Population Attributable Fraction of Diabetes on the Risk of Gastrointestinal Cancers: Overall and by Specific Cancer Sites in a Cohort Study.

Abstract

Background: This study aims to assess the Population Attributable Fraction (PAF) of diabetes on the gastrointestinal cancers overall and by specific cancer sites.

Methods: This study analyzed healthcare data from Taiwan (2006-2019) for 2,362,587 patients with and without diabetes. Gastrointestinal cancers were identified via cancer registry data. Poisson regression calculated incidence rate ratios (IRRs) and 95% confidence intervals (CIs), with propensity score-matched patients without diabetes as the reference. Population attributable fractions (PAFs) estimated cancer incidence attributable to diabetes by sites.

Results: 80,186 patients with diabetes (mean age, 63.3 year; 47.3% women) were matched with 152,323 patients without diabetes (62.7 year; 48.0% women). By the end of 2021, 2,659 out of 80,186 patients with diabetes (incidence rate: 3.89 per 1000 person-years) developed gastrointestinal cancers, compared to 4,150 out of 152,323 (incidence rate: 3.04 per 1000 person-years) in patients without diabetes. Diabetes was associated with a higher risk of gastrointestinal cancers (adjusted IRR of 1.24, 95%CI: 1.18-1.30; PAF: 4.4%, 95%CI: 3.1%-5.8%). The increased risk was primarily driven by pancreatic cancer (adjusted IRR: 1.77, 95%CI: 1.51-2.09; PAF: 12.9%, 95%CI: 7.9%-18.6%) and colorectal cancer (adjusted IRR: 1.28, 95%CI: 1.17-1.39; PAF: 5.1%, 95%CI: 3.0%-7.5%), with a borderline association for liver cancer (adjusted IRR: 1.08, 95%CI: 1.00-1.17; PAF: 1.5%, 95%CI: -0.3%-3.5%).

Conclusions: Diabetes is associated with an increased risk of overall gastrointestinal cancers, largely attributable to pancreatic and colorectal cancers.

Impact: Integrating cancer prevention into the objectives of optimal diabetes management is important, especially for cancers with limited screening options.