Background. Intermittent episodes of detectable human immunodeficiency virus (HIV) viremia (hereafter referred to as "blips") are generally not predictive of subsequent virologic failure. Limited data are available for patients treated with nonnucleoside reverse-transcriptase inhibitor (NNRTI)-based regimens. Methods. A retrospective cohort study evaluated patients receiving highly active antiretroviral therapy who were followed for ≥12 months, achieved an HIV RNA load of <50 copies/mL, and underwent evaluation every 2-3 months. A blip was defined as 1 HIV RNA load measurement of 50-1000 copies/mL that was preceded and followed by another HIV RNA load measurement of <50 copies/mL. The frequency and predictors of blips and virologic failure were studied. Results. There were 244 patients in the NNRTI group and 136 patients in the protease inhibitor (PI) group. Baseline characteristics between the 2 groups were similar. A total of 34% of patients in the NNRTI group and 33% in the PI group experienced viral blips (P = .855), with corresponding incidences of 19.2 and 19.7 blips, respectively, per 100 person-years. Median time to blips was 50.0 months after initiation of therapy in the NNRTI group (95% confidence interval [CI], 44.8-55.3 months) and 43.6 months in the PI group (95% CI, 33.7-53.6 months; P = .632, by the log-rank test). By Cox proportional hazards model analysis, only a history of antiretroviral therapy use (hazard ratio [HR], 2.01; P < .001) and a CD4 cell count of <200 cells/μL (HR, 1.70; P = .021) increased the risk for having a blip. During a median follow-up period of 23.5 months, 7.8% of patients in the NNRTI group and 8.1% in the PI group experienced virologic failure (P = .917). Cox proportional hazards model analysis showed that only a baseline CD4 cell count of <200 cells/μL predicted virologic failure (HR, 2.74; P = .032). Conclusions. There is no difference in the frequency or prognostic significance of viral blips between patients receiving NNRTI-based therapy and patients receiving PI-based therapy. Our results suggest that viral blips occur at a similar rate among patients receiving NNRTI-based regimens and patients receiving PI-based regimens and that they are not predictive of virologic failure. © 2005 by the Infectious Diseases Society of America. All rights reserved.