A comparison of central venous-arterial and mixed venous-arterial carbon dioxide tension gradient in circulatory failure

The arterial and mixed venous carbon dioxide tension gradient has been shown to increase when there is a decrease in cardiac output. Monitoring central venous gases is an attractive alternative to monitoring mixed venous gases in circulatory failure because central venous catheterisation is a less invasive procedure than pulmonary artery catheterisation. This study aims to evaluate the agreement between central venous-arterial carbon dioxide (CVA-CO2) and mixed venous-arterial carbon dioxide (SVA-CO2) tension gradients and assess whether CVA-CO2 tension gradient can be used to predict cardiac output in circulatory failure. Samples of arterial, central venous and mixed venous blood were obtained from 16 patients with circulatory failure at different inspired oxygen concentrations and cardiac indexes within 24 hours of study enrolment. CVA-CO2 and SVA-CO2 tension gradient were not interchangeable numerically (bias=0.14 mmHg, 95% limits of agreement: -3.0 to 3.2 mmHg). CVA-CO2 (Spearman correlation coefficient r=-0.385) and SVA-CO2 (r=-0.578) tension gradient were significantly correlated with the cardiac index but the cardiac index only accounted for 21% and 32% of the variability of CVA-CO2 and SVA-CO2 tension gradient, respectively. The ability of CVA-CO2 tension gradient (area under the ROC curve=0.77, 95% confidence interval [CI]: 0.49-0.99; P=0.08) to predict a low cardiac output state (cardiac index <2.5 l/min/m(2)) was lower than SVA-CO2 (area under the ROC curve=0.95, 95%CI: 0.88-0.99; P=0.003). The utility of CVA-CO2 and SVA-CO2 tension gradient appeared to be limited to their negative predictive value to exclude a low cardiac output state when CVA-CO2 or SVA-CO2 tension gradient was normal (<5 mmHg).