Abstract Body

Background: There are no comparisons of resistance after randomized virologic switch criteria. Methodology: PENPACT 1 was an international open label 2 by 2 factorial trial randomizing HIV infected children to start antiretrovirals with a protease inhibitor (PI) vs non nucleoside reverse transcriptase inhibitor (NNRTI) based regimen, and switch at RNA threshold of 1000 vs 30000 c/ml. Switch criteria were: not achieving the threshold by week 24, confirmed rebound above the threshold after week 24 (RNA tested 12 weekly), or CDC C event. Resistance testing was on the last sample ≥1000 c/ml before switch, or confirmed ≥1000 c/ml before re-suppression; to ensure fair comparison by 1000 and 30000 arms. Additional testing was on samples ≥1000 c/ml at 4 years and trial end. This secondary analysis tested differences in number of major IAS mutations by Poisson models. Results: Of 263 children, 67 started PIs and were assigned the 1000 threshold (PI 1000), 64 PI 30000, 67 NNRTI 1000 and 65 NNRTI 30000. Half on PIs started lopinavir/ritonavir, 62% on NNRTIs efavirenz (Table). The 1000 criteria were reached by 94 (36%) children during median 5.0 (IQR 4.2-6.0) years follow-up. Time to reaching the 1000 criteria was similar by class (log rank P 0.26). Most in the 1000 arm reaching the criteria switched soon after (PI vs NNRTI median 12 vs 8 wks, log rank P 0.60). Median weeks from 1000 to 30000 criteria was 80 for the 30000 arm; PI vs NNRTI 58 vs 80, log rank P 0.81, and 25% of children had switched by 44 weeks; PI vs NNRTI 63 vs 17, log rank P 0.16. 87/107 (81%) children with last sample ≥1000 c/ml had resistance test results. Median weeks from entry to last test was 96; 1000 vs 30000 72 vs 124, rank sum P<0.01. More NRTI resistance mutations were accumulated in NNRTI 30000 than the other arms (Table). More NNRTI than PI mutations were accumulated; NNRTI mutations were frequent in 1000 and 30000, and PI resistance was less common in both arms. For non-randomized NRTIs, 62 started ABC+3TC, 166 3TC+ZDV/d4T, and 35 other (25 ZDV+ddI). The ABC+3TC group had fewer mutations than the others. The 5 with mutations on ABC+3TC were on NNRTIs; all had M184V/I and 1 K65R, L74V, Y115F, while 22/39 3TC+ZDV/d4T were on NNRTIs; all had M184V/I and 10 had TAMs. Conclusions: Children starting PIs tended to switch later spending longer above 1000 but accumulated fewer mutations. Children starting NNRTIs switched soon after the 30000 criteria and accumulated more mutations, particular to NRTIs. An ABC+3TC backbone seemed most protective against NRTI resistance.