Background:  Clinical trials of ART for pre-exposure prophylaxis (PREP) are ongoing, with new drug candidates under evaluation. Even if clinical trials show efficacy for PREP, high drug costs and safety concerns could limit wide-scale introduction of PREP in developing countries.

Methods:  A MEDLINE search identified epidemiological studies evaluating the incidence of new HIV-1 infections worldwide. Incidence estimates were used to calculate the number of people to treat with PREP to prevent one HIV infection. Lowest published ART costs (in US$) were combined with HIV incidence estimates to calculate cost per HIV infection prevented by PREP. Efficacy of PREP was assumed to be 80% in the primary analysis. Lifetime costs of HIV treatment and care in sub-Saharan Africa were estimated at US$6000 to 10,000 per infected person.

Results:  The annual incidence of new HIV infections varied from 0.5% to 20% in 19 different epidemiological studies (median 4%). With an HIV incidence of 0.5%, 250 people would require PREP treatment to prevent 1 HIV infection, whereas at HIV incidence of 20%, 6 people require PREP per infection prevented. Lowest annual drug costs included $53 per person-year for lamivudine (3TC) ($26 if used intermittently), $300 for tenofovir (TDF), $500 for Kaletra, and $850 for TDF/3TC/Kaletra. The ART drug cost per HIV infection prevented, ranged from $208, using intermittent 3TC in a population with 15% annual HIV incidence, $2500 for standard dose TDF with 15% annual incidence, to $106,250, using TDF/3TC/Kaletra in a population with 1% HIV incidence.

Conclusions:  For target populations in developing countries with annual seroconversion rates of at least 1%, use of efficacious PREP treatments with annual costs under US$50 would be more cost-effective than treating patients lifelong once infected with HIV. However, the cost-effectiveness of PREP varies widely and depends on low manufacturing costs for ART plus careful targeting of populations at highest risk of HIV infection. Intermittent use of PREP (i.e. only at time of high-risk behavior) substantially improves overall cost-effectiveness.