Background: The potential impact of injectable hormonal contraception (HC) on C. trachomatis (CT) and N. gonorrhoeae (NG) acquisition in women is unclear. This is the first direct comparison of two commonly used methods of injectable HC in South Africa (depot medroxyprogesterone acetate [DMPA] and norethisterone enanthate [NET-EN]) on acquisition of these infections.
Methods: MTN-003 was a randomized trial of HIV chemoprevention in Africa. Women using protocol-defined effective contraception were eligible for enrollment. Testing and treatment for CT and NG occurred at baseline, annual and exit visits, and when clinically indicated. Andersen-Gill proportional hazards models were used to assess the association between injectable contraceptive type and incident CT and NG among 4,077 South African MTN-003 participants.
Results: Among the 3,246 (79.6%) participants who used injectable HC during follow-up, DMPA users were more likely, compared to NET-EN users, to be >25 years (43.1% vs. 34.0%, p<0.001), but had similar baseline prevalence of CT (14.3% vs. 14.2%, p=0.95) and NG (3.2% vs. 3.7%, p=0.49). During 3,761 person-years of follow-up, 514 cases of CT (14.2/100 person-years) and 118 of NG (3.1/100 person-years) were observed. Incidence did not differ between current DMPA and NET-EN users for CT (14.1/100 person-years [p-y] vs. 14.5/100 p-y, hazard ratio [HR] 0.95, 95% CI 0.79-1.16) or NG, (3.3/100 p-y vs. 3.8/100 p-y, HR 0.80, 95% CI 0.58-1.09). Adjustment for age, marriage, number of partners and condom use did not modify these inferences (CT: adjusted hazard ratio [aHR] 1.09, 95% CI 0.92-1.29; NG: aHR 0.92, 95% CI 0.64-1.32).
Conclusions: The risk of incident CT or NG did not differ between DMPA and NET-EN users. Lack of a non-HC comparator prevented estimating risk associated with either injectable method compared to non-use. Women choosing injectable HC should be counseled that risk of these infections does not differ between DMPA and NET-EN use.