| Summary: | <h4>Background</h4> <p>Cross-resistance after first-line antiretroviral treatment (ART) failure is expected to impair activity of nucleoside-reverse-transcriptase-inhibitors (NRTIs) in second-line therapy, but evidence for effect on virological outcomes is limited.</p> <h4>Methods</h4> <p>We performed an observational analysis of a randomised-controlled trial of second-line ART (EARNEST) in sub-Saharan Africa. 1277 HIV-infected adults/adolescents failing first-line ART (WHO criteria, virological confirmation) were randomised to a boosted protease-inhibitor (PI; standardised to lopinavir/ritonavir) with either 2/3 NRTIs (clinician-selected, without resistance testing; PI/NRTI); raltegravir (PI/RAL); or as monotherapy (PI-mono; discontinued after week 96). Predicted activity of prescribed second-line NRTIs was determined by genotypic resistance testing on stored baseline samples in PI/NRTI. Viral load was measured on stored samples in all patients obtained every 12-16 weeks.</p> <h4>Findings</h4> <p>Baseline genotypes were available in 391 (92%) in PI/NRTI. For the 230 (59%) taking no predicted-active NRTIs, there was a high rate of VL suppression (89%(176/198) <400 copies per mL at week-144), superior to PI/RAL (81%(312/383) at week 144; P=0.02) and PI-mono (61%(233/280) at week-96; P<0.0001). VL suppression was no better with 1 predicted-active NRTI (85%(95/112), P=0.3 vs no active NRTIs) and appeared worse with 2-3 predicted active NRTIs (77%(20/26), P=0.08 vs no active NRTIs). Over all follow-up, greater predicted NRTI activity was associated with worse VL suppression (global p=0.0004). </p> <h4>Interpretation</h4> <p>Genotypic resistance testing may not accurately predict NRTI activity in PI-based second-line ART. Our results do not support the introduction of routine resistance testing in ART programmes in low-income settings for the purpose of selecting second-line NRTIs.</p>
|
|---|