All-cause mortality prognostic factors in type 2 diabetes-associated ketosis and ketoacidosis

Abstract

Background: Diabetic ketoacidosis (DKA) represents a life-threatening acute diabetic complication attributed to type 1 diabetes. In the past decades, the presence of DK or even DKA has been increasingly recognized among patients with type 2 diabetes mellitus (T2DM). Aim of the Study: We aimed to analyze characteristics and mortality rate prognostic factors in patients with T2DM presenting with non-ketotic hyperglycemia (NKH), DK, and DKA. Methods: Population-based retrospective study at the Emergency Department of Clinical Hospital Center Sestre Milosrdnice, Zagreb, Croatia, was conducted. DKA was defined as plasma glucose >13.9 mmol/L, ketonuria >2+, and capillary blood bicarbonate levels <18 mmol/L or capillary blood pH <7.30. 137 patients with T2DM presenting with DKA and 137 age- and gender-matched patients with NKH and DK were included in the study. We analyzed general anthropometric characteristics, precipitating factors, signs and symptoms, medications, alcohol, and cigarette consumption. Results: Patients with DKA had significantly higher mortality rates when compared with DK (HR 2.55, 95% CI 1.65–3.92, P < 0.001) and NKH (HR 1.27, 95% CI 1.05–1.53, P = 0.01). Older age, higher blood urea nitrogen, and calcium channel antagonists use were adverse prognostic factors across all groups. Acetylsalicylic acid and insulin in DKA and furosemide use in DK group were associated with increased mortality. Conversely, alcohol and tobacco use was associated with lower mortality in DKA group. These variables predicted 1-year mortality with and accuracy of 81% in DK group and 88% in DKA group. Conclusion: Prognostic factors and all-cause mortality differ between patients with NKH, DK, and DKA. Further, prospective trials should evaluate whether the addressing prognostic factors could improve patient outcomes.