Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Antibiotic use and childhood body mass index trajectory

Abstract

Background/Objectives:

Antibiotics are commonly prescribed for children. Use of antibiotics early in life has been linked to weight gain but there are no large-scale, population-based, longitudinal studies of the full age range among mainly healthy children.

Subjects/Methods:

We used electronic health record data on 163 820 children aged 3–18 years and mixed effects linear regression to model associations of antibiotic orders with growth curve trajectories of annual body mass index (BMI) controlling for confounders. Models evaluated three kinds of antibiotic associations—reversible (time-varying indicator for an order in year before each BMI), persistent (time-varying cumulative orders up to BMIj) and progressive (cumulative orders up to prior BMI (BMIj-1))—and whether these varied by age.

Results:

Among 142 824 children under care in the prior year, a reversible association was observed and this short-term BMI gain was modified by age (P<0.001); effect size peaked in mid-teen years. A persistent association was observed and this association was stronger with increasing age (P<0.001). The addition of the progressive association among children with at least three BMIs (n=79 752) revealed that higher cumulative orders were associated with progressive weight gain; this did not vary by age. Among children with an antibiotic order in the prior year and at least seven lifetime orders, antibiotics (all classes combined) were associated with an average weight gain of approximately 1.4 kg at age 15 years. When antibiotic classes were evaluated separately, the largest weight gain at 15 years was associated with macrolide use.

Conclusions:

We found evidence of reversible, persistent and progressive effects of antibiotic use on BMI trajectories, with different effects by age, among mainly healthy children. The results suggest that antibiotic use may influence weight gain throughout childhood and not just during the earliest years as has been the primary focus of most prior studies.

Introduction

Antibiotics have long been recognized to promote growth in animals1, 2 and sub-therapeutic doses of antibiotics in animal feeds are commonly used for this purpose.3, 4, 5 There has been increasing interest in whether therapeutic antibiotics might produce this effect in humans. It has been hypothesized that antibiotics may impact metabolism and energy balance by altering intestinal microbial populations. Observations of differences in gut microbiota in obese vs lean human subjects6, 7, 8, 9, 10, 11 suggest that microbial changes in the human gut may be a cause of obesity.12

Disruption of the gut microbiota ecology13 with antibiotics adversely affects physiology.6, 14, 15, 16 Antibiotics can lead to pervasive changes in the gut microbiota;12, 17 after short-term use, microbiota can return to pre-treatment patterns,18 but the composition can remain changed for several years and long-term use can result in permanent alterations.17, 19, 20, 21, 22, 23 Antibiotic-induced changes to the microbiota in early life tend to be long-lasting, as early-established bacteria tend to out-compete newly introduced ones.24

A recent review of therapeutic antibiotic use and weight gain in humans reported that a large majority of studies observed such effects.12 Most studies were limited by small sample sizes (median=113) and narrow populations studied.25, 26 Of the 14 that focused on children, half examined exposure in children under 3 years and 4 were confined to infancy. The majority studied only one antibiotic class and individuals with serious infections or chronic diseases, with the possibility that weight gain reflected resolution of the infection.12

We used electronic health record data from a large health system to evaluate whether and how antibiotics may be associated with longitudinal body mass index (BMI) trajectories. This study addresses many of the limitations of prior studies in its evaluation of longitudinal associations between antibiotics and BMI trajectories, across childhood, using a large population-based sample of mainly healthy children without chronic diseases.

Materials and methods

Study population and design

The study population and design have been previously reported.27 Data were obtained from children and adolescents (hereafter 'children') with a primary-care provider in the Geisinger Health System. These children are representative of the general population in the study area of 37 counties in central and northeastern Pennsylvania.28, 29, 30, 31 The study was approved by Institutional Review Boards at both the Geisinger Health System and Johns Hopkins Bloomberg School of Public Health.

Data collection

Data collection using electronic health record data has been previously described.27 We collected data on 257 729 children ages 2–18 years between January 2001 and February 2012. After data cleaning, geocoding and exclusion of 2-year olds,30, 31 163 820 children were included in the analysis. Children can enter or leave care by the health system at any time as family circumstances change, so children in this cohort have varying periods of contact with the health system. Data were obtained on sociodemographics, outpatient, inpatient and emergency encounters, vital signs, laboratory tests, procedures and medication orders. Orders and encounters were accompanied by (ICD-9) International Classification of Diseases, 9th Revision, Clinical Modification diagnostic codes.

