Contraception use and HIV outcomes among women initiating dolutegravir-containing antiretroviral therapy in Kenya: a retrospective cohort study

Abstract

Introduction: The rollout of dolutegravir (DTG) in low- and middle-income countries was disrupted by a potential association reported with periconceptional DTG exposure among women living with HIV (WLHIV) and infant neural tube defects. This prompted countries to issue interim guidance limiting DTG use among women of reproductive potential to those on effective contraception. Data to understand the potential impact of such guidance on WLHIV are limited.

Methods: We conducted a retrospective cohort analysis of WLHIV 15-49 years initiating DTG-containing antiretroviral treatment (ART) in Kenya from 2017 to 2020. We determined baseline effective (oral, injectable or lactational amenorrhea) and very effective (implant, intrauterine device or female sterilization) contraception use among women who initiated DTG before (Group 1) or during (Group 2) the interim guideline period. We defined incident contraception use in each group as the number of contraceptive methods initiated ≤180 days post-guideline (Group 1) or post-DTG initiation (Group 2). We determined the proportions of all women who switched from DTG- to non-nucleoside reverse transcriptase inhibitor (NNRTI)- (efavirenz or nevirapine) containing ART ≤12 months post-DTG initiation, compared their viral suppression (<1000 copies/ml) and conducted multivariable logistic regression to determine factors associated with switching from DTG to NNRTI-containing ART.

Results: Among 5155 WLHIV in the analysis (median age 43 years), 89% initiated DTG after transitioning from an NNRTI. Baseline effective and very effective contraception use, respectively, by the group were: Group 1 (12% and 13%) and Group 2 (41% and 35%). Incident contraception use in each group was <5%. Overall, 498 (10%) women switched from DTG to an NNRTI. Viral suppression among those remaining on DTG versus switched to NNRTI was 95% and 96%, respectively (p = 0.63). In multivariable analysis, incident effective and very effective contraception use was not associated with switching.

Conclusions: Baseline, but not incident, effective contraception use was higher during the interim guideline period compared to before it, suggesting women already using effective contraception were preferentially selected to initiate DTG after the guideline was released. These findings reveal challenges in the implementation of policy which ties antiretroviral access to contraceptive use. Future guidance should capture nuances of contraception decision-making and support women's agency to make informed decisions.