Gates Open Res Gates Open Research 2572-4754 F1000 Research Limited London, UK 10.12688/gatesopenres.13671.1 Research Article Articles Association between intestinal bacterial carriage, biomarkers of environmental enteric dysfunction, and stunting in rural Malawian children

[version 1; peer review: 1 approved with reservations, 1 not approved]

 Chaima David Conceptualization Data Curation Formal Analysis Investigation Methodology Writing – Original Draft Preparation Writing – Review & Editing https://orcid.org/0000-0002-5480-7542 a 1 2 Hart John Data Curation Methodology Project Administration Supervision Writing – Review & Editing https://orcid.org/0000-0002-9637-7974 2 Pickering Harry Data Curation Formal Analysis Writing – Original Draft Preparation Writing – Review & Editing 2 Burr Sarah Conceptualization Funding Acquisition Methodology Project Administration Resources Supervision Writing – Original Draft Preparation Writing – Review & Editing 1 2 Maleta Kenneth Funding Acquisition Supervision Writing – Review & Editing 3 Kalua Khumbo Funding Acquisition Methodology Project Administration Supervision Writing – Review & Editing 4 5 Bailey Robin Funding Acquisition Methodology Project Administration Resources Supervision Writing – Review & Editing 2 Holland Martin Conceptualization Funding Acquisition Methodology Supervision Writing – Original Draft Preparation Writing – Review & Editing https://orcid.org/0000-0002-5828-6318 2 Department of Pathology, Kamuzu University of Health Sciences, Blantyre, Malawi Clinical Research Department, London School of Hygiene & Tropical Medicine, London, WC1E 7HT, UK School of Public Health and Family Medicine, Kamuzu University of Health Sciences, Blantyre, Malawi Blantyre Institute of Community Outreach, Blantyre, Malawi Department of Opthalmology, Kamuzu University of Health Sciences, Blantyre, Malawi dchaima@kuhes.ac.mw

No competing interests were disclosed.

 18 7 2022 2022 6 78 5 7 2022 Copyright: © 2022 Chaima D et al. 2022 This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Background. Available data from murine studies suggest that intestinal bacteria may have a role in modulating growth phenotypes in the host. We investigated the prevalence of four gut bacteria known in murine models to impair growth ( Bifidobacterium longum, Faecalibacterium prausnitzii, Dorea formicigenerans, and Akkermansia muciniphila), the level of fecal biomarkers of environmental enteric dysfunction (EED) and stunting in rural Malawian children.

Methods. DNA and protein were extracted from fecal samples of rural Malawian children (aged 1-59 months) at a baseline cross-sectional survey in the Mangochi district of Malawi conducted within the framework of the Macrolides Oraux pour Réduire les Décès avec un Oeil sur la Résistance (MORDOR) trial. Intestinal carriage of bacteria was measured by PCR. Neopterin (NEO), myeloperoxidase (MPO), and alpha-1 antitrypsin (AAT), biomarkers of EED, were measured by an enzyme-linked immunosorbent assay (ELISA) test. Height-for-age Z (HAZ) score <-2 defined stunting. Tests of proportions and regression models were used to explore the relationship between bacterial carriage, EED, and stunting.

Results. Fecal samples from 613 children were available for laboratory analyses. F. prausnitzii and D. formicigenerans were prevalent in over 70% of children while B. longum was the least prevalent. B. longum carriage in younger children was associated with elevated EED biomarkers. Two thirds of children had elevated NEO, 33% had elevated MPO, and 16% had elevated AAT. Stunting was found in 38%. No significant associations were found between EED biomarkers or intestinal bacteria carriage and stunting.

Conclusion. Intestinal carriage of these four bacteria was not associated with stunting in Malawian children. Carriage was also not associated with EED, nor EED biomarker levels associated with stunting. Further factors acting in concert are necessary to impact EED, perturb growth, and alter gut bacterial carriage.

 Biomarker environmental enteric dysfunction intestinal bacterial carriage qPCR stunting Gates Foundation OPP1032340 OPP1066930 This work was supported by the Gates Foundation [OPP1066930 and OPP1032340]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Introduction

According to emerging evidence from murine studies, gut bacteria have a role in modulating growth phenotypes in the host. A study that looked into whether a growth phenotype could be passed down through a combination of diet and gut bacteria reported links between a weight loss phenotype and a combination of a Malawian diet and a kwashiorkor microbiota 1 . Another study, which looked at the interaction between gut microbiota development and undernutrition, reported links between gut microbiota and growth characteristics in a murine model 2 . Bifidobacterium longum, Faecalibacterium prausnitzii, and Dorea formicigenerans were found to be related with weight gain in mice. Another study utilizing a mouse model reported an inverse relationship between intestinal carriage of Akkermansia muciniphila and body weight 3 . Further studies in humans have reported relationships between the gut microbiota and growth characteristics. In Bangladeshi, intestinal bacterial diversity was persistently reduced in malnourished children compared to healthy children. Also, relative abundance of Bifidobacterium was reported to be lower in malnourished children compared to healthy children 4 . Furthermore, A. muciniphila abundance was shown to be lower in overweight or obese Swedish pre-school children 5 .

The mechanisms by which the gut bacteria can mediate growth are not well known. One proposed mechanism is the preservation of intestinal integrity, which improves nutrient absorption after digestion. This hypothetical mechanism is better explained by studies of environmental enteric dysfunction (EED), a subclinical condition of the gut defined by structural alterations in the intestinal epithelium and underlying chronic inflammatory responses that may result in loss of gut integrity and influence growth 6– 8 . In EED, alterations in gut microbiota assembly and function cause structural changes in the intestinal epithelium, allowing bacteria or microbial antigens from the lumen to enter the systemic circulation and cause persistent immunological activation. This then impacts linear growth, either by suppressing growth hormones or redirecting nutrients away from growth and toward more immunological metabolism 9, 10 . A multi-site, longitudinal study that examined the relationship between the fecal biomarkers of EED (myeloperoxidase (MPO), neopterin (NEO), and alpha-1 anti-trypsin (AAT) and linear growth in infants living in low-income settings of South Asia, Sub-Saharan Africa and Latin America showed that high levels of these biomarkers predicted a decline in infant length-for-age z (LAZ) scores 11 . Similarly, in a case-control study assessing a relationship between MPO or AAT and growth impairment in rural Brazilian children (aged between 6-26 months), higher fecal levels of both biomarkers were associated with growth faltering 12 .

