Background: Randomized trials of oral PrEP in gay and bisexual men (GBM) have reported efficacy of close to 90%, with HIV infections occurring only in non-adherent participants. In the first 3700 EPIC-NSW participants followed for one year, incidence was 0.51/1000 person-years (PY) compared to an expected incidence of over 20/1000PY. Concern has been raised about long-term PrEP efficacy. We report HIV incidence in the expanded study cohort with extended follow-up
Methods: EPIC-NSW is a population implementation study of daily oral PrEP based in 31 clinics across NSW, Australia. HIV diagnoses were reported 1) as serious adverse events, 2) through electronic medical record systems, and 3) in consenting participants (80%), by linkage to the state HIV register. Participants contributed PY from date of enrolment to date of HIV diagnosis or to 31/12/2018 in those who remained HIV-negative. PY incidence rates and 95% confidence intervals (CIs) were calculated. Hazard ratios (HRs) were estimated using Cox regression
Results: 9708 participants (98.5% GBM) were enrolled between March 2016 and April 2018. Over 17,747PY of follow-up, there were 16 HIV diagnoses, with evidence of PrEP non-adherence in all cases. HIV incidence was 0.90/1000PY (95%CI 0.55-1.47). HIV incidence was higher in the younger (p= 0.007), reaching 2.4/1000PY in those aged < 25. Incidence was higher in those who at baseline had either a rectal sexually transmitted infection (STI) or recent methamphetamine use, and incidence was highest in those who had both (8.93/1000PY, HR 48.9 95%CI 10.2-236). Incidence was not related to country of birth (p=0.100) or residing in a Sydney suburb with >10% of males identifying as gay (p=0.615).
Conclusions: Over a mean approaching 2 years of follow-up per person, PrEP remained highly effective, with incidence remaining below 1/1000PY overall. However, annual HIV incidence was about 1% in those who at baseline used methamphetamine and had a rectal STI.