Accuracy of base excess--an in vitro evaluation of the Van Slyke equation

To evaluate the precision, bias and CO2 invariance of base excess as determined by the Van Slyke equation over a wide P(CO2) range at normal and low hemoglobin concentrations.

Prospective in vitro study.

University research laboratory.

Normal human blood, both undiluted and diluted with plasma.

Two experiments were conducted. In the first, blood unmodified or after adding HCl or sodium bicarbonate was rendered hypercarbic (P(CO2) >70 torr) by gas equilibration. Rapid Pco2 reduction in > or =10 steps to a final P(CO2) < or =20 torr was then performed. In the second experiment, blood unmodified or diluted to a hemoglobin concentration of approximately 4 G% was mixed anaerobically (9:1, vol:vol) with varying concentrations of lactic acid in saline (0-250 mmol/L).

In the first experiment, blood gas analysis at each step during the progressive P(CO2) reduction revealed that base excess remained nearly constant (SD all specimens < or =0.6 mmol/L) whereas P(CO2) changed by >80 torr. In the second experiment, simultaneous blood gas and plasma lactate analyses showed that changes in base excess correlated closely with changes in both plasma and whole blood lactate concentrations (r2 > or = 0.91) despite concurrent P(CO2) elevations as great as 200 torr. Quantification by base excess of change in whole blood lactate concentration was precise with slight negative bias (mean negative bias, 1.1+/-1.9 mmol/L) in both diluted and undiluted blood. There was significant underestimation of change in plasma lactate concentration in undiluted blood, presumably because base excess is a whole blood variable.

Base excess calculated using the Van Slyke equation accurately quantifies metabolic (nonrespiratory) acid-base status in blood in vitro. This accuracy is little affected by large simultaneous alterations in P(CO2), or by very low hemoglobin concentrations similar to that used to calculate standard base excess.