Impairment of kidney function and kidney cancer: A bidirectional Mendelian randomization study

Abstract Background Many observational epidemiology studies discovered that kidney cancer and impaired kidney function have a bidirectional relationship. However, it remains unclear whether these two kinds of traits are causally linked. In this study, we aimed to investigate the bidirectional causal relation between kidney cancer and kidney function biomarkers (creatinine‐based estimated glomerular filtration rate (eGFRcrea), cystatin C‐based estimated glomerular filtration rate (eGFRcys), blood urea nitrogen (BUN), serum urate, and urinary albumin‐to‐creatinine ratio (UACR)). Methods For both directions, single‐nucleotide polymorphisms (SNPs), as genetic instruments, for the five kidney function traits were selected from up to 1,004,040 individuals, and SNPs for kidney cancer were from 408,786 participants(1338 cases). In the main analysis, we applied two state‐of‐the‐art MR methods, namely, contamination mixture and Robust Adjusted Profile Score to downweight the effect of weak instrument bias, pleiotropy, and extreme outliers. We additionally conducted traditional MR analyses as sensitivity analyses. Summary‐level data of European ancestry were extracted from UK Biobank, Chronic Kidney Disease Genetics Consortium, and Kaiser Permanente. Results Based on 99 SNPs, we found that the eGFRcrea had a significant negative causal effect on the risk of kidney cancer (OR = 0.007, 95% CI:2.6 × 10−4–0.569, p = 0.041). After adjusting for body composition or diabetes, urate had a significant negative causal effect on kidney cancer (OR