Causal Pathways from Enteropathogens to Environmental Enteropathy: Findings from the MAL-ED Birth Cohort Study

1Traditionally, the evaluation of pathogenicity from enteric infections in the host has focused on the evaluation of defined diarrheal or acute gastrointestinal illnesses and the health outcomes associated with such illness perceived as binary, that is, either death or survival. It has been posited that the host response to frequent enteric infections alters the gut in a way that adversely affects the health status of the host even in the absence of diarrhea or acute gastrointestinal illness. The consequences of this altered host phenotype may have long term effects on child health and development potential. This condition is known as environmental enteropathy (EE). A proposed consequence of EE is reduced linear growth in children, and EE may explain the less than expected effectiveness of nutritional interventions to improve growth in developing populations. Several mechanisms have been proposed to explain how EE results in poorer nutritional status by reducing functional capacities of the gut. These include reduced absorptive capacity, increased permeability, and chronic intestinal and systemic inflammation with resulting metabolic changes that affect nutrient and micronutrient availability and utilization.

2Multiple physiological mechanisms by which enteropathogens can disrupt gut functioning have been identified although the long term consequences in settings where exposure to enteropathogens is intense and continuous are poorly understood. Populations in low- and middle-income countries are also subject to other causes of growth failure, including inadequate dietary intake and frequent overt illness, any of which may influence both EE biomarkers and observed growth outcomes.

3The collection of non-invasive biomarkers of EE is expanding, with different markers characterizing different aspects of gut physiology and integrity. The most widely used EE biomarker is the lactulose:mannitol (L:M) dual sugar test for intestinal permeability, which has been used to demonstrate that altered gut permeability is related to the risk of stunting and is prevalent in environments with poor sanitation. Other EE bioassays available include fecal markers of gut inflammation, intestinal growth factors, and plasma markers of bacterial translocation. Additionally, an increasing set of markers are becoming available that encompass systemic inflammation and amino acid and lipid metabolism. Aligning the pathways indicted by this expanding collection of biomarkers with enteric infections and growth in early infancy and childhood across different populations is the subject of considerable current effort.

4The Etiology, Risk Factors and Interactions of Enteric Infections and Malnutrition and the Consequences for Child Health and Development (MAL-ED) study was designed to assess the role of enteropathogens and other factors in growth faltering from birth to two years across eight sites. A central hypothesis of MAL-ED is that gut injury resulting in disruption of normal physiology is the key route by which enteropathogens contribute to malnutrition. Here, we use a causal systems model (a directed acyclic graph [DAG]) to test key theoretical pathways of the EE conceptual model and examine how enteropathogen infection results in impaired physical growth in infancy and early childhood (Fig. 1).

5The MAL-ED study, conducted in eight diverse sites on three continents: Bangladesh (Dhaka: BGD), India (Vellore: INV), Nepal (Bhaktapur: NEB), and Pakistan (Naushero Feroze: PKN) in Southern Asia; Brazil (Fortaleza: BRF) and Peru (Loreto: PEL) in Latin America; and South Africa (Venda: SAV) and Tanzania (Haydom: TZH) in Sub-Saharan Africa. The study design is described in detail elsewhere. In brief, children were enrolled within 17 days of birth, but excluded if they had a birth weight < 1500 g, were very ill, or were non-singleton; or if their mother was < 16 years of age. Approximately 10 children were enrolled per month per site, with the goal of retaining ≥ 200 per site at 24 months of age (after loss to follow up). Data collection methods are described elsewhere for illness and treatment, infant feeding, and stool microbiology. Each site obtained ethical approval from their respective institutions and written consent was obtained from participants.

