2981-LB: A Simplified Approach to Diagnosing Diabetic Ketoacidosis

Introduction and Objective: In a recent consensus statement from ADA and other national and international societies, the diagnosis of diabetic ketoacidosis (DKA) requires: 1) a diabetes diagnosis, 2) ≥2+ urinary ketones and/or a beta-hydroxybutyrate (βOHB) level ≥3.0 mM and 3) venous pH 7.30 and/or bicarbonate (HCO3) 18 mEq/L. We conducted a retrospective study to assess these criteria. Methods: We reviewed records from 13,470 encounters in 6091 patients in which admission βOHB and HCO3 had been measured, together with pH and urinary ketones when available. The relationship between βOHB and HCO3was evaluated by regression analysis. Mild, moderate and severe DKA were defined as HCO3 levels of 15-17, 10-14 and 10 mEq/L, respectively as stated in the guidelines. Results: In the whole cohort, an HCO3 of 18 mEq/L corresponded to a βOHB level of 3.9 mM. In a subset where urinary ketones were available, 56% and 29% of results were 2+ at βOHB levels of 3-5 and 4-8 mM, respectively. In mild, moderate and severe DKA, 57%, 24% and 4%, respectively had pH ≥7.30. Among individuals with βOHB ≥3.0 and admission pH ≥7.30, 62% had HCO3 ≥18 mEq/L. When βOHB was ≥3.0 mM and HCO3 ≥18 mEq/L, 86% had pH ≥7.30. In the group that had βOHB ≥3.0 mM, pH ≥7.30 and HCO3 ≥18 mEq/L (individuals who do not qualify for a DKA diagnosis by current criteria), βOHB was ≥4.0 mM in 56% and ≥5.0 mM in 17%. Conclusion: In mild to moderate DKA, discordance between urinary ketones and βOHB is frequent, as is discordance between pH and HCO3. A subset of patients with βOHB ≥4.0 mM do not qualify for a DKA diagnosis. Among these laboratory tests, βOHB is the most specific indicator of DKA, in part because its production is accompanied by equimolar production of hydrogen ions. In contrast, the urinary ketone assay is not quantitative, and both pH and HCO3 are non-specific because they can be affected by other acid-base processes that are frequent in DKA. These include respiratory compensation, hyperchloremic acidosis, and lactic acidosis. We propose simplified diagnostic criteria for DKA consisting of the presence of diabetes and a βOHB ≥3.9 mM. Disclosure G.Arias: None. D.Tilden: n/a. R.Mukherjee: None. A.Jain: None. K.Grdinovac: None. I.B.Hirsch: Consultant; Current; Abbott Diabetes, Roche Diabetes Care, Hagar, Research Support; Current; MannKind Corporation, Sequel Tech. J.M.Miles: None. Funding NIDDK K23DK143333 (DRT)