A comparative assessment of vidarabine and acyclovir in the treatment of varicella and disseminated zoster in immunosuppressed patients was undertaken. Thirty-eight immunosuppressed patients with varicella (N = 18) or disseminated zoster (N = 20) were treated intravenously with 10 mg/kg/day of vidarabine or 30 mg/kg/day of acyclovir for 5 days according to a preestablished code within each diagnosis group--varicella and disseminated zoster. Two deaths, although not directly related to VZV infection, were observed in the vidarabine-treated varicella group. The times to cessation of formation of new lesions and to the disappearance of fever were similar for vidarabine and acyclovir in each group. In the varicella group, VZV was isolated on day 5 in four out of five vidarabine patients versus one out of five acyclovir patients. No severe adverse effects were observed with either drug. Neutropenia present in patients of both drug groups was transitory and most often related to previous cytolytic chemotherapy. These data suggest that either vidarabine or acyclovir could be used in the treatment of severe VZV infections in immunosuppressed patients, although a larger number of patients would be required for definitive conclusion. Because of the large amount of solute required for vidarabine administration, acyclovir may be preferred when the risk of cardiorespiratory failure is high.