Reliability of transcutaneous oxygen monitoring of critically ill children in a general pediatric unit.

While transcutaneous continuous monitoring of partial pressure of oxygen (TcPO2) is a proven valuable diagnostic tool in intensive care of sick neonates, its use in older children is less well defined. The purpose of this study is to establish the accuracy and reliability of TcPO2 as a reflection of arterial partial pressure of oxygen (PaO2) in critically ill pediatric patients. Twenty-six children, age ranging from 4 days to 13 years, were studied. The main diagnostic categories were cardiac 54 per cent, respiratory 27 per cent, gastrointestinal 15 per cent, and neurologic 4 per cent. Forty three estimations of PaO2 (by radial arterial sampling) and TcPO2 (by cutaneous monitor) were obtained simultaneously. Their correlation coefficient, regression equation, and confidence limits were calculated by standard statistical methods with the aid of a microcomputer. Our data show that the high degree of correlation between TcPO2 and PaO2 is highly significant (r = 0.90, p less than 0.001) for the full range of PaO2 between 0 to 200 mm Hg. At the lower range of PaO2 (0-79 mm Hg), the degree of correlation (r = 0.91, p less than 0.001) is better than that (r = 0.77, p less than 0.001) at the higher range (80-200 mm Hg), although the difference between the two correlation coefficients is not significant (p greater than 0.05). It is important to note that TcPO2 consistently underestimates the PaO2 and the 95% confidence limits are rather wide. We conclude that TcPO2 is a reasonably accurate reflection of PaO2 in critically ill pediatric patients. Its main advantage is in predicting and showing trends in tissue oxygenation by noninvasive means. However, when critical assessment of the degree of hypoxemia is required, TcPO2 is no substitute for PaO2.