Background: An observational study was conducted in Maputo, Mozambique, to investigate trends in prevalence of HIV drug resistance (HIVDR) in antiretroviral (ART) naïve subjects initiating highly active antiretroviral treatment (HAART).
Methodology/principal findings: To evaluate the pattern of drug resistance mutations (DRMs) found in adults on ART failing first-line HAART [patients with detectable viral load (VL)]. Untreated subjects [Group 1 (G1; n=99)] and 274 treated subjects with variable length of exposure to ARV´s [6-12 months, Group 2 (G2;n=93); 12-24 months, Group 3 (G3;n=81); >24 months (G4;n=100)] were enrolled. Virological and immunological failure (VF and IF) were measured based on viral load (VL) and T lymphocyte CD4+ cells (TCD4+) count and genotypic resistance was also performed. Major subtype found was C (untreated: n=66, 97,06%; treated: n=36, 91.7%). Maximum virological suppression was observed in G3, and significant differences intragroup were observed between VF and IF in G4 (p=0.022). Intergroup differences were observed between G3 and G4 for VF (p=0.023) and IF between G2 and G4 (p=0.0018). Viral suppression (<50 copies/ml) ranged from 84.9% to 90.1%, and concordant VL and DRM ranged from 25% to 57%. WHO cut-off for determining VF as given by 2010 guidelines (>5000 copies/ml) identified 50% of subjects carrying DRM compared to 100% when lower VL cut-off was used (<50 copies/ml). Length of exposure to ARVs was directly proportional to the complexity of DRM patterns. In Mozambique, VL suppression was achieved in 76% of individuals after 24 months on HAART. This is in agreement with WHO target for HIVDR prevention target (70%).
Conclusions: We demonstrated that the best way to determine therapeutic failure is VL compared to CD4 counts. The rationalized use of VL testing is needed to ensure timely detection of treatment failures preventing the occurrence of TDR and new infections.