6Non-diarrheal stool samples collected monthly in the first year and quarterly in the second year were evaluated for > 40 pathogens using a standardized approach. In addition to analyzing total number of pathogens detected per stool, we also categorized pathogens into five groups based on pathophysiology. Group I included viruses that cause limited mucosal disturbances (rotavirus, adenovirus and astrovirus). Group II included bacteria that are enteroinvasive or cause extensive mucosal disruption (Campylobacter, Shigella, Salmonella, Plesiomonas, Yersinia, enteroaggregative E. coli (EAEC), enteropathogenic E. coli (EPEC), enteroinvasive E. coli (EIEC) and Aeromonas). Group III was enterotoxigenic E. coli (ETEC), which is a cause of secretory diarrhea with only limited mucosal changes. Cryptosporidium (Group IV) and Giardia (Group V) were considered independently as organisms have both been shown to be associated with linear growth failure and prolonged and persistent carriage.

7Three fecal biomarkers relating to aspects of gut inflammation and immunity (“local inflammation” in Fig. 1) were evaluated using the same non-diarrheal stool samples assayed for enteropathogens: (1) myeloperoxidase (MPO, ng/mL) as a marker of neutrophil activity in the intestinal mucosa (Alpco, Salem, NH, USA); neopterin (NEO, nmol/L) to indicate T-helper cell 1 activity (GenWay Biotech, San Diego, CA, USA); and alpha-1-antitrypsin (AAT, mg/g) to indicate protein loss and intestinal permeability (Biovendor, Candler, NC, USA). Because diarrhea leads to stool dilution, fecal biomarker values were excluded if proximate to diarrheal symptoms (within seven days prior). Similarly, stools collected the day of or the day following the L:M test were excluded as this test is an osmotic laxative.

8In addition to fecal biomarkers, urinary L:M testing (“gut permeability” in Fig. 1) was performed at three, six, nine, and 15 months, as described elsewhere. Urine samples were processed using high-performance liquid chromatography and pulsed amperometric detection or ion chromatography (depending on study site). The results were converted into a sample-based Z-score (LMZ) to minimize age and sex trends. Data from the BRF cohort were used as the internal reference standard.

9Finally, systemic inflammation was evaluated at seven, 15, and 24 months using alpha-1-acid glycoprotein (AGP) concentration in plasma. Incidence of acute lower respiratory infection (ALRI), diarrhea, fever (associated with neither ALRI nor diarrhea), and a composite for any of the three categorized illness episodes in the seven or 14 days preceding the blood collection were drawn from bi-weekly maternal reports. These were used to examine the influence of recent, non-diarrheal, overt illness on AGP concentration.

10Monthly length (cm) and weight (kg) measures were converted to Z-scores (LAZ, WAZ respectively) based on WHO 2006 standards. The change (Δ) in LAZ and WAZ for each child (final minus initial value for each period) served as the outcome in all analyses, controlling for the initial value. Intense quality assurance review procedures identified bias within the PKN length measures; therefore, these data were excluded from the system analysis. PKN biomarker data were however, included in the evaluation of associations between pathogens and biomarkers.

11First, to maximally leverage the large size of the MAL-ED dataset and to place our results in the context of previous studies, we analyzed relationships between pathogens and fecal biomarker concentrations, between pathogens and LMZ scores, between LMZ scores and changes in anthropometry, and among potential sources of systemic inflammation not associated with gut enteropathy.

12Linear mixed effects models were constructed to examine cross-sectional associations between individual pathogens and concentrations of each fecal biomarker. Specifically, the log concentrations of MPO, NEO, and AAT were modeled as functions of stool consistency (a categorical description of stool liquidity), linear and quadratic terms for child age (to capture age-related trends), the presence of individual pathogens (adjusting for the presence of other pathogens), and a random intercept for child nested in site.

13The same model structure was extended to evaluate associations between pathogen presence and LMZ scores, limiting the analyses to non-diarrheal stools collected at the same age as the L:M test. Additionally, changes in anthropometry (ΔLAZ and ΔWAZ) over three, six, and nine month windows starting at each L:M assay were evaluated as a function of the LMZ scores. Individual children nested within their respective site were treated as a random intercept to account for clustering at both the individual child and site levels.

14To determine whether the concentration of AGP was related to overt illness in the seven or 14 days preceding blood collection, another linear mixed effects model was constructed with log-transformed AGP concentration as a function of age and illness (i.e., the presence of diarrhea, fever, and ALRI). A random intercept for child nested in site was included.

