To determine whether naloxone infusion is efficacious in severe hyperdynamic septic shock, we conducted a prospective study of 22 patients randomly assigned to a naloxone or placebo group. Patients were treated 12 +/- 2 h (SEM) after the onset of shock, with a mean arterial pressure (MAP) of 63 +/- 3 mm Hg. All patients had clinical evidence of an infectious process and required dopamine 20 +/- 2 micrograms/kg.min. Five (46%) of 11 patients in the naloxone group and one (9%) of the other 11 patients in the placebo group responded clinically. The MAP among the five responders increased from 62 +/- 5 to 89 +/- 4 mm Hg within 20 min of naloxone treatment (p less than .01). This favorable hemodynamic response was sustained throughout the patients' clinical course. In contrast, the MAP did not change significantly in the nonresponders who received naloxone, nor did it change in the placebo group. More patients in the naloxone group than in the placebo group received steroids concurrently. Survival rate was 100% in those who responded to naloxone clinically. However, overall survival rate in each group was essentially the same. No adverse effects were observed, except for mild agitation in some of the patients receiving naloxone. We conclude that naloxone infusion is clinically efficacious in improving the hemodynamic profile of a subgroup of patients with severe early hyperdynamic septic shock, but does not appear to improve the overall survival rate.