A double-blind placebo controlled trial of foscarnet was conducted in 32 Human Immunodeficiency Virus (HIV) antibody positive male homosexuals with a presumed AIDS pneumonia. The study was designed to evaluate the importance of treating cytomegalovirus (CMV) as a possible lung pathogen of these patients and as a toxicity study of foscarnet. Trial subjects were randomised to receive either foscarnet or placebo as a continuous intravenous infusion for 2 weeks along with conventional therapy against Pneumocystis carinii (PC) pneumonia. Bronchoscopy or post-mortem showed PC to be present in 24 patients and the bronchoalveolar lavage fluid had early antigen foci of CMV in nine, five of these being double infections. The incidence of CMV infection in this group of patients was not sufficiently high to prove or disprove that treatment of CMV speeds recovery or improves prognosis in AIDS pneumonias. Overall however foscarnet was well tolerated with a slight increase in adverse reactions in the treated groups. This agent is therefore a relatively non-toxic drug to use in the treatment of established CMV disease.