In recent decades, obesity has surged to epidemic proportions in the developed world, and it has become crucial that obesity’s multifactorial pathophysiology be better understood, providing a means to prevent its increasing occurrence in children and adults.1 An initial observation, made in 2008, was that antibiotic use in infancy is associated with altered microbiota, possibly causing an increased risk for weight gain.2 Since then, numerous observational studies in which this hypothesis has been elucidated and 3 separate meta-analyses have been conducted, and authors have uniformly concluded that there is a weak but significant and positive association between exposure to antibiotics during infancy and weight gain during childhood.3,–5
In this issue of Pediatrics, Block et al6 use an extensive database of 362 550 children to report results that are in agreement with the findings of 3 recent meta-analyses,3,–5 further confirming a reported independent association between early-life antibiotics and greater weight. However, important questions remain: (1) Is there a causal relationship between early-life antibiotic use and future weight gain? (2) What is the clinical significance of this possible relationship?
The Bradford Hill criteria,7 established in 1965, are a group of principles used to evaluate whether observed epidemiological associations are causal.7 These 9 criteria (strength of association, consistency, specificity, temporality, biological gradient, plausibility, coherence, experiment, and analogy), together with the contemporary understanding of biological mechanisms, continue to be valid today.8 If the current research data are examined with these in mind, several aspects support causality between early-life antibiotic exposure and obesity.
The strength of association between early-life antibiotic exposure and weight gain has been weak. Block et al6 reported that the odds ratio for overweight and obesity was 1.05 (95% confidence interval, 1.03–1.07), similar to that obtained in the meta-analyses.3,–5 Because obesity is a multifactorial condition wherein both genetic and multiple environmental factors play roles in its pathogenesis, a relatively weak association does not necessarily exclude causality.
One marked consistency across numerous observational studies is the finding of a positive association between early-life antibiotic exposure and obesity.3,–5 The temporal relationship (the early-life exposure followed by weight gain) also supports causality. However, in retrospective studies, it is challenging to control for the multiple confounding factors that modulate risk of overweight. Additionally, the risk for bias is great, especially in infants, because of numerous essential but insufficiently controlled perinatal and neonatal conditions that could also impact obesity risk in later life.6,9,10 The Block et al6 study in this issue is, thus far, the best attempt at controlling for these hidden confounders. For example, changes in microbiota could also be attributed to the infections being treated. This was elucidated by Li et al,9 whose study revealed that weight gain was not associated with antibiotic exposure after accounting for common pediatric infections, suggesting that the antibiotics were the surrogates for underlying conditions. Although Block et al6 found the same, a weak association was shown even after adjusting for infections.6
Several researchers, including Block et al,6 have demonstrated a dose-dependent response (biological gradient) between antibiotic exposure and weight gain. Exposure to broad-spectrum antibiotics during infancy seems to favor greater fat mass during childhood compared with exposure to narrow-spectrum drugs.5 Furthermore, the most pronounced association was found with macrolides.10,11 All these observations, together with plausibility and coherence, or the idea that antibiotics are potent modulators of the gut microbiome that lead to altered metabolism and overweight, further support a causal association.
Human studies of the specific mechanisms in which antibiotic exposure is linked to weight gain are scarce. Mouse studies have revealed that altered gut microbiota could be involved in the development of obesity12; however, these studies cannot be straightforwardly extrapolated to humans because of the marked differences in energy metabolism and the microbiome. One mechanistic study in humans revealed an association between frequent macrolide exposure in infancy and increased weight gain11 together with the changes in gut flora previously associated with obesity risk.2 Although a causal connection between macrolide-altered microbiota and obesity risk could not be confirmed, the experimental mouse studies were used to endorse the results in humans.
In the light of existing data,3,–5 repeating retrospective observational studies cannot provide further insight into the roles of antibiotics in the pathogenesis of obesity. According to Hill,6,7 the strongest evidence for causality would be experimental (ie, prospective trials with interventions). Together with mechanistic studies in humans, a well-designed prospective follow-up study would provide invaluable new data. On the basis of sound observational data from numerous studies, the effects of antibiotic exposure on weight gain is likely causal but hardly clinically significant, and controlling antibiotic exposure will not provide a magic bullet to control childhood obesity. More importantly, prospective studies would be used to elucidate other potential consequences of modulating the gut microbiome in early childhood.
Opinions expressed in these commentaries are those of the authors and not necessarily those of the American Academy of Pediatrics or its Committees.
FUNDING: No external funding.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2018-0290.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.