Keywords
neonate - extracorporeal membrane oxygenation - congenital heart disease - open heart
surgery
Introduction
Extracorporeal membrane oxygenation (ECMO) is an important technology, which can be
used for severe heart failure.[1]
[2] Since the first successful ECMO application in a newborn in 1975,[3] more than 85,000 neonates treated with ECMO have been reported to Extracorporeal
Life Support Organization (ELSO) nowadays. In China, ECMO has been practiced for two
decades, but most patients are adults. As more children undergoing cardiac surgery
may require ECMO support for cardiorespiratory failure,[4] pediatric ECMO applications have increased and reported in recent years. However,
among the numerous studies investigating the utility and risk factors associated with
ECMO in pediatric cardiac patients, few specifically address patients with neonatal
heart disease. In this study, we conducted a retrospective review of neonates who
undergone ECMO at Shanghai Children's Medical Center (SCMC) and summarized the clinical
application of ECMO in the treatment of heart failure after neonatal cardiac surgery.
Materials and Methods
Patient Characteristics
The current study was approved by the Institutional Review Board of the Shanghai SCMC
affiliated to Shanghai Jiaotong University School of Medicine. As the study design
was a retrospective review of collected data, the need for informed consent was waived.
From January 2017 to June 2019, 9,354 cases with congenital heart disease (CHD) were
undergone open heart surgery in SCMC and 441 were neonates, 27 cases required support
postoperatively (26 venoarterial [VA]-ECMO cases and 1 ventricular-assist device case).
The ventricular assist device infant and three children supported with ECMO less than
24 hours were excluded, and 23 patients (15 males and 8 females) were analyzed. Mean
age was 9.08 ± 8.02 days (range: 0–25 days) and mean birth weight was 3,170 ± 535 g
(range: 2,300–4,500 g). The diseases were interrupted aortic arch/coarctation of the
aorta (n = 10), D-transposition of the great artery (n = 8), pulmonary atresia (n = 2), total anomalous pulmonary venous connection (n = 1), double outlet right ventricle (n = 1), and coronary artery–right ventricular fistula (n = 1).
Clinical Data
Successful weaning was defined as weaning without reinsertion of ECMO or death within
24 hours. Clinical variables were collected: indication of ECMO, intensive care unit
details, and complications. The lactate clearance time means that the time required
for the body's lactate value to fall from the maximum to <2 mmol/L.
Complications
Complications were categorized: (1) surgical site bleeding (cannulation or surgical
site bleeding requiring an intervention); (2) brain injury (central nervous system
hemorrhage, infarction causing ischemic injury); (3) gastrointestinal (GI) bleeding
(fecal occult blood test is positive, hematemesis, melena, or bloody stool); (4) mechanical
complication (oxygenator or pump failure, air in circuits); (5) infection (culture
positive infection); (6) residual malformation (residual shunt or obstruction, moderate
or severe reflux).
ECMO Details and Management
In our institution, for all subjects, we recorded indication for therapy (defined
as postcardiotomy shock, fail to transition from cardiopulmonary bypass [CPB], low
cardiac output syndrome [LCOS], extracorporeal cardiopulmonary resuscitation [ECPR],
or other). Neonates were cannulated for ECMO by pediatric cardiovascular surgeons
and underwent cannulation through the right atrium and common ascending aorta. Typically,
8 to 10 Fr cannulas were used for arterial cannulation and 14 to 16 Fr for venous
access.
The standard circuit includes a centrifugal pump and HI-LITE 800LT hollow-fiber membrane
oxygenator. The goal of ECMO treatment is to use sufficient flow to maintain the children's
systemic circulation. Therefore, it is necessary to gradually reduce the dose of positive
inotropic drugs in the initial phase, and maintain a high flow of 130 to 180 mL/kg/min
with the average arterial blood pressure at 40 to 60 mm Hg and the ventilator was
adjusted to rest settings. The doctors adjust the flow according to changes in the
heart function, hemodynamics, lactic acid level, blood gas results, and so on. When
urine output fell to < 1 to 2 mL/kg/h, diuretics were given to promote diuresis. If
neonates experienced severe renal impairment and low urine volume, continuous renal
replacement therapy (CRRT) which was connected to the ECMO circuit was started immediately,
and continuous venovenous hemodiafiltration treatment was used often.
