• View in gallery

    (A) Box plot of day 9 Simpson’s diversity by baseline age (years). (B) Box plot of day 9 Shannon’s diversity by baseline age (years).

  • 1.

    Ahmed T, Auble D, Berkley JA, Black R, Ahern PP, Hossain M, Hsieh A, Ireen S, Arabi M, Gordon JI, 2014. An evolving perspective about the origins of childhood undernutrition and nutritional interventions that includes the gut microbiome. Ann N Y Acad Sci 1332: 2238.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2.

    Smith MI et al. 2013. Gut microbiomes of Malawian twin pairs discordant for kwashiorkor. Science 339: 548554.

  • 3.

    Blanton LV et al. 2016. Gut bacteria that prevent growth impairments transmitted by microbiota from malnourished children. Science 351: aad3311.

  • 4.

    Gough EK et al. 2014. The impact of antibiotics on growth in children in low and middle income countries: systematic review and meta-analysis of randomised controlled trials. BMJ 348: g2267.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 5.

    Korpela K, Salonen A, Virta LJ, Kekkonen RA, Forslund K, Bork P, de Vos WM, 2016. Intestinal microbiome is related to lifetime antibiotic use in Finnish pre-school children. Nat Commun 7: 10410.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6.

    Mbakwa CA, Scheres L, Penders J, Mommers M, Thijs C, Arts ICW, 2016. Early life antibiotic exposure and weight development in children. J Pediatr 176: 105113.e2.

  • 7.

    Azad MB, Bridgman SL, Becker AB, Kozyrskyj AL, 2014. Infant antibiotic exposure and the development of childhood overweight and central adiposity. Int J Obes 38: 12901298.

    • Search Google Scholar
    • Export Citation
  • 8.

    Bailey LC, Forrest CB, Zhang P, Richards TM, Livshits A, DeRusso PA, 2014. Association of antibiotics in infancy with early childhood obesity. JAMA Pediatr 168: 10631069.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9.

    Trasande L, Blustein J, Liu M, Corwin E, Cox LM, Blaser MJ, 2013. Infant antibiotic exposures and early-life body mass. Int J Obes 37: 1623.

  • 10.

    Jernberg C, Löfmark S, Edlund C, Jansson JK, 2007. Long-term ecological impacts of antibiotic administration on the human intestinal microbiota. ISME J 1: 5666.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11.

    Dethlefsen L, Relman DA, 2011. Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation. Proc Natl Acad Sci USA 108: 45544561.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12.

    Doan T et al. 2017. Gut microbial diversity in antibiotic-naive children after systemic antibiotic exposure: a randomized controlled trial. Clin Infect Dis 64: 11471153.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Oldenburg CE et al. 2018. Effect of commonly used pediatric antibiotics on gut microbial diversity in preschool children in Burkina Faso: a randomized clinical trial. Open Forum Infect Dis 5: ofy289.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14.

    Sié A et al. 2018. Effect of antibiotics on short-term growth among children in Burkina Faso: a randomized trial. Am J Trop Med Hyg 99: 789796.

  • 15.

    Gough EK, Stephens DA, Moodie EEM, Prendergast AJ, Stoltzfus RJ, Humphrey JH, Manges AR, 2015. Linear growth faltering in infants is associated with Acidaminococcus sp. and community-level changes in the gut microbiota. Microbiome 3: 24.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16.

    Subramanian S et al. 2014. Persistent gut microbiota immaturity in malnourished Bangladeshi children. Nature 510: 417421.

  • 17.

    Zambruni M et al. 2019. Stunting is preceded by intestinal mucosal damage and microbiome changes and is associated with systemic inflammation in a cohort of Peruvian infants. Am J Trop Med Hyg 3: 10091017.

    • Search Google Scholar
    • Export Citation
  • 18.

    Sié A, Louis VR, Gbangou A, Müller O, Niamba L, Stieglbauer G, M, Kouyaté B, Sauerborn R, Becher H, 2010. The health and demographic surveillance system (HDSS) in Nouna, Burkina Faso, 1993–2007. Glob Health Action 3, doi: 10.3402/gha.v3i0.5284.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19.

    World Health Organization, UNICEF, 2006. Core Questions on Drinking-Water and Sanitation for Household Surveys. Available at: https://www.who.int/water_sanitation_health/monitoring/oms_brochure_core_questionsfinal24608.pdf. Accessed May 29, 2019.

