Associations of Blood Caffeine and Genetically Predicted Coffee Consumption with Anthropometric Measures of Obesity:...

In observational studies, caffeine has been associated with a lower risk of obesity. However, whether the associations are causal and apply to coffee, which is a mixture of chemical compounds is unclear.

Two sample Mendelian randomization study.

Genetic instruments predicting caffeine were extracted from an existing GWAS of serum metabolites in 1960 individuals of European descent. For coffee consumption up to 91,462 individuals of European ancestry with top SNPs followed-up in ~30,062 coffee consumers and up to 375,833 individuals of European ancestry were taken from two separate studies. Genetic associations with obesity classes (n= 263,407), waist-to-hip ratio (WHR) (n=210,086), waist circumference (WC) (n= 231,355), and hip circumference (HC) (n=211,117) were obtained from summary statistics of individuals of European ancestry from the Genetic Investigation of Anthropocentric Traits (GIANT).

The inverse-variance weighted method (IVW) was used as the main analysis. We also employed the weighted median approach (WM) and MR-Egger regression as sensitivity analyses. To gauge evidence of directional pleiotropy, we used Cochrane's Q test, and MR-PRESSO global test, as measures of heterogeneity between ratio estimates of variants.

There was little evidence to support an association between blood caffeine and any anthropometric measure of obesity in the primary and sensitivity analyses. However, genetically predicted coffee consumption was positively associated with higher class I obesity and WHR. Furthermore, this association was maintained after correction for multiple testing (P < 0.05/6 = 0.008). Results from the GWAS of coffee consumption were in tandem with results from the GWMA, but associations with class I obesity and waist to hip ratio (WHR) were not maintained after correction for multiple testing.

We found little evidence that caffeine or coffee consumption protects against obesity, adding to growing literature suggesting that previous observational studies may have been confounded. This study demonstrates the dangers of ignoring genetic testing for targeted interventions and basing dietary policy recommendations solely on observational studies restricted to specific populations.