To define the utility of high-dose barbiturate therapy following an episode of complete global cerebral ischemia, we investigated the effects of 60 mg/kg of thiopental given to cats five minutes after resuscitation from 12, 14, and 16 min of electrically induced ventricular fibrillation (VF). All aspects of the arrest, resuscitation, with post-arrest care were carefully controlled, with the EEG becoming isoelectric 20-25 s after the onset of VF, a 89-91 per cent rate of successful resuscitation, with an overall mean resuscitation time of 2.5 ± 0.2 (SEM) min. For any given duration of VF, there were no differences (control vs thiopental) in any pre- or post-arrest parameters (blood pressure, blood gases, electrolytes, etc.). A total of 68 resuscitated cats were entered into various treatment and control groups, and all but one group received 20-24 h of post-resuscitation paralysis, mechanical ventilation, and ICU support before being extubated. Cats received an additional six days of aggressive nursing care, and daily examinations were performed with the assignment of a neurologic deficit score (NDS) before 0 (normal) and 100 (brain dead). Autopsies were performed to determine the cause of death in animals which died before the end of the seven-day observation period. The early post-arrest period was marked by the occurrence of repetitive, rhythmic bursts of high-frequency electroencephalographic (EEG) activity (? seizures) in 38 per cent of control animals (16/42, all arrest times combined). Ten of these animals died as a result or severe neurologic injuries. By contrast, only 12 per cent of treated cats (3/26) developed similar EEG patterns (P < 0.05) and there were no neurologic deaths in the thiopental groups. The differences in the incidence of neurologic deaths (control vs. thiopental) was significant (P < 0.02). The change in overall mortality did not quite reach significance (36 per cent vs. 21 per cent), and treatment had no effect on the incidence of deaths due to cardiovascular causes (e.g., myocardial infarctions). In spite of the effects on mortality, treatment had no effect on the neurologic function of survivors (assessed by NDS). These findings suggest that thiopental improved survival rates by suppressing an unusual post-arrest EEG pattern (? anticonvulsant effect), but had no additional cerebral protective effects.