    • Search Google Scholar
    • Export Citation
  • 20.

    Jost L, 2007. Partitioning diversity into independent alpha and beta components. Ecology 88: 24272439.

Past two years Past Year Past 30 Days
Abstract Views 976 5 0
Full Text Views 482 378 13
PDF Downloads 301 218 14
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

Gut Bacterial Diversity and Growth among Preschool Children in Burkina Faso

View More View Less
  • 1 Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, California;
  • | 2 Centre de Recherche en Santé de Nouna, Nouna, Burkina Faso;
  • | 3 Heidelberg Institute of Global Health (HIGH), Heidelberg, Germany;
  • | 4 Africa Health Research Institute (AHRI), Somkhele, South Africa;
  • | 5 Department of Global Health and Population, Harvard T.H. Chan School of Public Health, Boston, Massachusetts;
  • | 6 Francis I Proctor Foundation, University of California, San Francisco, San Francisco, California;
  • | 7 The Carter Center Niger, Niamey, Niger;
  • | 8 Department of Ophthalmology, Francis I Proctor Foundation, University of California, San Francisco, San Francisco, California

ABSTRACT

There is a lack of empirical, prospective human data on the gut microbiome and its relationship with growth, especially in low- and middle-income countries. We prospectively assessed the association between gut microbial diversity and short-term growth in a cohort of preschool children in Burkina Faso to better characterize whether there is any evidence that changes in gut microbial diversity may affect growth. Data were obtained from a randomized controlled trial evaluating the effect of antibiotic administration on gut microbial diversity in preschool children. We followed up the enrolled children for 35 days, with anthropometric measurements at baseline and day 35 and microbial diversity measured at baseline and day 9 (analytic sample, N = 155). We estimated linear mixed-effects regression models with household random intercepts to assess the association of Simpson’s and Shannon’s alpha diversity with measures of change in anthropometry (e.g., ponderal growth since baseline) and absolute anthropometric measurements (e.g., day 35 weight). We did not find evidence that alpha gut microbial diversity was associated with growth or absolute anthropometric measurements after adjusting for confounding variables. Effect estimates were close to the null (P ≥ 0.15 for all fully adjusted comparisons), with the association between Simpson’s alpha diversity and day 35 height (cm) farthest from the null (coefficient = −0.03, 95% CI: −0.07, 0.01). The change in gut microbial diversity also was not associated with the change in anthropometry in crude or adjusted models. Future research is needed to explore whether gut diversity has an impact on growth over a longer time period, in both healthy and malnourished children.

INTRODUCTION

There is a lack of empirical, prospective human data on the gut microbiome and its relationship with growth, especially in low- and middle-income countries. The gut microbiome is hypothesized to impact a child’s risk for undernutrition.1 A study in Malawi demonstrated that it may be a causal factor in kwashiorkor after transplanting feces from discordant twin pairs into gnotobiotic mice.2 Another study found that mice displayed phenotypes of stunting and underweight on receiving microbiota from affected young children.3

A meta-analysis of randomized controlled trials has shown that antibiotics are growth promoting in children with pre-existing morbidity in low- and middle-income countries.4 Observational studies in high-income countries have corroborated associations between antibiotic exposure and weight gain.59 Antibiotic use is known to reduce diversity and alter the composition of gut microbiota.5,1012 These changes to the microbiome may be responsible for antibiotics’ growth-promoting effects, or such effects may be due to treatment of clinical or subclinical infections.4

Prior analyses from this randomized controlled trial on the effect of a short-course of antibiotics on gut microbial diversity and growth in children in Burkina Faso demonstrated that antibiotics decreased gut microbial diversity in azithromycin-treated children compared with placebo.13 Antibiotics (amoxicillin, in particular) increased growth velocity, weight-for-height z-score (WHZ), and weight-for-age z-score (WAZ).14 Given these findings that antibiotics decreased gut microbial diversity and increased growth, we hypothesize an inverse relationship between gut microbial diversity and growth (Supplemental Figure 1). There have been mixed findings on the effect of gut microbial diversity on growth in low- and middle-income countries.3,1517 Here, we prospectively assess the association between gut microbial diversity and short-term growth in our cohort to better characterize whether changes in gut microbial diversity may affect growth.

