Abstract
Background: Continuous glucose monitoring (CGM) can comprehensively reflect the dynamic changes of blood glucose varying with time. CGM data with more patterns of change represents glucose time series with higher complexity. However, the relationship between complexity of glucose dynamics and clinical outcomes of diabetes remains unknown. Objective: To investigate the association between complexity of glucose time series, quantified by multiscale entropy (MSE) analysis, and all-cause mortality in patients with type 2 diabetes mellitus (T2DM) in a prospective cohort study. Methods: Data of 6,001 adult inpatients with T2DM included from January 2005 to December 2015 in a single center was analyzed. Complexity of glucose index (CGI) of glucose time series from CGM was measured by MSE analysis at baseline. Participants were stratified by the tertiles of CGI: < 2.15,2.15∼2.99, and ≥ 3.00. Cox proportional hazards regression models were used to estimate the association between different levels of CGI and the risks of all-cause mortality. Results: During a median follow-up of 6.8 years, 799 deaths were identified. Multivariate Cox regression analysis revealed that the increase of CGI was associated with a lower risk of all-cause mortality (P for trend = 0.021) . The multivariable-adjusted HRs for all-cause mortality at different CGI levels [< 2.15 (reference group) , 2.15∼2.99, and ≥ 3.00] were 1.00, 0.49 (95%CI 0.30-0.79) , and 0.56 (95%CI 0.36-0.87) in patients with HbA1c < 7.0% (P for trend = 0.015) , while the association was nonsignificant in those with HbA1c ≥ 7.0%. And a significant interaction between HbA1c and CGI (P for interaction = 0.032) was noted in the study. Conclusion: Lower CGI is associated with an increased risk of all-cause mortality in T2DM achieving HbA1c target, which suggests that the improvement of CGI may further reduce the risk of long-term adverse outcomes in T2DM at an early stage or with good glycemic control.