INTRODUCTION:
Pre-exposure prophylaxis (PrEP) effectively prevents HIV infections, but low adherence among high-risk populations may limit its public health benefits. This study investigated factors associated with PrEP discontinuation in Brazil, with a focus on different markers of social vulnerability – race/ethnicity, sexual identity, homelessness, and sexual work – and their interactions.

METHODS:
We conducted a retrospective cohort study using individual-level data from Brazil’s nationwide PrEP panel (2018 - 2024). Fixed-effects multivariate regression models were used to estimate factors associated with PrEP discontinuation and re-initiation. To investigate time-to-discontinuity, we estimated multivariate proportional hazard Cox regressions, adjusting for the first year of PrEP initiation. We investigated intersectionality by estimating interaction effects between different markers of social vulnerability.

RESULTS:
Among 228,954 who initiated PrEP treatment, 50% discontinued at least once during follow-up. After adjusting for age, education, race/ethnicity, and sexual identity, people experiencing homelessness and sex workers were 11 and 7 percentage points more likely to discontinue. Time-to-discontinuation was significantly lower for people experiencing homelessness, sex workers, and prisoners. For all three vulnerable population groups, PrEP discontinuation was highest during the first year of treatment. Interactions between homelessness, sex work, prisoner status and sexual identity revealed substantially increased intersectional risks of discontinuation.

CONCLUSIONS:
PrEP discontinuation was substantially higher among vulnerable populations in Brazil. Targeted strategies are needed to support PrEP adherence in high-risk populations and to ensure effective HIV prevention.