MATERIALS AND METHODS

Study design.

This is a secondary analysis of data collected for a randomized controlled trial (clinicaltrials.gov NCT03187834) examining the effect of antibiotics on gut microbial diversity13 (primary outcome) and short-term growth (secondary outcome).14 The study took place in the Nouna Health and Demographic Surveillance Site (HDSS) in rural northwestern Burkina Faso from July to August 2017.18 Households were eligible for participation if they had two or three children between the ages of 6 and 59 months according to the most recent HDSS census. Households were randomized to a 5-day course of amoxicillin, azithromycin, co-trimoxazole, or placebo. Two children per household were randomly chosen to be sentinel children; anthropometry and rectal swabs were only collected for these children. In antibiotic households, one sentinel child randomly received antibiotics and the other received placebo (non-sentinel children also received placebo). In placebo households, all eligible children received placebo. Antibiotics were dosed according to weight.13,14

Although treatment was given in opaque syringes and treatment teams were not told the identity of the medications, it was impossible to fully mask the treatment team because of differences in taste and appearance of the drugs/placebo. However, examination teams were masked to treatment assignment, and laboratory personnel were masked to treatment assignment and time point.

Before randomization, caregivers completed a baseline (day 0) questionnaire, rectal swabs were collected from each child, and anthropometric measurements were obtained. Following this, the 5-day course of treatment commenced. On day 9 (5 days after the last treatment), we again collected rectal swabs. On day 35, we repeated the anthropometry assessment.

For this analysis, we analyzed complete cases, meaning we included only children with gut microbial diversity data at day 0 and day 9, plausible (see Outcomes) anthropometry data at day 0 and day 35, and all other covariate data.

Exposures.

To establish clear temporality, we designated gut microbial diversity (Simpson’s alpha diversity and Shannon’s alpha diversity) on day 9 as our exposure for this analysis. This was chosen so that when considering measures from day 0 to be potential confounders, there was no uncertainty about the temporal order between such measures and our exposure.

Rectal samples were collected in the field by inserting a swab 1–3 cm into the anus and rotating 360°. Between each participant, examiners changed gloves. Swabs were immediately placed in a Stool Nucleic Acid Collection and Transport Tube containing Norgen Stool Preservative (Norgen, Ontario, Canada). The transport media preserves DNA and RNA in the sample and prevents growth of organisms. Samples were stored at ambient temperature in the field, and then at −80°C until they were shipped to the United States, shipped on ice, and then stored at −80°C until processing. 16S library preparation and deep sequencing of the V3-V4 hypervariable regions of the 16S recombinant DNA gene to assess the gut bacterial community were performed as previously described.13 All laboratory personnel were masked by de-identifying samples in the field and then processing samples in a random order for library preparation and sequencing.

Samples were processed for all children who received antibiotics and all children in the placebo arm. In addition, they were processed for the sentinel siblings (who received placebo) of children who received azithromycin because of 1) an effect of antibiotics on gut microbial diversity present only in the azithromycin arm and 2) cost constraints.

Covariates.

All covariates included in the model were hypothesized a priori to be confounders and were measured at day 0. Demographics included the child’s gender and age at baseline. From a household survey asked of caregivers, we determined the presence of improved drinking water and an improved latrine according to the WHO criteria.19 We created a measure of wealth based on the number and value of livestock owned by the household (log wealth was included in the adjusted models). Households reported how many they had in total of each of the following categories: 1) chickens, guinea fowl, ducks, and poultry; 2) sheep, goats, and donkeys; and 3) bovines, horses, and camels. The value of each category was considered to be the mean cost (in West African CFA francs) of 1) chickens, 2) goats, and 3) cows (presumed to be the most commonly owned animal in that category) in the local market as determined by investigators in May 2019 when the analysis was initiated. From a survey asked to caregivers about each child, we created a continuous measure of dietary diversity by summing whether a child had eaten 11 food categories in the last 7 days. We also determined whether the child had recently visited a health facility and, if so, if the child had been treated with antibiotics after the visit. This was categorized as no clinic visit in the past 30 days, visited a clinic in the past 30 days but did not receive antibiotics, and visited a clinic in the past 30 days and received antibiotics. We also included the child’s treatment assignment as part of the randomized trial (whether a child got placebo, amoxicillin, azithromycin, or co-trimoxazole). Finally, we adjusted for baseline measures of anthropometry in some models as described in the following text.