Conceptual framework, hypotheses and antibiotic metrics

We evaluated associations of antibiotic exposure with BMI trajectories during childhood. Because this was a longitudinal study, we were able to evaluate three non-exclusive hypotheses (H1, H2, H3 below) about whether the associations of antibiotics were exclusively short-latency and short-term or there were longer-term implications (Figure 1). Antibiotic orders were used as a surrogate for antibiotic dosing because information on dose, duration and refills can be missing or difficult to discern in electronic health record data. Only one antibiotic order per day was counted. Relevant antibiotics were identified through the Medi-Span Generic Product Identifier Therapeutic Classification System.32

Figure 1
figure1

Schematic representation of reversible, persistent and progressive associations of antibiotic orders with body mass index (BMI) trajectories. All three association types were evaluated with time-varying antibiotic order variables. The cumulative counts for the persistent and progressive associations varied slightly. See text for details.

Using these data, we summarized the antibiotic history using three time-varying metrics:

  1. 1)

    H1—reversible effect: indicator (0/1) of any antibiotic order in the year before each BMI measurement (X1=recent exposure).

  2. 2)

    H2—persistent effect: cumulative number of antibiotic orders by the health system before BMIj (X2=cumulative exposure).

  3. 3)

    H3—progressive effect: cumulative number of antibiotic orders by the health system before the prior BMI (BMIj-1) (X3=lagged cumulative exposure).

Because cumulative number of antibiotic orders was skewed, X2 and X3 were modeled via dummy variables encoding categorical levels (0, 1, 2–3, 4–6 or 7+ orders).

The recent exposure coefficient (X1) encoded a reversible association; it allowed for an 'increment' (or decrement) to the next BMI if there was an antibiotic order in the prior year, which reverted to 0 when there was no such order. The cumulative exposure coefficient (X2) encoded a persistent association; once exposure had occurred to a given level, it continued to contribute at that level to subsequent BMIs (until more orders were added). The lagged cumulative exposure coefficient (X3) encoded a progressive association; it allowed past exposure to contribute to subsequent BMIs as a progression (or attenuation) of its original impact. Interactions with age evaluated whether associations varied by periods of childhood.

Variable creation

We computed BMIs by standard procedures using clinic-based measures.27 We analyzed untransformed BMI because this yields estimates that are more interpretable, precise and sensitive to factors that alter change when modeling trajectories compared with age-standardized metrics such as z-transformed BMI (BMI z-scores), because BMI z-scores are age-adjusted, cross-sectional deviations from national norms, which removes some of the longitudinal change across years within children.33, 34, 35 Age (years) was encoded as duration between the child’s date of birth and each BMI. For children with more than one BMI for an age-year, one was randomly selected to mitigate prevalent disease-sampling bias. Two primary comorbidities were examined because they were relatively common and not thought to be in the causal pathway from antibiotics to obesity; asthma (493. ×) and diabetes (250. ×) were considered present if there were two or more outpatient encounters or medication orders with ICD-9 codes up to 1 year after the last BMI. Medical Assistance for health insurance was used as a surrogate for low family socioeconomic status.27 Children were considered under care by the health system if they had any evidence of contact (for example, encounter, medication order, laboratory test) with the health system within or before the window of interest. First BMIs without a prior period during which the child was under care were excluded (n=49 587) because we could not determine antibiotic use prior to the BMI measurement.

Data analysis

We used mixed effects linear regression models to model growth trajectories of BMI by age.27 In initial analyses, outliers and distributional skewing were identified and crude longitudinal trends and bivariable relationships between BMI and predictor variables described. Models included fixed-effects terms for age, age2 and age3 to allow BMI to increase and decrease flexibly over time (age was grand mean-centered at 10.7 years), sex, race/ethnicity (African–American, Hispanic and other vs white) and Medical Assistance (ever vs never). Random intercept age and age2 terms also were included and allowed to covary with unstructured covariance. Fitting was accomplished by restricted maximum likelihood estimation. We allowed residual variances to vary by age groups (3–5, 5–8 and over 8 years). Models included cross-products of sex and Medical Assistance with all three age terms and of race/ethnicity with age and age2, all as fixed effects. Stata (StataCorp LP, College Station, TX, USA) was used for data analysis. All P-values were two-sided and there was no adjustment for multiple comparisons.

To evaluate antibiotic associations, we regressed BMI in a given year (BMIj) on antibiotic history up to that time through the antibiotic main effect terms (recent, cumulative and lagged cumulative orders) and their interactions with age (that is, with age, age2 and age3), added to models in stages. We undertook extensive model-checking and sensitivity analyses to ensure the validity of our findings, including the use of residual and partial residual plots.36 For the latter, we compared modeled relationships with a lowess-smooth37 of the plot; there was good agreement. To evaluate robustness of findings to model assumptions, final models were re-fit using a Huber–White estimator of coefficient variances. As there were no substantive differences only the results of the primary analyses are presented.

