These findings are consistent with earlier work carried out by other researchers. Lima et al. [12]. found that individuals Tacrolimus on boosted PI-based regimens were less likely to develop resistance than those on NNRTI-based regimens (AOR 0.42; 95% CI 0.28–0.62) and Riddler et al. [13] found that those on efavirenz-based regimens were more prone to the development of drug resistance mutations than those on lopinavir/ritonavir-based therapies (9 vs. 6%, respectively).

The two comparison drug classes were equally efficacious, as evidenced by proportions of participants who achieved virological suppression (plasma viral load <50 copies/mL) in the first year of therapy (66% for the NNRTI group and 67% for the boosted PI group). Such a similarity in virological response and other clinical outcomes has been documented in other studies [25,26]. This rate of response occurred despite lower adherence INNO-406 order in

the NNRTI group. This kind of response to NNRTI was also demonstrated by Nachenga et al., who found that moderate levels of adherence to these drugs often led to viral suppression among patients [27]. These results may also suggest that, despite adequate virological response, patients still remain at a greater risk of developing resistance to NNRTIs. The generalization of these findings to RLSs, where NNRTI-based ART is primarily used for first-line treatment, may be limited by the fact that this study was carried out in a developed country where most of the social demographic features are different from those in developing countries. Furthermore, most participants in this cohort had HIV-1 subtype B, which accounts for only 10% of HIV infections world-wide, and recent evidence suggests that different HIV genetic variants have different biological properties, including susceptibility and response to antiretroviral Pregnenolone drugs [28]. In addition, the way in which ART is managed in the face of drug resistance is very different in BC from

RLSs. However, we believe that concerns regarding NNRTI-induced resistance mutations require greater study in RLSs. The potential for the development of mutations is probably even greater in these settings, where individuals may have prolonged periods of uncontrolled viraemia prior to switching the class of their third drug. Our results suggest that evaluating the strategy of NNRTI- versus boosted PI-based HAART in RLSs should be a main priority. This should be coupled with documentation of the impact of these mutations on subsequent virological suppression and clinical outcomes among patients who are failing ART in RLSs. Advocacy targeted at reduction in prices for boosted PIs and licensing of generic products can help to increase the availability of these drugs in RLSs. The authors would like to thank the participants in the BC HIV/AIDS DTP and the nurses, physicians, social workers and volunteers who support them.