A comparison of two methods of estimating systemic carbon dioxide production during cardiopulmonary bypass

Introduction

The gold standard for assessing the efficacy of cardiopulmonary bypass (CPB) is systemic oxygen uptake (VO₂) as estimated by the Fick Principle [1]. Systemic carbon dioxide production (VCO₂) may also be determined in this way as another estimate of the aerobic efficacy. Both techniques, however, are error prone due to compounding of individual measurement errors during their arithmetical calculation. Also, they are mathematically coupled to predictor variables, such as flow rate, so invalidating statistical analysis of their relationships. VCO₂ may also be estimated as the oxygenator's CO₂ output which should be a robust and independent measure. The aim of this study was to compare the measurement of VCO₂ using the Fick Principle with that by oxygenator carbon dioxide production.

Methods

Blood was aspirated from the arterial and venous lines during CPB at initial cooling, stable hypothermia and rewarming (35°C). Samples were analysed on a blood gas analyser and CO₂ content was then estimated from its partial pressure using Kelman's algorithm [2]. The product of the arterio-venous difference in carbon dioxide content and pump flow rate was used to obtain VCO₂. VCO₂ was also estimated as the product of CO₂ concentration exhausting from, and fresh gas flow through, the oxygenator. Method comparison analysis was used to compare the techniques [3].

Results

A total of 186 measurements were made in 51 patients. The differences between results obtained by the two methods plotted against the mean of the paired results demonstrated poor agreement between the methods (limits -23, 105 ml/min). VCO₂ as predicted by the Fick Principle overestimated that determined by the oxygenator output technique with an average bias of 41 ml/min (95% CI32–50). The bias was proportional (r = 0.75, 95% CI0.55–0.95). Similar relationships between the techniques were found at each individual time point during CPB.

Discussion

Either of the methods could be responsible for the poor agreement. As discussed above, the Fick method is error prone. Also, our assumption that CO₂ can accurately be estimated from its partial pressure may also be invalid in this setting. Thermal gradients across the oxygenator caused by the heat exchanger will have altered the blood solubility of CO₂ so decreasing and increasing oxygenator CO₂ output during cooling and rewarming, respectively. Also, because of the respiratory quotient, the flow of gas exhausting from, will be less than the fresh gas entering, the oxygenator. In conclusion, the two methods of VCO₂ cannot be used interchangeably.