The causal effects of serum lipids and apolipoproteins on kidney function: multivariable and bidirectional Mendelian-randomization analyses

• Published in: International journal of epidemiology 2021-11, Vol.50 (5), p.1569-1579
• Main Authors: Rasheed, Humaira, Zheng, Jie, Rees, Jessica, Sanderson, Eleanor, Thomas, Laurent, Richardson, Tom G, Fang, Si, Bekkevold, Ole-Jørgen, Stovner, Endre Bakken, Gabrielsen, Maiken Elvestad, Skogholt, Anne Heidi, Romundstad, Solfrid, Brumpton, Ben, Hallan, Stein, Willer, Cristen, Burgess, Stephen, Hveem, Kristian, Davey Smith, George, Gaunt, Tom R, Åsvold, Bjørn Olav
• Format: Article
• Language: English
• Subjects: Apolipoproteins - genetics; Genome-Wide Association Study; Humans; Kidney; Lipids; Mendelian Randomization; Mendelian Randomization Analysis; Random Allocation; Triglycerides
• ISSN: 0300-5771; 1464-3685
• DOI: 10.1093/ije/dyab014

Abstract Background The causal nature of the observed associations between serum lipids and apolipoproteins and kidney function are unclear. Methods Using two-sample and multivariable Mendelian randomization (MR), we examined the causal effects of serum lipids and apolipoproteins on kidney function, indicated by the glomerular-filtration rate estimated using creatinine (eGFRcrea) or cystatin C (eGFRcys) and the urinary albumin-to-creatinine ratio (UACR). We obtained lipid- and apolipoprotein-associated genetic variants from the Global Lipids Genetics Consortium (n = 331 368) and UK Biobank (n = 441 016), respectively, and kidney-function markers from the Trøndelag Health Study (HUNT; n = 69 736) and UK Biobank (n = 464 207). The reverse causal direction was examined using variants associated with kidney-function markers selected from recent genome-wide association studies. Results There were no strong associations between genetically predicted lipid and apolipoprotein levels with kidney-function markers. Some, but inconsistent, evidence suggested a weak association of higher genetically predicted atherogenic lipid levels [indicated by low-density lipoprotein cholesterol (LDL-C), triglycerides and apolipoprotein B] with increased eGFR and UACR. For high-density lipoprotein cholesterol (HDL-C), results differed between eGFRcrea and eGFRcys, but neither analysis suggested substantial effects. We found no clear evidence of a reverse causal effect of eGFR on lipid or apolipoprotein traits, but higher UACR was associated with higher LDL-C, triglyceride and apolipoprotein B levels. Conclusion Our MR estimates suggest that serum lipid and apolipoprotein levels do not cause substantial changes in kidney function. A possible weak effect of higher atherogenic lipids on increased eGFR and UACR warrants further investigation. Processes leading to higher UACR may lead to more atherogenic lipid levels.