DMPA and HIV: do we need a trial?

Whether hormonal contraceptive use increases the risk of HIV acquisition is still a much debated question. In countries that endure the double challenges of high HIV prevalence and high maternal mortality, the need for safe contraceptives is enormous, yet the meaning of “safe” becomes impossible to parse because of competing risks. Effective contraceptive use in these settings clearly and dramatically reduces maternal mortality. If women — or clinicians or health systems — avoid some contraceptives due to fear of HIV, will overall maternal deaths increase or decrease? The answer may vary by country and population subgroup, and thus general advice for public health policy is paralyzed. Thus far, our understanding of the increased risks comes from observational studies and secondary outcomes from trials designed to answer other questions. The quandary is whether DMPA users (and perhaps NET-EN users) have an increased risk of HIV acquisition and, if increased, by how much. If an increased risk were certain, what should we do? How might the magnitude of this risk drive public policy in different settings? Will the answer be the same across settings that have different HIV prevalence, different maternal mortality and different contraceptive method mix? Jain's calculations, previously published in this journal, suggest that, under most assumptions (based on current assessment of risk), the withdrawal of DMPA is unwarranted [ , ].