LAUSR

OBJECTIVE To investigate the associations of sleep duration with all-cause mortality, glycemic control, and other clinical parameters of patients with type 2 diabetes. METHODS From April 2013 to December 2015, we conducted a retrospective cohort study. Study participants were divided into three groups according to their sleep duration. Multiple regression analysis and Cox proportional hazards analysis were performed to assess the independent associations of sleep duration with clinical parameters and all-cause mortality. RESULTS We enrolled 1233 patients who were then followed for 860 ± 264 days. During the follow-up period, 20 patients (1.6%) died. Sleep duration inversely associated with plasma B-type natriuretic peptide levels (β = -0.203, p = 0.012) in short (<7 h) sleepers, whereas it was positively associated with hemoglobin A1c levels (β = 0.156, p = 0.021) in long (≥9 h) sleepers. Moreover, Cox proportional hazard analysis revealed that short sleep duration was a significant predictor of all-cause mortality (hazard ratio = 0.473; confidence interval 0.248-0.905, p = 0.024). CONCLUSION Short sleep duration may serve as a prognostic indicator of mortality in Japanese patients with type 2 diabetes and may increase cardiovascular stress. Adequate sleep is essential for the management of type 2 diabetes.