Cortisol levels were measured before and after administration of naloxone-HCl in patients with affective disorder (n = 16) and normal control subjects (n = 8). On two consecutive days, 20 mg of naloxone-HCl or placebo was administered i.v. over 15 minutes in a double-blind crossover design. Blood samples were collected at 30, 15, and l minute(s) both before and after infusion. Cortisol rose from a mean baseline level of 14.8 microgram% to a mean peak level of 23.1 microgram% following the naloxone administration. Significant cortisol increases were found in both the 15- and 30-minute samples during the naloxone session. There were no differences between patient and normal subject samples or between diagnostic groups. A subgroup of manic patients who had responded to naloxone with a reduction of their manic behavior also had an attenuated cortisol response to naloxone. This proved to be an artifact secondary to variability in the cortisol response in these patients.