Formulation of a predicting index of survival following intracardiac surgery in children.

Fifteen infants and children were studied after intra-cardiac surgery as a pilot project, in an attempt to determine whether an index can be derived from the mixed venous oxygen tension (PvO2) that incorporates the duration and magnitude of the change form the normal PvO2. Two of the fifteen patients died (from apparent acute cardiac failure) both with atrio-ventricular defects. Pertinent data were recorded every thirty minutes for six hours, including cardiac index (CI), PvO2, arterial blood gases, and cardiac filling pressures. The PvO2 in all subjects were markedly inconsistent. The tow non-survivors did not demonstrate any significant differences between each of their thirty minute PvO2’s as compared with the means of the total group. Two survivors’ PvO2 measurements were less than thirty torr for part or all of the six hour study period. However, the means of the group as a whole ranged from 29.8 torr to 32.4 torr for each measurement. The CI was equally erratic. In two subjects, both the Total Anomalous Pulmonary Venous Return (TAPVR) the CI remained less than two liters per minute. At the time of the measurements, all other clinical parameters did not correlate with these low output measurements. An attempt was made to match the measurements of each subject with the prediction isobars developed by Parr, Blackstone, and Kirklin (1975). In particular, several subjects with low PvO2’s and CI survived against the probabilities defined by Parr, and on subject died, although the odds were in favor of survival. The mean arterial pressure (MAP) of each subject varied, but there was no correlation between MAP, and CI, PvO2 or survival. A positive correlation was noted between core temperature and PvO2. The possible or residual intracardiac shunt causing falsely elevated PvO2’s was considered but not verified. The author concludes that multiple variables may affect the oxygenation as reflected by the PvO2, and in order to devise a predictive index, a population of at least sixty subject must be studied in order to statistically account of each variable.