A total of 208 patients underwent brain biopsy for presumptive herpes simplex encephalitis and were randomized to receive either vidarabine, vira-A, at 15 mg/kg/day, or acyclovir, at 30 mg/kg/day for ten days. 69 patients (33%) had biopsy-proven disease; 37 received vira-A and 32 acyclovir. With the exception of age, patient populations were balanced for demographic characteristics. Overall survival for acyclovir recipients was 72% compared with 46% for vira-A-treated patients 18 months after therapy (p = 0.008). After adjustment for differences of age between treatment populations by multivariant regression analyses, acyclovir treatment remained superior to vidarabine therapy (p = 0.041). Mortality varied according to the level of consciousness at the onset of therapy. For lethargic, semicomatose and comatose patients, mortality was 42%, 46%, and 67%, respectively, for the vira-A-treated patients and 0%, 25% and 25%, respectively, for acyclovir-treated patients. Six months post-therapy morbidity assessments revealed five (14%) vira-A versus 12 (38%) acyclovir recipients who had returned to normal function, while eight (22%) and three (9%), respectively, had moderate debility. Outcome differences were significant (p = 0.02; Wilcoxon, 2-sample test) using an adapted scoring system. Age and Glasgow coma scale greater than 10 predicted the best outcome following acyclovir treatment. Disoriented patients who flex and respond by eye to pain had no mortality and 50% returned to normal. These data indicate that acyclovir is the treatment of choice for biopsy-proven herpes simplex encephalitis.