The impact of early life antibiotic use on atopic and metabolic disorders: Meta-analyses of recent insights

Abstract

Background and objectives: The impact of antibiotics use early in life on later-in-life morbidities has received substantial attention as explanations for atopic and metabolic disorders with a surge as modern lifestyle diseases. The objective of this study was to perform meta-analyses to determine if antibiotics administration during the first 2 years of infant life is associated with increased risks of atopic or metabolic disorders later in life.

Methodology: We screened more than 100 English-language prospective and retrospective studies published between January 2002 and March 2020 and assessed study quality using the Newcastle-Ottawa scale. We performed overall and subgroup meta-analyses on 31 high-quality comparable studies on atopic and 23 on metabolic disorders, involving more than 3.5 million children.

Results: Antibiotic exposure prenatally and during the first 2 years of life significantly impacts the risk of developing atopic and metabolic disorders. Exposure during the first 6 months of life appears most critical, consistent with this being the time when the microbiome is most susceptible to irreversible perturbations. The presence of dose-response associations and stronger impacts of broad- than narrow-spectrum antibiotics further point to effects being mediated by microbiota-induced changes.

Conclusions and implications: Our findings support that antibiotics use is a mismatch to modernity that can negatively affect the symbiotic associations we rely on for proper immune function and metabolism. Improving our understanding of these associations, the underlying proximate mechanisms and the impact of antibiotics use on future human-symbiont evolution will be important to improve human health.

Lay summary: The use of antibiotics in infancy has been suggested to increase the risks of atopic and metabolic disorders later in life. Through meta-analyses of more than 100 studies of >3.5 million children, we confirm these risks, and show that patterns are consistent with effects being due to microbiota-driven changes.

Keywords: atopies; coevolution; dysbiosis; immune systems; metabolism; microbiome.

Figure 1.

The process of normal establishment of a healthy microbiota (A) and potential consequences of early onset dysbiosis (B). The microbiota of infants starts as less diverse and with high inter-individual variability; upon reaching adulthood the microbiota stabilizes, presenting increased diversity and more similarities in composition between individuals. The introduction of early disturbers such as antibiotic use or pathogen infection during development can impact the infant microbiota and lead to dysbiosis, which ultimately may lead to a series of metabolic or atopic disorders.

Figure 2.

Main findings in overall and subgroup meta-analysis of immune-related studies given as pooled ORs with 95% CI; for the full results, see Supplementary Table S1. (A) Pooled OR risks for all three immune-related disorders across all available studies. Significance levels indicated with asterisks: **P < 0.0001. Dashed line indicates OR of 1; i.e. no difference in risk to controls. (B) Asthma studies and their outcomes (OR) with total studies and sub-analyses. The number of studies considered are indicated in brackets. Significance levels after post hoc analyses indicated with asterisks: *P = 0.01 and **P < 0.0001. Dashed line indicates OR of 1; i.e. no difference in risk to controls. (C) Dose−response meta-analysis of the association between early antibiotic exposure and childhood asthma (n = 6). Solid and dashed lines represent estimated pooled ORs and CIs, respectively.

Figure 3.

Main findings in overall and subgroup meta-analysis of metabolic-related studies. For the full results, see Supplementary Table S2. (A) Obesity and overweight studies and their outcomes (OR, 95% CI). The number of studies included are given in brackets. Significance levels are indicated with asterisks: *P < 0.05, **P < 0.001 and ***P < 0.00001. Dashed line indicates OR of 1; i.e. no difference in risk to controls. (B) Meta-analysis of BMI-z score studies represented as main difference (MD***) and 95% CI. Significance levels are indicated with asterisks: *P < 0.05 and ***P < 0.00001. Dashed line at MD = 0 indicates the point at which there is no difference to controls. (C) Dose−response meta-analysis of the association between early antibiotic exposure and childhood obesity (n = 4). Solid and dashed lines represent the estimated pooled ORs and CIs, respectively.