The standard anticoagulation is unfractionated heparin, ranging 5 to 40 units/kg/h
which was administered to keep activated clotting time (ACT) level at 160 to 200 seconds
and activated partial thromboplastin time (APTT) between 50 and 70 seconds. ACT and
APTT levels were monitored every 4 hours, and antithrombin (AT)III, platelet (PLT)
count, hematocrit (HCT) and fibrinogen levels were monitored every 24 hours, unless
otherwise indicated. The blood components were transfused if HCT < 35%, PLT count < 50 × 109/L, AT < 83%, and fibrinogen < 1.5 g/L.
Cardiac filling degree, ventricular systolic/diastolic status, ejection fraction,
and residual lesions were evaluated daily by transthoracic echocardiography. With
the improvement of the clinical condition, if the neonate's hemodynamic status was
stable at 50 mL/kg/min pump flow, decannulation was anticipated.
Statistical Analysis
Statistical analysis was performed using SPSS22 for windows. Quantitative variables
are presented as mean ± standard deviation when normally distributed, or as median
(interquartile range) if nonnormally distributed. Independent t-test and chi-square test were used to compare baseline characteristics between groups.
We used univariate analysis and multivariable logistic regression to evaluate the
risk factors with nonsurvivors. The p < 0.05 were considered statistically significant.
Results
Characteristics
Twenty-three neonates with heart failure after open heart surgery were undergone ECMO,
and two were premature infants and three were less than 2,500 g. Eighteen cases (78.26%)
were weaned off ECMO and 12 (52.17%) discharged successfully ([Table 1]).
Table 1
Clinical characteristics
|
Classification
|
Overall (N = 23), %
|
Survivors (N = 12), %
|
Nonsurvivors (N = 11), %
|
p-Value
|
|
Gender
|
|
Male
|
15
|
7 (58.33)
|
8 (72.73)
|
0.46
|
|
Female
|
8
|
5 (41.67)
|
3 (27.27)
|
|
Body weight (kg)
|
|
≤2.49
|
3 (13.04)
|
1 (8.33)
|
2 (18.18)
|
0.098
|
|
2.50–2.99
|
6 (26.09)
|
1 (8.33)
|
5 (45.45)
|
|
3–3.99
|
12 (52.17)
|
8 (66.66)
|
4 (36.36)
|
|
≥4
|
2(8.70)
|
2(16.67)
|
0
|
|
Age (d)
|
|
Premature neonate
|
2 (8.70)
|
1 (8.33)
|
1 (9.09)
|
0.14
|
|
<7
|
9 (39.13)
|
7 (58.33)
|
2 (18.18)
|
|
7–14
|
6 (26.09)
|
1 (8.33)
|
5 (45.45)
|
|
14–30
|
6 (26.09)
|
3 (25)
|
3 (27.27)
|
|
Cardiac defect
|
|
CoA/IAA/VSD
|
10 (43.48)
|
6 (50)
|
4 (36.36)
|
0.56
|
|
TGA
|
8 (34.78)
|
5 (41.67)
|
3 (27.27)
|
|
TAPVC
|
1 (4.35)
|
0
|
1 (9.09)
|
|
PA/IVS
|
2 (8.70)
|
1 (8.33)
|
1 (9.09)
|
|
DORV
|
1 (4.35)
|
0
|
1 (9.09)
|
|
Coronary artery–right ventricular fistula
|
1 (4.35)
|
0
|
1 (9.09)
|
|
Indication
|
|
LCOS
|
10 (43.48)
|
7 (58.33)
|
3 (27.27)
|
0.18
|
|
Unable to wean off CPB
|
5 (21.74)
|
1 (8.33)
|
4 (36.36)
|
|
ECPR
|
8 (34.78)
|
4 (33.33)
|
4 (36.36)
|
|
Hypoperfusion time (min)
|
|
≤30
|
4 (16.67)
|
4 (30.77)
|
0
|
0.005
|
|
> 30
|
4 (16.67)
|
0
|
4 (36.36)
|
|
ECMO duration (d)
|
|
1–3
|
8 (34.78)
|
5 (41.67)
|
3 (27.27)
|
0.01
|
|
3–5
|
9 (39.13)
|
7 (58.33)
|
2 (18.18)
|
|
> 5
|
6 (26.09)
|
0
|
6 (54.55)
|
|
Successful weaning
|
18 (78.26)
|
12 (100)
|
6 (54.55)
|
|
Abbreviations: COA, coarctation of the aorta; CPB, cardiopulmonary bypass; DORV, double
outlet right ventricle; ECMO, extracorporeal membrane oxygenation; ECPR, extracorporeal
cardiopulmonary resuscitation; IAA, interrupted aortic arch; LCOS, low cardiac output
syndrome; PA/IVS, pulmonary atresia with intact ventricular septum; TAPVC, total anomalous
pulmonary venous connection; TGA, transposition of the great arteries; VSD, ventricular
septal defect.
