Background:  The replacement of lysine (K) by asparagine (N) at position 103 in HIV-1 reverse transcriptase (RT) is frequent in a significant proportion of infected women after exposure to single-dose nevirapine (NVP) used in resource-poor settings to prevent mother-to-child transmission. Persistence of this resistance mutation has been proposed as 1 possible explanation for increases in virologic failure among single-dose NVP-exposed women on subsequent treatment with NVP-containing regimens. Because conventional genotyping fails to detect low frequency drug-resistance mutations, we used a real-time PCR assay to detect and quantify minority populations of 103N mutants among a cohort of women given single-dose NVP in South Africa.

Methods:  Amplicons generated from viral RNA or DNA were screened by real-time PCR for the presence of the wild type (lysine, K, AAA, or AAG) and mutant (asparagine, N, AAC, or AAT) codons at position 103 in HIV-1 RT. The relative frequency of each codon was calculated for samples collected from a cohort of HIV-1 subtype-C-infected mothers who received intra-partum single-dose NVP. Blood was collected at 6 weeks post-partum and at regular intervals thereafter over the course of a year.

Results:  Real-time PCR and direct population sequencing results for RNA samples at 6 weeks post-partum were compared. 103N mutants were detected in 16 of 18 (89%) samples by real-time PCR and in 9 of 18 (50%) by population sequencing. Generally only samples with ≥ 32% 103N by real-time PCR were detected by sequencing. Longitudinal samples from 13 women were examined to monitor the levels of 103N in RNA over time:  3 women had no evidence of 103N and amplified only wild type virus at all time-points tested; 4 had the 103N mutation at > 60% within 12 weeks of NVP exposure, which declined to 10 to 40% within 7 to 12 months; 6 had 103N at levels ranging from 5 to 40% at 12 weeks and in 4 cases declined to below the threshold of detection by 7 to 12 months. In a cross-sectional analysis, at 6 weeks post-partum, 103N was detected in DNA, but was lower than the levels found in RNA. At 1 year 103N was still detectable at low levels in 4 of 16 (25%) RNA samples and in none of 15 DNA samples tested.

Conclusions:  Real-time PCR shows greater sensitivity for the detection of 103N mutants than in population sequencing. 103N mutants faded over time in RNA among all but a subset of single-dose NVP-exposed women. We found no evidence for archiving of 103N mutations in DNA at 1 year post-single-dose NVP.

Keywords: MTCT; NVP; resistance