A systems approach identifies time-dependent associations of multimorbidities with pancreatic cancer risk.

Background: Pancreatic ductal adenocarcinoma (PDAC) is usually diagnosed in late adulthood; therefore, many patients suffer or have suffered from other diseases. Identifying disease patterns associated with PDAC risk may enable a better characterization of high-risk patients. Methods: Multimorbidity patterns (MPs) were assessed from 17 self-reported conditions using hierarchical clustering, principal component and factor analyses in 1705 PDAC cases and 1084 controls from a European population. Their association with PDAC was evaluated using adjusted logistic regression models. Time since diagnosis of morbidities to PDAC diagnosis/ recruitment was stratified into recent (<3 years) and long term (⩾3 years). The MPs and PDAC genetic networks were explored with DisGeNET bioinformatics-tool which focuses on gene-diseases associations available in curated databases. Results: Three MPs were observed: gastric (heartburn, acid regurgitation, Helicobacter pylori infection and ulcer), metabolic syndrome (obesity, type-2 diabetes, hypercholesterolemia and hypertension) and atopic (nasal allergies, skin allergies and asthma). Strong associations with PDAC were observed for ⩾2 recently diagnosed gastric conditions [odds ratio (OR), 6.13; 95% confidence interval CI 3.01-12.5)] and for ⩾3 recently diagnosed metabolic syndrome conditions (OR, 1.61; 95% CI 1.11-2.35). Atopic conditions were negatively associated with PDAC (high adherence score OR for tertile III, 0.45; 95% CI, 0.36-0.55). Combining type-2 diabetes with gastric MP resulted in higher PDAC risk for recent (OR, 7.89; 95% CI 3.9-16.1) and long-term diagnosed conditions (OR, 1.86; 95% CI 1.29-2.67). A common genetic basis between MPs and PDAC was observed in the bioinformatics analysis. Conclusions: Specific multimorbidities aggregate and associate with PDAC in a time-dependent manner. A better characterization of a high-risk population for PDAC may help in the early diagnosis of this cancer. The common genetic basis between MP and PDAC points to a mechanistic link between these conditions. [ABSTRACT FROM AUTHOR]