[Veno-arterial difference of carbon dioxide as a predictor of low cardiac output in an experimental pediatric model]

Cardiac output (CO) measurement is not a standard of care for critically ill children, but it can be estimated by indirect methods such as veno-arterial pCO2 difference (ΔVACO2).

To determine the correlation between CO and ΔVACO2 and evaluate the usefulness of ΔVACO2 in the diagnosis of low CO in an experimental pediatric model.

Thirty piglets weighing 4.8 ± 0.35 kg were anesthetized and monitored with transpulmonary thermodilution. Lung injury was induced with tracheal instillation of Tween 20®. Serial measurements of central venous and arterial blood gases, as well as CO, were obtained at baseline, 1, 2 and 4 h after lung injury induction. Low cardiac output (LCO) was defined as CO lower than 2.5 Llminlm².

There was an inverse correlation between CO and ΔVACO2 (r = -0.36, p < 0.01). ΔVACO2 was 14 ± 8 mmHg in LCO state and 8 ± 6 mmHg when this condition was not present (p < 0.01). Area under the receiver operating characteristic (ROC) curves of ΔVACO2 and LCO state was 0.78 (0.68-0.86). The best cut-point was 8.9 mmHg to determine LCO with a sensibility 0.78, specificity 0.7, positive predictive value 0.27 and negative predictive value 0.96.

In this model there was an inverse correlation between ΔVACO2 and CO. The best cutoff value to discard LCO was ΔVACO2 of 8.9 mmHg, indicating that under this value the presence of LCO is very unlikely.