To evaluate magnitudes of associations, we calculated the predicted difference in BMI at age 15 years for a child with an antibiotic order in the prior year and at least seven lifetime orders, compared with a child without antibiotic orders, matched on all other variables in our models. Predictions were developed incorporating reversible and persistent associations (Model 2c), and then with progressive association added (Model 3a), separately for four inclusions: (1) all antibiotics and all observed annual BMIs; (2) all antibiotics and the first six observed annual BMIs among children with at least six BMIs; (3) macrolides only and all observed annual BMIs; and (4) macrolides only and the first six observed annual BMIs among children with at least six BMIs. Analyses tailored to the first six observed annual BMIs sought to mitigate systematic under-counting of antibiotic orders among children with fewer observed BMIs. Because we could verify that these children were under care by the health system for at least 6 consecutive years, we would be less likely to under-count cumulative use over time compared with children under care for shorter periods of time. Standard errors, test statistics and 95% confidence intervals for the predictions were computed from the variance/covariance matrix of the component coefficients by standard formulas for statistical contrasts.38 The predicted BMI difference was converted to weight in kilograms using the average height in study children at 15 years.

In an additional sensitivity analysis we evaluated associations of cesarean delivery, a microbiome modifier39, 40 that is associated with childhood obesity,41, 42 with BMI trajectories in the sub-population of 12 629 children delivered by a Geisinger physician with at least a BMI at age 3 years. Deliveries were identified with ICD-9 V30-V39 codes in the child’s electronic health record (fifth digit of code). Model parameterization differed slightly due to the absence of observations from teenagers, longer and more complete follow-up within children, and a shift in the distribution of cumulative antibiotic orders to larger values. The age3 term was removed as was the allowance for residual variances to vary by age, both because of the smaller age range. Progressive antibiotic associations were not evaluated.

Results

Description of children and antibiotic use

Children were primarily white, contributing on average three age-year BMIs to the analysis (Table 1). Antibiotic orders were common, with over 59% of children with at least one order and an average of over four orders while under care (Table 1). Among all 163 820 children there were 683 821 antibiotic orders, and 475 275 BMIs from 142 824 children who were under care in the year before their first BMI. The mean (s.d.) duration in years from their first to their last BMI for children (n=142 824) whose first BMI was at age 3–6 years, 7–10 years, 11–14 years and 15–18 years was 3.8 (3.3), 4.6 (3.5), 3.3 (2.3) and 0.9 (1.0) years, respectively. As expected, children with at least 1 year of documented contact with the health system prior to their first BMI had a higher proportion with an antibiotic order and a higher mean cumulative number of orders than did all children (Table 1). Among children with contact with the Geisinger Health System in their first year of life, 49% had an antibiotic order in that year. Among children who received antibiotics, there were approximately two orders per year (Table 2). The orders (n (%)) by class were: cephalosporins (n=117 894 (17.2)), clindamycin (n=7110 (1.0)), macrolides (n=111 881 (16.4)), metronidazole (n=3116 (0.5)), penicillins (n=367 797 (53.8)), quinolones (n=4513 (0.7)), sulfonamides (n=44 942 (6.6)), tetracyclines (n=22 930 (3.4)) and other classes (n=3638 (0.6)).

Table 1 Summary statistics for two groups of children in analysis (age 3–18 years at any time between 2001 and 2012 with valid body mass index)
Table 2 Antibiotic orders by child age

Associations of antibiotic use with BMI trajectories

There was evidence of a reversible association of antibiotics with higher BMI (Table 3, Model 1). This association was modified by age; the short-term BMI gain was strongest in the mid-teen years. The cumulative order term was next added to the model to evaluate the persistent association while controlling for the reversible association (Table 3). Estimates increased across categories of cumulative orders (1, 2–3, 4–6 and 7+), consistent with a dose–response relation (Model 2a). Cross-products of level of cumulative orders with age, age2 and age3 were next added (Model 2b), with evidence of effect modification by age. To achieve parsimony while focusing on inference and prediction, this model was reduced to include only the linear age cross-products (Model 2c). Estimates increased across categories of cumulative orders and also in their cross-products with age, evidence that the persistent association was stronger with increasing age. Lagged cumulative orders were next added to evaluate progressive associations (Table 4). There was a trend of increasing beta coefficients across cumulative order groups, evidence of a dose–response relation (Model 3a). There was no consistent evidence that the progressive association varied by age (Model 3b) or by higher-order age terms (results not shown).

