PubMed1995
The p50 and derived indexes, calculated by using the Siggaard-Andersen algorithm from a single measurement of arterial blood gas tensions and hemoglobin-oxygen saturation, are used to assess tissue oxygen availability in critical illness. We tested the accuracy of the Siggaard-Andersen p50 algorithm over a wide range of pathophysiologic conditions. Blood gases, cooximetry, and calculation of standard and in vivo p50 were performed at multiple saturations, CO2 tensions, and H+ concentrations on blood with normal (standard p50 of 26.1 and 26.7 mm Hg), increased (19.0 and 25.4), and reduced (33.9 and 38.2) hemoglobin-oxygen affinity, as well as on high-affinity blood from two patients with diabetic ketoacidosis (16.7 and 20.8). Log p50 in vivo/pH plots were constructed to determine the Bohr effect. Except in the normal affinity specimens (coefficient of variation < 1.7%), standard p50 values showed high variability (coefficient of variation > 5.9%), with saturation-linked bias and distortion of the Bohr effect. Standard p50 was overestimated by up to 11 mm Hg as saturation approached 97%. Although base deficit correction of the stored specimens (6.9 < pH < 7.1) restored the Bohr effect and improved the accuracy of standard p50 calculations (coefficient of variation = 4.4% and 2.9%), saturation-linked bias persisted. We conclude that Siggaard-Andersen p50 calculations may be misleading when there are disturbances of hemoglobin-oxygen affinity and acid-base balance, owing to changes in shape of the hemoglobin-dissociation curve. When metabolic acidosis occurs with high hemoglobin-oxygen affinity, as can occur in critical illness, indexes derived by the Siggaard-Andersen algorithm on arterial blood may greatly overestimate oxygen availability.