Tuberculosis, bacillus Calmette–Guérin vaccination, and allergic disease: Findings from the International Study of Asthma and Allergies in Childhood Phase Two

Abstract

Some have suggested a protective effect of tuberculosis (TB) infection on allergic disease risk, but few studies have examined the association between the two. We therefore investigated whether TB disease and bacillus Calmette-Guérin (BCG) vaccination in early life protect against allergic disease. Information on allergic disease symptoms, past TB disease, and BCG vaccination as well as potential confounding factors was gathered by parental questionnaire from a randomly selected subset of 23,901 8- to 12-yr-old schoolchildren in 20 centers in both developed and developing countries. Children were also physically examined for flexural eczema and underwent skin prick testing. Pooled odds ratio (OR) estimates and corresponding 95% confidence intervals (CIs) across study centers were calculated, using random effects meta-analysis models. There were 245 (1.0%) reported cases of TB disease, and 66.3% (15,857) of all children received the BCG vaccine. Asthma, hay fever, and flexural eczema symptoms in the past year as well as flexural eczema on skin examination were all positively linked to a history of TB (adjusted pooled OR 'wheeze in the past year' = 2.27, 95% CI 1.52-3.41; adjusted pooled OR 'hay fever symptoms in the past year' = 2.23, 1.22-4.09; adjusted pooled OR 'flexural eczema symptoms in the past year' = 3.21, 2.01-5.12; adjusted pooled OR 'flexural eczema on skin examination' = 4.04, 1.71-9.56). Even higher risk estimates were seen for severe asthma and eczema symptoms [adjusted OR = 4.02 (2.17-7.47) and adjusted OR = 6.31 (2.19-18.17), respectively]. There was no significant association between past TB and skin prick test positivity (adjusted pooled OR = 1.32, 0.87-2.02). BCG vaccination during the first year of life was also not associated with any of the allergy outcomes. We found a uniform positive association between TB and all allergic disease outcomes, including eczema on skin examination. As this was a cross-sectional study, it is unclear whether this positive association is attributable to a causal relationship, and further longitudinal studies are required.