The indications were categorized into ECPR, inability to wean from CPB, and LCOS in
8, 5, and 10 cases. Of the ECPR neonates, four patients discharged (hypoperfusion
time ≤30 minutes), and four patients died finally (hypoperfusion time >30 minutes)
(p = 0.005). Five neonates, inability to wean from CPB, were weaned successfully, and
just one survived. Ten cases had LCOS, nine (90%) were weaned successfully, and seven
(70%) discharged.
The duration of ECMO support was between 38 and 456 hours. There were eight neonates
weaned off ECMO in 3 days and five survived; nine patients between 3 and 5 days and
seven survived; and six neonates above 5 days and all died (p = 0.01).
Outcomes
Compared with the survivors, the weights of the nonsurvivors were lighter (3.11 ± 0.67
vs. 3.43 ± 0.49 kg, p = 0.01), and ECMO duration (145.09 ± 90.38 vs. 78.92 ± 17.36 hours, p = 0.046) was significantly longer. The nonsurvivors had less time after surgery to
ECMO (7.01 ± 3.15 vs. 9.25 ± 4.28 hours, p = 0.03), higher the highest lactate levels (18.34 ± 5.13 vs. 13.87 ± 4.19 mmol/L,
p = 0.03), higher lactate value of ECMO 12 hours (9.24 ± 3.70 vs. 4.76 ± 2.49 mmol/L,
p = 0.005), 24 hours (5.36 ± 2.13 vs. 2.27 ± 1.05 mmol/L, p = 0.001), more time to lactate normalization (59.34 ± 21.15 vs. 25.01 ± 13.50 hours,
p = 0.036), higher peak creatinine values (151.29 ± 68.28 vs. 70.54 ± 30.66 mg/dL,
p = 0.03), and blood transfusion volume (275.86 ± 100.90 vs. 88.51 ± 30.24 mL/d, p = 0.04). The lowest PLT count was significantly lower in nonsurvivors than survivors
(p = 0.03). The relationship between surgical site bleeding, lactic acid >10 mmol/L
before ECMO, and residual malformation on nonsurvivors was statistically significant
(p = 0.002, 0.01, and 0.04, respectively). There was no statistical difference between
CPB time, clamp time, pump flow of ECMO 12 to 72 hours, ejection fraction of weaning
from ECMO, the peak alanine aminotransferase and aspartate aminotransferase values,
intracranial hemorrhage, GI bleeding, renal failure, and ECPR (p > 0.05) ([Table 2]).
