Clinical utility of serum biochemical variables for predicting acid-base balance in critically ill horses

Background: Profiles from serum biochemical analyzers include the concentration of strong electrolytes (including L-lactate), total carbon dioxide (tCO(2)), and total protein. These variables are associated with changes in acid-base balance. Application of physicochemical principles may allow predicting acid-base balance from serum biochemistry without measuring whole blood pH and pCO(2). Objectives: The purpose of the study was to determine if the acid-base status of critically ill horses could be accurately predicted using variables included in standard serum biochemical profiles. Methods: Two jugular venous blood samples were prospectively obtained from critically ill horses and foals. Samples were analyzed using a whole blood gas and pH analyzer (BG) and a serum biochemistry multi analyzer system (AMAS). Linear regression, Deming regression, and Bland-Altman plots were used for method comparison and P < .05 was considered significant. Results: Values from 70 horses and foals for Na, K, Cl, and total protein concentrations, and consequently the calculated variables used for acid base interpretation, were different between the AMAS and BG analyzer. Using physicochemical principles, BG results accurately predicted pH, whereas the AMAS results did not when a fixed value for pCO(2) was used. Conclusions: Measurement of pCO(2) is required in critically ill horses for accurate prediction of whole blood pH. Differences in the measured values of Na and Cl concentration exist when measured in serum by the AMAS and in whole blood or plasma by BG, indicating that the accurate prediction of whole blood pH is analyzer-dependent. Application of physicochemical principles to plasma or serum provides a practical method to evaluate analyzer accuracy.