Outcomes.

Outcomes of interest were day 35 height (cm), weight (kg), WHZ, WAZ, height-for-age z-score (HAZ), and growth velocity (g/kg/day). Height, weight, and mid-upper arm circumference were measured on day 0 and day 35. Recumbent length was recorded for children unable to stand and standing height for children who could stand (Shorrboard; Shorr Products, LLC, Olney, MD). Children were weighed on a digital scale standing if able or in the arms of a caregiver (Seca 874 flat floor scale; Seca GMBH & Co., Hamburg, Germany). Height and weight measurements were repeated three times, and the median of these was used for analysis. Z-scores were calculated based on the 2006 WHO standards.14

Children with implausible weight or height changes between the day 0 and day 35 measurements (e.g., gained or lost more than 5 cm or 2 kg) were assumed to be data entry errors and were excluded from analyses.

Statistical methods.

Simpson’s alpha diversity (inverse Simpson’s) was calculated at the genus level and was expressed as effective number.20 Shannon’s alpha diversity was calculated as a secondary analysis in the primary trial. Simpson’s and Shannon’s diversities were calculated in the R package “vegan.” Descriptive statistics of the cohort at baseline were calculated, and box plots of gut microbial diversity by age were created. We checked for nonlinearity of the relationships between gut microbial diversity and each outcome using scatterplots. Crude associations between Simpson’s alpha diversity and Shannon’s alpha diversity and anthropometry were estimated using linear mixed-effects regression models with household random intercepts to account for clustering of children within in the same household. Next, we estimated models adjusting for age, and, finally, for all the previously listed covariates. In the primary fully adjusted models, baseline anthropometry was included for the respective outcome (e.g., in the model for day 35 weight, day 0 weight was included). This defined the outcome of these fully adjusted analyses as growth (change from baseline) at day 35. Because growth velocity is a form of change score and includes baseline weight in the calculation (change in weight from baseline), baseline weight was not added as a covariate in growth velocity models. In addition, to assess the association between gut microbial diversity and absolute anthropometric measurements at day 35, rather than on growth (e.g., weight at day 35, rather than change in weight), we also estimated models with all the listed covariates, excluding baseline anthropometry (for all outcomes other than growth velocity).

Furthermore, we estimated the crude, age-adjusted, and fully adjusted models with the delta anthropometry as the outcome (e.g., difference between day 35 weight and day 0 weight) and the delta gut microbial diversity as the exposure (e.g., difference between day 9 Simpson’s diversity and day 0 Simpson’s diversity). This was performed to test a hypothesis that a change in gut microbial diversity was associated with in a change in growth. These models were not adjusted for baseline anthropometry, as the outcomes themselves already incorporated baseline anthropometry but included the covariates as listed earlier otherwise. The outcome of growth velocity remained the same (there was no baseline with which to compare it).

Finally, we explored post hoc whether interaction between age and gut microbial diversity was present. We added an interaction term to our primary models and compared the crude, age-adjusted with interaction, and fully adjusted with interaction models. Analyses were conducted in R, version 3.6.3 (The R Foundation for Statistical Computing, Vienna, Austria).

Ethical approval.

This study was reviewed and approved by the Comite ́Institutionnel d’Ethique at the Center de Recherche en Sante ́de Nouna in Nouna, Burkina Faso, and the Institutional Review Board at the University of California, San Francisco. Written informed consent was obtained from the caregiver of each participant.

RESULTS

In total, 248 children in 124 households were enrolled in the study. Of these, we chose not to process the rectal swabs of 62 children (sentinel siblings of children who received co-trimoxazole and amoxicillin) as described previously. These data are missing completely at random as treatment arms were randomly assigned to households and children were randomly selected within households. We excluded 14 children missing day 9 diversity data and two additional children missing day 0 diversity data. Of the 170 children with complete gut microbial diversity data, 10 did not have complete data on outcomes and covariates and another five had anthropometry measurements that indicated a data entry problem, leaving 155 children in the analytic sample.