Table 3 Evaluation of associations of antibiotics with BMI trajectories for children observed in the prior year, evaluating an antibiotic order in the prior year and cumulative orders up to the BMI measurementa (reversible and persistent effects)
Table 4 Evaluation of associations of antibiotics with BMI trajectories for children with at least three BMIs, evaluating an order in the prior year, cumulative antibiotic orders up to the BMI measurement and lagged cumulative antibiotic orders up to the prior BMI measurementa (reversible, persistent and progressive effects)

Sensitivity analyses

We did not observe all children across their entire childhoods, so our models 'spliced' data from different children at different ages. We evaluated age-stratified models of children within narrower age ranges to evaluate whether 'splicing' distorted inferences. When Model 2c without age3 was repeated in three age strata (3–8, 9–13, 14–18 years), inferences were substantively the same in each age stratum as in the all-ages model. In our primary models, we assumed that effects of all classes of antibiotics were similar. To evaluate this assumption, we re-estimated Model 3a separately for three classes with the most orders (penicillins, cephalosporins and macrolides, controlling for other antibiotic use, compared with children who had received none). Inferences were generally similar for all three classes but magnitude estimates were larger for macrolides (next section). To evaluate confounding by comorbid conditions, asthma and diabetes were added to Models 2c and 3a. Both were associated with higher BMI at the average age (results not shown). There were no consistent or substantive changes to antibiotic inferences when each was added alone; there was slight attenuation of antibiotic associations without change in inferences (results not shown) when both were added together. Because in prior work we observed that stimulant use was associated with BMI trajectories,27 we added duration of stimulant use (categorical) and cross-products with age to the antibiotic models. There were no substantive changes in antibiotic associations or inferences after these additions (results not shown).

The sub-population of 12 629 primary-care patients delivered by a Geisinger physician had improved observation over their lifetimes than did the entire population of 163 820 children, with all observed from birth to the last BMI with a minimum of three years of observation. These children, between the ages of 3 and 8 years at the end of observation, had mean (s.d., median) duration from birth to last BMI of 5.1 (1.4, 5.1) years and mean (s.d., median) cumulative antibiotic orders of 6.3 (6.4, 5.0). Of the 12 629 children, 4382 (34.7%) were delivered by cesarean. A strong association of cesarean delivery with BMI was found that increased with age, though we were unable to account for initial birth weight in the model due to limitations in the electronic health record data. The analysis also revealed a stronger antibiotic association than in the primary analysis appearing at a much earlier age. There was little evidence of effect modification by delivery mode on antibiotic associations.

Estimation of the magnitude of the antibiotic association

The magnitude of the predicted weight gain (compared with no antibiotic use) was calculated separately for each of the three most commonly used classes of antibiotics (penicillins, cephalosporins, macrolides). The magnitude of the predicted weight gain associated with antibiotic use ranged from 0.73 to 1.50 g among the various models (Table 5). The excess weight gain was larger after the addition of progressive associations to reversible and persistent, from all antibiotics to macrolides alone, and from all BMIs to the first six BMIs. Among the antibiotic classes the predicted weight gain at age 15 years was highest for macrolides, but the penicillin and cephalosporin classes use were also separately associated with weight gain.

Table 5 Average (95% CI) kilograms of additional weight at age 15 years compared with children who did not receive antibiotics, for boys and girls combined with the average height in our data, for both models 2c (persistent) and 3a (progressive)a

In the delivery mode subset analysis, in the highest group of cumulative antibiotic orders (12 or more), the antibiotic association was equivalent to 1.10 more kg at age 8 years (95% CI 0.68–1.52 kg), amounting to a 3.5% increase in overall mass over the average at that age. The cesarean delivery association at age 8 years of 0.80 kg (95% CI 0.54–1.06 kg) was approximately equivalent to the association with 5–7 cumulative antibiotic orders (0.75 kg, 95% CI 0.34–1.15 kg).

Discussion

In a large, longitudinal study of children representative of the general population in the region, there were associations between receipt of antibiotics and increases in BMI. Associations were complex; reversible associations as a function of antibiotic orders in the prior year that were stronger in the mid-teen years; persistent associations as a function of cumulative orders that got stronger with age; and progressive associations as a function of lagged cumulative orders that did not vary by age. Dose–response relations were observed for both persistent and progressive associations. Several of the observations have not been previously examined, including that antibiotics during childhood may influence BMI at any age, both recent and cumulative doses impact BMI, that antibiotics can impact BMI for long periods of time, and that some impacts strengthen with age. In a subgroup analysis, there was little evidence that delivery mode modified associations of antibiotics with BMI trajectories, but children delivered by cesarean did have higher BMIs at the average age and more rapid BMI growth with increasing age, consistent with prior studies.43 Despite recent declines in antibiotic prescribing in the ambulatory pediatric setting, over-use of broad-spectrum antibiotics continues.44, 45 The findings add to mounting evidence in favor of more judicious use of antibiotics in clinical practice.

The estimated magnitude of the effect predicted by our models ranged from 0.73 to 1.50 kg at age 15 years. However, we likely under-estimated these magnitudes, as our sensitivity analyses showed stronger effects among children who were under care for longer periods of time (that is, in subgroup using the first six annual BMIs, who were all under care for at least six consecutive years, and in the delivery mode analysis). Not all children were observed in the first 2 years of life when antibiotics may have a stronger influence on obesity risk.46 The effects were larger for certain antibiotic classes, and we did not account for more than seven lifetime orders because of the very skewed distribution. Finally, we relied on a surrogate (that is, physician orders) for antibiotic dose and duration, and we cannot be certain about complete ascertainment of antibiotic use given the possible receipt of care outside the health system.

