48 M.-H. Lee and T. Rosenthal: J Extra Corpor Technol 2026, 58, 43 – 50
Table 3. Average error in mmHg and % Error of the calculated in-line PaCO 2
|
Avg error( mmHg) |
SD( mmHg) |
Avg % Error(%) |
SD(%) |
Patients( n) |
1st gas series |
�3.6 |
6.4 |
�7.8 |
14.2 |
81 |
2nd gas series |
0.2 |
4.8 |
0.9 |
11.5 |
68 |
3rd gas series |
1.1 |
1.8 |
2.8 |
4.5 |
42 |
4th gas series |
4.4 |
5.4 |
11.5 |
15.7 |
77 |
5th gas series |
�0.7 |
3.5 |
�1.4 |
8.1 |
50 |
4th + 5th gas series |
3.9 |
6.8 |
10.5 |
18.6 |
50 |
Avg: average; SD: standard deviation.
Table 4. Error distribution of the calculated in-line PaCO 2 and the correlation coefficient( R) between the error in mmHg and patient weight.
Error distributions( mmHg) |
1st gas(%) |
2nd gas(%) |
3rd gas(%) |
4th gas(%) |
5th gas(%) |
4th + 5th(%) |
Error0 |
75.3 |
44.1 |
26.8 |
18.2 |
66.0 |
28.0 |
0 < Error5 |
18.5 |
42.6 |
73.2 |
46.8 |
28.0 |
40.0 |
5 < Error10 |
2.5 |
13.2 |
0 |
23.4 |
6.0 |
16.0 |
10 < Error15 |
2.5 |
0 |
0 |
5.2 |
0 |
6.0 |
15 < Error20 |
1.2 |
0 |
0 |
5.2 |
0 |
8.0 |
20 < Error25 |
0 |
0 |
0 |
1.3 |
0 |
2.0 |
R( error vs. weight) $ |
0.22 |
0.17 |
�0.10 |
�0.24 |
�0.14 |
�0.21 |
%: percentage of patients, see Table 3 to find total patient numbers for each blood gas series; $: R value of the linear correlation analysis between the error in mmHg and patient weight.
Table 5. Predicted FiO 2 to achieve the target PaO 2 of 250 mmHg and the predicted in-line PaO 2 corresponding to the predicted FiO 2.
Figure 3. The calculated in-line PaCO 2 of Quantum System drifts upward during the rewarming period, correlating with the temperature gradient. The temperature gradient is defined as the patient’ s nasopharyngeal temperature at the 4th blood gas series minus nadir T during CPB. A scattered XY plot was drawn for the temperature gradient( X-axis) and error in mmHg of the 4th blood gas series( Y-axis). The data was fitted into a linear regression line, demonstrating a moderate correlation. R value is shown in the upper right corner.
was 48.3 %, which exceeds the acceptable target( Table 1). Notably, the calculated in-line PaO 2 at the 1st blood gas series is almost perfectly correlated with FiO 2( Figure 1D), indicating other factors such as the flow, SaO 2, SvO 2, FeCO 2, temperature, and other variables have minimal impact on PaO 2 calculation prior to the first in vivo calibration.
Following the first in vivo calibration, the calculated in-line PaO 2 at the 2nd blood gas series became acceptable, although it
Weight( kg)
Target PaO 2( mmHg)
Predicted FiO 2(%)
Predicted in-line PaO 2( mmHg)
2 |
250 |
41 |
316 |
3 |
250 |
43 |
329 |
4 |
250 |
45 |
342 |
5 |
250 |
48 |
355 |
6 |
250 |
50 |
368 |
7 |
250 |
52 |
381 |
8 |
250 |
54 |
394 |
9 |
250 |
56 |
407 |
10 |
250 |
58 |
420 |
11 |
250 |
60 |
433 |
12 |
250 |
62 |
446 |
13 |
250 |
64 |
459 |
14 |
250 |
66 |
472 |
exhibited a high SD( Table 1). After the second in vivo calibration, the error and SD were reduced, demonstrating improved accuracy and precision with each in vivo calibration. This suggests that it is a better practice to use“ Capture All / Sync” each time with the blood gas analysis when using the Quantum System.
However, when rewarming began, the calculated in-line PaO 2 drifted upward, leading to unacceptable error in mmHg and % Error at two consecutive blood gas series( 4th and 5th in Table 1). In 51 patients, when two consecutive blood gas series were performed during the rewarming and rewarmed periods, the combined average error in mmHg was 82.3 mmHg and the average % Error was 50.5 %( Table 1). This suggests