Abstract
OBJECTIVE: To determine how Escherichia coli contamination of household water affects the probability of stunting and underweight in children younger than 5 years in 29 low- and middle-income countries and territories. METHODS: We used data describing health, nutrition, education, and water, sanitation and hygiene (i.e. E. coli testing) from the global Multiple Indicator Cluster Surveys. We conducted multiple linear regression analyses to estimate the effects of E. coli contamination on the growth outcomes of stunting and underweight in children, and to explore the underlying mechanisms. We also conducted subgroup analyses to examine heterogeneous effects at both the macro- and microlevels. FINDINGS: Three quarters of the children in our pooled sample (26 498/35 012) were living in households with drinking water contaminated with E. coli. We observed that these children had a 2.3 (95% confidence interval, CI: 0.006 to 0.039) and 1.8 (95% CI: 0.006 to 0.031) percentage point higher probability of experiencing stunting and underweight, respectively, than children living in households with uncontaminated water. Our heterogeneity analyses revealed significant effects of E. coli contamination in girls and in poorer households (microlevel); in low- and lower-middle-income countries and territories; and in the World Health Organization African Region and Region of the Americas (macrolevel). Finally, we identified diarrhoea as a potential mechanism through which E. coli contamination might adversely affect child growth. CONCLUSION: Our findings highlight the critical need to eliminate E. coli contamination from household water sources to improve both child health and growth outcomes; changing behaviours related to open defecation remains a key strategy.