Pregnancy in HIV Clinical Trials in Sub Saharan Africa: Failure of Consent or Contraception?

Abstract

Higher than expected pregnancy rates have been observed in HIV related clinical trials in Sub-Saharan Africa. We designed a qualitative study to explore the factors contributing to high pregnancy rates among participants in two HIV clinical trials in Sub-Saharan Africa. Methods: Female and male participants enrolled in one of two clinical HIV trials in south-west Uganda were approached. The trials were a phase III microbicide efficacy trial among HIV negative women using vaginal gel (MDP); and a trial of primary prevention prophylaxis for invasive cryptococcal disease using fluconazole among HIV infected men and women in Uganda (CRYPTOPRO). 14 focus group discussions and 8 in-depth interviews were conducted with HIV positive and negative women and their male partners over a six month period. Areas explored were their experiences about why and when one should get pregnant, factors affecting use of contraceptives, HIV status disclosure and trial product use. Results: All respondents acknowledged being advised of the importance of avoiding pregnancy during the trial. Factors reported to contribute to pregnancy included; trust that the investigational product (oral capsules/vaginal gel) would not harm the baby, need for children, side effects that led to inconsistent contraceptive use, low acceptance of condom use among male partners. Attitudes towards getting pregnant are fluid within couples over time and the trials often last for more than a year. Researchers need to account for high pregnancy rates in their sample size calculations, and consider lesser used female initiated contraceptive options e.g. diaphragm or female condoms. In long clinical trials where there is a high fetal or maternal risk due to investigational product, researchers and ethics committees should consider a review of participants contraceptive needs/pregnancy desire review after a fixed period, as need for children, partners and health status of participants may alter over time.