Objective: The objective of the study was to determine whether serum bicarbonate (HCO₃) concentration can accurately predict venous pH in the evaluation of diabetic ketoacidosis (DKA).
Methods: A retrospective review of patients who presented to a children's hospital emergency department and received an International Classification of Diseases, Ninth Revision code related to DKA or diabetes mellitus was performed. To be eligible for inclusion and data abstraction, patients had blood sampled simultaneously for venous blood gas and metabolic panel. A linear regression model was created using pH (dependent variable) and HCO₃ (predictor). The diagnostic performance and accuracy of HCO₃ to discriminate abnormal pH were evaluated using receiver operating characteristic curve analysis.
Results: Three hundred patients met the inclusion criteria. The linear relationship between pH and HCO₃ using the Pearson correlation coefficient was found to be R = 0.89 (confidence interval [CI], 0.83-0.95; R = 0.79). Receiver operating characteristic curve analysis that maximized sensitivity and specificity demonstrated that a HCO₃ 18.5 or less predicts pH less than 7.3 (area under the curve = 0.97; CI, 0.94-0.99; sensitivity, 93%; specificity, 91%), and a HCO₃ 10.5 or less predicts pH less than 7.1 (area under the curve = 0.97; CI, 0.95-0.99; sensitivity, 97%; specificity, 88%).
Conclusions: Serum bicarbonate accurately predicts abnormal venous pH in children with DKA. Venous pH determination may not be necessary for all patients being evaluated for DKA.