Effect of Paco2 variation on standard base excess value in critically ill patients

PurposeThe aim of this study was to investigate the impact of acute Paco2 temporal variation on the standard base excess (SBE) value in critically ill patients.MethodsA total of 265 patients were prospectively observed; 158 were allocated to the modeling group, and 107 were allocated to the validation group. Two models were developed in the modeling group (one including and one excluding Paco2 as a variable determinant of SBE), and both were tested in the validation group.ResultsIn the modeling group, the mathematical model including SIDai, SIG, l-lactate, albumin, phosphate, and Paco2 had a predictive superiority in comparison with the model without Paco2 (R2 = 0.978 and 0.916, respectively). In the validation group, the results were confirmed with significant F change statistics (R2 change = 0.059, P < .001) between the model with and without Paco2. A high correlation (R = 0.99, P < .001) and agreement (bias = −0.25 mEq/L, limits of agreement 95% = −0.72 to 0.22 mEq/L) were found between the model-predicted SBE value and the SBE calculated using the Van Slyke equation.ConclusionsAcute Paco2 temporal variation is related to SBE changes in critically ill patients.