Table 2
Comparison of clinical information between two groups
|
Classification
|
Overall (N = 23)
|
Survivors (N = 12)
|
Nonsurvivors (N = 11)
|
p-Value
|
|
Weight (kg)
|
3.17 ± 0.53
|
3.43 ± 0.49
|
3.11 ± 0.67
|
0.01
|
|
CPB time (min)
|
168.48 ± 93.70
|
155.45 ± 54.14
|
185.93 ± 134.37
|
0.23
|
|
Clamp time (min)
|
90.22 ± 53.92
|
88.91 ± 15.52
|
97.23 ± 74.28
|
0.55
|
|
ECMO duration (h)
|
109.48 ± 63.03
|
78.92 ± 17.36
|
145.09 ± 90.38
|
0.046
|
|
Ventilator duration (d)
|
12.92 ± 4.45
|
16.42 ± 4.74
|
10.82 ± 3.91
|
0.035
|
|
ICU duration (d)
|
18.04 ± 7.58
|
23.83 ± 8.72
|
10.39 ± 5.76
|
0.007
|
|
Length of hospitalization (d)
|
24.25 ± 9.79
|
34.83 ± 11.73
|
11.64 ± 7.63
|
0.01
|
|
Time after surgery to ECMO (h)
|
9.01 ± 3.80
|
9.25 ± 4.28
|
7.01 ± 3.15
|
0.03
|
|
The highest lactate (mmol/L)
|
15.03 ± 4.97
|
13.87 ± 4.19
|
18.34 ± 5.13
|
0.03
|
|
Lactate of ECMO 12 h (mmol/L)
|
6.34 ± 3.41
|
4.76 ± 2.49
|
9.24 ± 3.70
|
0.005
|
|
Lactate of ECMO 24 h (mmol/L)
|
3.33 ± 1.78
|
2.27 ± 1.05
|
5.36 ± 2.13
|
0.001
|
|
Time to lactate normalization (<2 mmol/L, h)
|
40.29 ± 18.46
|
25.01 ± 13.50
|
59.34 ± 21.15
|
0.036
|
|
Pump flow of ECMO 12–72 h (mL/kg/min)
|
157.29 ± 21.32
|
160.30 ± 19.00
|
155.77 ± 22.54
|
0.49
|
|
Ejection fraction of weaning from ECMO (%)
|
53.41 ± 13.47
|
57.36 ± 13.12
|
46.64 ± 14.82
|
0.06
|
|
Peak creatinine(mg/dL)
|
119.22 ± 50.49
|
70.54 ± 30.66
|
151.29 ± 68.28
|
0.03
|
|
Peak ALT (IU/L)
|
44.06 ± 30.15
|
36.20 ± 26.02
|
51.32 ± 38.52
|
0.12
|
|
Peak AST (IU/L)
|
201.83 ± 133.56
|
156.88 ± 121.71
|
238.27 ± 169.50
|
0.26
|
|
Lowest PLT count (× 109/L)
|
38.05 ± 18.77
|
48.10 ± 20.03
|
30.90 ± 18.14
|
0.03
|
|
Blood transfusion (mL/d)
|
193.77 ± 97.18
|
88.51 ± 30.24
|
275.86 ± 100.90
|
0.04
|
|
PLT transfusion (mL/d)
|
70.87 ± 36.28
|
55.01 ± 22.82
|
86.43 ± 42.2
|
0.31
|
|
FFP and plasma transfusion (mL/d)
|
168.56 ± 67.64
|
98.00 ± 55.76
|
247.14 ± 86.35
|
0.18
|
|
Surgical site bleeding, n (%)
|
16 (69.57)
|
5 (41.66)
|
11 (100)
|
0.002
|
|
Intracranial hemorrhage, n (%)
|
10 (43.48)
|
6 (50)
|
4 (36.36)
|
0.50
|
|
ECPR, n (%)
|
8 (34.78)
|
4 (33.33)
|
4 (36.36)
|
0.87
|
|
Pre-ECMO lactate <10 mmol/L, n (%)
|
5 (21.74)
|
4 (33.33)
|
1 (9.09)
|
0.10
|
|
Pre-ECMO lactate >10 mmol/L, n (%)
|
18 (78.26)
|
7 (58.33)
|
11 (100)
|
0.01
|
|
Surgical residual malformation, n (%)
|
6 (26.09)
|
1 (8.33)
|
5 (45.45)
|
0.04
|
|
Renal failure, n (%)
|
5 (21.74)
|
2 (16.67)
|
3 (27.27)
|
0.53
|
|
GI bleeding, n (%)
|
4 (17.39)
|
1 (8.33)
|
3 (27.27)
|
0.23
|
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; CPB,
cardiopulmonary bypass; ECMO, extracorporeal membrane oxygenation; ECPR, extracorporeal
cardiopulmonary resuscitation; FFP, fresh frozen plasma; GI, gastrointestinal; PLT,
platelet.
