Our objective was to evaluate the efficacy and safety of zidovudine (250 mg every 6 h) alone or in combination with acyclovir (800 mg every 6 h) as treatment for AIDS-related complex (ARC). A double-blind, controlled clinical trial of 6 months therapy was conducted at teaching hospital ambulatory clinics in eight European countries and Australia; 199 patients were studied. Time to development of AIDS-defining opportunistic infections (OI) and AIDS-associated neoplasms, survival, performance status, body weight and CD4+ cell counts were measured. During the study six (9%) zidovudine recipients, five (7%) combination recipients and 12 (18%) placebo recipients developed AIDS-defining OI; the probability of developing an OI was 0.23, 0.09 and 0.08 for the placebo, zidovudine and combination recipients, respectively. Four patients in the placebo group, three in the zidovudine group and one in the combination group died during the study. Patients receiving zidovudine with or without acyclovir had moderate increases in CD4+ cell counts compared with placebo recipients and serum HIV p24 antigen level decreased significantly in all those receiving zidovudine. Fourteen (21%) patients in the zidovudine group and 16 (24%) in the combination group experienced bone-marrow suppression compared with three (5%) placebo recipients. Red-cell transfusions were administered to 6, 19 and 13% of placebo, zidovudine and combination recipients, respectively. These data confirm the efficacy of zidovudine therapy after 4 weeks' treatment in the reduction of development of OI in patients with ARC and support the use of a maintenance dose of 250 mg zidovudine 6-hourly. Given the increased development of OI in the treated groups compared with placebo during the first 4 weeks of therapy, we cannot exclude an initial adverse effect of zidovudine and recommend caution in the use of a loading dose of zidovudine. At 6 months there was no apparent difference in efficacy between the combination of zidovudine and acyclovir compared with zidovudine alone. Moreover, the addition of high-dose acyclovir resulted in a minimal increase in the risk of toxicity.