OBJECTIVE: The current recommendation that patients infected with HIV-1 be treated early is based on little evidence. We examined whether the early initiation of antiretroviral treatment affects residual HIV-1 viraemia. METHODS: Viraemia was measured using an assay with a detection limit of 3 HIV-1 RNA copies/ml in drug-naive patients who started antiretroviral therapy at the time of primary HIV-1 infection (PHI) (n = 10), during chronic infection without immune suppression (CD4 cell counts > or = 500/mm3; median 577) (n = 10), or after immune suppression developed (CD4 cell counts < 500/mm3; median 113) (n = 21). RESULTS: In 249 samples collected 24 to 120 weeks after treatment initiation, the mean proportion of samples with HIV-1 RNA levels of less than 3 copies/ml was 75% for PHI patients compared with 32 and 8% for immunocompetent and immunosuppressed chronically infected patients, respectively. Fifty per cent of PHI patients, but none of the chronically infected patients, had persistently fewer than 3 HIV-1 RNA copies/mL. PHI patients had lower residual HIV-1 RNA levels than chronically infected patients, and immunocompetent patients had lower residual HIV-1 RNA levels than immunosuppressed patients (all pairwise, P< 0.001). The mean residual HIV-1 RNA level was independently associated with the initiation of therapy during PHI and baseline CD4 cell counts (P < 0.001 for both associations). CONCLUSION: Viraemia levels are associated with clinical progression and predict virological treatment failure. The initiation of antiretroviral therapy at the time of PHI and while CD4 cell counts are high results in lower residual viraemia. These results support early antiretroviral therapy in HIV-1-infected patients.