Background:  Previous analyses of tenofovir DF (TDF) therapy have shown decreased responses associated with certain thymidine analogue mutations (TAM) and K65R. The current analyses confirm and extend these analyses with a final multivariate analysis from 566 patients. 

Methods:  Studies 902 and 907 were placebo -controlled, randomized studies of TDF therapy added to stable ART in patients with HIV RNA >400 copies/mL. Previous baseline resistance analyses had included 333 patients (222 TDF, 111 placebo). Current analyses include 233 new patients (158 TDF, 75 placebo), of whom 94% had baseline nucleoside reverse transcriptase inhibitor (NRTI) mutations. In multivariate regression models, the effects of individual RT mutations, HIV RNA, CD4, and other baseline parameters were assessed for effect on week 24 HIV RNA response. Week 48 genotyping was conducted for resistance development analyses. 

Results:  In confirmation of earlier analyses, patients with baseline TAM showed a significant response to TDF therapy (–0.50 log₁₀ copies/mL, n=102) which was reduced, however, in comparison to those without TAM (–1.0 log₁₀ c/mL, n = 56). ≥3 TAM in the TAM pathway of M41L, L210W, and T215Y showed the greatest reduction in TDF response (–0.24 log₁₀ copies/mL, n = 45). Patients with M184V alone showed the strongest response (–1.12 log₁₀ copies/mL, n = 41). In univariate analyses of all patients (n = 566), 12 RT mutations, including L74V, were associated with decreased responses (p <0.05). In multivariate analyses, baseline HIV RNA, baseline CD4, and TDF treatment were significant response predictors. Baseline K65R was infrequent (n = 8) but independently reduced response. In contrast, baseline M184V improved response (p = 0.004). Adjusting for these co-variates, a step-wise selection model showed that 4 TAM (M41L, D67N, L210W, and T215Y) and L74V were the best predictors of TDF response. The negative effect of L74V was unexpected given the full susceptibility of this mutant to TDF in vitro. However, L74V is associated both with multiple TAM and development of K65R in this study, both of which contributed to poorer TDF treatment responses. Through wk 48, the majority of mutations that developed were additional TAM. K65R developed in 2 of 133 (1.5%) newly analyzed patients for a cumulative total of 16 of 566 (2.8%) patients. There was no development of Q151M.

Conclusions:  Final multivariate regression analyses of studies 902 and 907 demonstrate that K65R, L74V, or ≥3 mutations in the 41-210-215 TAM pathway are associated with reduced responses to TDF therapy while M184V is associated with improved TDF response. Development of K65R remains infrequent.