Optimal adherence is essential for successful antiretroviral therapy. We analyzed the relation between minimum plasma drug concentration (Cmin) and total drug exposure over 24 hr (AUC24) with virologic failure for therapy-adherent patients in the nevirapine (NVP) and efavirenz (EFV) groups of the double nonnucleoside study (2NN), which compared the efficacy of NVP and/or EFV together with stavudine and lamivudine. The objective was to find cutoff values of the Cmin and AUC24 below which the risk of virologic failure increased. The relation between Cmin and AUC24 with virologic failure (never a plasma viral load [pVL] < 50 copies/ml or a rebound to two consecutive pVL > 50 copies/ml) was analyzed with proportional hazard analyses. Data were censored at end of study or change of allocated treatment. The risk of virologic failure with NVP (n = 511) started to increase at a Cmin < 3.1 mg/L (hazard ratio [HR], 1.33; 95% confidence interval [CI], 0.89-1.97), but there was no cutoff value below which a statistically significant increased risk occurred. Neither was such a cutoff point identified for the AUC24. The risk of virologic failure with EFV (n = 312) was significantly increased at a Cmin < 1.1 mg/L (HR, 1.95; 95% CI, 1.08-3.54) and an AUC24 < 40 mg x hr x L1 (HR, 1.95; 95% CI, 1.07-3.54). Both cutoff values represent the median values for adherent patients. These associations were driven by patients from Thailand. Adjusting for geographical region made the association between Cmin and AUC24 with virologic failure statistically nonsignificant. The sensitivity of the Cmin values was too low (29% for NVP, 64% for EFV) to be an adequate predictor for virologic failure. We conclude that identifying the Cmin value for the sole purpose of predicting virologic failure in patients who report to be adherent to NVP or EFV is questionable because of the absence of a concentration-response relation (NVP) or the low sensitivity for such a cutoff value (NVP and EFV).