Genome-wide association study of body mass index in 23 000 individuals with and without asthma

Abstract

Background: Both asthma and obesity are complex disorders that are influenced by environmental and genetic factors. Shared genetic factors between asthma and obesity have been proposed to partly explain epidemiological findings of co-morbidity between these conditions.

Objective: To identify genetic variants that are associated with body mass index (BMI) in asthmatic children and adults, and to evaluate if there are differences between the genetics of BMI in asthmatics and healthy individuals.

Methods: In total, 19 studies contributed with genome-wide analysis study (GWAS) data from more than 23 000 individuals with predominantly European descent, of whom 8165 are asthmatics.

Results: We report associations between several DENND1B variants (P = 2.2 × 10(-7) for rs4915551) on chromosome 1q31 and BMI from a meta-analysis of GWAS data using 2691 asthmatic children (screening data). The top DENND1B single nucleotide polymorphisms(SNPs) were next evaluated in seven independent replication data sets comprising 2014 asthmatics, and rs4915551 was nominally replicated (P < 0.05) in two of the seven studies and of borderline significance in one (P = 0.059). However, strong evidence of effect heterogeneity was observed and overall, the association between rs4915551 and BMI was not significant in the total replication data set, P = 0.71. Using a random effects model, BMI was overall estimated to increase by 0.30 kg/m(2) (P = 0.01 for combined screening and replication data sets, N = 4705) per additional G allele of this DENND1BSNP. FTO was confirmed as an important gene for adult and childhood BMI regardless of asthma status.

Conclusions and clinical relevance: DENND1B was recently identified as an asthma susceptibility gene in a GWAS on children, and here, we find evidence that DENND1B variants may also be associated with BMI in asthmatic children. However, the association was overall not replicated in the independent data sets and the heterogeneous effect of DENND1B points to complex associations with the studied diseases that deserve further study.