Univariate and Logistic Regression Analysis Results
The univariate analysis showed that nonsurvivors were related to the factors: higher
lactate value of ECMO 12 and 24 hours (9.24 ± 3.70 vs. 4.76 ± 2.49 mmol/L, p = 0.008 and 5.36 ± 2.10 vs. 2.27 ± 1.05 mmol/L, p = 0.001, respectively), longer time to lactate normalization (59.34 ± 21.13 vs. 25.01 ± 13.50 hours,
p = 0.001), lower weight (3.11 ± 0.67 vs. 3.43 ± 0.49 kg, p = 0.001), longer ECMO time (145.02 ± 90.25 vs. 78.92 ± 17.31 hours, p = 0.005), lower the lowest PLT count (30.90 ± 18.14 vs. 48.10 ± 20.03 × 109/L, p = 0.001), more surgical site bleeding (11 vs. 5, p = 0.001) and surgical residual malformation (5 vs. 1, p = 0.04). Further logistic regression analysis revealed that higher lactate value
of ECMO 24 hours (p = 0.003), longer ECMO duration (p = 0.015), and more surgical site bleeding event (p = 0.025) were independent risk factors for nonsurvivors ([Table 3]).
Table 3
Univariate analysis and multivariate logistic regression results
|
Classification
|
Survivors
|
Nonsurvivors
|
Univariate analysis
|
Multivariate logistic regression
|
|
F
|
p-Value
|
OR
|
95% CI
|
p-Value
|
|
Weight (kg)
|
3.43 ± 0.49
|
3.11 ± 0.67
|
7.90
|
0.01
|
|
|
|
|
ECMO duration (h)
|
78.92 ± 17.31
|
145.02 ± 90.25
|
9.62
|
0.005
|
0.75
|
0.29–0.94
|
0.015
|
|
Time after surgery to ECMO (h)
|
9.25 ± 4.28
|
7.01 ± 3.15
|
1.46
|
0.24
|
|
|
|
|
The highest lactate (mmol/L)
|
13.87 ± 4.19
|
18.34 ± 5.13
|
1.83
|
0.19
|
|
|
|
|
Lactate of ECMO 12 h (mmol/L)
|
4.76 ± 2.49
|
9.24 ± 3.70
|
8.55
|
0.008
|
2.55
|
0.42–10.33
|
0.38
|
|
Lactate of ECMO 24 h (mmol/L)
|
2.27 ± 1.05
|
5.36 ± 2.10
|
101.41
|
0.001
|
13.87
|
2.33–18.84
|
0.003
|
|
Time to lactate normalization (<2 mmol/L, h)
|
25.01 ± 13.50
|
59.34 ± 21.13
|
22.75
|
0.001
|
3.88
|
0.96–4.06
|
0.07
|
|
Peak creatinine (mg/dL)
|
70.54 ± 30.66
|
151.29 ± 65.28
|
2.98
|
0.10
|
|
|
|
|
Lowest PLT count (× 109/L)
|
48.10 ± 20.03
|
30.90 ± 18.14
|
42.51
|
0.001
|
3.62
|
0.92–1.90
|
0.06
|
|
Blood transfusion (mL/d)
|
88.51 ± 30.24
|
275.86 ± 100.90
|
5.11
|
0.05
|
|
|
|
|
Surgical site bleeding, n (%)
|
5 (41.66)
|
11 (100)
|
16.13
|
0.001
|
7.41
|
1.28–43.41
|
0.03
|
|
Pre-ECMO lactate >10 mmol/L, n (%)
|
7 (58.33)
|
11 (100)
|
3.35
|
0.08
|
|
|
|
|
Surgical residual malformation, n (%)
|
1 (8.33)
|
5 (45.45)
|
5.12
|
0.04
|
|
|
|
Abbreviations: CI, confidence interval; ECMO, extracorporeal membrane oxygenation;
OR, odds ratio; PLT, platelet.
Causes of Death
The mortality rate was 47.83% (11/23), six neonates were weaned successfully but died
finally. Among the six patients, five had residual anatomical problems, even the heart
function recovered during ECMO, but follow-up was not sustainable. Three cases involved
GI bleeding with unstable circulatory function. Two cases failed to undergo further
correction for renal failure, medical cost, and heart failure reasons, no further
interventions were provided, and eventually both of them failed to wean.
