Population-based cohort study of diabetes mellitus and mortality in gastric adenocarcinoma.

Abstract

Gastric adenocarcinoma is a common cause of cancer death globally. It remains unclear whether coexisting diabetes mellitus influences survival in patients with this tumour. A cohort study was conducted to determine whether coexisting diabetes increases mortality in gastric adenocarcinoma. This nationwide population-based cohort study included all patients diagnosed with gastric adenocarcinoma in Sweden between 1990 and 2014. Cox proportional hazards regression and competing risks regression were used to assess the influence of coexisting diabetes on disease-specific mortality in gastric adenocarcinoma with adjustment for sex, age, calendar year and co-morbidity (Charlson Co-morbidity Index score excluding diabetes). Among 23 591 patients with gastric adenocarcinoma, 2806 (11·9 per cent) had coexisting diabetes. Overall, patients with diabetes had a moderately increased risk of disease-specific mortality after diagnosis of gastric adenocarcinoma compared with those without diabetes, as shown by both Cox regression (hazard ratio (HR) 1·17, 95 per cent c.i. 1·11 to 1·22) and competing risks regression (sub-HR 1·08, 1·02 to 1·13). The HRs for disease-specific mortality were notably increased in diabetic patients without other co-morbidity (HR 1·23, 1·15 to 1·32) and in diabetic patients who had surgery with curative intent (HR 1·27, 1·16 to 1·38). These findings indicate a worse prognosis in patients with gastric adenocarcinoma and coexisting diabetes compared with those without diabetes.