We studied the effect of thiopental loading during resuscitation of 53 patients following cardiopulmonary arrest and compared the outcome with that found in 54 patients treated conventionally in the 30 previous months. Thiopental therapy (10 mg/kg i.v.) was begun within 30 min of the arrest once hemodynamic stability had been established. Thiopental infusion (20 mg/kg over 6 h) was followed by phenobarbital sodium (125 mg every 12 h), tolerance to the initial dose having been assessed. The in-hospital mortality rate for both groups was similar. In patients with ischemic heart disease, the mortality rate within the first 6 h was significantly higher in the thiopental group (p less than 0.05), although for the remaining patients there were more survivors among the thiopental treated patients (p less than 0.05). Excluding the patients who died within the first 6 h, 61% of the patients in the thiopental group survived cardiac arrest with normal cerebral performance, whereas only 37% in the standard treatment group showed normal functional outcome (p less than 0.03). These results suggest a favorable neurologic effect of thiopental loading during resuscitation of patients without ischemic heart disease. In patients with ischemic heart disease, an initial hemodynamic deterioration may contribute to minimising the beneficial effect of barbiturate therapy.