The Inadequate Oxygen Delivery Index and Its Correlation with Venous Saturation in the Pediatric Cardiac Intensive Care Unit.

Abstract

Continuous monitoring software, T3, has an integrated index called the inadequate oxygen delivery index 50% (IDO2-50) which displays a probability that the mixed venous saturation is below a user-selected threshold of 30-50%. The primary aim of this study was to determine the correlation of the IDO2-50 with a measured venous saturation. The secondary aim of this study was to characterize the hemodynamic factors that contributed to the IDO2-50. This single-center, retrospective study aimed to characterize the correlation between IDO2-50 and inferior vena cava (IVC) saturation. A Bayesian Pearson correlation was conducted to assess the correlation between the collected variables of interest, with a particular interest in the correlation between the IDO2-50 and the IVC saturation. Receiver operator curve (ROC) analysis to assess the ability of the IDO2-50 to identify when the venous saturation was less than 50%. Bayesian linear regression was done with the IDO2-50 (dependent variable) and other independent variables. A total of 113 datasets were collected across 15 unique patients. IDO2-50 had moderate correlation with the IVC saturation (correlation coefficient - 0.569). The IDO2-50 had a weak but significant correlation with cerebral near-infrared spectroscopy (NIRS) values, a weak but significant correlation with heart rate, and a moderate and significant correlation with arterial saturation. ROC analysis demonstrated that the IDO2-50 had a good ability to identify a venous saturation below 50%, with an area under the curve of 0.797, cutoff point of 24.5 with a sensitivity of 81%, specificity of 66%, positive predictive value of 44%, and negative predictive value of 91%. Bayesian linear regression analysis yielded the following model: 237.82 + (1.18 × age in months) - (3.31 × arterial saturation) - (1.92 × cerebral NIRS) + (0.84 × heart rate). The IDO2 index has moderate correlation with IVC saturation. It has good sensitive and negative predictive value. Cerebral NIRS does appear to correlate better with the underlying venous saturation than the IDO2 index.