[The diagnostic value of serum beta-hydroxybutyrate in diabetic ketosis or diabetic ketoacidosis].

ABSTRACT

OBJECTIVE: To determine the diagnostic value of serum beta-hydroxybutyrate (betaOHB) in diabetic ketosis or diabetic ketoacidosis. METHODS: We conducted a retrospective review of clinical data, in West China Hospital from May 2011 to May 2013, of 1 209 patients with non-ketosis diabetics (DM group), 262 patients with diabetic ketosis or diabetic ketoacidosis (DK/DKA group), and 480 healthy people undergoing routine medical examinations (normal control group). Logistic regression analyses and ROC curves were performed in determining the diagnostic value of betaOHB for DK/DKA. RESULTS: The level of serum betaOHB was much higher in the DK/DKA patients than that of the participants in the DM group and normal control group (P < 0.01). The serum betaOHB turned negative earlier than urine ketone (P < 0.01) in the DK/DKA patients. The logistic regression analysis indicated that betaOHB was one of the independent risk factors for DK/DKA. The betaOHB had an area under of 0.975 in ROC curve, with 1 mmol/L [sensitivity (Sen.) 85.1%, specificity (Spe.) 95. 3%, positive predictive value (PV+) 80.36%, negative predictive value (PV-) 96.89%] as a diagnostic point for DK/DKA and 0.66 mmol/L (Sen. 95%, Spe. 89.2%, PV+ 66.41%, PV- 99.9%) as a screening point. CONCLUSION: Diabetic patients with a level or higher than 1 mmol/L serum betaOHB can accurately predict DK/DKA. Patients with a level or lower than 0.66 mmol/L serum betaOHB are unlikely to have DK/DKA.