The objectives of this study were to evaluate the prognostic value of a single serum HIV-RNA measurement 6-24 months after HIV seroconversion, and to investigate whether any differences in outcome related to the clinical course of primary infection could be explained by presumed steady-state HIV-RNA levels at 6-24 months after seroconversion. Disease progression was analyzed by life tables and Cox proportional hazard models. A total of 93 HIV seroconverters followed for a median of 76 months (range, 19-143 months) were included in the study. The main outcome measures were development of AIDS stratified by age, the year of seroconversion, serum HIV-RNA levels, CD4 cell counts, and duration of primary illness. The proportion of patients who developed AIDS was 36% (95% confidence interval, 22%-50%) at 8 years and 50% (26%-75%) at 10 years. In the unadjusted analyses, clinical progression was significantly associated with serum HIV-RNA levels, CD4 cell counts, and duration of primary illness. The adjusted analyses indicated that HIV-RNA was the strongest predictor. Patients with long-lasting symptoms associated with primary infection had significantly higher serum HIV-RNA levels than those with less severe presentations (median counts, 11,660 vs. 2880 copies/ml, p = 0.001). It is concluded that the serum HIV-RNA level in early HIV infection is a strong independent predictor of clinical progression. Patients with long-lasting primary illnesses reach a higher viral load in steady state after seroconversion, which is probably the main reason for the poorer prognosis observed in this group of patients.