Plasma samples collected at 14 weeks of age were available for 151 (68.6%) of the 220 HIV-infected infants who received either extended NVP or extended NVP+ZDV prophylaxis in PEPI-Malawi. Sixty-nine infants did not have a 14-week sample available for testing (36 samples were used up in previous testing, the parents of 16 infants refused to participate in the trial after enrollment, 10 infants died, four infants were lost to follow-up, one infant missed the 14-week study visit, one infant relocated, and one infant's sample could not be located); four of the 151 available 14-week samples were not tested because the mother initiated ARV treatment prior to the 14-week study visit. Resistance test results (ViroSeq and LigAmp, see Methods) were obtained for 108 (73.5%) of the 147 samples that were tested (53 in the extended NVP arm, 55 in the extended NVP+ZDV arm); 39 samples failed analysis (no amplification).
Table 1 shows characteristics of the 108 infants who had resistance test results (included in the sub-study) and of the 112 infants who did not. There were no statistically significant differences between infants in these two groups of infants for the following variables: the proportion of infants in the extended NVP vs. extended NVP+ZDV study arms, the duration of prophylaxis received prior to HIV diagnosis, the median maternal CD4 cell count at study enrollment, the proportion of infants whose mothers received sdNVP, or the proportion of infants with in utero HIV infection. The median maternal HIV log10 viral load at enrollment was lower for infants included in this sub-study (Table 1).
At 14 weeks of age, 82 (75.9%) of 108 infants had NVP resistance detected with the ViroSeq assay and 78 (72.2%) had K103N and/or Y181C detected with the LigAmp assay (P=0.45, McNemar's test, ). None of the infants had ZDV resistance detected with the ViroSeq assay. There was no significant difference in the proportion of infants in the two study arms who had NVP resistance detected with the ViroSeq or LigAmp assay at 14 weeks of age (P=0.12 for ViroSeq, P=0.13 for LigAmp, Fisher's exact test, ). We also did not find a significant association between detection of NVP resistance at 14 weeks of age using the LigAmp assay and any of the clinical or laboratory variables described above (Table 1).
Figure 1 Resistance test results are shown for all infants tested (no shading), infants in the extended NVP arm (dark shading), and infants in the extended NVP+ZDV arm (light shading). (A) Panel A shows results obtained for samples collected at 14 weeks of age (more ...)
We next used the LigAmp assay to analyze persistence of K103N and Y181C in infants who had those mutations detected at 14 weeks of age. Sixty-eight (87.2%) of the 78 infants who had NVP resistance at 14 weeks of age were still alive at 6 months; 55 of those 68 infants had a 6-month sample available for analysis (13 samples were used up in other testing). Three of the 55 samples were not included in the analysis because the mother started ARV treatment before the 6-month visit and six samples failed analysis (no amplification). The remaining 46 samples were analyzed. Thirty-eight (82.6%) of the 46 infants still had K103N and/or Y181C detected at 6 months of age. There was no significant difference in detection of these mutations in the two study arms (P=1.0, Fisher's exact test, ). Five infants had NVP resistance mutations other than K103N and Y181C detected by ViroSeq at 14 weeks (V106A, Y188C/L, G190A) in the absence of K103N or Y181C. Four of the five infants had a 6-month sample available for analysis with ViroSeq. Three of the four infants did not have any NVP resistance mutations detected in the 6-month sample; in one infant, G190A was detected.
A similar analysis was performed to analyze persistence of NVP resistance at 12 months of age. Twenty-six (68.4%) of the 38 infants who had NVP resistance at 6 months of age were still alive at 12 months; 21 of those 26 infants had a 12-month sample available for analysis (four samples were used up in previous testing, one infant relocated). Two of the 21 samples were not included in the analysis because ARV treatment was initiated before the 12-month visit in both of the mothers and in one of the infants. We also analyzed 12-month samples from infants who had NVP resistance detected at 14 weeks, but did not have a 6-month sample available for analysis. Among 19 infants in that group, 14 were still alive at 12 months and 10 of those were evaluable (one infant relocated, the parents of one infant refused further participation in the trial, one infant missed the 12-month study visit, and one mother started ARV treatment before the 12-month visit). Therefore, a total of 29 12-month samples were tested (19 samples from infants who had resistance detected at 6 months of age, and 10 samples from infants who had resistance detected at 14 weeks of age who did not have a resistance testing result from the 6-month visit). Nineteen (66.5%) of the 29 infants who were included in the analysis still had K103N and/or Y181C detected at 12 months of age (). There was no significant difference in the percentage of infants in the extended NVP vs. extended NVP+ZDV study arms who had detectable K103N or Y181C at either 6 or 12 months (P=1.0 at 6 months, P=0.43 at 12 months, Fisher's exact test, ). The LigAmp assay provides a quantitative measure of the percentage of the viral population in a sample that has the mutation of interest. There was no significant difference in the level of K103N or Y181C detected in the infants in these two groups at the 6- or 12-month visits (data not shown).