The Journal of ExtraCorporeal Technology No 57-2 | Page 10

A. G. Beshish et al.: J Extra Corpor Technol 2025, 57, 59 – 65 61
Table 1. Patient demographics and clinical characteristics of entire VV-ECLS cohort stratified by median PaO 2 levels in the first 48 h in to non-hyperoxia group( PaO 2 122 mmHg) and hyperoxia group( PaO 2 > 122 mmHg).
Variables
Total cohort( n = 110)
Non-hyperoxia group( PaO 2 122 mmHg)( n = 87)
Hyperoxia group( PaO 2 > 122 mmHg)( n = 23) p-value
Age( months)
4.9( 0.1, 105.4)
1.6( 0.0, 81.7)
105.4( 3.4, 154.4)
0.001
Age group
0.003
Neonatal
63( 57.3 %)
56( 64.4 %)
7( 30.4 %)
Pediatrics
47( 42.7 %)
31( 35.6 %)
16( 69.6 %)
Weight( kg)
5.4( 3.4, 35.0)
4.3( 3.2, 22.0)
28.3( 5.0, 81.2)
0.002
Height( cm)
69.0( 51.0, 139.0)
56.0( 51.0, 127.0)
128.0( 73.4, 170.0)
0.005
BSA( m 2)
0.4( 0.2, 1.2)
0.3( 0.2, 1.1)
1.4( 0.4, 2.0)
0.005
Race
0.230
Black
57( 51.8 %)
41( 47.1 %)
16( 69.6 %)
White
38( 34.5 %)
32( 36.8 %)
6( 26.1 %)
Hispanic
12( 10.9 %)
11( 12.6 %)
1( 4.3 %)
Other
3( 2.7 %)
3( 3.4 %)
0( 0.0 %)
Sex
0.194
Female
58( 53.2 %)
43( 50.0 %)
15( 65.2 %)
Male
51( 46.8 %)
43( 50.0 %)
8( 34.8 %)
ECLS indication
0.001
Cardiac
8( 7.3 %)
4( 4.6 %)
4( 17.4 %)
ECPR
8( 7.3 %)
3( 3.4 %)
5( 21.7 %)
Pulmonary
94( 85.5 %)
80( 92.0 %)
14( 60.9 %)
Time from admission to ECLS Initiation
39.0( 3.0, 116.0)
24.0( 2.5, 102.0)
40.0( 3.0, 116.0)
0.560
Initial ECLS flow( L / min)
0.5( 0.4, 2.8)
0.4( 0.4, 1.8)
2.9( 0.7, 4.5)
0.001
Mand PaO 2( SD)
97.7( 46.64)
79.2( 23.4)
167.5( 46.8)
< 0.001
Median PaO 2( mmHg)
86.6( 62.9, 114.6)
77.4( 59.9, 100.5)
159.9( 129.7, 202.2)
< 0.001
PaO 2 range( mmHg)
37.2 – 277.8
37.2 – 120.8
122.5 – 277.8
< 0.001
Number of PaO 2 samples per patient Mean( SD)
13.3( 4.21)
12.5( 3.97)
16.2( 3.87)
< 0.001
Median( IQR)
12.0
12.0
16.0
< 0.001
Duration of ECLS run( hours)
140.5( 98.0, 287.0)
146.0( 98.0, 289.0)
136.0( 89.0, 205.0)
0.549
ECLS complications Cardiovascular
40( 40.8 %)
28( 37.3 %)
12( 52.2 %)
0.203
Hemorrhagic
35( 35.7 %)
19( 25.3 %)
16( 69.6 %)
0.0001
Mechanical
57( 58.2 %)
43( 57.3 %)
14( 60.9 %)
0.764
Renal
67( 68.4 %)
48( 64.0 %)
19( 82.6 %)
0.093
Neurologic
14( 14.3 %)
11( 14.7 %)
3( 13.0 %)
0.846
Metabolic
20( 20.4 %)
15( 20.0 %)
5( 21.7 %)
0.856
Infection
3( 3.1 %)
3( 4.0 %)
0( 0.0 %)
0.330
Reason for coming off Died or poor prognosis
21( 19.1 %)
11( 12.6 %)
10( 43.5 %)
0.001
ECLS complication
6( 5.5 %)
6( 6.9 %)
0( 0.0 %)
Expected recovery
80( 72.7 %)
69( 79.3 %)
11( 47.8 %)
VAD
3( 2.7 %)
1( 1.1 %)
2( 8.7 %)
AKI stage II / III
61( 64.9 %)
50( 66.7 %)
11( 57.9 %)
0.474
Mortality
29( 26.4 %)
17( 19.5 %)
12( 52.2 %)
0.002
Results depicted in n(%) and Median( interquartile range / IQR). ECLS: Extracorporeal Life Support; VV: Veno-veno; VAD: Ventricular Assist Device; AKI: Acute Kidney Injury.
PaO 2 to mortality and other outcomes. We describe the mean, median and range as well as number of samples for outcomes such as mortality, ECLS complications and Stage II or III AKI in Supplemental Table 1.
Cut-point analysis
Using ROC analysis, PaO 2 > 122 mmHg had the optimal discriminatory ability for operative mortality with a sensitivity of 41 %, and specificity of 86 % and was defined as hyperoxia for this study population( Figure 1). Area under the curve for
PaO 2 to predict mortality was 0.53( 95 % CI: 0.38, 0.68, p = 0.962). Patients in the hyperoxia group were older [ 105.4 months( IQR: 3.4, 154.4) vs. 1.6( IQR: 0.0, 81.7), p = 0.001 ], more likely to be in the pediatric group vs. neonates 69.6 % vs. 30.4 %, p = 0.003, weighed more 28.3 kg( IQR: 5.0, 81.2) vs. 4.3( IQR: 3.2, 22.0), p = 0.002, and had higher BSA [ 1.4 kg / m 2( IQR: 0.4, 2.0) vs. 0.3( 0.2, 1.1), p = 0.005 ]. Additionally, patients with hyperoxia had a higher rate of hemorrhagic complications while on ECLS 69.6 % vs. 25.3 %, p = 0.0001, and had a higher mortality rate 52.2 % vs. 19.5 %, p = 0.002( Table 1, Figure 2).