IntroductionOnce-daily tenofovir/emtricitabine-based pre-exposure prophylaxis (PrEP) can reduce HIV acquisition in men who have sex with men (MSM), by 44% in the iPrEx trial, and reaching up to 99% with high adherence. We examined the potential population-level impact and cost-effectiveness of different PrEP implementation strategies.MethodsWe developed a dynamic, stochastic compartmental model of HIV transmission among the estimated 57,400 MSM in Toronto, Canada. Parameterization was performed using local epidemiologic data. Strategies examined included (1) uniform PrEP delivery versus targeting the highest risk decile of MSM (with varying coverage proportions); (2) increasing PrEP efficacy as a surrogate of adherence (44% to 99%); and (3) varying HIV test frequency (once monthly to once yearly). Outcomes included HIV infections averted and the incremental cost ($CAD) per incremental quality-adjusted-life-year (QALY) gained over 20 years.ResultsUse of PrEP among all HIV-uninfected MSM at 25, 50, 75 and 100% coverage prevented 1970, 3427, 4317, and 4581 infections, respectively, with cost/QALY increasing from $500,000 to $800,000 CAD. Targeted PrEP for the highest risk MSM at 25, 50, 75 and 100% coverage prevented 1166, 2154, 2816, and 3012 infections, respectively, with cost/QALY ranging from $35,000 to $70,000 CAD. Maximizing PrEP efficacy, in a scenario of 25% coverage of high-risk MSM with PrEP, prevented 1540 infections with a cost/QALY of $15,000 CAD. HIV testing alone (Q3 months) averted 898 of infections with a cost savings of $4,000 CAD per QALY.ConclusionsThe optimal implementation strategy for PrEP over the next 20 years at this urban centre is to target high-risk MSM and to maximize efficacy by supporting PrEP adherence. A large health benefit of PrEP implementation could come from engaging undiagnosed HIV-infected individuals into care.