Background:  DART is a randomized trial comparing clinical versus clinical plus laboratory monitoring in 3316 HIV-1-infected patients initiating ART with CD4 <200 cells/mm³ in Uganda and Zimbabwe. In common with many resource-poor environments, there is no contemporaneous virological monitoring, and switch to second-line treatment follows clinical or immunological failure. The evolution of resistance to triple ART has rarely been studied under such circumstances. We previously demonstrated that viral load suppression (<50 copies/mL) at 24 and 48 weeks was 59 and 61%, respectively, (intent to treat) on combinavir (zidovudine/lamivudine; CBV) + tenofovir (TDF)

Methods:  Viral load measurements at baseline, weeks 24 and 48 were undertaken retrospectively by Roche Amplicor 1.5 in a specific subset  of  300 patients initiating CBV+TDF. Plasma virus from samples with >1000 copies/mL were sequenced in the pol gene

Results:  Genotype results are reported for 26 of 43 (60%) and 35 of 64 (55%) samples with viral load >1000 copies/mL at weeks 24 and 48, respectively, for which samples were available and sequencing was successful. The prevalence of specific mutations at 24 and 48 weeks were:  M184V (65%, 77%), T215F/Y (31%, 51%), D67G/N (38%, 60%), K70R (31%, 51%), and K65R (12%, 14%). The proportion of patients at 24 and 48 weeks with thymidine analog mutations (TAM) were:  0 TAM (42%, 26%), 1 to 3 TAM (54%, 37%), 4 to 6 TAM (4%, 37%). At 48 weeks, there were more mutations in those 12 individuals tested (of 23 in total) who had viral load >1000copies/mL at 24 weeks (mean 4.1, range 1 to 6) than in those 19 tested (of 41 in total) with viral load <1000 copies/mL at 24 weeks (mean 3.1, range 0 to 6) (p = 0.17). In 6 individual patients for whom resistance data were available at both 24 and 48 weeks, a mean of 2.5 (range 1 to 4) new nucleoside reverse transcriptase inhibitors (NRTI) mutations emerged between these time points. No clear differences were observed in patterns of emerging mutations between HIV-1 subtypes A, C, and D. Of 300 baseline samples, we tested 91 (30%):  8 (9%) demonstrated ≥1 key resistance mutations (5 with non-NRTI [NNRTI] resistance, 5 with M184V, 2 with M41L).

Conclusions:  Virological monitoring within ART rollout in the resource-poor world will be limited. Without viral load–guided treatment switch, extensive resistance evolves over 48 weeks in those with viremia who continue to receive first-line CBV/TDF. This population may still benefit from 2 major classes of drugs. However, we note the prevalence of baseline resistance to NNRTI and NRTI, which probably reflects prior undisclosed therapy.