceptors. The average ventilatory response (delta VE/delta SaO2) increased significantly from 0.51 +/- SD 0.26 to 0.65 +/- 0.42 L/min/% (p less than 0.05) after naloxone infusion, whereas there were no significant changes between two tests with normal saline in the control study (n = 7). Because there was considerable interindividual variation in the response to naloxone administration, we selected "high responders" (n = 8) who showed larger increases with naloxone than the upper limit of the 95% confidence interval for the change with the second saline in the control study. They showed greater delta VE/delta SaO2 (p less than 0.01), respiratory frequency (p less than 0.01), and mean inspiratory flow (p less than 0.01) during hypercapnic hypoxia before naloxone infusion than did the other subjects. There was no significant change in the withdrawal response before and after naloxone infusion, even in such high responders. We conclude that endogenous opioids participate in the control of breathing in normal adults during hypercapnic hypoxia. This may be particularly true for those subjects who exhibit greater chemosensitivity to hypercapnic hypoxia. Endogenous opioids appear to act centrally rather than peripherally.