Abstract
RATIONALE & OBJECTIVE: Additional investigations into the causal effects of kidney function on various metabolites, particularly lipoprotein lipids in detailed subfractions of lipoprotein particles, in the general population are warranted. STUDY DESIGN: Integrated cross-sectional observational and Mendelian randomization (MR) analyses. SETTING & PARTICIPANTS: We included 157,541 participants aged 40-69 years from the UK Biobank study, a population-scale prospective cohort. Genetic instruments for estimated glomerular filtration rate (eGFR) were developed from the Chronic Kidney Disease Genetics genome-wide association study meta-analysis results, which comprised 567,460 individuals of European ancestry. EXPOSURE: eGFR for observational analysis and genetically predicted eGFR for MR analysis. OUTCOMES: Each of the 178 metabolites from recently updated metabolomics data, including detailed lipoprotein components within 14 subclasses of lipoprotein particles. ANALYTICAL APPROACH: Observational analysis was performed using multivariate linear regression adjusted for various clinicodemographic characteristics. A 2-sample MR analysis was performed using the random-effects inverse-variance weighted method as the main MR method. RESULTS: In the integrated results of the observational and MR analyses, 25 metabolites were causally associated with eGFR. A lower eGFR causally decreased lipoprotein components of high-density lipoprotein and several of its subclasses, particularly medium-sized high-density lipoprotein. Conversely, a lower eGFR causally increased triglyceride levels in smaller-sized very low-density lipoprotein and intermediate-density lipoprotein, as well as increased lipoprotein particle concentrations and total lipids in small very low-density lipoprotein. Additionally, a lower eGFR causally increased the ratio of monounsaturated fatty acids to total fatty acids and that of apolipoprotein B to apolipoprotein A-1. LIMITATIONS: Possibility of false-negative findings when integrating observational and MR analyses. CONCLUSIONS: Decreased kidney function causally aggravates lipoprotein lipid profiles; therefore, clinicians should closely monitor the lipid profiles of individuals with impaired kidney function.