We tested the effect of human leukocyte interferon on early localized herpes zoster infections in three placebo-controlled, randomized double-blind trials involving 90 patients with cancer. There were no significant differences in pretreatment severity of infection or nature of underlying disease in the groups. Higher dosages of more purified interferon in the second and third trials produced a significant (P less than or equal to 0.01) decrease in cutaneous dissemination. No dissemination occurred in those receiving the highest dosage (5.1 x 10(5) U per kilogram per day) (P less than or equal to 0.025). The number of days of new-vesicle formation in the primary dermatome decreased (mean, 2.3 days, P less than or equal to 0.05) in this group. Treated patients had a trend toward less acute pain, and significantly (P less than or equal to 0.05) diminished severity of post-herpetic neuralgia, at the two highest dosage levels. Visceral complications were six times less frequent in interferon recipients. High-dosage interferon appeared effective in limiting cutaneous dissemination, visceral complications and progression within the primary dermatome.