ABSTRACT:
The transmission prevalence of drug-selected HIV-1 variants may increase or decrease over time owing to changing use of antiretroviral drugs, changes in prevalence of virological drug failure, changes in risk behavior, or changes in infectiousness of treated subjects. From August 1996 to December 1998, initial plasma viral RNA from subjects referred to San Francisco General hospital within 12 months of HIV-1 seroconversion was analyzed for evidence of antiretroviral drug resistance. Genetic markers of resistance were assessed by population sequencing the HIV-1 protease and RT coding regions using high-density probe arrays and cycle sequencing.
From 1996-1997 to 1998, the prevalence of genetic markers of NRTI resistance decreased from 9 of 42 (21.4%) to 1 of 47 (2.1%), P=0.005. The greatest reduction was observed for RT M184V (9.5% to 0; P=0.04) and RT T215Y/D (9.5% to 0; P=0.04). The frequency of mutations in the protease-coding region and NNRTI binding region did not change over time. One subject with multiple protease resistance mutations and poor virological response to PI-containing regimens was identified in November 1997 and another in December 1998. Phenotypic testing confirmed decreased lamivudine susceptibility in viruses with RT M184V. Two- to sixfold decreased susceptibility to NNRTIs was detected in 11/36 (31%) in 1996-97, and was not associated with substitutions in the NNRTI binding region, except in one subject. In drug-naïve subjects, NNRTI drug susceptibility varied more than the other drug classes, suggesting that minimal NNRTI resistance may represent natural variation rather than drug selection. The prevalence of genetic markers of NRTI resistance decreased from 1996-97 to 1998 in recently infected persons in San Francisco. The use of combination antiretroviral regimens became widespread in San Francisco during this time, and may have decreased infectiousness in populations who previously transmitted NRTI-resistant HIV-1.