The median age of the participants was 38 months, and slightly more than half (56%) were male (Table 1). The majority of respondents (67%) lived in households with improved drinking water, but less than one-third (30%) lived in households with an improved latrine. Only 10% of participants had visited a clinic in the past 30 days and received antibiotics; another 7% had visited a clinic and not received antibiotics. Wasting was rare in this population (7%), although underweight (12%) and stunting (19%) were slightly more common. Approximately half of the children (51%) received placebo, 15% received amoxicillin, 18% received azithromycin, and 16% received co-trimoxazole in the trial.

Table 1

Baseline characteristics

N = 155
Age (months), median (IQR)38 (25.00, 48.50)
Gender, N (%)
 Male87 (56)
 Female68 (44)
Improved drinking water, N (%)
 Improved water104 (67)
 Unimproved water51 (33)
Improved latrine, N (%)
 Improved latrine47 (30)
 Unimproved latrine108 (70)
Wealth (worth of livestock), median (IQR)646,250.00 (245,625.00, 1,345,000.00)
Clinic visit and antibiotic use, N (%)
 No clinic visit in the past 30 days129 (83)
 Visited clinic/no antibiotics11 (7)
 Visited clinic/antibiotics15 (10)
Weight (kg), median (IQR)12.75 (10.93, 14.70)
Height (cm), median (IQR)91.60 (85.30, 99.50)
Weight-for-height z-score, median (IQR)−0.35 (−1.10, 0.21)
Weight-for-age z-score, median (IQR)−0.88 (−1.48, −0.35)
Height-for-age z-score, median (IQR)−1.01 (−1.60, −0.42)
Wasted, N (%)
 Wasted11 (7)
 Not wasted144 (93)
Underweight, N (%)
 Underweight19 (12)
 Not underweight136 (88)
Stunted, N (%)
 Stunted29 (19)
 Not stunted126 (81)
Treatment, N (%)
 Placebo79 (51)
 Amoxicillin23 (15)
 Azithromycin28 (18)
 Co-trimoxazole25 (16)
Simpson’s diversity (day 9), median (IQR)8.21 (6.29, 10.16)
Shannon’s diversity (day 9), median (IQR)13.52 (10.45, 16.60)

IQR = interquartile range.

Baseline characteristics of the children excluded from the sample due to missing data are presented in Supplemental Table 1. However, they were on average similar to the analytic sample, they were more likely to be female (61%) and were slightly more likely to be wasted (15%), underweight (16%), and stunted (25%).

Median Simpson’s alpha diversity at day 9 was 8.2; median Shannon’s alpha diversity at day 9 was 13.5. However, gut microbial diversity increased with age (Figure 1A and B).

Figure 1.
Figure 1.

(A) Box plot of day 9 Simpson’s diversity by baseline age (years). (B) Box plot of day 9 Shannon’s diversity by baseline age (years).

Citation: The American Journal of Tropical Medicine and Hygiene 103, 6; 10.4269/ajtmh.20-0059

On day 35, the median weight increase was 0.35 kg, representing a median growth velocity of 0.7 g/kg/day. The median height increase was 1 cm.

In crude analyses, both Simpson’s and Shannon’s alpha diversity were positively associated with weight, height, and WHZ (Tables 2 and 3). They were negatively associated with HAZ and unassociated with WAZ and growth velocity. Adjusting for age attenuated nearly all effects. After controlling for all hypothesized, observed confounders and baseline anthropometry in the fully adjusted model, there was no association of Simpson’s or Shannon’s alpha diversity with any measures of growth. Excluding baseline anthropometry to assess the association of diversity with day 35 absolute anthropometric measurements (rather than growth) adjusted for confounding variables did not meaningfully change the results (Supplemental Tables 2 and 3).

Table 2

Association of Simpson’s diversity and day 35 anthropometry

CrudeAge adjustedFully adjusted*
Coefficient95% CIP-valueCoefficient95% CIP-valueCoefficient95% CIP-value
Weight (kg)0.240.11, 0.36< 0.010.03−0.05, 0.100.51−0.01−0.03, 0.020.64
Height (cm)0.740.30, 1.17< 0.01−0.12−0.32, 0.070.23−0.03−0.07, 0.010.20
Weight-for-height z-score0.040.00, 0.090.070.04−0.01, 0.080.140.00−0.02, 0.030.82
Weight-for-age z-score−0.01−0.05, 0.040.690.00−0.05, 0.050.96−0.01−0.02, 0.010.49
Height-for-age z-score−0.07−0.12, −0.020.01−0.05−0.10, 0.000.07−0.01−0.02, 0.000.15
Growth velocity (g/kg/day)−0.04−0.09, 0.010.12−0.02−0.07, 0.030.46−0.01−0.07, 0.040.60

N = 155, households = 108 for all models.

