In this double-blind, placebo-controlled trial of HIV-infected asymptomatic haemophiliacs, the efficacy of 2-year zidovudine therapy (1000 mg daily in two divided doses) in preventing progress of HIV infection was prospectively evaluated. Drug tolerance was also studied. 143 haemophiliacs from five European countries and Australia with p24 antigenaemia and/or CD4 cell counts of 0.1-0.4 x 10(9)/l were enrolled. The main measures of outcome were progression to AIDS, CDC group IV disease, symptomatic HIV-related disease, and a decrease in CD4+ T-lymphocyte count to fewer than 0.2 x 10(9)/l. There were no significant treatment differences in the proportion of patients progressing to AIDS, CDC group IV or symptomatic disease. Analysis of time to CD4+ counts less than 0.2 x 10(9)/l showed a non-significant trend in favour of zidovudine. Haemoglobin concentrations were less than 8 g/dl in 4% of zidovudine recipients; neutropenia was less than 0.75 x 10(9) cells/l in 5% of zidovudine recipients; alanine aminotransferase levels were greater than 10 times the upper normal limit in 3% of zidovudine recipients, but also in 4% of placebo recipients. Hence there was a very low prevalence of side-effects in haemophiliacs, despite the use of a higher zidovudine dosage than is currently widely used.