Study objectives: To evaluate the return of spontaneous circulation (RSC) rates in human victims of cardiac arrest treated with standard doses of epinephrine (SDE) or high-dose epinephrine (HDE). Design: Prospective case series. Setting: A university hospital emergency department during 1987 through 1989. Participants: Forty-nine adult victims of nontraumatic cardiac arrest. Interventions: At the discretion of the treating physician, patients received epinephrine in bolus doses ranging from 1 to 15 mg. HDE was defined as a dose of at least 0.2 mg/kg; smaller doses were defined as SDE. Patients were grouped as + RSC if they developed a sustained spontaneous palpable pulse or blood pressure and as - RSC if they did not develop a pulse or blood pressure. Measurements: Patients were grouped as + RSC if they developed a sustained spontaneous palpable pulse or blood pressure and as - RSC if they did not develop a pulse or blood pressure. Patients were also grouped by their presenting rhythm. Potentially perfusing rhythm was electro-mechanical dissociation or ventricular tachycardia. Nonperfusing rhythm was asystole or ventricular fibrillation. Rates were analyzed using the Fisher exact test and the two-tailed unpaired t test. HDE improved the rate of initial resuscitation (P = .008). The effect was greatest in patients with nonperfusing rhythms (P = .014) and disappeared when evaluating patients with potentially perfusing rhythms. No patient survived to hospital discharge. Conclusion: High-dose epinephrine improves initial resuscitation rates in human victims of cardiac arrest. Its greatest effect is in patients with a nonperfusing rhythm.