Objective.To evaluate changes in serum HIV p24-antigen levels in a subset of patients who participated in a European/Australian double-blind, placebo-controlled trial evaluating the efficacy of zidovudine (250 mg every 6 h) alone or in combination with acyclovir (800 mg every 6h) in patients with AIDS, AIDS-related complex (ARC) or Kaposi's sarcoma (KS). Design.Double-blind, placebo-controlled randomized clinical trial of ≤6 months' therapy. Setting.Samples were obtained from patients attending teaching hospital outpatient clinics in seven European countries and Australia. Subjects.One hundred and ninety-seven HIV-infected patients (60 with AIDS and 137 with ARC or KS). Main outcome measures.Serum HIV p24-antigen levels measured using the Abbott HIV solid-phase enzyme immunoassay. Results.Of 76 ARC/KS patients who were initially HIV p24-antigen-positive, one out of 25 randomized to placebo, eight out of 23 to zidovudine and 11 out of 28 to the zidovudine/acyclovir combination became antigen-negative. The proportion of patients who became antigen-negative was significantly higher in both the zidovudine group (P = 0.016) and the zidovudine/acyclovir group (P = 0.004), compared with the placebo group. There were no statistical differences between the zidovudine and the zidovudine/acyclovir groups. During the trial p24-antigen levels in the zidovudine-treated patients reached their minimum after 4–8 weeks of therapy, and tended to increase gradually thereafter. Disease progression occurred irrespective of whether p24-antigen levels declined during therapy. No association between p24-antigen responses to therapy and baseline disease stage, Karnofsky score or baseline CD4 cell count was detectable. Conclusion.Acyclovir does not potentiate the effect of zidovudine on p24-antigen levels. Change in antigen level in response to antiviral therapy needs further investigation before it is used as a surrogate marker for clinical efficacy of antiviral therapy.