20th International AIDS Conference - Melbourne, Australia


THAC0503 - Oral Abstract Session

Hormonal contraception and HIV infection: results from a large individual participant data meta-analysis

Presented by Charles Morrison (United States).

C. Morrison1, P.-L. Chen2, C. Kwok2, A. Bernholc2, N. Low3, for the HC-HIV IPD Meta-Analysis Study Group

1FHI 360, Clinical Sciences, Durham, United States, 2FHI 360, Biostatistics, Durham, United States, 3University of Bern, Institute of Social and Preventive Medicine, Bern, Switzerland

Background: Effective contraception is essential for women''s reproductive health. Observational study results conflict about whether hormonal contraception (HC), particularly depot-medroxyprogesterone acetate (DMPA), increases the risk of HIV acquisition. There is an urgent need to resolve this question.
Methods: We conducted an individual participant data (IPD) meta-analysis of studies from sub-Saharan Africa to compare the incidence of HIV infection in women aged 15-49 years using combined oral contraceptives (COC) or the injectable progestins DMPA or norethisterone enanthate (NET-EN) with women not using HC. Eligible studies measured HC exposure and incident HIV prospectively, recorded >15 incident HIV infections, and measured pre-specified covariates. Our primary analysis estimated the hazard ratio (HR) using two-stage random effects meta-analysis, controlling for region, marital status, age, number of sex partners and condom use. We conducted sensitivity analyses to assess whether results were influenced by risk of methodological bias in component studies, HIV incidence, pregnancy status, or limiting person-time to periods with no condom use.
Results: We included 18 studies which followed 37,124 women for 43,613 person-years with 1,830 incident HIV infections. Relative to non-users, the pooled adjusted HR for HIV acquisition was 1.50 95%CI:1.24-1.83 for DMPA, 1.24 95%CI:0.84-1.82 for NET-EN, and 1.03 95%CI:0.88-1.20 for COC. Between study heterogeneity was mild (I2 < 50%). There was no modification of the HC effect on HIV risk by either age or HSV-2 infection status. The risk of bias in component studies affected the results (p=0.003). Studies at lower risk of bias showed lower HRs: DMPA (1.22; 95%CI:0.99-1.50), NET-EN (0.67; 95%CI:0.47-0.96), COC (0.91; 95%CI:0.73-1.14) than those at higher risk of bias: DMPA (HR 1.73; 95%CI:1.39-2.16), NET-EN (HR 1.50; 95%CI:1.14-1.96) and COC (HR 1.16; 95%CI:0.93-1.45). Other sensitivity analyses found very similar results to the primary findings.
Conclusions: We found evidence that DMPA but not NET-EN or COC use increased women''s risk of HIV. However, the estimated risks associated with HC use were substantially lower in studies at less risk of methodological bias, highlighting the limitations of observational data. Randomized clinical trials are required to conclusively determine whether HC, especially DMPA, increases the risk of HIV acquisition compared to alternative contraceptive methods.

Back to the Programme-at-a-Glance