Adjusted for gender, mean-centered age, improved drinking water, improved sanitation, log wealth, dietary diversity, health facility visit/receipt of antibiotics, study treatment, and day 0 anthropometry (except growth velocity). These fully adjusted models are for the outcome of change in anthropometry (growth).

Table 3

Association of Shannon’s diversity and day 35 anthropometry

CrudeAge adjustedFully adjusted*
Coefficient95% CIP-valueCoefficient95% CIP-valueCoefficient95% CIP-value
Weight (kg)0.160.08, 0.24< 0.010.02−0.03, 0.070.370.00−0.02, 0.010.58
Height (cm)0.490.20, 0.78< 0.01−0.06−0.19, 0.070.37−0.02−0.05, 0.010.24
Weight-for-height z-score0.030.00, 0.060.060.03−0.01, 0.060.120.00−0.01, 0.020.86
Weight-for-age z-score0.00−0.03, 0.030.930.00−0.03, 0.040.780.00−0.01, 0.010.52
Height-for-age z-score−0.04−0.07, −0.010.02−0.03−0.06, 0.010.14−0.01−0.01, 0.000.18
Growth velocity (g/kg/day)−0.02−0.06, 0.010.19−0.01−0.05, 0.030.62−0.01−0.04, 0.020.59

N = 155, households = 108 for all models.

Adjusted for gender, mean-centered age, improved drinking water, improved sanitation, log wealth, dietary diversity, health facility visit/receipt of antibiotics, study treatment, and day 0 anthropometry (except growth velocity). These fully adjusted models are for the outcome of change in anthropometry (growth).

There was no association of the change in gut microbial diversity with the change in anthropometry in crude or adjusted models (Supplemental Tables 4 and 5). There was some evidence of an interaction between age and gut microbial diversity (Supplemental Tables 6 and 7), suggesting that gut microbial diversity may improve growth for older children. Although these results should be interpreted with caution as they were not a prespecified analysis and have not been adjusted for multiple comparisons, there is a hypothesis-generating pattern apparent in the fully adjusted models. At the mean age, Simpson’s diversity and weight are unassociated, yet the interaction term has a small positive association (0.002, 95% CI: 0.0002, 0.003, P = 0.03). This is repeated for weight-for-age (0.001, 95% CI: 0.0003, 0.002, P = 0.01) and growth velocity (0.004, 95% CI: 0.0005, 0.007, P = 0.03). For example, although for children at the mean age of 37.3 months, a unit increase in Simpson’s diversity was nonsignificantly associated with 0.0059 unit lower WAZ, the interaction coefficient (0.0012) indicates that for a 48.5-month-old child (75th percentile in our sample), a unit increase in Simpson’s diversity would be associated with about a 0.0075 unit increase in WAZ. Shannon’s diversity had a significant interaction with age only for weight-for-age (0.0006, 95% CI: 0.0000, 0.0012, P = 0.04). Given this pattern, the relationship between gut microbial diversity and ponderal growth may change with age, although associations were small during this short follow-up time.

DISCUSSION

Neither gut microbial diversity nor change in gut microbial diversity was associated with short-term linear or ponderal growth in preschool children in Burkina Faso. This finding is in contrast to results from a small sample of children from cohorts in Malawi and Bangladesh, where it was found that Simpson’s diversity at baseline was lower in children who were stunted at follow-up visits than healthy controls.15 In our data, gut microbial diversity and height were positively associated in crude analyses, although diversity was negatively associated with HAZ in crude analyses. Neither of these associations, however, persisted after adjustment for age. Subramanian et al.16 similarly found no significant correlation between the Shannon Diversity Index and WHZ z-score in an age-adjusted cross-sectional analysis among children aged 18 months. However, healthy children appeared to have greater diversity (as measured by age-adjusted Shannon Diversity Index) than children with severe acute malnutrition, both before and after treatment. Thus, alpha diversity may matter for severe acute malnutrition; we were unable to assess this association in our population because of the rarity of severe acute malnutrition.

