Abstract
BACKGROUND: The association between coffee and pancreatic cancer risk has reported inconsistent results. Therefore, a Mendelian randomization (MR) study was undertaken to investigate the association between coffee and pancreatic cancer from a genetic perspective. METHODS: In East Asian and European populations, independent genetic variants strongly associated with coffee were chosen as instrumental variables (IVs) from relevant genome-wide association studies (GWASs). GWAS data for pancreatic cancer were obtained from the JENGER (Japanese Encyclopedia of Genetic Associations by Riken) project and GWAS catalog database. Two-sample (TSMR) and multivariable Mendelian randomization (MVMR) analyses were conducted to investigate the genetically predicted causal relationship between coffee consumption and pancreatic cancer. A fixed-effect meta-analysis was employed to aggregate estimates from the two populations to reveal the overall association. RESULTS: Both in East Asian and European populations, an increase in coffee intake of a cup per day was not associated with pancreatic cancer risk, regardless of coffee type (including caffeine drinks, instant coffee, decaffeinated coffee, ground coffee, etc.). The results aligned with the findings of the meta-analysis (OR = 1.100, 95%CI = 0.862-1.403, p = 0.450). Also, for coffee intake with positive results in the TSMR analysis (OR = 1.739, 95%CI 1.104-2.739, p = 0.017), consistent negative results were observed after adjusting for potential confounders (smoking traits, drinking, type 2 diabetes, body mass index) in the MVMR analyses. CONCLUSIONS: This study found no genetically predicted causal relationship between coffee consumption and pancreatic cancer risk.