Since prevalent cohorts may be biased by the duration of human immunodeficiency virus (HIV) infection (onset bias), it is useful to assess the potential predictive value of markers in incident cohorts of HIV-positive subjects for whom the date of seroconversion is known or can reliably be estimated. Of 131 homosexual men with HIV-1 seroconversion from New York City and Washington, DC, who were evaluated annually beginning in 1982, 60 developed acquired immunodeficiency syndrome (AIDS) by the end of 1989. The prognostic significance of immunologic markers (proportion of CD4+ T-lymphocytes, neopterin, ß2-microglobulin, serum interferon, and anti-p24 antibody) and of a virologic marker (HIV p24 antigen) was determined using measurements made at defined time intervals after the known or estimated date of HIV seroconversion. When measurements made 3 years after seroconversion were used, all markers except anti-p24 antibody were found to be significant estimators of AIDS risk in univariate analyses. In multivariate Cox regression modeling, the maximum information was obtained by including neopterin, interferon, and the CD4+ T-lymphocyte proportion. The predictive value of markers after HIV seroconversion could change considerably from one interval to another. Elevated levels of ß2-microglobulin and neopterin significantly predicted the development of Kaposi‘s sarcoma. These two markers were highly correlated (r=0. 74). The authors conclude that immunologic markers can be important for an HIV staging system for estimating prognosis and facilitating early therapeutic intervention in HIV-positive patients.