Objective: To determine markers that are associated with the durability of virologic response to therapy with HIV protease inhibitors in HIV-infected individuals. Design: This study encompassed two retrospective analyses of the duration of virologic response to protease inhibitor therapy. The first analysis included 29 patients receiving either monotherapy or combination therapy with the protease inhibitor ritonavir whose plasma HIV RNA levels rebounded from the point of greatest decline with mutations associated with resistance to ritonavir. The second analysis included a cohort of 102 patients who initially responded to randomized treatment with either monotherapy with ritonavir or combination therapy with ritonavir and zidovudine. Methods: Durability of response was defined as the time from the initiation of therapy to the point at which plasma HIV RNA displayed a sustained increase of at least 0.6 log10 copies/ml from the nadir value. In the first analysis, durability of response was analyzed with respect to baseline HIV RNA, HIV RNA at the nadir, and the drop in HIV RNA from baseline to the nadir. In the second analysis, time to rebound was examined using Kaplan-Meier analysis, stratifying by either baseline HIV RNA or HIV RNA at the nadir. Results: In both analyses, the durability of response was not highly associated with either baseline RNA or the magnitude of RNA decline from baseline. Instead, a strong relationship was observed between the durability of response and the nadir plasma HIV-1 RNA value (P < 0.01). The nadir in viral load was generally reached after 12 weeks of randomized therapy. Conclusions: Viral RNA determinations at intermediate timepoints may be prognostic of impending virologic failure of protease inhibitor therapy. Therapeutic strategies that allow intensification of initial antiretroviral regimens in the subset of patients with incomplete virological response before the emergence of high level resistance should be investigated.