We conducted a placebo-controlled, double-blind study of acyclovir therapy for acute herpes zoster in immunocompromised patients. Of the 94 patients enrolled in the study, 52 had localized skin lesions at entry, and 42 had disseminated cutaneous zoster. A one-week course of intravenous acyclovir (1500 mg per square meter of body-surface area per day) halted progression of zoster in both groups, as determined by development or progression of cutaneous dissemination, development of visceral zoster, or proportion of cases deemed treatment failures. Significantly fewer patients treated with acyclovir within the first three days after the onset of exanthem had complications of zoster, as compared with patients treated with placebo (P = 0.02 by Fisher's exact test), but acyclovir also stopped progression of zoster in patients treated after three days of rash (P = 0.05 by Fisher's exact test). Acyclovir recipients with disseminated cutaneous zoster had a significantly accelerated rate of clearance of virus from vesicles, as compared with placebo recipients (P = 0.05 by the Breslow test).