Limitations in predicting oxyhemoglobin saturation of dog blood from PO2

This study compared oxyhemoglobin saturation (SO2) and O2 content (CO2) estimated from O2 tension (PO2) by the Rossing-Cain nomogram (J. Appl. Physiol. 21: 195-201, 1966) with SO2 and CO2 estimated by a galvanometric O2 analyzer in blood samples from eight dogs. The nomogram consistently and significantly overestimated SO2 over the range of 20-60%. The greatest absolute difference, which averaged 10% saturation, was between 40 and 59% saturation. Between 30 and 39% saturation, the difference averaged 30% of SO2 estimated galvanometrically. CO2, calculated as the product of SO2, hemoglobin concentration (cyanmethemoglobin method), and hemoglobin O2 capacity, was significantly overestimated by the nomogram by as much as 1.2 ml/dl between 2 and 9.9 ml/dl. Between 14 and 21.9 ml/dl, the nomogram underestimated CO2 by as much as 1.2 ml/dl. We conclude that because coronary venous SO2 and CO2 values normally lie in the range of greatest error, estimates of these values based on PO2 are particularly unsuited for studies of myocardial O2 usage.