We conducted a double-blind, placebo-controlled trial to assess the value of vidarabine therapy for the prevention of complications from herpes zoster in immunocompromised patients. Of 121 patients with localized herpes zoster of 72 hours' duration or less, 63 received vidarabine and 58 received the placebo. Populations were matched for pertinent characteristics. Therapy accelerated cutaneous healing and decreased the rates of cutaneous dissemination (from 24 per cent [14 patients] to 8 per cent [5 patients]) (P = 0.014); and of zoster-related visceral complications (from 19 per cent [11 patients] to 5 per cent [3 patients]) (P = 0.015). Therapy also decreased the total duration of post-herpetic neuralgia (P = 0.047). Patients with lymphoproliferative cancers and those 38 years of age or older were at greatest risk for complications and benefited most from therapy. There was no serious drug toxicity. We conclude that vidarabine therapy, when started within the first three days, is valuable for the reduction of complications related to herpes zoster. (N Engl J Med. 1982; 307:971–5.) Herpes zoster, a common infection in immunocompromised patients, can be life-threatening if the viscera become involved.1 2 3 4 As a consequence, herpes-zoster infections have become targets for experimental antiviral trials. In 1976, we reported the results of a blind, controlled, crossover study of vidarabine therapy for herpes-zoster infections in the immunocompromised host, which demonstrated accelerated cutaneous healing,5 especially in patients who were 38 years of age or less, had lymphoproliferative cancers, and received therapy within six days of the onset of disease.6 Because of the crossover design of the study, the value of vidarabine for the prevention of complications (such as post-herpetic. © 1982, Massachusetts Medical Society. All rights reserved.