Abstract
We used Mendelian randomization (MR) to examine the relationship between smoking, various categories of blood lipids, and bladder cancer (BLCA). Data for this study were drawn from the genome-wide association studies of the GSCAN consortium (~1.2 million participants), a subset of the UK Biobank (~120,000 participants), and the FinnGen consortium (2,072 cases and 307,082 controls). Initially, we utilized inverse variance weighted (IVW), complementary and sensitivity analyses, multivariable MR, and meta-analysis to confirm the association between blood lipids and BLCA. We then performed mediation MR to elucidate the relationship between smoking, blood lipids, and BLCA. Our analysis identified five lipids, including triglycerides in very large HDL, cholesterol in small VLDL, free cholesterol in very large HDL, total free cholesterol, and apolipoprotein B, as having strong and inverse associations with BLCA. These lipids demonstrated no heterogeneity or pleiotropy and exhibited consistent direction and magnitude across IVW, weighted median, and MR-Egger analyses. Our mediation MR further revealed that triglycerides in very large HDL and cholesterol in small VLDL could reduce the impact of smoking on BLCA, mediating -4.3% and -4.5% of the effect, respectively. In conclusion, our study identified five lipids exhibiting a robust inverse relationship with BLCA, two of which can buffer the impact of smoking on BLCA.