Other studies have also failed to find an association between growth and alpha gut microbial diversity. Zambruni et al.17 found differences in beta diversity between stunted and non-stunted children that preceded the development of clinical stunting in a cohort of children followed up for 6 months. No such association was found for alpha diversity. In a study which transplanted fecal samples from healthy and underweight children into germ-free mice, the mice’s growth reflected the phenotypes of the fecal donors. However, there was no relationship between the Shannon index 28 days post-fecal transplant and growth.3

It follows that other aspects of the microbiome, such as maturity of the microbiome16 or the relative abundance of specific genera,15,17 may be more important for growth than alpha diversity. However, most of these studies on alpha diversity, including our own, followed up children for only a short period of time. Thus, if gut microbial diversity does have any effect on growth, however limited, it may be that we simply could not detect it in the amount of time we followed up the children.

Other limitations of our study include the small sample size, although our CIs are fairly narrow around the null, indicating the precision of our estimates was not unduly impaired. Still, our study was not powered specifically to detect associations of gut microbial diversity with growth. Although these analyses were in the context of a randomized controlled trial with standardized measures, this specific research question did not use the study’s randomization and may suffer from residual confounding, such as diet or exposure to pathogens. Finally, the generalizability of these results may be relatively limited outside of similar settings.

In this cohort of preschool children in Burkina Faso, antibiotics decreased alpha gut microbial diversity and increased ponderal growth. However, we did not find evidence that gut microbial diversity was associated with growth. Given these null findings, it appears unlikely that decreased gut microbial diversity mediates the relationship between antibiotics and ponderal growth. In exploratory analyses, we detected an interaction between gut microbial diversity and age that was associated with multiple measures of ponderal growth. Future research is needed to determine how age may influence the relationship between gut microbial diversity and growth, as well as to examine whether gut diversity has an impact on growth over a longer time period, in both healthy and malnourished children.

Supplemental figure and tables

REFERENCES

  • 1.

    Ahmed T, Auble D, Berkley JA, Black R, Ahern PP, Hossain M, Hsieh A, Ireen S, Arabi M, Gordon JI, 2014. An evolving perspective about the origins of childhood undernutrition and nutritional interventions that includes the gut microbiome. Ann N Y Acad Sci 1332: 2238.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2.

    Smith MI et al. 2013. Gut microbiomes of Malawian twin pairs discordant for kwashiorkor. Science 339: 548554.

  • 3.

    Blanton LV et al. 2016. Gut bacteria that prevent growth impairments transmitted by microbiota from malnourished children. Science 351: aad3311.

  • 4.

    Gough EK et al. 2014. The impact of antibiotics on growth in children in low and middle income countries: systematic review and meta-analysis of randomised controlled trials. BMJ 348: g2267.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 5.

    Korpela K, Salonen A, Virta LJ, Kekkonen RA, Forslund K, Bork P, de Vos WM, 2016. Intestinal microbiome is related to lifetime antibiotic use in Finnish pre-school children. Nat Commun 7: 10410.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6.

    Mbakwa CA, Scheres L, Penders J, Mommers M, Thijs C, Arts ICW, 2016. Early life antibiotic exposure and weight development in children. J Pediatr 176: 105113.e2.

  • 7.

    Azad MB, Bridgman SL, Becker AB, Kozyrskyj AL, 2014. Infant antibiotic exposure and the development of childhood overweight and central adiposity. Int J Obes 38: 12901298.

    • Search Google Scholar
    • Export Citation
  • 8.

    Bailey LC, Forrest CB, Zhang P, Richards TM, Livshits A, DeRusso PA, 2014. Association of antibiotics in infancy with early childhood obesity. JAMA Pediatr 168: 10631069.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9.

    Trasande L, Blustein J, Liu M, Corwin E, Cox LM, Blaser MJ, 2013. Infant antibiotic exposures and early-life body mass. Int J Obes 37: 1623.

  • 10.

    Jernberg C, Löfmark S, Edlund C, Jansson JK, 2007. Long-term ecological impacts of antibiotic administration on the human intestinal microbiota. ISME J 1: 5666.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11.

