Water insecurity (WI) is a serious and worsening problem worldwide, but its role in health outcomes among people living with HIV or pregnant women is unknown. We assessed experiences of WI in a cohort of 323 pregnant Kenyan women of mixed HIV status. The majority (77.7%) had at least one experience of WI in the previous month; it was associated with negative economic, nutrition, disease, and psychosocial outcomes. A standardized cross-culturally valid household WI scale would facilitate assessment of the prevalence and consequences of WI, and increased attention to WI could reveal an overlooked, but modifiable, cause of adverse HIV outcomes.

HIV infection was associated with lower weight, fat mass, fat-free mass, TSF, AFA, and MUAC values during pregnancy and lactation, while food insecurity was intermittently associated with body composition. This suggests that pregnant and lactating women living with HIV and food insecurity could benefit from nutritional support.

Food insecurity (FI) is well established to be harmful among children, adolescents, and adults, but little is known about either its prevalence or covariates during pregnancy. Given the importance of maternal health for the lifelong well‐being of children, a greater understanding of FI may provide insight into potentially modifiable determinants of maternal and child health. Therefore, we examined (1) the population prevalence of food insecurity among all pregnant women attending 7 urban and rural antenatal clinics in Nyanza province, Kenya, and (2) assessed the predictors of FI at baseline among women enrolled in Pith Moromo, an ongoing observational pregnancy cohort study in this region.Cohort participants (n=363) were enrolled at <30 weeks gestational age, and were purposively sampled to achieve equal proportions of HIV‐infected and –uninfected women across low, moderate, and severe tertiles of food insecurity. FI was measured by clinic‐based study staff using an individually focused food insecurity access scale (IFIAS, Natamba et al., 2014). Staff assessed depression using the Center for Epidemiologic Studies Depression Scale (CES‐D), and assessed stress using the Perceived Stress Scale (PSS). An asset score was created by assessing patterns of asset ownership, and weighted using principal component analysis. Linear regression was used to evaluate predictors of FI at enrollment into the cohort, including maternal age, gestational age, HIV status, marital status, perceived stress, and urban/peri‐urban/rural locality. In the population‐based sample (n=1,188), the mean score for FI was 12.9 (SD 7.4), 39.44%, 34.54%, and 26.01% of women were in the low, moderate, and severe FI groups. Among participants in the cohort, the mean score for FI was 13.6 (SD 7.1). In multivariate linear models among cohort participants, CES‐D score was positively associated with FI (β 0.11, 95% confidence interval 0.00 to 0.21, p<0.05), and greater perceived stress was also associated with higher FI (β 0.19 score, 95% confidence interval 0.00 to 0.39, p<0.05). Greater asset index scores were associated with lower FI (β −1.43, 95% confidence interval −2.02 to −0.85, p<0.001). Maternal HIV and marital status were not significantly associated with FI.In sum, FI is experienced by many pregnant women in this region. In our cohort, we observed significant positive correlations between FI and psychosocial indicators, including probable maternal depression and perceived stress. Understanding the relationships between mental health and food insecurity may help inform the development of psychosocial health interventions to improve mental health and/or food security among pregnant women.Support or Funding InformationSupport: SLY NIH K01 MH098902