Role of low-frequency HIV-1 variants in failure of nevirapine-containing antiviral therapy in women previously exposed to single-dose nevirapine

Abstract

In the OCTANE/A5208 study of initial antiretroviral therapy (ART) in women exposed to single-dose nevirapine (sdNVP) ≥ 6 mo earlier, the primary endpoint (virological failure or death) was significantly more frequent in the NVP-containing treatment arm than in the lopinavir/ritonavir-containing treatment arm. Detection of NVP resistance in plasma virus at study entry by standard population genotype was strongly associated with the primary endpoint in the NVP arm, but two-thirds of endpoints occurred in women without NVP resistance. We hypothesized that low-frequency NVP-resistant mutants, missed by population genotype, explained excess failure in the NVP treatment arm. Plasma samples from 232 participants were analyzed by allele-specific PCR at study entry to quantify NVP-resistant mutants down to 0.1% for 103N and 190A and to 0.3% for 181C. Of 201 women without NVP resistance by population genotype, 70 (35%) had NVP-resistant mutants detected by allele-specific PCR. Among these 70 women, primary endpoints occurred in 12 (32%) of 38 women in the NVP arm vs. 3 (9%) of 32 in the lopinavir/ritonavir-containing arm (hazard ratio = 3.84). The occurrence of a primary endpoint in the NVP arm was significantly associated with the presence of K103N or Y181C NVP-resistant mutations at frequencies >1%. The risk for a study endpoint associated with NVP-resistant mutant levels did not decrease with time. Therefore, among women with prior exposure to sdNVP, low-frequency NVP-resistant mutants were associated with increased risk for failure of NVP-containing ART. The implications for choosing initial ART for sdNVP-exposed women are discussed.

Fig. 1.

Association of primary endpoints (virological failure or death) with NVP resistance at study entry determined by standard genotype analysis. (A) Number of subjects with primary endpoint stratified by baseline NVP resistance. The percent of subjects in each treatment arm in each category is given in parentheses. (B) Proportion of primary endpoints by treatment arm and detection of NVP resistance (black bar) or not (gray bar) by standard genotype.

Fig. 2.

Population frequency of NVP-resistant mutants among subjects with NVP resistance detected by ASP. Each symbol represents a different subject. Bars indicate median frequency. Open symbols represent subjects identified as genotype-positive and ASP-positive.

Fig. 3.

Three categories of baseline mutation frequencies and their association with primary study endpoints, NVP arm only. Mutation frequencies were calculated by summing all three codons. VF, virological failure.

Fig. 4.

Proportion reaching a primary endpoint among subjects with NVP resistance and no NVP resistance detected by ASP in the NVP arm and stratified by time since exposure to sdNVP. In the >2-y category, there were no failures in the No NVP/Resistance group.