A prospective, randomized study using either bretylium tosylate (BT) or lidocaine (L) as the first-line antiarrhythmic for patients in refractory ventricular fibrillation was conducted using the Milwaukee County Paramedic System. If the patient did not respond to the initial American Heart Association protocol, BT (10 to 30 mg/kg total) or L (2 to 3 mg/kg total) was given randomly as the first antiarrhythmic. If the patient failed to convert, the alternate antiarrhythmic was given. In the L group, 81% (39/48) of the patients obtained an organized electrical rhythm and 56% (27/48) converted to a rhythm with a pulse. The resuscitation rate (admission to an emergency department with pulse) was 23% (11/48), and the save rate was 10.4% (5/48). In the BT group, 74% (32/43) obtained an organized electrical rhythm, 35% (15/43) were converted, 23% (10/43) were resuscitated, and 5% (2/43) were saved. The only significant difference in outcome was that L converted patients better than did BT (P less than .05). Of the 24 patients known to be on digitalis preparations prior to arrest, 41% (5/12) in the L group were resuscitated and 16% (2/12) were resuscitated in the BT group. Data were analyzed for witnessed arrest outcome and for patients given multiple antiarrhythmics.