15In addition to these linear regressions and given that disease systems composed of different interacting pathways lend themselves to causal graphical modeling (e.g., Fig. 1) we constructed a DAG model to test hypothetical pathways between the presence of enteropathogens, biomarker concentrations, and changes in LAZ and WAZ. Combining all factors into a single system allowed for the explicit partition of associations into direct and indirect pathways.

16Variables within this system were represented as conditionally independent, multivariate, generalized linear mixed models such that the probability of observing a given value for each variable was a function of other variables connected within the system (indicated by arrows in Fig. 1). To account for heterogeneity between sites, random effects for both site and child were added at every node. The DAG analysis focused on two time periods, 4 ≤ months ≤ 11 (Age 1) and 12 ≤ months ≤ 21 (Age 2), using the data collection schedule shown in Fig. 2 to capture temporal associations between events marked by the biomarkers. Specifically, pathogen data was coincident with collection of the fecal biomarkers (MPO, NEO, and AAT). Their collection preceded collection of L:M as alterations in gut permeability are a hypothesized result of inflammation. Measures of systemic inflammation then followed. The temporal window then extended beyond biomarker collection to assess associations with subsequent growth.

17The net effects of both direct and indirect pathways were simulated from the fitted DAG. Sensitivities of ΔLAZ and ΔWAZ to changes in each biomarker were examined by fixing each biomarker to its observed mean concentration as well as one standard deviation higher or lower. The ΔLAZ and ΔWAZ were then simulated and the difference between their mean values when biomarkers were raised or lowered relative to when they were held at mean concentration were estimated.

18The model was run in JAGS (version 3.4.0) to perform Markov Chain Monte Carlo simulations. Further details are given in the Appendix.

19The Bill & Melinda Gates Foundation did not play any role in the writing of the manuscript nor did the funders have of the study had any role in the study design, data collection, analysis, or interpretation of study results. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

20Data from the entire cohort were included in the linear analyses for associations between pathogens and fecal biomarkers and LMZ, the changes in anthropometry following the L:M test, and between illness and AGP (Table 1, Table 2). Among the over 20,000 non-diarrheal stools tested for concentrations of MPO, NEO, and AAT, there was a trend for the concentration of each biomarker to decrease with increasing age. Within individual children; however, biomarker concentrations were highly variable across time, with intra-class correlations (ICC) of 0·07, 0·03, and 0·06 for MPO, NEO, and AAT respectively. Pearson correlation coefficients between the biomarkers were low (≤ 0·2) suggesting they measured different physiological insults.

21A comparison of associations between enteropathogens and the fecal biomarkers revealed that pathophysiological groups tended to show similar trends, with some prevalent pathogens being associated with either higher or lower biomarker concentrations (e.g., Campylobacter and EAEC were associated with higher concentrations of MPO and AAT, while Giardia was associated with lower concentrations of all three fecal biomarkers). The associations were more pronounced in some of the rarer pathogens (e.g., Yersinia enterocolitica was strongly associated with increased MPO and decreased NEO concentrations the few times it was detected) (Fig. 3). These exploratory models assumed additive effects of pathogens, though in many instances more than one pathogen was detected.

22LMZ tended to be higher, indicating increased permeability, in children with pathogens detected, especially those with positive tests for Cryptosporidium (0·34 ± 0·08 for mean ± standard deviation, N = 134) and Giardia (0·20 ± 0·05, N = 457) (Fig. 3). However, there was no statistical support for LMZ relating to changes in either LAZ or WAZ (Table 3).

23With respect to systemic inflammation, approximately half the measures of AGP were elevated (> 100 mg/dL), indicating common subclinical inflammation. Across the two time points when AGP was measured, high concentrations were dispersed between children rather than focused in an identifiable subset (ICC 0·16). Across 4257 AGP assays, neither symptoms of ALRI (which was uncommon at the time of the two blood draws) nor diarrhea were associated with concentrations of AGP (Table 4). Maternally-reported fever in the 14 days prior to the blood draw was associated with increased AGP concentration. Only the subset of MAL-ED children with complete data (i.e. with no missing observations across all variables) was included in the DAG (Table 1).