The present study investigated the prevalence of Akkermensia muciniphila, Bifidobacterium longum, Dorea formicigenerans, and Faecalibacterium prausnitzii in rural Malawian children. We investigated the link between intestinal carriage of these bacteria, fecal biomarkers of EED, and stunting in rural Malawian children to understand if the growth mediating effects of these bacteria observed in mouse models are also observed in the human population.

 Methods Ethical considerations

The MORDOR trial was conducted in accordance with the Declaration of Helsinki. It was approved by the London School of Hygiene and Tropical Medicine Ethics Committee (UK) (reference number 6500) and the College of Medicine Research Ethics Committee (Malawi) (P.02/14/1521). Information and consent forms were translated into local languages (Yao and Chichewa) prior to their approval by the local ethics committee. Consent was first obtained at the community level through discussions with the village chief and community elders who then verbally indicated whether the community was willing to participate. Written, informed consent (by thumbprint or signature) was then obtained from the parent or legal guardian of each child for inclusion before they participated in the study. All parents and guardians were informed of their freedom to withdraw their child from the study at any time without a requirement to give a reason(s).

 Study design

This study made use of fecal samples, and demographic and anthropometry data gathered during a baseline assessment of the prevalence of carriage of macrolide resistant enteropathogens undertaken as part of a multi-country randomized controlled trial named “MORDOR” (NCT02047981). The trial study design and protocol have been previously published 13, 14 . Briefly, a total of 1,090 children residing in 30 clusters within the Mangochi district were enrolled into the baseline survey. All enrolled children were aged 1 to 59 months and weighed at least 3.8kg. Sample and data collection took place between May and July 2015. Details of fecal sample collection have been described in previous studies 15, 16 . This study is based off the first named author’s PhD thesis 17 .

 Laboratory methods

Total, genomic DNA was isolated from 250mg fecal sample using the commercially available PowerSoil DNA Isolation Kit (MO BIO Laboratories Inc, Carlsbad, CA, now a part of Qiagen, Germany). As a quality control measure, DNA yield was quantified in 10% of samples using the Qubit dsDNA HS assay kit (Invitrogen, CA, USA). Endpoint qPCR was performed to detect the presence of the following bacteria: Akkermensia muciniphila, Bifidobacterium longum Dorea formicigenerans, and Faecalibacterium prausnitzii. Each PCR reaction used a 1:10 diluted DNA sample to negate any unwanted effects of high DNA yield. The qPCR was performed using a Rotorgene-Q instrument (Qiagen, Hilden, Germany) where samples were run in a 72-well rotor format. The qPCR assays were designed and performed using the microbial DNA qPCR assay platform (Qiagen). The total volume of a single PCR reaction was 12.5µl and each reaction contained 6.25µl of microbial qPCR master mix (Qiagen), a commercial microbial DNA qPCR assay containing primers (10µM), and a 5’-hydrolysis probe (5 µM) targeting the bacterial 16S rRNA gene and 3µl template DNA in an aqueous solution. A no template control and a microbial DNA positive control (Qiagen) were included on each run. Thermal cycling conditions for the assay were 95°C for 10 minutes followed by 40 cycles of 95°C for 15 seconds and 60°C for 2 minutes. The endpoint qPCR results were classified as positive or negative using a method by Pickering et al., 18

To quantify fecal biomarkers of EED, commercially available enzyme-linked immunosorbent assay (ELISA) kits for MPO (Immundiagnostik, Germany), AAT (Immundiagnostik), and NEO (Genway, CA, USA) were used for the quantification of each biomarker 16 .

 Height/length measurements

The SECA Leicester height measure (Chasmors Ltd., UK) was used to measure height for children who were ≥2 years of age and able to stand whereas the SECA 417 infantometer (Chasmors Ltd.) was used to measure recumbent length for children who were <2 years of age or unable to stand unaided. Both height and length were measured to within 0.1 cm. All measurements were taken in triplicate by field nurses.

 Statistical analysis

To describe the prevalence of carriage of defined bacteria, the proportions of fecal samples with a positive qPCR result for each of the defined bacteria were calculated. The biomarkers of EED were assessed individually or were combined to form a composite EED score as described by Kosek et al., 11 with minor modifications as detailed elsewhere 16 . The distribution of the biomarkers of EED was determined using the median (IQR) concentration of the fecal biomarkers. Published cut-off values for each of the three biomarkers 11, 19 were used to determine proportions of fecal samples with elevated biomarkers. The World Health Organization (WHO) 2006 Child Growth Standards were used for the calculation of height-for-age Z (HAZ) scores 20 . Participants with out-of-range HAZ scores, (>6 or <-6) were not included in the subsequent analyses as such scores are biologically implausible. HAZ scores were used in analyses as continuous variables or categorized into not stunted or stunted, defined as having a Z score of <-2. Fisher’s exact or Pearson’s chi-square tests were used to explore associations between bacterial carriage and sex, age, biomarkers of EED, and stunting. Age and sex adjusted regression models were also used to explore the relationship between bacterial carriage, biomarkers of EED, and stunting.

 Results Baseline characteristics of participants

The baseline survey of prevalence of carriage of macrolide resistant enteropathogens enrolled 1,090 children, of which only 709 (65%) returned fecal samples. Of the 709 fecal samples, 613 fecal samples had sufficient volume for the analyses of intestinal bacterial carriage while 523 of the 613 fecal samples had sufficient remaining volume for the analyses of fecal biomarkers of EED. Characteristics of all children who were included in the final analyses of this study and those not included are shown in Table 1. Participant sex was comparable between all participants who were not included the analyses and those included in the bacterial carriage and EED biomarker analyses whilst differences were seen for age and height between children included in the analyses of bacterial carriage and EED biomarkers and all participants not included in the analyses.

