[The estimation of p50 by the algorithm proposed by Siggard-Andersen. Its experimental assessment in critical...

To determine the ability of O. Siggaard-Andersen algorithm in the estimation of the "in vivo" p50 and standard p50 values from a single blood sample with sO2% less than or equal to 97.

comparison between measured and calculated standard p50 values.

Intensive care unit.

thirteen cardio-pulmonary critical ill patients. Mean age of seventy-four years (range 53-84 years).

The experimental measurement of p50 standard (p50st sper) was performed tonometering the venous blood samples (60 specimens) using an IL-237 tonometer at 37 degrees C, with two different gas mixtures to obtain pCO2 at 5.33 kPa (DS = 0.06), and pO2 at levels to achieve sO2% values close to 50%. The gases's complete equilibration was not deemed important. The pO2 values were corrected to a pH of 7.40 using a Bohr factor = -0.48 and the p50 was taken by simple interpolation of points on the sO2%/pO2 diagram. Calculated standard p50 (p50st calc) and calculated "in vivo" p50 on the venous specimens (No. 60) and the correspondent arterial specimens with sO2% less than or equal to 97 (No. 40) were obtained by Siggaard-Andersen's computerized algorithm. Blood specimen analysis was performed by means of an ABL3 Radiometer gas analyzer and an OSM3 Radiometer oximeter. Statistical analysis was made by Anova test for liner regression.

There was excellent correlation between the 60 experimental p50st determined by Siggaard-Andersen's oxygen dissociation curve on the same blood samples. The regression equation was: p50st sper = -0.79 + 1.21 x p50st calc, r = 0.90, R2 = 81.1%; with F = 249.5 and less than 10(-5). No good correlation was found between p50st and standards p50 calculated on arterial specimens (p50st calc art): p50 = 1.38 + 0.52 x p50st calc art, r = 0.52, R2 = 26.6%, F = 14 e P less than 10(-3). Regression of in vivo P50 calculated on correspondent venous samples (p50 ven) was: p50 ven = 0.79 = 0.77 x p50 art, r = 0.93, R2 = 87.2%, F = 256 and P less than 10(-5).

Our results suggest that the curve describes the curve also at high saturation when it is not longer linear. Accurate measurement (including dishemoglobin percentage) and sO2% less than or equal to 97 are necessary. We did not perform experimental measurements of "in vivo" p50 but we postulate that as the p50st was well calculated so too would be the p50 "in vivo" at 37 degrees C.