Abstract Body

Background: Injectable contraception constitutes a large proportion of modern contraceptive use globally. Several observational studies have reported an increased risk of HIV acquisition among women using depot medroxyprogesterone acetate (DMPA) compared to women using non-hormonal methods. Previous analyses have used a comparator group not on hormonal contraception (HC), and few have disaggregated use of norethisterone enanthate (NETEN), another commonly used injectable method in South Africa.

Methodology: MTN-003 was a multi-site, randomized trial of oral and topical tenofovir-based HIV chemoprevention in women. Participants using effective contraception (HC, intrauterine device or sterilization) were eligible for enrollment. Contraceptive use and pregnancy and HIV status were assessed monthly during follow-up. Using a Cox proportional hazards model stratified by study site and censoring for pregnancy, we directly compared rates of HIV acquisition between women using DMPA and those using NET-EN at 11 South African MTN-003 sites.

Results: Among 4,077 participants, 3,167 women used injectable contraception during the study, of which 1,801 (56.9%) exclusively used DMPA, 1,109 (35.0%) exclusively used NET-EN and 257 (8.1%) used both DMPA and NET-EN during follow-up. Although DMPA and NET-EN users had similar rates of sexually transmitted infection at baseline, compared to NET-EN users, DMPA users were more likely to be over 25 (55.7% vs. 48.4%; p<0.001), married (5.7% vs. 3.7%; p<0.04) and have ≥ one child (93.5% vs. 69.8%; p<0.001). Distribution of injectable method types and HIV incidence differed by site (proportion of DMPA, 26.9% to 73.4%; HIV incidence, 1.7 to 9.3/100 person-years [p-yrs]). Among users of injectable contraception, 246 acquired HIV for an overall rate of 6.6 events/100 p-yrs. Rates of HIV-1 acquisition among women currently using DMPA and NET-EN during follow-up were 7.86 and 5.34/100 p-yrs, respectively (unadjusted hazard ratio [HR] 1.31, 95% CI 0.96 to 1.79, p=0.092; adjusted for age and marriage [aHR] 1.44, 95% CI 1.05 to 1.98, p=0.024). Restricting analysis to women under 25 found a similar but non-significant estimate (aHR 1.55, 95% CI 0.99 to 2.32, p=0.055).

Conclusions: Among women willing to use effective contraception while participating in an HIV prevention trial, there was an increased risk of HIV acquisition among women currently using DMPA compared to those using NET-EN. However, lack of a non-HC comparator group prevents estimation of risk associated with either method compared to non-use. Site differences in distribution of methods and HIV incidence may have impacted these results. Our findings support current WHO recommendations that women using injectable progestin contraception should be advised to use condoms.