Characteristics of participants who were included in the analyses as compared to all participants not included in the analyses.
Variable All participants excluded in the analyses Participants included in bacterial carriage analyses P value Participants included in EED + biomarkers analyses P value
Number of participants 469 613 523
Male sex N (%) 222 (47.33) 299 (48.78) 0.664 264 (50.5) 0.24¥
¥
Height, cm 81.90 (12.79) 84.5 (12.1) 0.007 * 85.0 (11.6) 0.005 *
Age, months 27.74 (16.21) 31.1 (16.4) 0.0008 * 32.0 (16.0) <0.001 *

Data are presented as mean (SD) unless otherwise stated.

*Denotes P values obtained from Student’s t-test while ¥denotes P values obtained from sample proportion test.

+ Environmental enteric dysfunction (EED).

 Prevalence of bacterial carriage and biomarker distribution

F. prausnitzii was the most prevalent bacterium, appearing in 98% of the samples ( Table 2). D. formicigenerans was present in 79% of the samples, A. muciniphila was found in 43% of the samples, and B. longum was present in just over one third of the fecal samples. Fecal samples were stratified by sex and age to compare bacterial carriage. There was no difference in the prevalence of F. prausnitzii, B. longum and A. muciniphila between male and female children, however, a higher proportion of fecal samples from male children were positive for D. formicigenerans compared to the female children ( Table 2). B. longum was most prevalent in fecal samples from the youngest age group while D. formicigenerans, F. prausnitzii, and A. muciniphila were more prevalent in the older children ( Table 2).

Prevalence of bacterial carriage across all samples, by age, and by sex.
Variable Carriage in all samples (N=613) Carriage by age in months Carriage by sex
1-12 (n=104) 13-24 (n=134) 25-36 (n=120) 37-48 (n=144) 49-60 (n=111) P value Female (n=314) Male (n=299) P value
B. longum positive % (n) 29 (179) 82 (85) 46 (61) 17 (17) 8 (12) 4 (4) <0.001 a 31 (96) 28 (83) 0.5 a
A. muciniphila positive % (n) 43 (264) 26 (27) 48 (64) 53 (63) 42 (60) 45 (50) 0.001 a 42 (132) 44 (132) 0.65 a
D. formicigenerans positive % (n) 79 (483) 42 (44) 81 (109) 88 (105) 85 (122) 93 (103) <0.001 a 75 (235) 83 (248) 0.019 a
F. prausnitzii positive % (n) 98 (603) 95 (99) 99 (132) 100 (120) 99 (143) 98 (109) 0.06b 98 (307) 99 (296) 0.34 b

aDenotes P values obtained from Pearson’s chi-square test,

bDenotes P values obtained from Fisher’s Exact test

Of the 523 fecal samples that were available for the measurement of fecal biomarkers of EED, 488 had sufficient sample volume for MPO ELISA, 421 for NEO, and 495 for AAT. Four hundred and twenty one samples had results for all the three markers. A larger proportion of children [77% (324/421)] had elevated NEO levels relative to published reference values (NEO <=70 nmol/L, AAT <=0.27 mg/g and MPO <=2000 ng/ml) 11, 19 . Approximately, one third of the children had elevated MPO levels and levels of fecal AAT were elevated in 16% of the children ( Table 3). The comparison of biomarker concentrations by sex did not show any differences in MPO and AAT concentrations; however, NEO concentration was higher in female children compared to male children ( Table 3). There were significant differences in biomarker distribution among the five age groups. Median fecal concentration for all the three biomarkers was higher in younger children compared to older children ( Table 3).

Biomarker distribution across all samples, age, and sex.
Variable *Proportion with elevated biomarkers in all samples, % (n) Age in months Sex
1-12 13-24 25-36 37-48 49-60 P value Female Male P value
MPO + (ng/ml) 22 (106/488) 4255 (13165) 971 (1778) 580 (1077) 402 (757) 290 (357) <0.001 a 610 (1383) 518 (1629) 0.91 b
N=68 N=104 N=86 N=114 N=86 N=242 N=246
NEO + (nmol/L) 76 (324/425) 2101 (3103) 402 (1379) 133 (295) 134 (348) 63 (153) <0.001 a 0.06 (0.14) 0.08 (0.15) 0.21 b
N=55 N=99 N=85 N=108 N=81 N=244 N=251
AAT + (mg/g) 16 (81/495) 0.25 (0.64) 0.1 (0.2) 0.06 (0.11) 0.06 0.10) 0.03 (0.05) <0.001 a 140 (505) 201 (1044) 0.02 b
N=66 N=105 N=90 N=117 N=85 N=210 N=211

* Normal values for the biomarkers were based on values reported in the literature (NEO <=70 nmol/L, AAT <=0.27 mg/g and MPO <=2000 ng/ml) 11, 19, 21

Biomarker concentration are presented as median (IQR)

aDenotes P value obtained from Kruskal-walis test

bDenotes P values obtained from Wilcoxon rank sum test

+ Neopterin (NEO), myeloperoxidase (MPO), and alpha-1 antitrypsin (AAT).

Association between bacterial intestinal carriage and biomarkers of EED

Logistic regression analysis examining the relationship between bacterial carriage and biomarker concentration did not find any association between fecal carriage of A. muciniphila or F. prausnitzii and any of the three individual biomarkers of EED. However, fecal carriage of B. longum was associated with increased odds of having elevated biomarker concentrations (vs normal concentrations), whilst fecal samples that were positive for D. formicigenerans were 70% less likely to have elevated MPO concentration ( Table 4).

