Population Attributable Fraction of Diabetes on the Risk of Gastrointestinal Cancers: Overall and by Specific Cancer Sites in a Cohort Study
Menée à partir de données de santé taïwanaises 2006-2019 portant sur 2 362 587 personnes, cette étude estime la part des cancers gastro-intestinaux attribuable au diabète
This study aims to assess the population attributable fraction (PAF) of diabetes on the gastrointestinal cancers overall and by specific cancer sites.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 (IRR) and 95% confidence intervals (CI), with propensity score–matched patients without diabetes as the reference. PAFs estimated cancer incidence attributable to diabetes by sites.80,186 patients with diabetes (mean age, 63.3 years; 47.3% women) were matched with 152,323 patients without diabetes (62.7 years; 48.0% women). By the end of 2021, 2,659 of 80,186 patients with diabetes (incidence rate: 3.89 per 1,000 person-years) developed gastrointestinal cancers compared with 4,150 of 152,323 (incidence rate: 3.04 per 1,000 person-years) 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% to 3.5%).Diabetes is associated with an increased risk of overall gastrointestinal cancers, largely attributable to pancreatic and colorectal cancers.Integrating cancer prevention into the objectives of optimal diabetes management is important, especially for cancers with limited screening options.
Cancer Epidemiology, Biomarkers & Prevention , article en libre accès 2025