Prior studies have reported a range of findings regarding the impact and timing of antibiotic exposure on BMI and obesity risk. Antibiotics given to children in the first 6 months of life were found to be associated with subsequent risk of over-weight47, 48 but those given from 6–23 months were not related to later BMI.25 Erythromycin was associated with weight gain in preterm neonates with feeding intolerance;49 antibiotics improved weight gain in children with malnutrition.50, 51 Vancomycin, macrolides and tetracyclines (and penicillin) were associated with weight gain, whereas penicillin was associated with weight loss in one study in children in Guatemala.52 A recent large study using electronic health record data reported that antibiotic use in the first 23 months of life was associated with increased risk of obesity at 24–59 months, especially for broad-spectrum antibiotics.46 The differences in findings across studies are likely due, in part, to the range of specific disease populations studied (for example, cystic fibrosis, endocarditis, peptic ulcer), differences in the antibiotics under study and differences in the age groups studied.

We found that associations between antibiotic exposure and BMI change were reversible, persistent and progressive. These findings are consistent with previous studies of the complex association between antibiotics and the microbiome.53 Some human studies have reported recovery of the microbiota after exposure to antibiotics, supporting our reversible association.53 In other studies, changes resulting from the antibiotic exposure have lasted for as long as four years, consistent with our observed persistent association.23, 53, 54 Our finding that associations differed by age is biologically plausible. Gut microbiota can change throughout childhood and even during adulthood.55 The changes in the proportion of calories from carbohydrates with age may be another explanation,56 as microbiota facilitate lipogenesis when presented with high-carbohydrate diets, thus promoting positive energy balance and gains in body mass.57

Increasing evidence points to the possibility that obesity is the consequence of selective environmental exposures during sensitive developmental periods that alter basic programming of metabolic parameters.58, 59, 60 The finding of progressive effects implies that antibiotics may alter the growth trajectories of children in ways that become amplified over time. This has important life course implications; whereas the magnitude of weight increase attributable to antibiotics may be modest by the end of childhood, the finding of progressive effects raises the possibility of accumulating and accelerating effects compounding over the life course.

Our study has several limitations. We did not observe complete childhood trajectories, but rather, a series of exposures and outcomes over a portion of childhood typically spanning 3–5 years for each child. Importantly, we lacked prior antibiotic history on children before they entered care with the Geisinger Health System, so that cumulative counts may be more severely biased in children who entered the health system at later ages. We could not ascertain antibiotic prescribing outside of the health system or compliance with antibiotic orders. We may have assessed exposure more accurately in sicker children with more frequent contact with the health-care system. We were unable to include information on maternal prenatal antibiotic use or child birth weight because electronic health record data on adults were not included in the study’s data and birth weight was not routinely recorded in an easily retrievable way. We believe most of these are all likely to bias our results towards the null and to reduce the magnitude of the observed associations. It is possible that our inability to control for maternal factors influencing development such as prenatal antibiotic use could confound our results. For maternal factors during pregnancy to account for our findings would require these to be highly correlated with cumulative antibiotic use over many years in children 3–8 years later. We have been unable to find support in the literature for such a correlation. However, it is known that some factors during pregnancy are associated with postnatal antibiotic use in children.25

In conclusion, in the largest study to date of the association between antibiotics and longitudinal BMI trajectories in children, we observed a combination of reversible, persistent and progressive associations across several antibiotic classes. We studied a population-based sample of children mainly without chronic diseases and included a broad age range, allowing us to study whether effects differed by age. The results suggest that antibiotic use may influence weight gain throughout childhood and not just during the earliest years as has been the primary focus of most prior studies.

References

  1. 1

    Moore PR, Evenson A, Luckey TD, McCoy E, Elvehjem CA, Hart EB . Use of sulfasuxidine, streptothricin, and streptomycin in nutritional studies with the chick. J Biol Chem 1946; 165: 437–441.

    CAS  PubMed  Google Scholar 

  2. 2

    Stokstad ELR, Jukes TH, Pierce J, Page J, Franklin AC . AL. The multiple nature of the animal protein factor. J Biol Chem 1949; 180: 647–654.

    CAS  PubMed  Google Scholar 

  3. 3

    Jukes TH . Antibiotics in animal feeds and animal production. Bioscience 1972; 22: 526.

    CAS  Article  Google Scholar 

  4. 4

    Kiser JS . Perspective on use of antibiotics in animal feeds. J Anim Sci 1976; 42: 1058–1071.

    CAS  Article  PubMed  Google Scholar 

  5. 5

    Butaye P, Devriese LA, Haesebrouck F . Antimicrobial growth promoters used in animal feed: effects of less well known antibiotics on gram-positive bacteria. Clin Microbiol Rev 2003; 16: 175–188.

