Mortality in patients with diabetes mellitus and Addison's disease: a nationwide, matched, observational cohort study

Objective: Our hypothesis was that patients with diabetes mellitus obtain an additional risk of death if they develop Addison's disease (AD). Design and methods: Nationwide, matched, observational cohort study cross-referencing the Swedish National Diabetes Register with Inpatient, Cancer and Cause of Death Registers in patients with diabetes (type 1 and 2) and AD and matched controls with diabetes. Clinical characteristics at baseline, overall, and cause-specific mortality were assessed. The relative risk of death was assessed using a Cox proportional hazards regression model. Results: Between January 1996 and December 2012, 226 patients with diabetes and AD were identified and matched with 1129 controls with diabetes. Median (interquartile range) follow-up was 5.9 (2.7-8.6) years. When patients with diabetes were diagnosed with AD, they had an increased frequency of diabetes complications, but both medical history of cancer and coronary heart disease did not differ compared with controls. Sixty-four of the 226 patients with diabetes and AD (28%) died, while 112 of the 1129 controls (10%) died. The estimated relative risk increase (hazard ratio) in overall mortality in the diabetes and AD group was 3.89 (95% confidence interval 2.84-5.32) compared with controls with diabetes. The most common cause of death was cardiovascular in both groups, but patients with diabetes and AD showed an increased death rate from diabetes complications, infectious diseases and unknown causes. Conclusions: Patients with the rare combination of diabetes and AD showed a markedly increased mortality and died more frequently from infections and unknown causes than patients with diabetes alone. Improved strategy for the management of this combination of metabolic disorders is needed.