Evaluation of common genetic variants identified by GWAS for early onset and morbid obesity in population-based samples

Abstract

Background: Meta-analysis of case-control genome-wide association studies (GWAS) for early onset and morbid obesity identified four variants in/near the PRL, PTER, MAF and NPC1 genes.

Objective: We aimed to validate association of these variants with obesity-related traits in population-based samples.

Design: Genotypes and anthropometric traits were available in up to 31 083 adults from the Fenland, EPIC-Norfolk, Whitehall II, Ely and Hertfordshire studies and in 2042 children and adolescents from the European Youth Heart Study. In each study, we tested associations of rs4712652 (near-PRL), rs10508503 (near-PTER), rs1424233 (near-MAF) and rs1805081 (NPC1), or proxy variants (r (2)>0.8), with the odds of being overweight and obese, as well as with body mass index (BMI), percentage body fat (%BF) and waist circumference (WC). Associations were adjusted for sex, age and age(2) in adults and for sex, age, age group, country and maturity in children and adolescents. Summary statistics were combined using fixed effects meta-analysis methods.

Results: We had 80% power to detect odds ratios of 1.046 to 1.092 for overweight and 1.067 to 1.136 for obesity. Variants near PRL, PTER and MAF were not associated with the odds of being overweight or obese, or with BMI, %BF or WC after meta-analysis (P>0.15). The NPC1 variant rs1805081 showed some evidence of association with %BF (β=0.013 s.d./allele, P=0.040), but not with any of the remaining obesity-related traits (P>0.3).

Conclusion: Overall, these variants, which were identified in a GWAS for early onset and morbid obesity, do not seem to influence obesity-related traits in the general population.

Figure 1

Associations of variants near PRL (A), near PTER (B), near MAF (C) and in NPC1 (D) with the odds of being overweight and obese compared with non-overweight (control group), calculated using logistic regression in individual studies and after fixed effects meta-analysis of ORs and 95% CI. Associations were adjusted for age, age and sex in adults and for age, age-group, sex, country and maturity in children and adolescents. Odds ratios (OR) represent the change in outcome for each additional risk allele under an additive model, with risk alleles defined as the obesity-susceptibility increasing alleles in the discovery study.

Figure 2

Associations of variants near PRL (A), near PTER (B), near MAF (C) and in NPC1 (D) with BMI, percentage body fat (%BF) and waist circumference (WC), calculated using linear regression in individual studies and after fixed effects meta-analysis of betas and standard errors. Associations were adjusted for age, age and sex in adults and for age, age-group, sex, country and maturity in children and adolescents. Associations with WC were additionally adjusted for height. As all traits were inverse normally transformed, betas represent the change in outcome for each additional risk allele under an additive model, expressed in SD units/ allele. Risk alleles are defined as the obesity-susceptibility increasing alleles in the discovery study.