A randomized double-blind, placebo-controlled, multicenter investigation assessed the usefulness of acyclovir in the treatment of immunosuppressed children with chickenpox. Twelve patients received placebo and eight received acyclovir. If the event of clinical deterioration, patients could be removed from the study to receive acyclovir. Eighteen patients had skin lesions within 96 hours of admission to the study. Nineteen patients had malignancies. The two groups of patients were similar in age, in concomitant or preceding immunosuppressive therapy, in status of malignancy, and in presenting granulocyte and lymphocyte counts. Zoster immune globulin or plasma had been given to 50% of the placebo group but to only 25% of the acyclovir group. One patient in each group had pneumonitis at entry. Of the patients without pneumonitis at entry, five of the 11 placebo patients compared with none of the seven acyclovir patients developed pneumonitis during treatment (P = 0.054). No evidence of toxicity related to acyclovir was observed.