CKD as a Mediator of the Associations between Adiposity and Cardiovascular Diseases

Background: Adiposity is associated with atherosclerotic and non-atherosclerotic cardiovascular diseases (CVDs). These associations are not fully explained by traditional cardiovascular risk factors. Genetic epidemiological methods can be used to estimate the mediating effects of risk factors including chronic kidney disease (CKD) on cardiovascular outcomes. Methods: Data from 288,984 UK Biobank participants were used to estimate the relevance of genetically-predicted waist-to-hip ratio (WHR, central adiposity) and body-mass index (BMI, general adiposity) to risk of CVDs, both atherosclerotic (aCVD) and non-atherosclerotic (nCVD). Genetic analyses employed Mendelian Randomisation (MR) and data from 394 WHR and 773 BMI-associated loci. To explore how adiposity causes CVDs, multivariable MR models were constructed adjusting for the genetic associations between these loci and mediators including blood pressure (BP), lipids, diabetes status (DM), inflammatory markers and CKD. Results: 21,759 (7.5%) participants had aCVD and 9,790 (3.4%) participants had nCVD. Genotype-predicted WHR and BMI were both associated with increased risk of aCVD and nCVD (Fig 1). TGs, diastolic BP, and DM were the top mediators for aCVD explaining ~90% of both BMI and WHR associations (Fig 1). Adding CKD to this model did not significantly increase the explained effect. TGs, CKD and inflammation (IL-6) were the top mediators for nCVD, together explaining 86% of WHR associations and 52% of BMI associations (Fig 1). Conclusion: MR suggests that any mediating effect of CKD on the effects of adiposity on risk of aCVD is small (particularly when compared to the mediating effects of TGs, BP and DM). Conversely, CKD is likely to be an important mediator of effects of central adiposity (WHR) on risk of nCVD. The effects of general adiposity (BMI) on nCVD appear to be only partly explained by CKD, inflammation, BP and lipids.