Nine of 16 immunocompromised patients with laboratory-documented cytomegalovirus (CMV) disease were given intravenous acyclovir in a randomized, placebo-controlled, double-blind treatment study. Fifteen of the 16 patients were organ transplant recipients, 11 of whom had a renal allograft. Patients given acyclovir (500 mg/m2 three times daily for seven days) experienced no major side effects. Acyclovir recipients had a significantly faster rate of improvement (p = 0.0437, Breslow test; p = 0.0595, Mantel-Cox), and a more rapid rate of defervescence (p = 0.0338, Breslow test; p = 0.0208, Mantel-Cox). The median day of improvement was 7 for the acyclovir recipients, and at least 31 for the placebo recipients. The median day afebrile was 13 for the acyclovir group and at least 31 for the placebo group. Three of the nine acyclovir recipients died, two of whom were semicomatose and septic with Bacteroides fragilis before the first dose of acyclovir was given. Four of the seven placebo recipients died. Although CMV could be recovered from throat and urine of acyclovir recipients throughout the study, viremia ceased after the first day of drug infusion. We conclude that acyclovir may be useful for treatment of CMV disease in certain immunosuppressed patients especially renal allograft recipients. Because the clinical response to acyclovir was not uniform and the number of patients studied was small, we are conducting another larger controlled trial of acyclovir for febrile renal transplant patients with CMV infections to extend and clarify the results of the first study. Patients randomized to acyclovir will receive 500 mg/m2 three times daily for two weeks. The protocol includes in vitro testing of patients' CMV isolates for susceptibility to acyclovir and monitoring plasma concentrations of the drug.