Are women who take hormonal contraceptives at an increased risk of acquiring HIV? If so, do some contraceptives put women at higher risk than others?
This week, the influential health journal, PLOS Medicine, published the results of a large individual-participant data meta-analysis, authored by FHI 360 and collaborators, that seeks to answer these questions.
While this issue matters to the field of reproductive health, it is especially critical to women in East and Southern Africa. In these regions, women potentially have a double risk factor: high rates of HIV and high use of hormonal contraception, particularly depot-medroxyprogesterone acetate (DMPA), a type of contraceptive that is injected every three months. So far, the evidence on DMPA shows that it is the hormonal contraceptive that has the most potential to increase HIV acquisition; however, the evidence is inconclusive.
FHI 360’s meta-analysis combines the results of 18 prospective studies, including more than 37,000 women, of whom more than 1,800 became infected with HIV. We found that women who used DMPA had a 50 percent increased risk of HIV acquisition compared with women who did not use hormonal contraceptives. We found no significant increase in HIV risk among women using combined oral contraceptives (COCs) or norethisterone enanthate (Net-En), a contraceptive injected every two months. Women using DMPA also had an increased HIV risk when compared directly with COC or Net-En users.
Importantly, those studies that we judged to have a lower risk of methodological bias had lower estimates for the risks of HIV infection associated with DMPA. This suggests that some of the increased HIV risk we found to be associated with DMPA use could be due to the methodological shortcomings of the studies.
Our results support previous research that suggests that DMPA, but not COCs or Net-En, may increase the risk of HIV infection. However, the results of our meta-analysis are not likely to change most experts’ opinions about whether hormonal contraception, and in particular DMPA, is biologically associated with an increased HIV risk.
On the one hand, those who suspect there is a true biological association between DMPA and HIV acquisition will have additional evidence for backing this view. On the other hand, those with doubts will likely point to the stratified analysis related to the methodological bias as evidence that the association is an artifact of methodological shortcomings — chiefly that all the research is observational and thus subject to bias.
More than ever, a well-conducted randomized trial is needed to provide the most robust evidence to answer this critical reproductive health question.