Investigating drug-drug interactions and contraceptive efficacy among women living with HIV initiating progestin-based hormonal contraception in a setting with high tuberculosis prevalence: analysis of FP-ART cohort data from Lilongwe, Malawi

In sub-Saharan Africa, approximately 16 million women are living with HIV, many of whom are of reproductive age. Concomitant use of antiretroviral therapy, progestin-based hormonal contraception, and other drugs (such as isonicotinic acid hydrazide [isoniazid] for the prevention of tuberculosis disease) is essential for reducing maternal morbidity and mortality and the prevention of vertical HIV transmission. In this setting, the risk of drug-drug interactions that may compromise contraceptive efficacy demands attention. As an example, there is strong evidence of a clinical drug-drug interaction between efavirenz and the levonorgestrel implant due to cytochrome P4503A4 (CYP3A4) induction by efavirenz that results in decreased levonorgestrel plasma concentrations and increased contraceptive failure during combined use. In contrast, data from in-vitro experiments suggests that isoniazid may inhibit CYP3A4, which could lead to higher progestin concentrations and potentially serve as a protective factor against contraceptive failure. In this dissertation, we used prospective cohort data from the Family Planning and Antiretroviral Therapy (FP-ART) study to investigate the potential effect of drug-drug interactions among women living with HIV in Malawi who were using either depot medroxyprogesterone acetate (DMPA) or the levonorgestrel implant – the two most commonly used hormonal contraceptive methods in Malawi. First, we investigated the effect of isoniazid use on contraceptive failure among participants with documented adherence to either DMPA or the implant. Next, we estimated the effect of ART regimen type (efavirenz or dolutegravir-based) and ART adherence on pregnancy incidence among women who initiated DMPA when they enrolled in the FP-ART cohort. We found higher-than-expected pregnancy incidence during FP-ART follow-up: among visits with recently documented adherence to either DMPA or the levonorgestrel implant, 2.3% had a pregnancy event in the 6-month post-visit period. Further, among the participants who enrolled in FP-ART after a DMPA injection and who were followed until first evidence of DMPA discontinuation, the pregnancy rate was 5.3 per 100 person-years. We found a slightly higher incidence of 6-month pregnancy for FP-ART study visits in which current isoniazid use (versus non-use) was reported, though the risk difference was not statistically significant (risk difference, [95% confidence interval (CI)]: 0.7%, [-0.7%, 2.1%]), and we found a higher pregnancy rate among DMPA users on efavirenz versus dolutegravir-based therapy (incidence rate ratio, [95% CI]: 1.54, [0.71, 3.33]).