    Dethlefsen L, Relman DA, 2011. Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation. Proc Natl Acad Sci USA 108: 45544561.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12.

    Doan T et al. 2017. Gut microbial diversity in antibiotic-naive children after systemic antibiotic exposure: a randomized controlled trial. Clin Infect Dis 64: 11471153.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Oldenburg CE et al. 2018. Effect of commonly used pediatric antibiotics on gut microbial diversity in preschool children in Burkina Faso: a randomized clinical trial. Open Forum Infect Dis 5: ofy289.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14.

    Sié A et al. 2018. Effect of antibiotics on short-term growth among children in Burkina Faso: a randomized trial. Am J Trop Med Hyg 99: 789796.

  • 15.

    Gough EK, Stephens DA, Moodie EEM, Prendergast AJ, Stoltzfus RJ, Humphrey JH, Manges AR, 2015. Linear growth faltering in infants is associated with Acidaminococcus sp. and community-level changes in the gut microbiota. Microbiome 3: 24.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16.

    Subramanian S et al. 2014. Persistent gut microbiota immaturity in malnourished Bangladeshi children. Nature 510: 417421.

  • 17.

    Zambruni M et al. 2019. Stunting is preceded by intestinal mucosal damage and microbiome changes and is associated with systemic inflammation in a cohort of Peruvian infants. Am J Trop Med Hyg 3: 10091017.

    • Search Google Scholar
    • Export Citation
  • 18.

    Sié A, Louis VR, Gbangou A, Müller O, Niamba L, Stieglbauer G, M, Kouyaté B, Sauerborn R, Becher H, 2010. The health and demographic surveillance system (HDSS) in Nouna, Burkina Faso, 1993–2007. Glob Health Action 3, doi: 10.3402/gha.v3i0.5284.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19.

    World Health Organization, UNICEF, 2006. Core Questions on Drinking-Water and Sanitation for Household Surveys. Available at: https://www.who.int/water_sanitation_health/monitoring/oms_brochure_core_questionsfinal24608.pdf. Accessed May 29, 2019.

    • Search Google Scholar
    • Export Citation
  • 20.

    Jost L, 2007. Partitioning diversity into independent alpha and beta components. Ecology 88: 24272439.

Author Notes

Address correspondence to Catherine E. Oldenburg, Department of Ophthalmology, Francis I Proctor Foundation, and Department of Epidemiology and Biostatistics, University of California, San Francisco, 513 Parnassus Ave., Rm. S334, San Francisco, CA 94143. E-mail: catherine.oldenburg@ucsf.edu

Financial support: C. E. O. and T. D. were supported by a Research to Prevent Blindness Career Development Award.

Authors’ addresses: Jean Digitale and Medellena Maria Glymour, Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, CA, E-mails: jean.digitale@ucsf.edu and maria.glymour@ucsf.edu. Ali Sié, Boubacar Coulibaly, Lucienne Ouermi, Clarisse Dah, and Charlemagne Tapsoba, Centre de Recherche en Santé de Nouna, Nouna, Burkina Faso, E-mails: alisie@yahoo.fr, boubacar@fasonet.bf, ouermil@yahoo.fr, n.clarissedah@yahoo.fr, and charlemagnetapsoba@gmail.com. Till Bärnighausen, Heidelberg Institute of Global Health (HIGH), Heidelberg, Germany, Africa Health Research Institute (AHRI), Somkhele, South Africa, and Department of Global Health and Population, Harvard T.H. Chan School of Public Health, Boston, MA, E-mail: till.baernighausen@uni-heidelberg.de. Elodie Lebas, Francis I Proctor Foundation, University of California, San Francisco, San Francisco, CA, E-mail: elodie.lebas@ucsf.edu. Ahmed M. Arzika, The Carter Center Niger, Niamey, Niger, E-mail: mamaneahmed@yahoo.fr. Jeremy D. Keenan and Thuy Doan, Department of Ophthalmology, Francis I Proctor Foundation, University of California, San Francisco, San Francisco, CA, E-mails: jeremy.keenan@ucsf.edu and thuy.doan@ucsf.edu. Catherine E. Oldenburg, Department of Ophthalmology, Francis I Proctor Foundation, University of California, San Francisco, San Francisco, CA, and Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, CA, E-mail: catherine.oldenburg@ucsf.edu.

Save