AIDS Clinical Trials Group (ACTG) 359 was a randomized, partially double-blinded factorial study of 6 antiretroviral regimens, all including saquinavir, among HIV-infected persons in whom prior therapy had failed (n = 258). Counts of remaining saquinavir capsules were determined between weeks 0 and 4; at weeks 4, 8, and 16, self-reported adherence was estimated from 2-day report of doses skipped, therapeutic coverage, and percent of doses taken were determined by electronic monitoring devices applied to saquinavir bottles, and the saquinavir 24-hour area under the curve (AUC) was estimated. Relationships were evaluated among these 4 adherence measures and the primary endpoint of week 16 HIV RNA change. Thirty percent of 254 subjects had HIV RNA < or =500 copies/mL at week 16. Only self-reported adherence and saquinavir AUC were significantly associated with week 16 HIV RNA change (P = 0.019 and 0.023, respectively), and these measures were higher in subjects with week 16 HIV RNA < or =500 copies/mL (P = 0.03 and 0.008, respectively). The ability to detect a correlation between electronically monitored adherence and virologic response was limited by the small sample size. Self-reported adherence and saquinavir AUC were significant predictors of virologic response, in this evaluation. These findings provide insight into methods of assessing and improving adherence to antiretroviral regimens.