INTRODUCTION
Asthma is one of the most frequent chronic disease in children.1,2 Type 1 diabetes mellitus (T1DM) occurs at an incidence of 8-20 per 100,000 each year,3 and is one of the most common immune-related diseases in children. The incidence and prevalence of both childhood asthma and T1DM have been increasing over the past decades,1,3 suggesting common susceptibility factors.4
A meta-analysis published in 2003 suggested a weak inverse association between childhood asthma and T1DM.5 However, majority of the studies included in the meta-analysis are case-control studies. In addition, cohort studies included in the meta-analysis are highly heterogeneous in disease definition and age spectrum. In addition, a positive association between the two diseases , which now attracts growing interest, has also been reported in several retrospective cohort studies with stronger quality of evidence.6-8
Since exaggerated immune response and inflammation underlie both asthma and T1DM, reciprocal cause-effect relationship between the two diseases is likely. Mendelian randomization (MR) analysis uses genetic variants as instrumental variables (IVs) for investigating potential cause-effect relationships between different human traits.9 With increasing number of genome-wide association studies (GWAS), MR has been increasingly used. Bidirectional MR analysis is an extension to the basic MR approach that could evaluate whether an “exposure” causes the “outcome” or vice versa.10 We conducted an updated meta-analysis of original studies that examined the association between childhood asthma and T1DM. Data were analyzed using bidirectional MR analysis to examine the possible cause-effect relationship.