24Within the systems model, which included the pathogen groups, all biomarkers, and both growth outcomes, the invasive bacteria (Group II) showed positive associations with MPO for both ages (Fig. 4). Additionally, ETEC (Group III) infection was associated with lower concentrations of AAT at Age 1 and higher concentrations of MPO in older Age 2 children. At Age 1, AAT concentration was positively associated with LMZ, while MPO was negatively associated with LMZ (Fig. 4). In the older age period, MPO concentration was positively associated with AGP, while NEO was negatively associated with AGP.

25Regarding associations between biomarkers and the two growth outcomes, ΔLAZ and ΔWAZ, the systems model revealed direct negative relationships between MPO and both ΔLAZ and ΔWAZ during Age 1 and negative associations between AAT and ΔWAZ in both age groups (Fig. 4) (i.e., the higher the concentration of MPO or AAT, the more restricted the growth over the respective period). There was no statistical support for an association between LMZ and either ΔLAZ or ΔWAZ in the system model (or for lactulose excretion, data not shown). Higher AGP concentrations were associated with decreased ΔLAZ at both ages and with ΔWAZ during Age 1 (Fig. 4). Because enteropathogens, particularly those associated with enteroinvasion or mucosal disruption (Group II), were directly related to increased MPO, they were thereby indirectly related to reduced growth. Giardia was associated directly with reduced growth, but not with the fecal biomarkers.

26The effects of higher or lower biomarker concentrations relative to their observed means on ΔLAZ and ΔWAZ over each age period are shown in Fig. 5. None of the biomarkers had large mean effects on the average ΔLAZ or ΔWAZ. Of the effects on ΔLAZ, changing the log(AGP) concentration by ± 1SD produced the greatest difference (± 0·05 mean ΔLAZ), but little effect on the mean ΔWAZ (± 0·02 ΔWAZ) relative to the observed concentration of log(AGP). Changing the MPO concentration by ± 1SD had the second largest impact on ΔLAZ (+ 0·04 or − 0·03 for a decrease or increase in log(MPO) concentration respectively). The effect of increasing MPO concentration on ΔWAZ was marginally larger (− 0·05 ± 0·03), after which AAT had the second biggest impact on ΔWAZ (+ 0·04 when decreased and − 0·06 when log(AAT) was increased by 1SD).

27The systems model presented here is the first effort to explicitly combine enteropathogen exposure and different EE biomarkers (i.e., for gut permeability, gut inflammation, and systemic inflammation) into a single system and examine prospectively the pathways through which they relate to growth, thus providing an important proof of principal for a poorly defined condition. The temporal nature of the data allows for the exploration of causality and for the natural history of the biologic processes being measured to be interpreted with rigor. Our model is informed by a substantial quantity of structured longitudinal data on enteropathogens in non-diarrheal stools, illness history, intestinal inflammation and permeability, systemic inflammation, and growth across seven populations. The number of repeated measures over time and measures within the same individual is a particular strength of these data. Enteropathogen presence changes the concentration of contemporary biomarkers of gut immunity (NEO), inflammation (MPO), and permeability (AAT and LMZ). We find additional evidence that both enteropathogens and gut inflammation relate to systemic inflammation.

28Our results yield several important findings. First, children living in these eight countries with differing epidemiologic settings have consistently and strikingly high concentrations of MPO (nearly 5–10 times the those seen in the USA) and nearly twice the concentration of AAT. They also have notable elevations in AGP and intestinal permeability (L:M) that are two to three times higher than values in healthy populations. Additionally, there was a high burden of enteropathogens detected even in the absence of overt diarrhea episodes. But most importantly, related MAL-ED analyses have demonstrated that, in the diverse epidemiologic settings of the study, growth velocities were low leading to an increased prevalence of stunting by two years. The evidence we present here suggests that EE as measured by these markers contributes to, but does not appear to be the predominant driver of growth faltering.

