Association between antibiotic exposure and childhood atopic dermatitis: a systematic review and meta-analysis

Abstract

Background: Antibiotics constitute a prevalent class of medications prescribed for pediatric patients. Concerns have arisen regarding the impact of early-life antibiotic use on the risk of atopic illnesses, including childhood atopic dermatitis (AD). AD affects approximately 2% of adults but is twice as prevalent in children, with rates ranging up to 4%. Therefore, the potential correlation between antibiotic exposure and childhood AD is of significant concern.

Methods: Relevant literature was searched from English databases (PubMed, Web of science, Scopus, and The Cochrane Library) and Chinese databases (CNKI database, VIP Database, Wanfang Database) to May 1, 2025. Random-effects meta-analyses were conducted to assess the association between antibiotic exposure and the risk of childhood AD. This study is registered with PROSPERO (CRD42024535141).

Findings: A total of 39 literature sources were included, encompassing 7,487,925 children. Meta-analysis showed that antibiotic exposure during pregnancy and childhood is associated with an increased risk of childhood AD (OR = 1.22, 95% CI: 1.17-1.28; heterogeneity: I² = 98.06%). The subgroup analyses showed that: (1) Children exposed to antibiotics during childhood had a higher risk of developing AD compared to those exposed to antibiotics during pregnancy; (2) the increased risk of AD following antibiotic exposure was significantly affected by the diagnostic criteria and race, but not living areas; (3) antibiotic exposure frequency and type of antibiotics may contribute to the increased risk of AD following antibiotic exposure; (4) the increased risk of AD following antibiotic exposure seems to be positive related with age of outcome measure but not the age of exposure.

Interpretation: Antibiotic exposure during pregnancy or early childhood enhances AD risk, with childhood exposure posing a higher risk than prenatal exposure. Exposure time (during pregnancy and childhood) and the age at outcome measurement are important influencing factors for this association. Diagnostic criteria, race, frequency of antibiotic exposure, and type of antibiotics may also affect this relationship. While the observed statistical significance may be associated with the increased statistical power afforded by the large sample size, the clinical implications of these findings warrant cautious interpretation. Further comprehensive and high-quality studies are warranted to clarify these findings.

Funding: This work was supported by the Natural Science Foundation of Zhejiang (LYY22H310002), the Medical Science and Technology Project of Zhejiang Province (2024KY1170), Special Fund for the Incubation of Young Clinical Scientist, The Children's Hospital of Zhejiang University School of Medicine (CHZJU2023YS006).

Keywords: Antibiotic; Atopic dermatitis; Children; Eczema; Meta-analysis.

Fig. 1

Flow diagram of the study selection process.

Fig. 2

Forest plot of analysis based on antibiotic exposure. OR: Odds ratio; aOR: Adjusted odds ratio; RR: Risk ratios; aRR: Adjusted risk ratios; aHR: Adjusted hazard ratios; CI, confidence interval; #: externally adjusted OR.

Fig. 3

Forest plot of subgroup analysis based on exposure time (pregnancy or childhood); OR: Odds ratio; aOR: Adjusted odds ratio; RR: Risk ratios; aRR: Adjusted risk ratios; aHR: Adjusted hazard ratios; #: externally adjusted OR.

Fig. 4

Forest plot of subgroup analysis based on diagnostic criteria. OR: Odds ratio; aOR: Adjusted odds ratio; RR: Risk ratios; aRR: Adjusted risk ratios; aHR: Adjusted hazard ratios; CI, confidence interval, NR: unclear diagnostic criteria group; ICD: Irritant Contact Dermatitis; ISAAC: The International Study of Asthma and Allergies in Childhood; #: externally adjusted OR.

Fig. 5

Forest plot of subgroup analysis based on races. OR: Odds ratio; aOR: Adjusted odds ratio; RR: Risk ratios; aRR: Adjusted risk ratios; aHR: Adjusted hazard ratios; CI, confidence interval; #: externally adjusted OR.

Fig. 6

Forest plot of subgroup analysis based on living environment. OR: Odds ratio; aOR: Adjusted odds ratio; RR: Risk ratios; aRR: Adjusted risk ratios; aHR: Adjusted hazard ratios; CI, confidence interval; #: externally adjusted OR.

Fig. 7

Forest plot of subgroup analysis based on antibiotic exposure frequency. OR: Odds ratio; aOR: Adjusted odds ratio; RR: Risk ratios; aRR: Adjusted risk ratios; aHR: Adjusted hazard ratios; CI, confidence interval; #: externally adjusted OR.

Fig. 8

Forest plot of subgroup analysis based on type of antibiotics. OR: Odds ratio; aOR: Adjusted odds ratio; RR: Risk ratios; aRR: Adjusted risk ratios; aHR: Adjusted hazard ratios; CI, confidence interval; #: externally adjusted OR.

Fig. 9

Random-effects meta-regression of log (odds ratios) against (A) age of outcome measure (B) age of exposure, and (C) NOS score for the risk of childhood AD after antibiotic exposure. OR: Odds ratio.