LAUSR

OBJECTIVE To investigate the nature of the relationship between HbA1c and sepsis among individuals with type 2 diabetes, and to assess the association between sepsis and all-cause mortality in such patients. RESEARCH DESIGN AND METHODS We included 502,871 individuals with type 2 diabetes recorded in the Swedish National Diabetes Register and used multivariable Cox regression and restricted cubic spline analyses to assess the association between time-updated HbA1c values and sepsis occurrence between 1 January 2005 and 31 December 2015. The association between sepsis and death was examined using multivariable Cox regression analysis. RESULTS Overall, 14,534 (2.9%) patients developed sepsis during the study period. On multivariable Cox regression analysis, compared with an HbA1c of 48-52 mmol/mol (6.5-6.9%), the adjusted hazard ratio for sepsis was 1.15 (95% CI 1.07-1.24) for HbA1c <43 mmol/mol (6.1%), 0.93 (0.87-0.99) for HbA1c 53-62 mmol/mol (7.0-7.8%), 1.05 (0.97-1.13) for HbA1c 63-72 mmol/mol (7.9-8.7%), 1.14 (1.04-1.25) for HbA1c 73-82 mmol/mol (8.8-9.7%), and 1.52 (1.37-1.68) for HbA1c >82 mmol/mol (9.7%). In the cubic spline model, a reduction of the adjusted risk was observed within the lower HbA1c range until 53 mmol/mol (7.0%), with a hazard ratio of 0.78 (0.73-0.82) per SD; it increased thereafter (P for nonlinearity <0.001). As compared with patients without sepsis, the adjusted hazard ratio for death among patients with sepsis was 4.16 (4.03-4.30). CONCLUSIONS In a nationwide cohort of individuals with type 2 diabetes, we found a U-shaped association between HbA1c and sepsis and a fourfold increased risk of death among those developing sepsis.