Faecal contamination of the environment and child health: a systematic review and individual participant data meta-analysis

Abstract

Background: Exposure to faecal contamination is believed to be associated with child diarrhoea and possibly stunting; however, few studies have explicitly measured the association between faecal contamination and health. We aimed to assess individual participant data (IPD) across multiple trials and observational studies to quantify the relationship for common faecal-oral transmission pathways.

Methods: We did a systematic review and meta-analysis of IPD from studies identified in an electronic search of PubMed, Web of Science, and Embase on May 21, 2018. The search was done in English, but full texts published in French, Portuguese, and Spanish were also reviewed. Eligible studies quantified (1) household-level faecal indicator bacteria concentrations along common faecal-oral transmission pathways of drinking water, soil, or food, on children's hands or fomites, or fly densities in food preparation areas; and (2) individual-level diarrhoea or linear growth measures for children younger than 5 years in low-income and middle-income countries. For the diarrhoea analysis, all definitions of diarrhoea were eligible but studies were excluded if they used a recall period longer than 7 days. For the linear growth analysis (using height-for-age Z scores [HAZ]), cross-sectional studies were excluded, because of the absence of longitudinal environmental contamination data measured before the growth outcomes. We used multilevel generalised mixed-effects models to estimate the odds ratio (OR) for diarrhoea and the difference in HAZ scores for individual studies associated with a 1-log₁₀ higher measure of faecal contamination. Estimates from each study were combined under a random-effects meta-analysis framework. The study protocol was pre-registered with PROSPERO (CRD42018102114).

Findings: From 72 eligible studies, we included IPD for 20 studies in the meta-analyses, totalling 54 225 diarrhoea or linear growth observations matched to faecal indicator bacteria concentrations in drinking water, and a further 35 010 observations with faecal contamination data for the other transmission pathways. Child diarrhoea was associated with 1-log₁₀ higher faecal indicator bacteria concentrations in drinking water (OR 1·09, 95% CI 1·04 to 1·13; p=0·0002, I²=34%, 95% CI 0 to 62) and on children's hands (1·11, 1·02 to 1·22; p=0·021, I²=0%, 0 to 71). Lower HAZ scores were associated with 1-log₁₀ higher median faecal indicator bacteria concentrations in drinking water (HAZ -0·04, 95% CI -0·06 to -0·01; p=0·0054; I²=19%, 95% CI 0 to 63) and on fomites (-0·06, -0·12 to 0·00; p=0·044, I²=57%, 0 to 90).

Interpretation: Although summary measures from individual studies often report little or no effect of measured faecal contamination on child health, this multi-study IPD analysis indicates that household faecal indicator bacteria concentrations are associated with important adverse health outcomes in young children. Improved direct measures of exposure and enteric pathogens could help to better characterise the relationship and inform intervention design in future studies.

Funding: None.

Figure 1

Study selection IPD=individual participant data.

Figure 2

Summary of analyses for diarrhoea (A) and difference in height-for-age Z scores (linear growth; B). Data are for 1-log higher faecal indicator bacteria concentrations in drinking water, on children's hands and on fomites, in food and in soil, and for 1-log higher concentration in kitchen fly density. OR=odds ratio. HAZ=height-for-age Z-scores.

Figure 3

Forest plot of diarrhoea analyses Data are for 1-log higher faecal indicator bacteria concentrations in (A) drinking water, (B) on children's hands, (C) on fomites, and (D) for 1-log higher concentration in kitchen fly density. OR=odds ratio.

Figure 4

Forest plot of linear growth analyses Data are differences in HAZ scores for 1-log higher median faecal indicator bacteria concentrations in (A) drinking water, (B) on children's hands, (C) and on fomites, and (D) for 1-log higher concentration in median kitchen fly density. HAZ=height-for-age Z scores.