P20.01
Background: After four randomized clinical trials showing efficacy in different population, Truvada was approved for PrEP use. Public-health authorities need to carefully consider the plausibility of increasing spread of drug resistance following mass PrEP use. We investigate the current knowledge on drug resistance due to PrEP and its implications for decision making.
Methods: We surveyed a panel of expert virologists to evaluate epidemiological relevance of key biological assumptions regarding drug-resistance due to PrEP use. We identified major points of disagreement and studied how different beliefs about resistance propagate into the PrEP impact projected by mathematical models. We assumed the use of 90% efficacious PrEP by 50% of the population under epidemic conditions representative for South Africa.
Results: Virologists disagreed on how fast resistance emerges in infected PrEP users (20–180 days) and both, positive or negative, correlations with adherence were thought possible. They differ on what level of protection PrEP retains against resistant HIV (25%-90%) and on the likelihood to transmit resistance (10%-75%). Based on the virologists' opinion, we project 1% to 8% of the infected individuals with detectable resistance 10 years after the introduction of PrEP and 5% to 35% of the infected individuals at risk to fail ART as a result of past PrEP use. The rate of resistance emergence in infected PrEP users and the rate of reversion back to wild type after PrEP is interrupted are most influential on these resistance projections with larger levels of resistance outcomes associated with imperfect adherence.
Conclusions: Infected individuals who develop resistance on PrEP may have no detectable levels of resistance after they interrupt PrEP use, but could be at risk to fail ART when initiated. Thus, the levels of detectable resistance due to PrEP, frequently used in modelling studies, may underestimate the potential resistance problem when PrEP is implemented in regions with overlapping ART regimens.
