Arterial blood gas derived variables as estimates of intrapulmonary shunt in critically ill children.

Oxygen transport data, prospectively collected from 52 critically ill children, were analyzed to determine whether any derived variable accurately estimated intrapulmonary shunt (Qsp/Qt). Arterial hemoglobin saturation was more closely correlated with Qsp/Qt than was PaO2, alveolar-arterial oxygen gradient, arterial mixed venous oxygen difference (C[a-v]O2), arterial/alveolar oxygen ratio, and the ratio of PaO2 to inspired oxygen (FIO2) (r = 0.8, p less than .0001). When C(a-v)O2 was normal, hemoglobin saturation became a very accurate (r = 0.96) assessment of Qsp/Qt. We conclude that various arterial blood gas derived variables do not accurately reflect Qsp/Qt in critically ill children. In these patients, a pulmonary artery catheter is needed to accurately assess intrapulmonary shunt.