    CAS  Article  PubMed  Google Scholar 

  6. 6

    Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL, Duncan A, Ley RE et al. A core gut microbiome in obese and lean twins. Nature 2009; 457: 480–484.

    Article  PubMed  Google Scholar 

  7. 7

    Ley RE, Turnbaugh PJ, Klein S, Gordon JI . Microbial ecology: human gut microbes associated with obesity. Nature 2006; 444: 1022–1023.

    CAS  Article  Google Scholar 

  8. 8

    Santacruz A, Marcos A, Warnberg J, Marti A, Martin-Matillas M, Campoy C et al. Interplay between weight loss and gut microbiota composition in overweight adolescents. Obesity (Silver Spring) 2009; 17: 1906–1915.

    Article  Google Scholar 

  9. 9

    Zhang H, DiBaise JK, Zuccolo A, Kudrna D, Braidotti M, Yu Y et al. Human gut microbiota in obesity and after gastric bypass. Proc Natl Acad Sci USA 2009; 106: 2365–2370.

    CAS  Article  Google Scholar 

  10. 10

    Angelakis E, Merhej V, Raoult D . Related actions of probiotics and antibiotics on gut microbiota and weight modification. Lancet Infect Dis 2013; 13: 889–899.

    CAS  Article  PubMed  Google Scholar 

  11. 11

    Schwiertz A, Taras D, Schafer K, Beijer S, Bos NA, Donus C et al. Microbiota and SCFA in lean and overweight healthy subjects. Obesity (Silver Spring) 2010; 18: 190–195.

    Article  Google Scholar 

  12. 12

    Million M, Lagier JC, Yahav D, Paul M . Gut bacterial microbiota and obesity. Clin Microbiol Infect 2013; 19: 305–313.

    CAS  Article  PubMed  Google Scholar 

  13. 13

    Macfarlane S . Antibiotic treatments and microbes in the gut. Environ Microbiol 2014; 16: 919–924.

    CAS  Article  PubMed  Google Scholar 

  14. 14

    Riley LW, Raphael E, Faerstein E . Obesity in the United States - dysbiosis from exposure to low-dose antibiotics? Front Public Health 2013; 1: 69.

    Article  PubMed  Google Scholar 

  15. 15

    Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER, Gordon JI . An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 2006; 444: 1027–1031.

    Article  Google Scholar 

  16. 16

    Harris K, Kassis A, Major G, Chou CJ . Is the gut microbiota a new factor contributing to obesity and its metabolic disorders? J Obes 2012; 2012: 879151.

    PubMed Central  PubMed  Google Scholar 

  17. 17

    Dethlefsen L, Huse S, Sogin ML, Relman DA . The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16 S rRNA sequencing. PLoS Biol 2008; 6: e280.

    Article  PubMed  Google Scholar 

  18. 18

    De La Cochetiere MF, Durand T, Lepage P, Bourreille A, Galmiche JP, Dore J . Resilience of the dominant human fecal microbiota upon short-course antibiotic challenge. J Clin Microbiol 2005; 43: 5588–5592.

    CAS  Article  PubMed  Google Scholar 

  19. 19

    Robinson CJ, Young VB . Antibiotic administration alters the community structure of the gastrointestinal micobiota. Gut Microbes 2010; 1: 279–284.

    Article  PubMed  Google Scholar 

  20. 20

    Robinson CJ, Bohannan BJ, Young VB . From structure to function: the ecology of host-associated microbial communities. Microbiol Biol Rev 2010; 74: 453–476.

    CAS  Article  Google Scholar 

  21. 21

    Dethlefsen L, Relman DA . Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation. Proc Natl Acad Sci USA 2011; 108: 4554–4561.

    CAS  Article  PubMed  Google Scholar 

  22. 22

    Jernberg C, Lofmark S, Edlund C, Jansson JK . Long-term impacts of antibiotic exposure on the human intestinal microbiota. Microbiology 2010; 156: 3216–3223.

    CAS  Article  PubMed  Google Scholar 

  23. 23

    Jernberg C, Lofmark S, Edlund C, Jansson JK . Long-term ecological impacts of antibiotic administration on the human intestinal microbiota. ISME J 2007; 1: 56–66.

    CAS  Article  PubMed  Google Scholar 

  24. 24

    Martin MA, Sela DA. Infant gut microbiota: developmental influences and health outcomes. In: Clancy KBH, Hinde K, Rutherford JN (eds). Building Babies: Primate Development in Proximate and Ultimate Perspective. Spring Science+Business Media: New York, NY, USA, 2011; 233–256.

    Google Scholar 

  25. 25

    Trasande L, Blustein J, Liu M, Corwin E, Cox LM, Blaser MJ . Infant antibiotic exposures and early-life body mass. Int J Obes (Lond) 2013; 37: 16–23.

