Background: For maximum effect pre-exposure prophylaxis should be targeted to the subpopulations that account for the largest proportion of infections (population-attributable fraction [PAF]) and for whom the number needed to treat (NNT) to prevent infection is lowest. We aimed to estimate the PAF and NNT of participants in the iPrEx (Pre-Exposure Prophylaxis Initiative) trial. Methods: The iPrEx study was a randomised controlled efficacy trial of pre-exposure prophylaxis with coformulated tenofovir disoproxil fumarate and emtricitabine in 2499 men who have sex with men (MSM) and transgender women. Participants aged 18 years or older who were male at birth were enrolled from 11 trial sites in Brazil, Ecuador, Peru, South Africa, Thailand, and the USA. Participants were randomly assigned (1:1) to receive either a pill with active pre-exposure prophylaxis or placebo, taken daily. We calculated the association between demographic and risk behaviour during screening and subsequent seroconversion among placebo recipients using a Poisson model, and we calculated the PAF and NNT for risk behaviour subgroups. The iPrEx trial is registered with ClinicalTrials.gov, NCT00458393. Findings: Patients were enrolled between July 10, 2007, and Dec 17, 2009, and were followed up until Nov 21, 2010. Of the 2499 MSM and transgender women in the iPrEx trial, 1251 were assigned to pre-exposure prophylaxis and 1248 to placebo. 83 of 1248 patients in the placebo group became infected with HIV during follow-up. Participants reporting receptive anal intercourse without a condom seroconverted significantly more often than those reporting no anal sex without a condom (adjusted hazard ratio [AHR] 5·11, 95% CI 1·55-16·79). The overall PAF for MSM and transgender women reporting receptive anal intercourse without a condom was 64% (prevalence 60%). Most of this risk came from receptive anal intercourse without a condom with partners with unknown serostatus (PAF 53%, prevalence 54%, AHR 4·76, 95% CI 1·44-15·71); by contrast, the PAF for receptive anal intercourse without a condom with an HIV-positive partner was 1% (prevalence 1%, AHR 7·11, 95% CI 0·70-72·75). The overall NNT per year for the cohort was 62 (95% CI 44-147). NNTs were lowest for MSM and transgender women self-reporting receptive anal intercourse without a condom (NNT 36), cocaine use (12), or a sexually transmitted infection (41). Having one partner and insertive anal sex without a condom had the highest NNTs (100 and 77, respectively). Interpretation: Pre-exposure prophylaxis may be most effective at a population level if targeted toward MSM and transgender women who report receptive anal intercourse without a condom, even if they perceive their partners to be HIV negative. Substance use history and testing for STIs should also inform individual decisions to start pre-exposure prophylaxis. Consideration of the PAF and NNT can aid in discussion of the benefits and risks of pre-exposure prophylaxis with MSM and transgender women. Funding: National Institute of Allergy and Infectious Diseases and the Bill & Melinda Gates Foundation. © 2014 Elsevier Ltd.