BACKGROUND: Clinical stability has been observed with continued antiretroviral therapy (ART) in the setting of partial virological suppression. The optimal time to switch treatment in patients with low but detectable HIV-1 RNA is not known. METHODS: Subjects on stable ART with HIV-1 RNA 200-10,000 copies/ml were randomized to an immediate treatment switch, or to a delayed switch when HIV-1 RNA increased to > or = 10,000 copies/ml or CD4+ T-cell count decreased by 20%. The primary outcome measures were immune activation (proportion of CD8+ T-cells expressing CD38 at week 48) and evolution of genotypic drug resistance. RESULTS: The study failed to fully accrue the originally planned 108 subjects. Only 47 subjects were randomized to immediate- or delayed-switch arms. Of the subjects in the delayed-switch arm, 10/23 (43%) met the criteria for ART switch during the study (median follow-up 82 weeks). After 48 weeks of observation, the level of immune activation was comparable in the two arms. New resistance mutations were observed in 3/17 and 8/19 subjects in the immediate- and delayed-switch groups, respectively. The loss of future treatment options, however, was comparable in the delayed- and immediate-switch groups. CONCLUSIONS: Individuals with partial viral suppression tend to remain immunologically stable, however, the accumulation of drug resistance mutations is an ongoing risk. Delayed switch in ART may be a reasonable short-term strategy for individuals with very limited treatment options.