The Journal of ExtraCorporeal Technology No 57-3 | Page 21

A. G. Beshish et al.: J Extra Corpor Technol 2025, 57, 129 – 136 131
Table 1. Patient demographics and clinical characteristics of entire VA-ECLS cohort stratified by mean PaO 2 levels in the first 48 h into nonhyperoxia group( PaO 2 233 mmHg) hyperoxia group( PaO 2 > 233 mmHg).
Variables
Total Cohort
Non-hyperoxia group
Hyperoxia group
p-value
( n = 229)
( PaO 2 233 mmHg)( n = 161)
( PaO 2 233 mmHg)( n = 68)
Age( months)
2.5( 0.3, 19.0)
1.5( 0.2, 17.2)
4.6( 0.4, 27.8)
0.019
Age Group
0.345
Neonatal
168( 73.4 %)
121( 75.2 %)
47( 69.1 %)
Pediatrics
61( 26.6 %)
40( 24.8 %)
21( 30.9 %)
Weight( kg)
4.4( 3.2, 10.7)
3.9( 3.1, 9.5)
5.5( 3.3, 11.9)
0.178
Height( cm)
54.0( 49.0, 80.0)
53.0( 49.0, 75.5)
61.8( 50.0, 84.3)
0.115
BSA( m 2)
0.3( 0.2, 0.5)
0.2( 0.2, 0.4)
0.3( 0.2, 0.5)
0.155
Race
0.204
Black
104( 45.4 %)
75( 46.6 %)
29( 42.6 %)
White
87( 38.0 %)
55( 34.2 %)
32( 47.1 %)
Hispanic
26( 11.4 %)
21( 13.0 %)
5( 7.4 %)
Other
12( 5.2 %)
10( 6.2 %)
2( 2.9 %)
Sex
0.410
Female
115( 50.2 %)
78( 48.4 %)
37( 54.4 %)
Male
114( 49.8 %)
83( 51.6 %)
31( 45.6 %)
ECLS indication
0.0004
Cardiac
112( 48.9 %)
71( 44.1 %)
41( 60.3 %)
ECPR
65( 28.4 %)
42( 26.1 %)
23( 33.8 %)
Pulmonary
52( 22.7 %)
48( 29.8 %)
4( 5.9 %)
Time from admission to ECLS Initiation
78.5( 14.0, 356.0)
72.5( 14.0, 411.0)
100.0( 7.0, 306.0)
0.997
Initial ECLS flow( mL / kg / min)
114( 94, 116)
109( 100, 1212)
128( 97, 131)
0.018
Duration of ECLS run( hours)
111.5( 65.5, 184.5)
111.5( 65.5, 195.5)
110.5( 66.5, 161.5)
0.412
ECLS Complications: Cardiovascular
87( 45.8 %)
51( 39.8 %)
36( 58.1 %)
0.018
Hemorrhagic
86( 45.3 %)
58( 45.3 %)
28( 45.2 %)
0.984
Mechanical
75( 39.5 %)
58( 45.3 %)
17( 27.4 %)
0.018
Renal
103( 54.2 %)
68( 53.1 %)
35( 56.5 %)
0.666
Neurologic
46( 24.2 %)
28( 21.9 %)
18( 29.0 %)
0.280
Metabolic
29( 15.3 %)
18( 14.1 %)
11( 17.7 %)
0.509
Infection
9( 4.7 %)
7( 5.5 %)
2( 3.2 %)
0.495
Reason for coming off
0.049
Died or Poor Prognosis
48( 21.1 %)
26( 16.4 %)
22( 32.4 %)
ECLS Complication
16( 7.0 %)
11( 6.9 %)
5( 7.4 %)
Expected Recovery
153( 67.4 %)
116( 73.0 %)
37( 54.4 %)
Unknown
1( 0.4 %)
1( 0.6 %)
0( 0.0 %)
VAD
9( 4.0 %)
5( 3.1 %)
4( 5.9 %)
Missing
2
2
0
AKI Stage II or III
160( 75.8 %)
111( 74.5 %)
49( 79.0 %)
0.483
Mortality
102( 44.5 %)
65( 40.4 %)
37( 54.4 %)
0.050
Results Depicted in n(%), and Median( Interquartile Range / IQR). ECLS: Extracorporeal Life Support; VA: Veno-Arterial; VAD: Ventricular Assist Device; AKI: Acute Kidney Injury. Results
During the study period, 229 VA-ECLS runs occurred in 229 patients. The median age of this cohort was 2.5 months( IQR 0.3, 19.0) with a weight of 4.4 kg( IQR 3.2, 10.7), and an even distribution of males( 49.8 %) and females( 50.2 %). The majority were neonates( 73.4 %), with primary cardiac failure being the most common indication( 48.9 %) for ECLS. The median time from admission to cannulation was 78.5 h( IQR 14, 356) with a median run duration of 111.5 h( IQR 65.5, 184.5). The overall mortality rate was 44.5 %( Table 1). There was a total of 3664 PaO 2 samples obtained for the entire cohort. The range of samples per patient was 11 – 32, and the median number of samples per patient was 17( IQR 14, 21).
Cut-point analysis
Using ROC analysis, PaO 2 > 233 mmHg had the optimal discriminatory ability for mortality with a sensitivity of 36 %, and specificity of 76 %, and was defined as hyperoxia for this study population( Fig. 1). AUC for PaO 2 to predict mortality was 0.55( 95 % CI [ 0.47 – 0.62 ]; p = 0.212). Approximately one-third of the patients were in the hyperoxia group( 68 / 229) with the remainder in the non-hyperoxia group( 161 / 229)( Fig. 2). The two cohorts were similar in demographic variables [ i. e., weight, height, body surface area( BSA), sex, and race / ethnicity ] except for age( Table 1). Patients in the hyperoxia group were older [ 4.6 months( IQR 0.4, 27.8) vs 1.5 months( IQR 0.2, 17.2), p = 0.019 ] and had higher median