29As hypothesized, pathogen presence in non-diarrheal stool samples was associated with higher fecal MPO; this finding was also noted for AAT, although the AAT findings were evident only in younger children (aged 4˗11 months). When fecal biomarkers were evaluated for associations with individual pathogens, stronger signals of inflammation were noted in response to the rarely detected Shigella and Yersinia and the ubiquitous Campylobacter. The differential importance of pathogens in inducing EE is an area that requires additional attention, especially given the availability of newer technologies that allow for broad coverage and quantification of enteropathogens, and other strategies for determining which microbial taxa are inducing mucosal response in the gut. Such studies will enhance our understanding of the possible impact of etiology-specific interventions on EE.

30Each of the three fecal biomarkers in this analysis have previously been associated with an elevated risk of stunting in the first year of life. Similar results have been found in other studies. Here, we also find support for direct effects of gut inflammation (i.e., as measured by the three fecal biomarkers) on ΔLAZ or ΔWAZ. The evidence for an effect of systemic inflammation (i.e., as measured by AGP) is more compelling than that of the EE biomarkers. We found no evidence for an association between augmented intestinal permeability (i.e., as measured by LMZ) and future growth. Despite the popular use of L:M as a test of EE in cross sectional studies, its predictive power for future growth is less clear. While many studies have shown the L:M ratio to be associated with the contemporaneous LAZ, others have found (as we did) that it was not predictive of change in LAZ of children. Given that AAT was associated with both gut permeability and subsequently to changes in growth (especially WAZ, both Age 1 and 2), it is possible that AAT more reliably reflects changes (or more severe changes) in permeability than the L:M test. The polar surface area of AAT greatly exceeds that of lactulose, and the size of the permeability defect may be important to delineate in greater detail using alternate probes. It is worth noting that other MAL-ED data (i.e., non-growth data) indicate the L:M test is predictive of impaired efficacy of oral polio vaccine. As such, the L:M test may capture other domains of EE not examined in these analyses.

31Our results suggest that reductions in the exposure to pathogens, in particular to invasive bacteria that increase MPO, could reduce systemic inflammation (AGP). Recent evidence from another cohort study similarly highlighted the importance of systemic inflammation. AGP interacts with bacterial lipopolysaccharide (LPS), which is an indicator of bacterial translocation, and LPS bound to AGP is more rapidly cleared from the body than unbound LPS. Consequently, it was surprising to see that neither LMZ nor AAT were associated with AGP, as permeability leading to bacterial translocation has been a principal hypothesized component of EE. Other markers of bacterial translocation are needed to confirm these findings.

32Although our model indicates that specific enteropathogens alter gut inflammation and permeability, the effect sizes of the pathways from biomarkers to growth were smaller than anticipated. Based on available evidence, the contribution of EE to growth deficits as captured in these biomarkers is small relative to the accrued growth deficits in these populations. The use of additional biomarkers or combinations of biomarkers and growth phenotypes or inclusion of metabolic markers may help to identify stronger support for links between the interactions of enteropathogen pressure and undernutrition on growth failure. Additionally, on-going work in this study population to understand links between the microbiome, host metabolism, and growth will expand our understanding of how microbial populations affect the nutritional status in these populations.

33The analytical framework described here is the first attempt to explicitly examine causal pathways of EE. Our results demonstrate that exposure to enteropathogens results in abnormal gut permeability, inflammation, systemic immune activation, and growth failure, but suggest that additional work incorporating other critical features of host metabolic status and the microbiome are needed to explain the gap between the insults attributable to EE and the observed cumulative acquired deficits of growth in children in these populations.

34Margaret N. Kosek and Benjamin J.J. McCormick devised the model in discussion with Richard L. Guerrant, Laura E. Caulfield, Tahmeed Ahmed, Zulfiquar Bhutta, Gagandeep Kang, Aldo Lima, Eric Houpt, and James Platts-Mills. Benjamin J.J. McCormick ran the analyses. Margaret N. Kosek, Benjamin J.J. McCormick, Gwenyth Lee, and Jessica C. Seidman participated in the interpretation of the results and drafted the manuscript. The MAL-ED Investigators participated in the design, conduct, and analysis of the MAL-ED study and its results.

35We declare that we have no conflicts of interest.