Association between bacterial carriage and individual biomarkers of environmental enteric dysfunction (EED).
Variable MPO + AAT + NEO +
Normal Elevated OR (95%CI) P value a Normal Elevated OR (95%CI) P value a Normal Elevated OR (95%CI) P value a
B. longum, n (%) 60/345 (17) 59/104 (57) 2.6 (1.4,4.6) 0.002 72/380 (19) 47/74 (64) 2.6 (1.4,5.1) 0.004 7/122 (6) 98/295 (33) 4.7 (2.1,12.3) 0.0004
A. muciniphila, n (%) 159/345 (46) 41/104 (39) 0.9 (0.5,1.5) 0.619 176/380 (46) 27/74 (37) 0.8 (0.4,1.3) 0.354 61/122 (50) 133/295 (45) 0.8 (0.5,1.3) 0.407
D. formicigenerans, n (%) 307/345 (89) 66/104 (64) 0.3 (0.2,0.6) 0.001 328/380 (86) 48/74 (65) 0.7 (0.4,1.4) 0.298 112/122 (92) 244/295 (83) 0.7 (0.3,1.5) 0.427
F. prausinitzii, n (%) 340/345 (99) 102/104 (98) 0.9 (0.1,7.6) 0.89 374/380 (98) 73/74 (99) 2.5 (0.3,53.6) 0.454 121/122 (99) 291/295 (99) 0.5 (0.03,4) 0.585

CI= confidence interval, OR=odds ratio.

n (%) = Number of fecal samples (proportion) positive.

aOdds ratio (95% CI) and P values were obtained with logistic regression after adjusting for age and sex. Using the Bonferroni correction for multiple testing, associations with a P value less than 0.0125 were considered statistically significant.

+ Neopterin (NEO), myeloperoxidase (MPO), and alpha-1 antitrypsin (AAT).

B. longum carriage in children younger than 24 months was associated with fecal concentrations of MPO and NEO but not in the older children ( Table 5). Elevated MPO concentration was found in fecal samples of children aged between 0-12 months who were positive for B. longum while children aged between 12-24 months who were positive for B. longum were more likely to have elevated NEO concentration. D. formicigenerans carriage was not associated with any individual biomarker in any of the age groups ( Table 6).

Association between <italic toggle="yes">B. longum</italic> carriage and biomarkers of environmental enteric dysfunction (EED) stratified by age.
MPO + AAT + NEO + MPO + AAT + NEO +
Age (months) Normal Elevated Normal Elevated Normal Elevated OR (95% CI) P value OR (95% CI) P value OR (95% CI) P value
0 – 12 *n (%) ¥ n=23 n=44 n=34 n=31 n=4 n=50 7.4 (1.9,34.5) 0.006
13 (56) 40 (91) 23 (68) 27 (87) 3 (75) 40 (80) 3.02 (0.81,13.1) 0.112 1.4 (0.02,14) 0.804
12 – 24 n (%) n=72 n=30 n=79 n=24 n=17 n=80 1.7 (0.7,4.3) 0.268
29 (40) 16 (53) 30 (38) 16 (67) 1 (6) 41 (51) 3.1 (1.2,8.6) 0.023 17.5 (3.26,324.8) 0.007
24 – 36 n (%) n=73 n=10 n=79 n=8 n=28 n=53 2.1 (0.3,11.8) 0.427
7 (10) 2 (20) 6 (8) 3 (38) 0 7 (13) 11.9 (1.6,121) 0.019 2.8 (0.03,14.2) 0.991
36 – 60 n (%) n=177 n=20 n=188 n=11 n=73 n=112 0.68 (0.04,11.8) 0.727
11 (6) 1 (5) 13 (7) 1 (9) 3 (4) 10 (9) 1.28 (0.06,8.5) 0.822 2.4 (0.7,11.3) 0.204

*n (%) = Number (proportion) positive for B. longum

¥Number with normal or elevated biomarker concentration

CI= confidence interval, OR=odds ratio. aOdds ratio (95% CI) and P values were obtained with logistic regression after adjusting for and sex. Using the Bonferroni correction for multiple testing, associations with a P value less than 0.0125 (highlighted in bold) were considered statistically significant.

+ Neopterin (NEO), myeloperoxidase (MPO), and alpha-1 antitrypsin (AAT).

Association between <italic toggle="yes">D. formicigenerans</italic> carriage and biomarkers of environmental enteric dysfunction (EED) stratified by age.
MPO + AAT + NEO + MPO + AAT + NEO +
Age (months) Normal Elevated Normal Elevated Normal Elevated OR (95% CI) P value OR (95% CI) P value OR (95% CI) P value
0 – 12 *n (%) ¥ n=23 n=44 n=34 n=31 n=4 n=50 0.27 (0.09,0.77) 0.018 0.34 (0.12,0.92) 0.037 0.38 (0.02,3.23) 0.42
16 (70) 17 (39) 21 (62) 11 (35) 3 (75) 27 (54)
12 – 24 n (%) n=72 n=30 n=79 n=24 n=17 n=80 0.91 (0.28,3.28) 0.88 0.98 (0.29,3.97) 0.97 0.32 (0.02,1.86) 0.30
61 (85) 25 (83) 66 (84) 20 (83) 16 (94) 67 (84)
24 – 36 n (%) n=73 n=10 n=79 n=8 n=28 n=53 1.04 (0.15,21.03) 0.97 0.86 (0.13,17.4) 0.90 1.08 (0.20,4.83) 0.92
66 (90) 9 (90) 71 (90) 7 (88) 25 (89) 48 (91)
36 – 60 n (%) n=177 n=20 n=188 n=11 n=73 n=112 0.23 (0.07,0.84) 0.019 1.12 (0.19,21.38) 0.92 0.85 (0.25,2.54) 0.77
164 (93) 15 (75) 170 (90) 10 (91) 68 (93) 102 (91)

*n (%) = Number (proportion) positive for D. formicigenerans

¥Number with normal or elevated biomarker concentration

CI= confidence interval, OR=odds ratio. aOdds ratio (95% CI) and P values were obtained with logistic regression after adjusting for sex. Using the Bonferroni correction for multiple testing, associations with a P value less than 0.0125 were considered statistically significant.

+ Neopterin (NEO), myeloperoxidase (MPO), and alpha-1 antitrypsin (AAT).

Bacterial carriage and composite EED score were analyzed by age and sex adjusted logistic regression. Composite EED scores were categorized into low and high scores centered around the mean score (1.73). B. longum, but not the other three bacteria species, was associated with the composite EED score. Fecal samples that were positive for B. longum were more likely to have a high composite EED score ( Table 7).

