Systematic instrumental errors between oxygen saturation analysers in fetal blood during deep hypoxemia

During a study of artificially produced deep hypoxemia in fetal cord blood, systematic errors of three different oxygen saturation analysers were evaluated against a reference CO oximeter.

The oxygen tensions (PO2) of 83 pre-heparinized fetal blood samples from umbilical veins were reduced by tonometry to 1.3 kPa (10 mm Hg) and 2.7 kPa (20 mm Hg). The oxygen saturation (SO2) was determined (n=1328) on a reference CO oximeter (ABL625, Radiometer Copenhagen) and on three tested instruments (two CO oximeters: Chiron865, Bayer Diagnostics; ABL700, Radiometer Copenhagen, and a portable blood gas analyser, i-STAT, Abbott). The CO oximeters measure the oxyhemoglobin and the reduced hemoglobin fractions by absorption spectrophotometry. The i-STAT system calculates the oxygen saturation from the measured pH, PO2, and PCO2. The measurements were performed in duplicate. Statistical evaluation focused on the differences between duplicate measurements and on systematic instrumental errors in oxygen saturation analysis compared to the reference CO oximeter.

After tonometry, the median saturation dropped to 32.9% at a PO2=2.7 kPa (20 mm Hg), defined as saturation range 1, and to 10% SO2 at a PO2=1.3 kPa (10 mm Hg), defined as range 2. With decreasing SO2, all devices showed an increased difference between duplicate measurements. ABL625 and ABL700 showed the closest agreement between instruments (0.25% SO2 bias at saturation range 1 and -0.33% SO2 bias at saturation range 2). Chiron865 indicated higher saturation values than ABL 625 (3.07% SO2 bias at saturation range 1 and 2.28% SO2 bias at saturation range 2). Calculated saturation values (i-STAT) were more than 30% lower than the measured values of ABL625.

The disagreement among CO oximeters was small but increasing under deep hypoxemia. Calculation found unacceptably low saturation.