The association of new-onset diabetes with subsequent diagnosis of pancreatic cancer—novel use of a large administrative database

Abstract

Abstract Background Screening options for pancreatic ductal adenocarcinoma (PDAC) are limited. New-onset type 2 diabetes (NoD) is associated with subsequent diagnosis of PDAC in observational studies and may afford an opportunity for PDAC screening. We evaluated this association using a large administrative database. Methods Patients were identified using claims data from the OptumLabs® Data Warehouse. Adult patients with NoD diagnosis were matched 1:3 with patients without NoD using age, sex and chronic obstructive pulmonary disease (COPD) status. The event of PDAC diagnosis was compared between cohorts using the Kaplan–Meier method. Factors associated with PDAC diagnosis were evaluated with Cox’s proportional hazards modeling. Results We identified 640 421 patients with NoD and included 1 921 263 controls. At 3 years, significantly more PDAC events were identified in the NoD group vs control group (579 vs 505; P < 0.001). When controlling for patient factors, NoD was significantly associated with elevated risk of PDAC (HR 3.474, 95% CI 3.082–3.920, P < 0.001). Other factors significantly associated with PDAC diagnosis were increasing age, increasing age among Black patients, and COPD diagnosis (P ≤ 0.05). Conclusions NoD was independently associated with subsequent diagnosis of PDAC within 3 years. Future studies should evaluate the feasibility and benefit of PDAC screening in patients with NoD.