Abstract
While some retrospective studies have reported that certain dietary habits may affect pancreatic cancer, the variety of dietary habits studied has been limited. A 2-sample Mendelian randomization (MR) analysis was performed using exposure data from the UK Biobank data (n = 10,019,305) and outcome data from the FinnGen database (731 pancreatic cancer cases and 314,193 controls). Instrumental variables were selected based on genome-wide significance, followed by linkage disequilibrium clumping. Sensitivity analyses, including Steiger filtering and MR-PRESSO, were conducted to ensure the robustness of the causal inferences. Following this process, we extracted and filtered the data, ultimately selecting 30 dietary phenotypes from the original 231. Our findings revealed that alcohol intake frequency (raw P-value =.002, odds ratio (OR) = 1.090, 95% CI: 1.032-1.151) and average weekly spirits intake (raw P-value = .005, OR = 1.793, 95% CI: 1.196-2.688) were risk factors for pancreatic cancer using data from 706 single nucleotide polymorphisms derived from FinnGen database. Notably, cooked vegetable intake (raw P-value = .029, OR = 0.704, 95% CI: 0.514-0.965) was found to have a protective effect against pancreatic cancer. Our MR analysis provides evidence that specific dietary habits significantly influence pancreatic cancer risk, highlighting the potential of dietary interventions within public health strategies for pancreatic cancer management.