Figure 2.

Due to almost complete overlap in the populations with thymidine analogue mutations (TAMs) present and the K65R mutation absent, fully adjusted analyses were not possible. Multivariable analyses that adjusted for TAMs revealed little change in the unadjusted estimates presented. Multivariable analyses that adjusted for K65R revealed additional confounding with nonnucleoside reverse transcriptase inhibitor (NNRTI) use prior to study entry. *P < .05. Abbreviations: CI, confidence interval; EFV, efavirenz; HIV-1, human immunodeficiency virus type 1; HR, hazard ratio.