Operation Everest II: Comparison of four instruments for measuring blood O₂ saturation

The bias and precision of four different methods for determining O₂ saturation (SO₂) were evaluated during a study of hypobaric hypoxia conducted with seven male subjects exposed progressively over a 40-day period to simulated altitudes from sea level (760 Torr) to 8,840 m (240 Torr). SO₂ of arterial and mixed venous blood samples were measured with the Instrumentation Laboratory 282 CO-oximeter (CO-OX), the Radiometer ABL-300 (ABL), and the Lex-O₂-Con-K (LEX). Noninvasive measurements of arterial SO₂ were made with a Hewlett-Packard 47201A ear oximeter (EAR-OX). The CO-OX method was used as a secondary standard for comparison with the other methods because it has been validated against the classical Van Slyke method over a wide physiological range (Maas et al., Clin. Chim. Acta 29: 303-309, 1970). The LEX results most closely approximated but consistently underestimated those of the CO-OX: LEX = 0.93 CO-OX -0.86, standard error of the estimate (SEE) = 5.17, r = 0.98, n = 670. The ABL method appeared to combine two linear trends: for SO₂ > 75%, ABL = 0.84 CO-OX +14.4, SEE = 1.77, r = 0.97, n = 369; < 75%, ABL = 0.98 CO-OX +5.9, SEE = 4.44, r = 0.97, n = 486. The EAR-OX results were found to approximate those of the CO-OX at SO₂ values only > 65%: EAR-OX = > 1.07 CO-OX -6.12, SEE = 7.71, r = 0.98, n = 326. In a separate experiment, the CO-OX, LEX, and ABL methods were used to determine the SO₂ of human blood previously tonometered with eight gases of known O₂ concentrations to produce saturations from 13.3 to 99.7%. This test confirmed the CO-OX as being the most accurate and precise of the three methods. Of the other two, the ABL was the most precise. The LEX although of lesser precision had the better accuracy, approximating that of the CO-OX.