Population-level correlation between incidence of selected sexually transmitted infections and HIV-1 among women participating in HIV pre-exposure prophylaxis trials in Africa


BACKGROUND: Biomedical strategies for HIV pre-exposure prophylaxis (PrEP) have been developed and many African countries have incorporated access to daily oral PrEP into their national guidelines. Because uptake of effective PrEP will influence the design and interpretation of future clinical trials of novel PrEP agents, surrogate indicators of HIV incidence in HIV prevention trials may prove to be valuable. For men who have sex with men, the incidence of rectal gonorrhea has been proposed as a strong surrogate for HIV incidence. While sexually transmitted infections have long been recognized as risk factors for HIV acquisition at the individual level, whether their incidence at the population level correlates with HIV incidence is not established. The objective of this analysis was to assess whether incidence of sexually transmitted infections was correlated with HIV incidence among African heterosexual women.
METHODS: Data was included from the placebo arms of two HIV PrEP based trials among African women: MTN-003/VOICE and MTN-020/ASPIRE. These were multi-site, double-blind, placebo-controlled randomized trials enrolling healthy sexually active HIV-uninfected women at risk of HIV acquisition. Women were screened and treated for STIs (Neisseria gonorrhoeae, Chlamydia trachomatis, and Trichomonas vaginalis) at baseline and throughout study participation. HIV serologic testing was performed monthly. Regression modelling was used to assess the correlation between incident STIs and incident HIV, with each site for each study representing a data point.
RESULTS: In total 3,314 women were enrolled across 18 sites in 4 countries (Malawi, South Africa, Uganda, Zimbabwe). Overall STI and HIV incidences were high: HIV 4.9, Neisseria gonorrhoeae 5.2, Chlamydia trachomatis 14.5, and Trichomonas vaginalis 7.1 per 100 person-years. There was limited correlation between HIV incidence and the incidence of individual STIs with R-squared values generally low: Neisseria gonorrhoeae (R2=0.001), Chlamydia trachomatis (R2=0.256), and Trichomonas vaginalis (R2=0.011). HIV incidence was greatest in South Africa; however, analysis limited to South African sites similarly did not show a strong correlation between site-level HIV and STI incidence.
CONCLUSIONS: The incidence of sexually transmitted infections did not reliably predict HIV incidence at the population level amongst at-risk African women participating in two large HIV PrEP trials.