    CAS  Article  Google Scholar 

  26. 26

    Lane JA, Murray LJ, Harvey IM, Donovan JL, Nair P, Harvey RF . Randomised clinical trial: helicobacter pylori eradication is associated with a significantly increased body mass index in a placebo-controlled study. Aliment Pharmacol Ther 2011; 33: 922–929.

    CAS  Article  PubMed  Google Scholar 

  27. 27

    Schwartz BS, Bailey-Davis L, Bandeen-Roche K, Pollak J, Hirsch AG, Nau C et al. Attention deficit disorder, stimulant use, and childhood body mass index trajectory. Pediatrics 2014; 133: 668–676.

    Article  PubMed  Google Scholar 

  28. 28

    Liu AY, Curriero FC, Glass TA, Stewart WF, Schwartz BS . The contextual influence of coal abandoned mine lands in communities and type 2 diabetes in Pennsylvania. Health Place 2013; 22: 115–122.

    Article  PubMed  Google Scholar 

  29. 29

    Casey JA, Curriero FC, Cosgrove SE, Nachman KE, Schwartz BS . High-density livestock operations, crop field Aapplication of manure, and risk of community-associated methicillin-resistant staphylococcus aureus infection in pennsylvania. JAMA Int Med 2013; 173: 1980–1990.

    Article  Google Scholar 

  30. 30

    Casey JA, Cosgrove SE, Stewart WF, Pollak J, Schwartz BS . A population-based study of the epidemiology and clinical features of methicillin-resistant Staphylococcus aureus infection in Pennsylvania, 2001-2010. Epidemiol Infect 2013; 141: 1166–1179.

    CAS  Article  PubMed  Google Scholar 

  31. 31

    Schwartz BS, Stewart WF, Godby S, Pollak J, Dewalle J, Larson S et al. Body mass index and the built and social environments in children and adolescents using electronic health records. Am J Prev Med 2011; 41: e17–e28.

    Article  PubMed  Google Scholar 

  32. 32

    Medi-Span Master Drug Data Base Documentation Manual. Medi-Span: Indianapolis, IN, USA, 2007.

  33. 33

    Berkey CS, Colditz GA . Adiposity in adolescents: change in actual BMI works better than change in BMI z score for longitudinal studies. Ann Epidemiol 2007; 17: 44–50.

    Article  PubMed  Google Scholar 

  34. 34

    Cole TJ, Faith MS, Pietrobelli A, Heo M . What is the best measure of adiposity change in growing children: BMI, BMI %, BMI z-score or BMI centile? Eur J Clin Nutr 2005; 59: 419–425.

    CAS  Article  Google Scholar 

  35. 35

    Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Mei Z et al. 2000 CDC Growth Charts for the United States: methods and development. Vital Health Stat 11 2002: 1–190.

    Google Scholar 

  36. 36

    Bandeen-Roche K, Hall CB, Stewart WF, Zeger SL . Modelling disease progression in terms of exposure history. Stat Med 1999; 18: 2899–2916.

    CAS  Article  PubMed  Google Scholar 

  37. 37

    Cleveland WS . Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 1979; 74: 829–836.

    Article  Google Scholar 

  38. 38

    Seber GAF Linear Regression Analysis. Wiley: New York, NY, USA, 1977.

  39. 39

    Munyaka PM, Khafipour E, Ghia JE . External influence of early childhood establishment of gut microbiota and subsequent health implications. Front Pediatr 2014; 2: 109.

    Article  PubMed  Google Scholar 

  40. 40

    Goedert JJ, Hua X, Yu G, Shi J . Diversity and composition of the adult fecal microbiome associated with history of cesarean birth or appendectomy: analysis of the American Gut Project. EBioMedicine 2014; 1: 167–172.

    Article  PubMed  Google Scholar 

  41. 41

    Mueller NT, Whyatt R, Hoepner L, Oberfield S, Dominguez-Bello MG, Widen EM et al. Prenatal exposure to antibiotics, cesarean section and risk of childhood obesity. Int J Obes (Lond) 2015; 39: 665–670.

    CAS  Article  Google Scholar 

  42. 42

    Blustein J, Attina T, Liu M, Ryan AM, Cox LM, Blaser MJ et al. Association of caesarean delivery with child adiposity from age 6 weeks to 15 years. Int J Obes (Lond) 2013; 37: 900–906.

    CAS  Article  Google Scholar 

  43. 43

    Li HT, Zhou YB, Liu JM . The impact of cesarean section on offspring overweight and obesity: a systematic review and meta-analysis. Int J Obes (Lond) 2013; 37: 893–899.

    Article  Google Scholar 

  44. 44

    Hersh AL, Jackson MA, Hicks LA . Principles of judicious antibiotic prescribing for upper respiratory tract infections in pediatrics. Pediatrics 2013; 132: 1146–1154.

