Associations Between Obesity, Diabetes Mellitus, and Pancreatic Cancer in the Japanese Population: A Mendelian Randomization Study

The association between body mass index (BMI) and pancreatic cancer remains elusive in Asian populations. Clarifying the causal relationship between type 2 diabetes (T2D) and pancreatic cancer warrants triangulation of evidence, with Mendelian Randomization (MR) providing a robust approach. A two-sample, multivariable MR analysis was performed to investigate the associations between BMI, T2D, and pancreatic cancer risk in the Japanese population. Single nucleotide polymorphisms (SNPs) associated with BMI and T2D were selected from the Genome-Wide Association Study (GWAS) Catalog, and summary statistics for pancreatic cancer were derived from our previous GWAS in the Japanese population. MR analyses were performed using a random-effects inverse variance-weighted (IVW) method as well as sensitivity analyses. Genetic predisposition to T2D, assessed using 891 SNPs, was associated with increased pancreatic cancer risk (IVW odds ratio [OR]: 1.16; 95% confidence interval [CI]: 1.08-1.25 per 1-unit increase in the ln odds of T2D). MR-Egger and weighted median MR analyses yielded similar results. By contrast, genetically indexed BMI showed no association with pancreatic cancer risk (IVW OR: 1.00; 95% CI: 0.95-1.04 per 1 kg/m² increase). The association between T2D and pancreatic cancer remained significant after adjusting for BMI (OR: 1.15; 95% CI: 1.07-1.24), while BMI was not significant in univariable or multivariable analyses. Our MR study supports a modest causal association of T2D, independent of BMI, with pancreatic cancer risk.