Background: Personalized risk assessments can assist individuals and providers in determining who may most benefit from HIV prevention interventions, such as Pre-exposure Prophylaxis (PrEP). We developed an HIV risk assessment tool for men who have sex with men (MSM) and sought to validate it among a modern cohort of Black MSM, the group with the highest HIV incidence in the US.
Methods: Two efficacy trials of MSM in the US from 1998 through 2003, (VAX004 vaccine trial and EXPLORE behavioral trial) were used to develop and validate the risk model. The HIV Prevention Trials Network (HPTN) 061, which enrolled Black MSM from 2009-2010, was used as a contemporary validation cohort. Self-reported sexual risk behavioral data by partner serostatus, updated at six month visits, included non-condom receptive anal intercourse (ncRAI), non-condom insertive anal intercourse (ncIAI), and receptive anal intercourse with a condom (cRAI). Other self-reported risk factors included numbers of HIV-negative anal sex partners, drug and heavy alcohol use, and sexually transmitted infections (STIs). A total of 561 HIV seroconversions were identified among 8,950 participants in EXPLORE and VAX004. The final model includes age (<35, ≥35), race/ethnicity, numbers of ncRAI, ncIAI, and cRAI contacts, numbers of HIV-negative anal sex partners, having 1 HIV-negative partner only, and indicators for heavy alcohol use, methamphetamine and poppers, and self-report of STIs.
Results: There were 28 HIV seroconversions among 1,164 initially HIV-negative Black MSM in HPTN 061 who were eligible for follow-up. The cross-validated C-statistic in the development cohort, EXPLORE, was 79.5, while external validation C-statistics were 73.7 in VAX004 and 73.2 in HPTN 061, reflecting good model fit. All 28 seroconversions in HPTN 061 occurred among men in the top four deciles of risk predicted by the model (Figure). Model variables of age <35 (OR=4.69; 95%CI 1.89-11.6), heavy alcohol use (OR=4.34; 95%CI 1.95-9.65), and reporting an STI (OR=4.62; 95%CI 1.56-13.7) were associated with incident HIV infection in HPTN 061.
Conclusions: We developed and validated a risk assessment model for MSM which can be used to provide individualized feedback on HIV risk. The model retained good ability to distinguish levels of HIV risk in a recent cohort of Black MSM, and may be useful for improving awareness of HIV risk among MSM, risk stratification by providers, and uptake of HIV prevention interventions including PrEP.