Association between bacterial carriage and composite environmental enteric dysfunction (EED) score.
Variable Composite EED score
Low (n=175) High (n=214) OR (95%CI) P value a
B. longum, n (%) 9 (5) 90 (42) 7.0 (3.3,16.3) <0.001
A. muciniphila, n (%) 83 (47) 96 (45) 1.0 (0.6,1.6) 0.98
D. formicigenerans, n (%) 162 (93) 169 (79) 0.7 (0.3,1.5) 0.06
F. prausinitzii, n (%) 172 (98) 212 (99) 1.4 (0.2,12.6) 0.73

CI= confidence interval, OR=odds ratio.

n (%) = Number of fecal samples (proportion) positive.

a Odds ratio (95% CI) and P values were obtained with logistic regression after adjusting for age and sex. Using the Bonferroni correction for multiple testing, associations with a P value less than 0.0125 were considered statistically significant.

 Bacterial carriage and biomarkers of EED were not associated with stunting

Height measurements and demographic (age and sex) data were available for all participants whose fecal samples were assayed for bacterial carriage (n=613) and 488 of those whose samples were assayed for individual biomarkers of EED. For the complete data set, mean (SD) HAZ score was -1.6 (1.5) with 38% (229/607) of the children stunted. There was no relationship between bacterial carriage and HAZ or stunting. Similarly, none of the individual EED biomarkers or the composite EED score were associated with HAZ or stunting ( Table 8).

Association between bacterial carriage, biomarkers of environmental enteric dysfunction (EED) and height-for-age Z (HAZ) or stunting.
Variable HAZ Stunting
Coefficient (95%CI) 1 P value 1 OR (95% CI) 2 P value 2
B. longum -0.26 (-0.64,0.12) 0.175 0.99 (0.62,1.57) 0.97
A. muciniphila -0.24 (-0.53,0.03) 0.083 1.21 (0.86,1.69) 0.27
D. formicigenerans -0.08 (-0.44,0.28) 0.656 1.18 (0.76,1.85) 0.47
F. prausnitzii -0.01 (-1.1, 1.08) 0.985 1.13 (0.30, 5.38) 0.86
MPO + -0.34 (-0.68,-0.01) 0.046 1.06 (0.65,1.77) 0.81
AAT + -0.22 (-0.60,0.16) 0.26 0.73 (0.42,1.26) 0.26
NEO + -0.19 (-0.51,0.13) 0.25 1.01 (0.64,1.58) 0.97
Composite EED score -0.07 (-0.38,0.24) 0.67 0.89 (0.56,1.40) 0.61

CI= confidence interval, OR=odds ratio

1Coefficients (95% CI) and P values obtained from a linear regression analysis after adjusting for age and sex.

2OR (95% CI) and P values obtained from a logistic regression analysis after adjusting for age and sex

+ Neopterin (NEO), myeloperoxidase (MPO), and alpha-1 antitrypsin (AAT).

* Of all the 613 participants for qPCR, six had out of range HAZ scores while of the 488 participants whose samples were assayed for individual biomarkers of EED, 5 had out of range HAZ scores. These were excluded in this analysis.

Discussion

The current study examined the prevalence of four bacteria previously linked with growth in mouse models, as well as the severity of intestinal inflammation and permeability in rural Malawian children. It also investigated the links between bacterial carriage, increased biomarker concentrations, and stunting. F. prausnitzii was found in practically all children while B. longum was more common in younger children and A. muciniphila and D. formicigenerans were more common in older children. Fecal concentrations of MPO, NEO, and AAT reduced with age.

The high prevalence of carriage of F. prausnitzii reported in the current study is consistent with earlier findings demonstrating that this bacterium is one of the most prevalent and prolific intestinal bacteria in the human gut 22 . This bacterium is present at low levels in infants throughout the first six months of life, increasing slowly in abundance from seven months of life and doubling during the second year of life 23 . Furthermore, F. prausnitzii was identified as an important age discriminatory bacterium and proved its low abundance in early life in studies that employed a machine-learning-based approach to define a healthy gut microbiota of Bangladeshi or Malawian children 2, 4 . The increase in D. formicigenerans carriage with age in our study population is comparable with earlier studies that used molecular approaches to demonstrate intestinal carriage 2, 4, 23 . The higher prevalence of B. longum in younger children is compatible with breast feeding as a source of B. longum 24– 26 , which is present in breast-milk 27, 28 .

Our data show that younger children had elevated levels of fecal AAT, MPO, and NEO, which is comparable to published findings from resource limited settings 19, 21 . The reasons for the negative association between age and fecal levels of AAT, MPO, and NEO are not unknown; however, some authors have proposed that breastfeeding may be a factor because AAT and NEO are present in minute concentrations in breast milk 19, 29, 30 whereas MPO is thought to be produced from the activation of infant mucosal neutrophils by alpha-lactalbumin and tryptophan as an immune response. Both alpha-lactalbumin and tryptophan are constituents of breastmilk 31, 32 . Nonetheless, there is insufficient evidence to determine whether the association between these biomarkers and age is physiological or pathological.

Our study also showed a positive relationship between fecal carriage of B. longum and raised biomarkers of EED, which is uncommon given that this bacterium has been demonstrated to have several benefits in the humans. This positive correlation could be attributed to confounding by other factors such as breastfeeding and enteric infections, which were not evaluated. Breastfeeding and enteric infections have been associated with fecal levels of AAT, NEO, and MPO, and MPO and AAT respectively in Bangladeshi infants 19 ; therefore more research is needed to investigate the effect of breastmilk, its constituents, and enteric infections on biomarker concentration, gut bacteria, and implications for gut health.

We found no relationship between bacterial carriage or fecal biomarkers of EED and stunting. Some of the important aspects of EED are bacterial composition, intestinal inflammation, and permeability and prior studies have reported or predicted links between high levels of MPO, AAT, or NEO and growth impairment 11, 12, 33 . Inconsistent findings between these biomarkers and growth outcomes have been previously reported by George et al., 34 who did not find a link between raised baseline fecal levels of AAT, MPO, or NEO and stunting, or between baseline composite EED score and stunting. Furthermore, Campbell and colleagues 35 did not find any relationship between composite EED score and HAZ score at 18 months of follow-up. The majority of studies that have demonstrated links between the biomarkers of EED and growth were longitudinal, used bigger sample sizes, and were conducted in children under the age of two 11, 19, 33 . There is currently no evidence demonstrating a link between bacterial carriage, assessed by qPCR, and stunting in humans. Given that there are many other factors that influence child growth, the relationship between bacterial carriage, intestinal inflammation or permeability, and stunting may be obscured by other factors such as, food insecurity, enteric infections, diarrheal diseases, social-economic status, maternal education, or other growth determinants.

