ABSTRACT
Background:
Coffee is the most commonly consumed beverage in the world after water, however the debate as to whether coffee consumption is beneficial or detrimental to health continues. Current evidence of the link between coffee and health outcomes is predominately observational, thus subject to methodological issues such a confounding and reverse causation.
Methods:
This Mendelian randomisation phenome-wide association study (MR-PheWAS) used information from up to 333,214 participants of White-British ancestry in the UK Biobank to examine the causal association between genetically instrumented habitual coffee consumption and the full range of disease outcomes. We constructed a genetic risk score for habitual coffee consumption and screened for associations with disease outcomes across 963 case-control series. All signals under false discovery rate controlled threshold (5.2 × 10−4) were followed by Mendelian randomisation (MR) analyses, with replication in independent data sources where possible.
Results:
The initial phenome-wide association analysis identified signals for 10 outcomes representing five distinct diseases. The strongest signal was seen for gout (P = 1.3 × 10−11), but there was notable pleiotropy (Pdistortion <0.001) and MR analyses did not support an association with habitual coffee consumption (inverse variance weighted MR OR 0.34, 95% CI 0.06 to 2.05, P = 0.24). Support for a possible causal relationship between habitual coffee consumption was only obtained for three distinct disease outcomes, including an increased odds of osteoarthrosis (OR 1.26, 95% CI 1.12 to 1.43) and obesity (OR 1.28, 95% CI 1.06 to 1.54), and a lower odds of postmenopausal bleeding (OR 0.74, 95% CI 0.65, 0.85). Evidence for an association between habitual coffee consumption and these three diseases was also supported by phenotypic associations with self-reported coffee consumption.
Conclusions:
This large-scale MR-PheWAS provided little evidence for notable harm or benefit with respect to higher habitual coffee consumption. The only evidence for harm was seen with respect to osteoarthrosis and obesity.
