Background:  PACTG 1022 was a phase III, randomized open-label study of PI-sparing or -including regimens in ARV-naïve women initiating ART during pregnancy. A total of 38 women with HIV RNA >1000 copies/mL were randomized to zidovudine (AZT), lamivudine (3TC), and either nelfinavir (NFV) or nevirapine (NVP) to compare virologic suppression and tolerance in pregnancy. The trial closed early due to maternal toxicity. The rate of HIV genotypic resistance was compared by treatment group in women stopping study ART post partum, to test the hypothesis that stopping ART would select resistance.

Methods:  Specimens from women who stopped the study ART and had viral loads >400 copies/mL were tested for drug-resistance by oligonucleotide ligation assay (OLA) of HIV amplified from plasma and peripheral blood mononuclear cells and by consensus sequence of virus from plasma using the ViroSeq Kit. When drug resistance was detected, specimens obtained prior to therapy and during treatment were also evaluated. Resistance by treatment arm was compared using Fisher’s exact test.

Results:  Preliminary analysis of 19 women randomized to NFV (n = 10) and NVP (n = 9) detected M184V in 6 of 18 (33%) and K103N in 3 of 19 (16%) by OLA. OLA was indeterminate at M184 in 1 of 19 with subtype CRF02_AG. M184V was selected primarily in women randomized to NFV (n = 5 of 6), and K103N only with NVP (n = 3 of 3). Neither K103N nor M184V were detected at study entry. A trend was observed for greater M184V in NFV women (5 of 10 vs 1 of 8; p = 0.15). M184V was first detected after cessation of ART in 3 of 6, while 3 experienced viremia during ART with M184V detected. K103N was detected only after discontinuation of ART. Consensus genotyping of plasma did not detect M184V and K103N in 2 women positive by OLA, suggesting low-level mutants. One or more major resistance mutations were detected after ART in 5 of 10 (50%) assigned to NVF and 4 of 9 (44%) to NVP.

Conclusions:  The rate of HIV drug resistance was high in women given NVP and NFV ART during pregnancy for prophylaxis of mother-to-child transmission (pMTCT). Selection of 3TC resistance during and following the cessation of NFV ART suggests that NFV pharmacokinetics in pregnancy may be suboptimal, and that staggered discontinuation of ARV may be warranted to reduce selection of 3TC resistance. Comparative studies of resistance during and following ART given for pMTCT and stopped postpartum are needed to identify regimens that minimize the risks of selecting HIV-1 drug resistance in women.