One of the limitations for the current study was the availability of fewer fecal samples for analysis, which limited the generalizability of results. Furthermore, the lack of data for some demographic factors reduced the number of potential confounders that might have been used in the analysis. Data on socio-economic status, breastfeeding, water quality, sanitation, and hygiene indicators may have enriched the analysis.

 Conclusions

In conclusion, our findings demonstrate a significant prevalence of F. prausnitzii and D. formicigenerans in fecal samples from older rural Malawian children, which is consistent with previous studies. The current study also found elevated levels of AAT, MPO, and NEO in the age group matching when children are breastfed. Future studies should determine whether the association between age and the fecal levels of these biomarkers in children is normal or pathological. We found an association between B. longum carriage and biomarkers of EED and future studies should corroborate this by sequencing B. longum strains carried by children in this population to determine if there are any genetic differences with the strains that have been identified as beneficial in other settings. The current study found no relationship between bacterial carriage, intestinal inflammation, intestinal permeability and stunting, implying that the selected bacteria do not contribute to EED or growth in this Malawian population. More studies are needed to determine other factors that influence growth and their association with EED. Such studies could look at risk factors for EED by identifying enteropathogens in food, water, and fecal samples and examining how the enteropathogens affect the gut microbiota, gut function and subsequent growth. Furthermore, such studies might look into how exclusive breastfeeding and complementary feeding alter gut microbiota composition, gut function, and ultimately linear growth.

 Data availability Underlying data

Figshare: Association between intestinal bacterial carriage, biomarkers of environmental enteric dysfunction and stunting in rural Malawian children.

https://doi.org/10.6084/m9.figshare.20178065.v1 36

This project contains the following underlying data:

Intestinal bacterial carriage_EED_stunting_Chaima et al associated dataset.xlsx. (description of data in file).

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).

 Acknowledgements

We would like to acknowledge the support of the field staff at the Blantyre Institute of Community Outreach, Blantyre, Malawi. This article is based off the author’s PhD thesis 17 .

