The safety and efficacy of cidofovir gel for treatment of acyclovir-unresponsive herpes simplex virus infections in AIDS patients was evaluated in a randomized, double-blind, multicenter trial. Cidofovir (0.3% or 1%) or placebo gel was applied once daily for 5 days. Ten of 20 cidofovir-treated and none of 10 placebo-treated patients had complete healing or >50% decreased area (P = .008); 30% of cidofovir-treated patients versus 0 placebo recipients had complete healing (P = .031). Viral shedding ceased in 13 (87%) of 15 cidofovir-treated and 0 of 9 placebo-treated patients (P = .00004). For cidofovir-treated patients, median time to complete or good response was 21 days, and median time to negative viral culture was 2 days (P = .025, P = .0001, respectively). Median lesion area decreases were 58% for cidofovir-treated versus 0 for placebo-treated patients (P = .005), and mean pain score changes were -1.84 versus -0.34 (P = .042). Application site reactions occurred in 25% of cidofovir-treated and 20% of placebo-treated patients; none was dose-limiting. Cidofovir therapy provided significant benefits in lesion healing, virologic effect, and pain reduction.