Water Source Correlates with E. Coli Contamination and Markers of Environmental Enteric Dysfunction in Rural Ugandan Infants (P10-123-19)

Abstract

To examine associations among water source, E. coli contamination, and environmental enteric dysfunction (EED) in a cohort of infants participating in the Uganda Birth Cohort Study (UBCS).Infants (n = 729) from 16 sub-counties in rural northern and southwestern Uganda were included in this cross-sectional secondary analysis. Water source and other covariate data were derived from the UBCS 6-month visit, which included a compartment bag test (CBT, Aquagenx, Chapel Hill, NC) to measure E. coli contamination of drinking water. Safe water was defined as no E. coli detected. EED was assessed at age 6 months using serum concentrations of antibodies to the bacterial components flagellin and lipopolysaccharide (LPS) via ELISA. EED biomarkers were natural-log transformed; associations were assessed using cluster-robust adjusted linear regression models.Half of infants were female; 95% were breastfeeding at 6 months of age. Most households (72.8%) relied primarily on an improved water source; i.e., piped water, public tap, tubewell/borehole, or protected well. While 2.6% relied primarily on rain water harvesting, 45.1% practiced it to some degree. Per CBT results, 33.5% of households had safe water, with large variations in quality observed by source [source (% with safe water): piped (50.0%), public tap (17.1%), tubewell/borehole (40.5%), protected well (20.8%), unprotected well (33.0%), rain water (68.4%), surface water (31.6%)]. In adjusted linear regression models (controlling for child sex, current breastfeeding status, and household wealth index), infants from households relying primarily on rain water had significantly lower anti-flagellin IgG levels compared to those relying on improved sources, including piped water (P < 0.001), public tap (P = 0.05), or tube well/borehole (P = 0.03). Furthermore, infants from households that practiced any rain water harvesting had significantly lower EED biomarkers compared to those that did not [anti-flagellin IgG: -0.08 (-0.13, -0.02, P = 0.008); anti-LPS IgG: -0.07 (-0.11, -0.03, P = 0.003)].In rural Uganda, E. coli contamination and infant EED differed significantly by water source; intervention trials involving rain water harvesting may be warranted.Feed the Future Innovation Lab for Nutrition at Tufts University, supported by USAID; CD also supported by NIH.