Epidemiology, microbiology, and diagnosis of infection in diabetic ketoacidosis and hyperosmolar hyperglycemic syndrome: A multicenter retrospective observational study

Aims We investigated the characteristics of infection and the utility of inflammatory markers in diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic syndrome (HHS). Methods A multicenter, retrospective observational study in 21 acute-care hospitals was conducted in Japan. This study included adult hospitalized patients with DKA and HHS. We analyzed the diagnostic accuracy of markers including C-reactive protein (CRP) and procalcitonin (PCT) for bacteremia. Multiple regression models were created for estimating bacteremia risk factors. Results A total of 771 patients, including 545 patients with DKA and 226 patients with HHS, were analyzed. The mean age was 58.2 (SD, 19.3) years. Of these, 70 tested positive for blood culture. The mortality rates of those with and without bacteremia were 14 % and 3.3 % (P-value < 0.001). The area under the curve (AUC) of CRP and PCT for diagnosis of bacteremia was 0.85 (95 %CI, 0.81 to 0.89) and 0.76 (95 %CI, 0.60 to 0.92), respectively. Logistic regression models identified older age, altered level of consciousness, hypotension, and higher CRP as risk factors for bacteremia. Conclusions The mortality rate was higher in patients with bacteremia than patients without it. CRP, rather than PCT, may be valid for diagnosing bacteremia in hyperglycemic emergencies.Trial Registration: This study is registered in the UMIN clinical trial registration system (UMIN000025393, Registered December 23, 2016)