Associations between breastfeeding intention, practice, and determinants among pregnant and postpartum women living with HIV and without HIV in western Kenya

Background: Infant morbidity and mortality remain high in sub-Saharan Africa (SSA). Not exclusively breastfeeding (BF) during the first six months of life contributes to 1.5 million child deaths per year . However, the prevalence of exclusive BF during the first six months is only 53.5% in East Africa (between 2010-2015) and 61% in Kenya specifically (2016). Many of these same settings with high infant morbidity and mortality and low rates of exclusive BF also have a high burden of human immunodeficiency virus (HIV). Understanding potential factors that may influence BF intention and practice can inform future interventions to better support women’s desires to BF and for as long as they wish. To help address these gaps, we leveraged existing data collected as part of the Measuring Adverse Pregnancy and Newborn Congenital Outcomes (MANGO) study between September 2020 and March 2024 and conducted a secondary analysis to investigate associations of HIV status with BF intention during pregnancy and practice during the postpartum period in western Kenya. We also examined whether these associations are modified by employment status or parity. Methods: For the current analyses, we used data from Cohort 1 (C1) of the MANGO study which is a multifaceted prospective and retrospective cohort study (with two sub-studies) conducted to strengthen the pregnancy outcomes surveillance system of the current Kenya Ministry of Health among women living with and without HIV. The total analytic population consisted of 1003 participants. The primary exposure was having a positive HIV status while the two co-primary outcomes were BF intent and exclusive BF practice during the first six months of life. We used unadjusted and adjusted (for age, education, employment, education, marital status and parity) logistic regression models with robust standard errors to estimate odds ratios (ORs) and corresponding 95% confidence intervals (CIs). We also fitted stratified logistic regression models (described above) to examine potential effect modification by employment status and parity. In addition, we fitted interaction models that included exposure, outcome, adjustment, and effect modifier variables, as well as an interaction term for exposure and effect modifier. Results: Among study participants, 506 (50.45%) were pregnant women living with HIV (WLWHIV) while 497 (49.55 %) were pregnant women not living with HIV (WNLWHIV). A little over a third of the participants were self-employed and around 40% were not employed across both groups. About 22% of participants in the WLWHIV group and 27.6% in the WNLWHIV group were nulliparous. Slightly higher percentage of WLWHIV intended to exclusively BF (N: 379; 50.8%) compared with WNLWHIV (356; 47.8%). WLWHIV had a statistically insignificant 55% higher odds of exclusive BF intent compared with WNLWHIV in the adjusted model (OR: 1.55; 95% CI 0.41, 5.83). WLWHIV had a statistically insignificant 20% higher odds of exclusive BF practice compared to WNLWHIV in the adjusted model (OR: 1.20; 95% CI 0.72, 2.00). The HIV status-exclusive BF intent associations were stronger among employed participants (vs. unemployed participants) while the HIV status-exclusive BF practice associations were stronger among unemployed participants (vs. employed participants) and nulliparous (vs. parous) participants. However, none of the associations (and interaction p-values) were statistically significant. Conclusion: Our findings suggest that WLWHIV had higher odds of intending to exclusively BF and practicing exclusive BF compared to WNLWHIV. Our findings also suggest that employment may modify association of HIV status with exclusive BF intent while employment and parity may modify associations of HIV status with exclusive BF practice. Future, larger studies are warranted to better understand relationships between HIV status and BF, and related determinants. Knowledge gained from these studies can influence practice and policy related to promoting BF in SSA.