M.-H. Lee and T. Rosenthal: J Extra Corpor Technol 2026, 58, 43 – 50 47
before the first in vivo calibration. Almost a perfect linear correlation exists between the calculated in-line PaO 2 and FiO 2 at the 1st blood gas series( R = 0.96; Figure 1D). However, this correlation diminished significantly following the first in vivo calibration, showing low correlation values in the subsequent blood gas series, specifically R = 0.11 at the 2nd series and R = 0.18 at the 3rd series( Data not shown).
Utilizing this very strong linear correlation, we derived a formula to predict FiO 2 to achieve a predicted in-line PaO 2 at the 1st blood gas series:
Predicted FiO 2 ¼ 0:16 predicted inline PaO 2 � 9:
Then, by using our predicted in-line PaO 2 formula, we can derive another formula to predict FiO 2 to achieve a target PaO 2 at the 1st blood gas series:
Predicted FiO 2 ¼ 0:16 ðtarget PaO 2 þ 13 Weight þ 40Þ�9 ¼ 0:16 target PaO 2 þ 2:1 Weight � 3:
For example, to achieve a target PaO 2 of 250 mmHg with a 10 kg patient at the 1st blood gas series, the predicted FiO 2 is 58 %, which is expected to result in the calculated in-line PaO 2 of approximately 420 mmHg and the measured PaO 2 of around 250 mmHg.
Calculated in-line PaO 2 drifts upward significantly during the rewarming and rewarmed periods
We showed that after in vivo calibrations, the calculated inline PaO 2 became acceptable during the cooling and cooled periods. However, the errors were significantly overestimated again during the rewarming and rewarmed periods. At least two blood gases( the 4th and 5th blood gas series) of 51 patients were measured during the rewarming and rewarmed periods. The errors in mmHg at the 4th and 5th blood gases were calculated separately and combined. The combined errors of calculated in-line PaO 2 were over-estimated in 84 % of the patients( Table 2) with the average % Error of 50.5 ± 59.5 %( Table 1), which is beyond the acceptable target. These large % Error and SD indicate that the calculated in-line PaO 2 drifted upward significantly during the rewarming and rewarmed periods.
The average combined error in mmHg was 82.3 mmHg and SD was ± 77.6 mmHg( Table 1). Over-estimation higher than 100 mmHg occurred in 43 % of the patients( Table 2). As shown in Figure 2, there is a moderate linear correlation( R = 0.46) between the combined error in mmHg at the 4th and 5th blood gas series and patient weight.
Calculated in-line PaCO 2 is acceptable without in vivo calibration during the cooling and cooled periods
We found that the error of calculated in-line PaCO 2 is acceptable at the 1st blood gas series without in vivo calibration and at the 2nd and 3rd blood gas series during the cooling and cooled periods(± 5 mmHg or ± 8 % greater; Table 3). This is likely due to the fact that the in-line PaCO 2 calculation is
Figure 2. The calculated in-line PaO 2 of Quantum System drifts upward during the rewarming and rewarmed periods, showing a moderate correlation to patient weight. A scattered XY plot was drawn for patient weight( X-axis) and combined error in mmHg of the 4th and 5th blood gas series( Y-axis). The data was fitted into a linear regression line, showing a moderate correlation. R value is shown in the upper right corner.
largely dependent on the actual measurement of FeCO 2 of the oxygenator. Nevertheless, SD is largest at the 1st blood gas series and becomes smaller at the 2nd and 3rd blood gas series with each in vivo calibration( Table 3). Errors between 0 and 5 mmHg were observed in 73.2 % of the patients at the 3rd blood gas series, compared to 18.5 % and 42.6 % at the 1st and 2nd blood gas series, respectively( Table 4).
Calculated in-line PaCO 2 drifted upward during the rewarming period, correlating with the temperature gradient
While the calculated in-line PaCO 2 remained acceptable during the cooling and cooled periods, it drifted upward during the rewarming period. At the 4th blood gas series, which is the first blood gas measured after the rewarming was initiated, the error of the calculated in-line PaCO 2 was 4.4 ± 5.4 mmHg or 11.5 ± 15.7 %( Table 3). At the 4th blood gas series, 35 % of the patients had errors higher than 5 mmHg compared to 0 % at the 3rd blood gas series( Table 4). The error became acceptable at the 5th blood gas series following the in vivo calibration after the 4th blood gas series( Table 3).
Notably, the errors of calculated in-line PaCO 2 showed very weak or weak correlations with patient weight across all blood gas series( see R values in Table 4). However, a moderate correlation was identified between the error in mmHg and temperature gradient, defined as the difference between the patient’ s nasopharyngeal temperature at the 4th blood gas series and nadir T during CPB, with an R value of 0.46( Figure 3).
Discussion
In this retrospective study, we demonstrated that the calculated in-line PaO 2 of the Quantum System with the FX05 oxygenator is unacceptable at the 1st blood gas series. The average error was 117 mmHg, and the average % Error