A Prediction Model of Mortality in Patients Hospitalized with Diabetic Ketoacidosis in a Tertiary Referral Hospital in Surabaya, Indonesia

Background: The mortality rate of diabetic ketoacidosis (DKA) remains high in Indonesia. Differentpatterns were contributing to DKA mortality from previous studies. We aim to identify clinical and laboratoryfindings as mortality predictors of DKA.Methods: This retrospective study was conducted in a tertiary referral hospital in Surabaya. Medical recordsof DKA patients from January 2016 to December 2018 were analyzed. Clinical and laboratory data wereobtained from medical records for three years period. Clinical outcome was defined as either dischargedhome or death.Results: Fifty-six among 116 DKA patients did not survive during hospitalized. Age of 60 or older, thedepressed mental state with GCS below 13, potassium serum less than 3.5 mmol/L, and bicarbonate serumless than 15 mmol/L are found to be significant as independent factors of mortality in DKA patients.Conclusion: The mortality rate of DKA patients in this study is still high. It independently associatedwith the age of 60 years or older, depressed mental state, hypokalemia, and low bicarbonate level. Earlystratification of these predictors would help to treat patients accordingly.