 Smith MI Yatsunenko T Manary MJ : Gut microbiomes of Malawian twin pairs discordant for kwashiorkor. Science. 2013;339(6119):548–554. 23363771 10.1126/science.1229000 3667500 Blanton LV Charbonneau MR Salih T : Gut bacteria that prevent growth impairments transmitted by microbiota from malnourished children. Science. 2016;351(6275). 26912898 10.1126/science.aad3311 4787260 Schneeberger M Everard A Gómez-Valadés AG : Akkermansia muciniphila inversely correlates with the onset of inflammation, altered adipose tissue metabolism and metabolic disorders during obesity in mice. Scientific reports. 2015;5:16643. 26563823 10.1038/srep16643 4643218 Subramanian S Huq S Yatsunenko T : Persistent gut microbiota immaturity in malnourished Bangladeshi children. Nature. 2014;510(7505):417–421. 24896187 10.1038/nature13421 4189846 Karlsson CL Onnerfält J Xu J : The microbiota of the gut in preschool children with normal and excessive body weight. Obesity (Silver Spring). 2012;20(11):2257–2261. 22546742 10.1038/oby.2012.110 Keusch GT Denno DM Black RE : Environmental enteric dysfunction: pathogenesis, diagnosis, and clinical consequences. Clin Infect Dis. 2014;59 Suppl 4(Suppl 4):S207–212. 25305288 10.1093/cid/ciu485 4481570 Fagundes-Neto U Viaro T Wehba J : Tropical enteropathy ( environmental enteropathy) in early childhood: a syndrome caused by contaminated environment. J Trop Pediatr. 1984;30(4):204–209. 6492234 10.1093/tropej/30.4.204 Campbell DI Murch SH Elia M : Chronic T cell-mediated enteropathy in rural west African children: relationship with nutritional status and small bowel function. Pediatr Res. 2003;54(3):306–311. 12788978 10.1203/01.PDR.0000076666.16021.5E Prendergast AJ Humphrey JH Mutasa K : Assessment of Environmental Enteric Dysfunction in the SHINE Trial: Methods and Challenges. Clin Infect Dis. 2015;61 Suppl 7:S726–S732. 26602300 10.1093/cid/civ848 4657593 Lunn PG : The impact of infection and nutrition on gut function and growth in childhood. Proc Nutr Soc. 2000;59(1):147–154. 10828184 10.1017/s0029665100000173 Kosek M Haque R Lima A : Fecal markers of intestinal inflammation and permeability associated with the subsequent acquisition of linear growth deficits in infants. Am J Trop Med Hyg. 2013;88(2):390–396. 23185075 10.4269/ajtmh.2012.12-0549 3583335 Guerrant RL Leite AM Pinkerton R : Biomarkers of Environmental Enteropathy, Inflammation, Stunting, and Impaired Growth in Children in Northeast Brazil. PLoS One. 2016;11(9):e0158772. 27690129 10.1371/journal.pone.0158772 5045163 Hart JD Samikwa L Sikina F : Effects of Biannual Azithromycin Mass Drug Administration on Malaria in Malawian Children: A Cluster-Randomized Trial. Am J Trop Med Hyg. 2020;103(3):1329–1334. 32342841 10.4269/ajtmh.19-0619 7470590 Doan T Arzika AM Hinterwirth A : Macrolide Resistance in MORDOR I - A Cluster-Randomized Trial in Niger. N Engl J Med. 2019;380(23):2271–2273. 31167060 10.1056/NEJMc1901535 6518950 Chaima D Pickering H Hart JD : Up to Four Biannual Administrations of Mass Azithromycin Treatment Are Associated with Modest Changes in the Gut Microbiota of Rural Malawian Children. Preprints. 2020;2020100138. 10.20944/preprints202010.0138.v1 Chaima D Pickering H Hart JD : Fecal biomarkers of environmental enteric dysfunction and the gut microbiota of rural Malawian children: an observational study. medRxiv. 2021;2021.2001.2019.21250144. 10.1016/j.heliyon.2021.e08194 Chaima D : Characterization of the fecal microbiota of rural Malawian children, associations with biomarkers of environmental enteric dysfunction and the impact of a mass drug administration program.Doctoral thesis, University of London, London School of Hygiene and Tropical Medicine. 2020. 10.17037/PUBS.04657527 Pickering H Holland MJ Last AR : Evaluation of a Chlamydia trachomatis-specific, commercial, real-time PCR for use with ocular swabs. Parasit Vectors. 2018;11(1):102. 29463279 10.1186/s13071-018-2686-y 5819642 McCormick BJJ Lee GO Seidman JC : Dynamics and Trends in Fecal Biomarkers of Gut Function in Children from 1-24 Months in the MAL-ED Study. Am J Trop Med Hyg. 2017;96(2):465–472. 27994110 10.4269/ajtmh.16-0496 5303054 de Onis M Garza C Onyango AW : [WHO growth standards for infants and young children]. Archives de pediatrie: organe officiel de la Societe francaise de pediatrie. 2009;16(1):47–53. 10.1016/j.arcped.2008.10.010 Naylor C Lu M Haque R : Environmental Enteropathy, Oral Vaccine Failure and Growth Faltering in Infants in Bangladesh. EBioMedicine. 2015;2(11):1759–1766. 26870801 10.1016/j.ebiom.2015.09.036 4740306 Qin J Li R Raes J : A human gut microbial gene catalogue established by metagenomic sequencing. Nature. 2010;464(7285):59–65. 20203603 10.1038/nature08821 3779803 Hopkins MJ Macfarlane GT Furrie E : Characterisation of intestinal bacteria in infant stools using real-time PCR and northern hybridisation analyses. FEMS Microbiol Ecol. 2005;54(1):77–85. 16329974 10.1016/j.femsec.2005.03.001 Khonsari S Suganthy M Burczynska B : A comparative study of bifidobacteria in human babies and adults. Biosci Microbiota Food Health. 2016;35(2):97–103. 27200263 10.12938/bmfh.2015-006 4858883 Odamaki T Kato K Sugahara H : Age-related changes in gut microbiota composition from newborn to centenarian: a cross-sectional study. BMC Microbiol. 2016;16:90. 27220822 10.1186/s12866-016-0708-5 4879732 Yatsunenko T Rey FE Manary MJ : Human gut microbiome viewed across age and geography. Nature. 2012;486(7402):222–227. 22699611 10.1038/nature11053 3376388 Martín R Jiménez E Heilig H : Isolation of Bifidobacteria from Breast Milk and Assessment of the Bifidobacterial Population by PCR-Denaturing Gradient Gel Electrophoresis and Quantitative Real-Time PCR. Appl Environ Microbiol. 2009;75(4):965–9. 19088308 10.1128/AEM.02063-08 2643565 Peirotén A Arqués JL Medina M : Bifidobacterial strains shared by mother and child as source of probiotics. Benef Microbes. 2018;9(2):231–238. 29488411 10.3920/BM2017.0133 Chowanadisai W Lönnerdal B : Alpha(1)-antitrypsin and antichymotrypsin in human milk: origin, concentrations, and stability. Am J Clin Nutr. 2002;76(4):828–833. 12324297 10.1093/ajcn/76.4.828 Matsubara Y Gaull GE : Biopterin and neopterin in various milks and infant formulas. Am J Clin Nutr. 1985;41(1):110–112. 3966413 10.1093/ajcn/41.1.110 Tsopmo A Diehl-Jones BW Aluko RE : Tryptophan Released From Mother's Milk Has Antioxidant Properties. Pediatr Res. 2009;66(6):614–618. 19730158 10.1203/PDR.0b013e3181be9e7e Elisia I Tsopmo A Friel JK : Tryptophan from human milk induces oxidative stress and upregulates the Nrf-2-mediated stress response in human intestinal cell lines. J Nutr. 2011;141(8):1417–1423. 21677072 10.3945/jn.111.139451 Kosek MN MAL-ED Network Investigators : Causal Pathways from Enteropathogens to Environmental Enteropathy: Findings from the MAL-ED Birth Cohort Study. EBioMedicine. 2017;18:109–117. 28396264 10.1016/j.ebiom.2017.02.024 5405169 George CM Oldja L Biswas S : Geophagy is associated with environmental enteropathy and stunting in children in rural Bangladesh. Am J Trop Med Hyg. 2015;92(6):1117–1124. 25918214 10.4269/ajtmh.14-0672 4458812 Campbell RK Schulze KJ Shaikh S : Biomarkers of Environmental Enteric Dysfunction Among Children in Rural Bangladesh. J Pediatr Gastroenterol Nutr. 2017;65(1):40–46. 28644348 10.1097/MPG.0000000000001557 5492885 Chaima D : Association between intestinal bacterial carriage, biomarkers of environmental enteric dysfunction and stunting in rural Malawian children. Figshare. Dataset. 2022. http://www.doi.org/10.6084/m9.figshare.20178065.v1 10.21956/gatesopenres.14956.r33712 Reviewer response for version 1 Donowitz Jeffrey R 1 Referee https://orcid.org/0000-0003-4816-4728 Division of Infectious Diseases and International Health, Division of Pediatric Infectious Diseases, Children's Hospital of Richmond at Virginia Commonwealth University, University of Virginia, Charlottesville, Virginia, USA

Competing interests: No competing interests were disclosed.

 21 7 2023 Copyright: © 2023 Donowitz JR 2023 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation reject

Summary:

Overall, this is a well written manuscript detailing an interesting study investigating the association of four intestinal bacteria with EED biomarkers and growth. For the most part the methods are well described and I was particularly impressed with the attention to detail in the laboratory methods. The statistical analysis is atypical with unclear rational for discretizing the data differently in different analyses.  It is actually unclear why the authors discretized these variables at all. Other, more minor, comments are outlined below. 

