Gut Microbes, Enteropathy and Child Growth: The Role of the Microbiota in the Cycle of Diarrhea and Undernutrition in Peru

Background. The cycle between diarrhea and undernutrition continues to underlie a vast proportion of under-five mortality and is the primary driver of long-term disability among children living in lower and middle-income countries (LMIC). Interventions aimed at reducing childhood stunting have not achieved desired results, highlighting the need for novel research and strategies to target this problem. There is increasing evidence that the gut microbiota are implicated in growth acquisition and sustaining intestinal barrier integrity in a manner that impacts immunity to and consequences of disease; however, these relationships have not yet been examined in large-scale studies of children living in LMIC. Objective. To evaluate relationships between the gut microbial community, child growth, diarrhea and enteric infections (Campylobacter spp) in a birth cohort of 271 children aged 0-24 months in Iquitos, Peru. Methods. Analyses were conducted on children participating in the multi-site cohort study entitled ‘The Interactions of Malnutrition & Enteric Infections: Consequences for Child Health and Development (MAL-ED).’ Data were contributed over two years by mothers and children living in a peri-urban riverine community in Amazonian Peru. Regular home-visits were conducted to ascertain anthropometric indices, illness history, and dietary habits. Length-for-age (LAZ) and weight-for-length (WLZ) Z-scores below - 2 were used to classify stunting and wasting, respectively. Fecal specimens were collected during routine surveillance visits at monthly intervals (N=6004) and additionally during each maternal report of diarrheal symptoms (N=2436). Culture methods, immunoassays and amplification methods were employed according to a unified MAL-ED protocol to identify a panel of over 40 protozoa, bacteria and viruses of public health importance. Microbiota in fecal samples contributed at 6, 12, 18 and 24 months were analyzed by polymerase chain reactions using primers to identify variable regions of bacterial 16S ribosomal RNA genes at the Gordon Laboratory at Washington University. Members were binned into operational taxonomic units (OTU) sharing ≥97% nucleotide sequence identity, producing bacterial communities differentiated at the species level which were then used to generate metrics of maturity (microbiota-for-age Z score; MAZ), diversity (Shannon, Simpson indices) and richness (CHAO1, Faith’s Phylogenetic Diversity). Multivariable regression was used to detect and describe population-averaged associations between microbial metrics, growth acquisition, illness and infection with a generalized estimating equations approach to adjust for within-child correlations over time. Indicator species analysis (ISA) was employed to identify particular gut taxa whose presence and abundance was statistically indicative of phenotypes of interest. Results. Two-thirds of children (67%) were stunted and 9% of children experienced wasting before their 2nd birthday. Microbial diversity and richness increased significantly with age and weaning, and were suppressed by breastmilk exposure. In the first two years of life, we detected a suggestive relationship between microbial maturity and WLZ, but did not observe evidence of associations between microbial maturity, diversity or richness with LAZ in the full cohort. LAZ at birth was significantly associated with MAZ score throughout follow-up (β=0.10, p=0.012) and children born stunted had significantly lower gut microbial diversity and richness (Shannonβ=-0.19, CHAO1 =- 9.75; p-values <0.05) from birth to two years of age. In this subgroup, we additionally observed that children weaned before 24m of age experienced significantly pronounced deficits in microbial diversity and richness acquisition relative to those with continued breastfeeding. Nearly all children (96%) experienced diarrhea during follow-up. Odds of being severely stunted increased by 8% with each additional diarrheal episode throughout the first two years of life (OR=1.08; p<0.001). Cumulative diarrheal frequency, duration and severity were associated with significant reductions in microbial indices (p<0.05), and we observed evidence of enduring deficits beyond 1m after exposure. Children who were born stunted experienced greater insults to microbial diversity per diarrheal episode than those children who were not (Interaction terms: Shannon β =-0.04, p=0.037; Simpson β =-0.01, p=0.032). Time elapsed since last diarrheal episode was associated with recovery of Shannon (β =0.02, p=0.03) and phylogenetic diversity (β =0.11, p<0.01) and we detected evidence that this regeneration process was significantly slower among severely stunted children. Lower diversity and richness were associated with increased subsequent diarrheal incidence; a 1-unit increase in the Shannon and Simpson’s Diversity scales at 6m corresponding to a mean reduction of 1.3 and 3.4 diarrheal episodes from 6-24m of age, respectively. By two years of age, 251 (93%) of all children in the cohort had Campylobacter present in asymptomatic stools, and 221 (82%) experienced infection with diarrhea. Asymptomatic infection was associated with reduced LAZ concurrently and at 3, 6, and 9m thereafter (β=0.02, p