Antibiotics and asthma medication in a large register-based cohort study - confounding, cause and effect.

BACKGROUND

An association between asthma and antibiotic usage has been demonstrated, and the issue of reverse causation and confounding by indication is much debated.

OBJECTIVE

Our aim was to study the association between different classes of antibiotics and prescription of asthma medication in a register-based cohort of all Swedish children, born between July 2005 and June 2009, ever treated with antibiotics.

METHODS

Data on dispensed prescriptions of antibiotics (ATC-codes J01) and asthma medication (ATC-codes R03A-D) were requested from the Prescribed Drug Register. The association between dispensed prescriptions of different classes of antibiotics and asthma medication was analysed with Cox regression and a descriptive sequence symmetry analysis.

RESULTS

In total, 211 192 children had received prescriptions of antibiotics. There was a strong association between prescription of antibiotics and prescription of asthma medication. The hazard ratios (HRs) for asthma medication associated with prescription of amoxicillin, penicillin, cephalosporin and macrolides (Gram-positive infections) were stronger than HRs associated with prescription of sulphonamides, trimethoprim and quinolones (urinary tract infections) and flucloxacillin (skin and soft tissue infections), e.g. first year HR = 2.27 (95% confidence intervals 2.17-2.37) as compared with HR = 1.04 (0.78-1.40). The HR associated with broad spectrum antibiotics was significantly higher than the narrow spectrum.

CONCLUSIONS AND CLINICAL RELEVANCE

Our data suggest that the association between antibiotics and asthma is subject to either reverse causation or confounding by indication due to respiratory tract infections. This implies that careful consideration is required as to whether or not symptoms from the respiratory tract in early childhood should be treated with antibiotics or asthma medication.