Abstract:

Authors state that Carriage was not associated with EED but B. longum was.

Methods:

Authors state that EED score was modified from the original Kosek, et al. method and reference #16.  I recommend they detail this formula and modification keeping the references. 

Authors excluded children with HAZ >6 or <-6.  Please state how many children were excluded based on this criteria. 

The lack of covariates/confounders investigated here is significant and greatly diminishes interpretability of findings. Did the MORDOR study really collect such limited metadata? Breastfeeding is a major one but there are many others one could think to include in these analyses.

It is unclear why the authors choose to use percent carriage rather than CT values or a burden calculation. This decision needs to be justified. 

Why did the authors discretize EED biomarkers and EED score in some analyses and not in others? This feels like p value fishing although I am not sure that was the motivation. The thought process behind the analysis plan needs to be made clear. Ideally a link to a clinicaltrials.gov record with an a priori analysis plan would be available. It would make the most sense to keep both PCR data (i.e. CT values or burden) and the biomarker data continuous. 

Results:

In table 1, comparison of height seems irrelevant given age was significantly different between the two groups. HAZ would be a better measure. 

Table 1: ensure all symbols in the table are described in the footnotes. 

When discussing prevalence of bacterial carriage, please provide actual data (i.e. % carriage) rather than vague statements like “just over one third”.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Partly

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

Partly

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Clinical pediatric infectious diseases, environmental enteric dysfunction, small intestine bacterial overgrowth and enteric dysbiosis in EED, stunting and neurodevelopment in LMICs

I confirm that I have read this submission and believe that I have an appropriate level of expertise to state that I do not consider it to be of an acceptable scientific standard, for reasons outlined above.

 10.21956/gatesopenres.14956.r32669 Reviewer response for version 1 Manges Amee R. 1 1 Referee https://orcid.org/0000-0003-2462-7249 University of British Columbia, Vancouver, BC, Canada

Competing interests: No competing interests were disclosed.

 25 10 2022 Copyright: © 2022 Manges AR 2022 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

Summary:

The authors conducted a cross-sectional study to examine the relationship between four members of the intestinal microbiome (as measured by qPCR), which have been previously associated with growth deficits in mouse models ( Bifidobacterium longum, Faecalibacterium prausnitzii, Dorea formicigenerans, and Akkermansia muciniphila), and environmental enteric dysfunction (EED) (neopterin, myeloperoxidase, and alpha-1 antitrypsin) biomarkers, and child growth in 613 Malawian children ages 1-59 months from the MORDOR Trial. The authors confirm that these bacterial species are prevalent, and that EED markers are higher at early ages. Surprisingly, B. longum was associated with higher myeloperoxidase and neopterin levels in the 1-12 and 13-24 age groups, respectively. The authors did not identify any meaningful associations between the EED biomarkers, the four intestinal bacterial species and linear growth in MORDOR Trial children at baseline.

Major comments:

This study replicates results from other recent studies that suggest that current biomarkers used to define EED are not associated with poor linear growth in children. Previous studies have also supported a link between the microbiome and stunting, although this has also been disputed. The study size is a strength. The manuscript is clearly written and well organized. However, there are several limitations to the study design and analysis.

Constraining the study to these 4 bacterial species by PCR detection is a limitation, as some studies, but not all, have linked various microbiome members to stunting. Another study of the microbiome, undernutrition and EED in mice identified E. coli and Bacteroides sp. to be associated with EED and poor ponderal growth (Brown et al., 2015 1 ); this study may not have been considered when planning the set of bacterial species to target in the MORDOR children.

The 1-12 month age category is much too broad. 

There is a greater range of F. prausnitzii abundance in this age group (very low at 1 month and much higher by 12 months), by collapsing these children together, the authors are obscuring important variation by age for this taxon. Anaerobes are expanding significantly in the gut microbiome over the first 12 months as infants mature. Similarly, D. formicigenerans are more abundant in older children, and differences in carriage may again be obscured by using such a wide age range, especially in the earlier age categories.

It is also possible that a significant fraction of the poor linear growth is baked in before 12 months of age, making the early measurements more important.

The authors do not have data on breast feeding patterns, which are associated with EED measures and microbiome composition.  Finer stratification in the 1-12 month groups would better capture the period of exclusive breast feeding (although age is not a perfect proxy for breast feeding status), partially removing the effect of confounding by breast feeding.

There are very large differences in EED biomarker values, especially between the 1-12 and 13-24 month age groups. A more careful analysis of these data by age would be useful, there are strong correlations between microbiome compositional changes, EED markers and child growth, especially during early life. It is challenging to untangle these effects, and estimate direct effects of either bacterial species on growth or EED markers on growth, separate from age.  

The reviewer acknowledges that stratifying this early group further, given that there are only 104 children, is also a big challenge.

Table 5, did the regression analyses within age categories also include adjustment for age? It is not clear from the table legend.

As indicated by the authors, the absence of information on enteric infections or asymptomatic pathogen carriage is a major weakness; pathogen carriage and infection are related to both EED marker levels and linear growth and could be a major confounder in these analyses.

The authors state that “implying that the selected bacteria do not contribute to EED or growth in this Malawian population”; this should be rephrased as it is not clear, due to temporality issues, whether EED markers vary because of these specific bacterial species, or whether these bacterial species lead to differences in EED markers.

Could the authors explain why they only used presence/absence and not quantitative measures of these 4 bacterial species?

It is not clear from methods which primers were used for the qPCR assay.

There are multiple tables, these could be consolidated or simplified.

Is the work clearly and accurately presented and does it cite the current literature?

Partly

If applicable, is the statistical analysis and its interpretation appropriate?

Partly

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Partly

Are sufficient details of methods and analysis provided to allow replication by others?

Partly

Reviewer Expertise:

Epidemiology, molecular epidemiology, human microbiome research

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

References : Diet and specific microbial exposure trigger features of environmental enteropathy in a novel murine model. Nat Commun .2015;6: 10.1038/ncomms8806 7806 26241678 10.1038/ncomms8806