Vasopressin administration has been suggested during cardiopulmonary resuscitation, and a previous clinical trial has suggested that vasopressin is most effective when administered with epinephrine. Adult subjects (n = 325) who received > or =1 dose of intravenous epinephrine during cardiopulmonary resuscitation for nontraumatic, out-of-hospital cardiac arrest were randomly assigned to receive 40 IU of vasopressin (n = 167) or placebo (n = 158) as soon as possible after the first dose of epinephrine. The rate of return of pulses was similar between the vasopressin and placebo groups (31% vs 30%), as was the presence of pulses at the emergency department (19% vs 23%). No subgroup appeared to be differentially affected, and no effect of vasopressin was evident after adjustment for other clinical variables. Additional open-label vasopressin was administered by a physician after the study drug for 19 subjects in the placebo group and 27 subjects in the vasopressin group. Results were similar if these subjects were excluded or were assigned to an actual drug received. Survival duration for subjects admitted to the hospital did not differ between groups. In conclusion, vasopressin administered with epinephrine does not increase the rate of return of spontaneous circulation.