    Article  PubMed  Google Scholar 

  45. 45

    Hersh AL, Shapiro DJ, Pavia AT, Shah SS . Antibiotic prescribing in ambulatory pediatrics in the United States. Pediatrics 2011; 128: 1053–1061.

    Article  PubMed  Google Scholar 

  46. 46

    Bailey LC, Forrest CB, Zhang P, Richards TM, Livshits A, DeRusso PA . Association of Antibiotics in Infancy With Early Childhood Obesity. JAMA Pediatr 2014; 168: 1063–1069.

    Article  PubMed  Google Scholar 

  47. 47

    Ajslev TA, Andersen CS, Gamborg M, Sorensen TI, Jess T . Childhood overweight after establishment of the gut microbiota: the role of delivery mode, pre-pregnancy weight and early administration of antibiotics. Int J Obes (Lond) 2011; 35: 522–529.

    CAS  Article  Google Scholar 

  48. 48

    Ray K . Gut microbiota: adding weight to the microbiota's role in obesity—exposure to antibiotics early in life can lead to increased adiposity. Nat Rev Gastroenterol Hepatol 2012; 9: 615.

    Article  PubMed  Google Scholar 

  49. 49

    Mansi Y, Abdelaziz N, Ezzeldin Z, Ibrahim R . Randomized controlled trial of a high dose of oral erythromycin for the treatment of feeding intolerance in preterm infants. Neonatology 2011; 100: 290–294.

    Article  PubMed  Google Scholar 

  50. 50

    Trehan I, Goldbach HS, LaGrone LN, Meuli GJ, Wang RJ, Maleta KM et al. Antibiotics as part of the management of severe acute malnutrition. N Engl J Med 2013; 368: 425–435.

    CAS  Article  PubMed  Google Scholar 

  51. 51

    Smith MI, Yatsunenko T, Manary MJ, Trehan I, Mkakosya R, Cheng J et al. Gut microbiomes of Malawian twin pairs discordant for kwashiorkor. Science 2013; 339: 548–554.

    CAS  Article  PubMed  Google Scholar 

  52. 52

    Guzman MA, Scrimshaw NS, Monroe RJ . Growth and development of Central American children. I. Growth responses of rural Guatemalan school children to daily administration of penicillin and aureomycin. Am J Clin Nutr 1958; 6: 430–438.

    CAS  Article  PubMed  Google Scholar 

  53. 53

    Lozupone CA, Stombaugh JI, Gordon JI, Jansson JK, Knight R . Diversity, stability and resilience of the human gut microbiota. Nature 2012; 489: 220–230.

    CAS  Article  PubMed  Google Scholar 

  54. 54

    Jakobsson HE, Jernberg C, Andersson AF, Sjolund-Karlsson M, Jansson JK, Engstrand L . Short-term antibiotic treatment has differing long-term impacts on the human throat and gut microbiome. Plos One 2010; 5: e9836.

    Article  PubMed  Google Scholar 

  55. 55

    Agans R, Rigsbee L, Kenche H, Michail S, Khamis HJ, Paliy O . Distal gut microbiota of adolescent children is different from that of adults. FEMS Microbiol Ecol 2011; 77: 404–412.

    CAS  Article  PubMed  Google Scholar 

  56. 56

    U.S. Department of Agriculture Agricultural Research Service. Energy Intakes: Percentages of Energy from Protein, Carbohydrate, Fat, and Alcohol, by Gender and Age, What We Eat in America, NHANES 2009–2010. 2012; www.ars.usda.gov/ba/bhnrc/fsrg (accessed 8 August 2014).

  57. 57

    Arora T, Sharma R . Fermentation potential of the gut microbiome: implications for energy homeostasis and weight management. Nutr Rev 2011; 69: 99–106.

    Article  PubMed  Google Scholar 

  58. 58

    Barker DJ . Obesity and early life. Obes Rev 2007; 8: 45–49.

    Article  PubMed  Google Scholar 

  59. 59

    Huang TT, Drewnosksi A, Kumanyika S, Glass TA . A systems-oriented multilevel framework for addressing obesity in the 21st century. Prev Chronic Dis 2009; 6: A82.

    PubMed Central  PubMed  Google Scholar 

  60. 60

    Ludwig DS, Gortmaker SL . Programming obesity in childhood. Lancet 2004; 364: 226–227.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

The project described was supported by Grant Number U54 HD-070725 from the Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD). The project is co-funded by the NICHD and the Office of Behavioral and Social Sciences Research (OBSSR). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author information

Affiliations

Authors

Corresponding author

Correspondence to B S Schwartz.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Schwartz, B., Pollak, J., Bailey-Davis, L. et al. Antibiotic use and childhood body mass index trajectory. Int J Obes 40, 615–621 (2016). https://doi.org/10.1038/ijo.2015.218

Download citation

Further reading

Search

Quick links