HIV treatment, preexposure prophylaxis, and drug resistance: reconciling conflicting predictions from mathematical models.

TO THE EDITOR—We read with interest the article by Abbas et al, in which they model the potential impact of preexposure prophylaxis (PrEP) on human immunodeficiency virus (HIV) transmission and drug resistance in South Africa [1]. They predict that rolling out Truvadabased PrEP will increase the transmission and prevalence of resistance. Notably, their results conflict with those from 2 previous modeling studies [2, 3]. In one study, Supervie et al [2] predict that rolling out Truvada-based PrEP in San Francisco will decrease both the transmission and prevalence of resistance. In the other study, they predict that rolling out Truvada-based PrEP in Botswana will increase the transmission and decrease the prevalence of resistance [3]. The conflicting predictions regarding PrEP and resistance are illustrated in Figure 1A. Mathematical models can be useful health policy tools. However, they can, as in this case, cause confusion. To reconcile the results from the 3 studies, we examine their underlying assumptions. They all assume that PrEP can lead to the development and transmission of resistance—in other words, they assume that (1) resistance can develop if HIV-infected individuals begin taking PrEP and/or uninfected individuals become infected while receiving PrEP, and (2) these drugresistant strains could be transmitted. In addition, they all assume that (3) HIV treatment programs are in place when PrEP is rolled out and (4) HIV treatment can lead to the development and transmission of resistance. Resistance due to HIV treatment has been observed in resource-rich countries since 1987 and has reached moderate-to-high levels [4]. In contrast, resistance in resource-constrained countries is fairly low, because treatment has only recently become available [5, 6]. Each of the modeling studies predicts the potential effect of PrEP on the transmission and prevalence of resistance [1–3]. The important difference between the 3 studies is that each study makes different assumptions—which drive their results— regarding the initial level of resistancewhen PrEP is rolled out. Abbas et al make the intuitive prediction that rolling out Truvada-based PrEP in South Africa (whereHIV treatment is available)will increase resistance (Figure 1A) [1]. Notably, this result will only hold true