The Association of Early Life Exposure to Antibiotics and the Development of Asthma, Eczema and Atopy in a Birth Cohort: Confounding or Causality?
Wickens K, Ingham T, Epton M, et al; New Zealand Asthma and Allergy Cohort Study Group. Clin Exp Allergy. 2008;38(8):1318–1324
PURPOSE OF THE STUDY. The goal was to examine the association between antibiotic exposure in infancy and the development of asthma, eczema, and atopy in early childhood. A secondary goal was to determine whether the association is secondary to confounding chest infections in infancy.
STUDY POPULATION. Expectant mothers were recruited from a random sample of midwives in 2 major New Zealand cities between 1997 and 2001; full details on the nonresponding mothers are incomplete.
METHODS. This was a birth cohort study that collected reported antibiotic exposure before 3 months and before 15 months, along with outcomes (wheeze, asthma, eczema, rash, and inhaler use) at 15 months (N = 1011) and 4 years (N = 986). Questionnaires were administered by study nurses at recruitment and 3, 15, 24, 36, and 48 months of age, in home visits at 3 and 15 months and subsequently by telephone. Outcome measures were collected by using identical questions at 15, 24, 36, and 48 months, covering the period since birth or the previous visit. Analyses were limited to outcomes at 15 months (covering the recall period from birth to 15 months) and at 4 years (covering the recall period from 3 to 4 years). Atopy was defined as >1 positive skin-prick test result at 15 months of age with a panel of common inhalant and food antigens.
RESULTS. Antibiotic exposure before 3 months was significantly associated with asthma developing between birth and 15 months (odds ratio [OR]: 2.32 [95% confidence interval [CI]: 1.5–3.7]; P = .0004); however, with adjustment for chest infections (univariate analysis), this association was reduced (OR: 1.6 [95% CI: 0.96–2.60]) and only trended toward statistical significance (P = .07). Multivariate analysis (with adjustment for gender, ethnicity, family history, parity, otitis media, and antibiotic use between 15 months and 4 years) further decreased this association (OR: 1.3 [95% CI: 0.8–2.2]; P = .4). Similarly, although the association of antibiotics with atopy initially trended toward statistical significance (OR: 1.44 [95% CI: 0.96–2.14]; P = .08), the association was reduced after adjustment for chest infections (OR: 1.36 [95% CI: 0.91–2.05]; P = .14). There was no effect of antibiotic exposure before 15 months on asthma developing after 15 months and remaining present between 3 and 4 years (OR: 1.4 [95% CI: 0.9–2.1]; P = .20). Antibiotic exposure before 3 months was not significantly associated with eczema and rash developing between 0 and 15 months, but exposure before 15 months was significantly associated with both eczema (OR: 1.8 [95% CI: 1.1–3.1]; P = .02) and rash (OR: 1.6 [95% CI: 1.02–2.53]; P = .04) developing after 15 months and remaining present at 4 years; however, these associations also lost statistical significance with both univariate and multivariate analyses.
CONCLUSIONS. There is a statistically significant association between antibiotic exposure in infancy and the subsequent presence of asthma and eczema; however, these associations lose statistical significance with adjustment in univariate and multivariate analyses. The effect of antibiotics on respiratory disease may be a result of confounding by chest infections at an early age when asthma may be indistinguishable from infection.
REVIEWERS COMMENTS. Increases in both asthma prevalence and use of antibiotics in recent years have led some to postulate connections between the 2. Retrospective studies in general have shown strong associations between early antibiotic use and the symptoms of asthma, although these associations have been weaker in prospective studies. Reverse causation as an explanation has been suggested by some (ie, people with asthma may tend to have more respiratory infections that require treatment with antibiotics). This study was limited by a lack of data regarding the reason for early antibiotic use, but it suggests that much of the effect of antibiotics on respiratory disease is a result of the confounding effects of early chest infection. More prospective studies are needed to better illuminate this complex association.
- Copyright © 2009 by the American Academy of Pediatrics