We performed a randomized, double-blind, placebo-controlled trial to assess the efficacy and safety of inosine pranobex (Isoprinosine) [corrected] in the treatment of patients with human immunodeficiency virus (HIV) infection but without manifest acquired immunodeficiency syndrome (AIDS). A total of 866 patients were enrolled in 21 centers in Denmark and Sweden. The patients were stratified in three groups according to their CD4+ cell count and randomly assigned to receive either inosine pranobex (1 g three times a day) (n = 429) or matching placebo (n = 437) for 24 weeks. Of the 831 patients who could be evaluated, AIDS developed in 17 in the placebo group as compared with 2 in the inosine pranobex group (P less than 0.001; odds ratio, 8.6 [95 percent confidence limits, 2.2 and 52.6]). There were no significant differences between the groups with respect to changes in CD4+ cell count or the development of other HIV-related conditions, with the exception of thrush, which developed in fewer patients in the inosine pranobex group (P = 0.05). No serious side effects were observed. We conclude that treatment with inosine pranobex delays progression to AIDS in patients with HIV infection. The duration of this beneficial effect, the optimal dose, and the mode of action of inosine pranobex remain to be clarified.