Complications
Surgical Site Bleeding
Major bleeding requiring re-exploration occurred in 16 (69.57%) patients. Eleven (68.75%)
patients were explored in the first 24 hours. Five had no initial major bleeding but
required exploration later in the ECMO run for “late bleeding.” Survival rate in the
explored group was five (31.25%). Survival rates for early bleeding versus late bleeding
were 25 and 6.25% (p = 0.51) ([Table 2]).
Nervous System Injury
Ten (43.48%) infants had central nervous system injury, but only six survived. Overall
neurological injury included intracranial bleeding (n = 8, 34.78%) and ischemic brain injury (n = 2, 8.70%). Five cases had neurological bleeding and surgical site bleeding simultaneously.
Renal Failure
The incidence of renal failure requiring CRRT was five (21.74%), only two patients
survived. None of the survivors required long-term dialysis after hospital discharge
(p = 0.53).
GI Bleeding
GI bleeding was a serious complication. Among four neonates, only one survived. Two
(50%) neonates needed laparotomies, but failed to undergo further correction.
Other Complications
Other complications included cardiac tamponade (n = 5, 21.74%), arrhythmia (n = 4, 17.39%), liver function damage (n = 4, 17.39%), infection (n = 3, 13.04%), and clots in the ECMO circuit (n = 2, 8.70%).
Residual Anatomical Malformation
Another serious issue was residual anatomical problems (26.09%, n = 6). Only one patient who undergone reoperation survived. Three patients involved
residual ventricular septal shunt and residual aortic arch obstruction, and two patients
had postoperative aortic valve insufficiency. Although several patients weaned successfully,
they died finally due to recurrent heart failure.
Discussion
ECMO is commonly used in pediatric patients with CHD, particularly in the setting
of low-output failure, arrhythmia, cardiopulmonary arrest, or inability to wean from
CPB.[5] However, despite significant advances in ECMO techniques and management over the
past several years, prognosis remains poor.[6] The underlying causes of death vary from cardiovascular events to organ failure,
including GI, renal, neurologic, coagulation, and it carries a significant cost burden.[7] Neonates are special population, whose cardiomyocytes are softer with less contractile
tissue and lower energy reserves. Longer CPB time leads to poor myocardial cell compliance
and insensitivity to drugs, and these lead to heart failure controlled by drugs difficultly
after open heart surgery. According to the latest statistics from ELSO 2017, the neonatal
survival rate of CHD is 47% and discharge rate is 39%.[8] In our study, although the survival rate of neonate is as high as 52.17%, there
are many risk factors that affect the outcomes.
Metabolic Acidosis
Studies have shown that hyperlactic acidosis is an effective biomarker, and the severity
of pre-ECMO arterial acidosis is related to outcomes.[9] In our study, five patients with lactic acid <10 mmol/L before ECMO were discharged,
while in 18 patients with lactic acid >10 mmol/L, only 7 survived. This suggests that
higher lactate levels before ECMO is an independent risk factor for poor outcomes.
There was a significant difference in the highest lactate, the levels of ECMO 12 and
24 hours and the lactate clearance time between two groups (p = 0.03, 0.005, 0.001, and 0.036, respectively). The results are similar to those
findings of the reports.[10]
[11] Fux et al[12] analyzed VA-ECMO patients and found that ischemic heart disease and arterial lactate
were independent predictors of 90-day mortality. The 90-day survival rate of lactic
acid >10 mmol/L was lower than patients with lactic acid <10 mmol/L before ECMO (13
and 55%, p < 0.001). If the lactate remained at 3 mmol/L after 48 hours, the 30-day mortality
rate is 52%. Other articles[13]
[14] suggested that persistent metabolic acidosis after ECMO reflects the severity of
ischemia and hypoxia and confirmed that the peak lactate level affects the survival
rate. Therefore, early application to reverse poor perfusion and prevention high lactate
are critical factors for successful outcomes following ECMO.
ECMO Duration
Our findings are consistent with the results reported in the literature on ECMO support
after cardiac surgery[14]: Compared with the survivors, ECMO duration were significantly longer in nonsurvivals
(p = 0.046). ECMO duration is directly linked to the other complications and survival
rate. The ELSO data analysis also showed that[15]: The highest survival rate for the ECMO weaning is at the fourth day, and the shorter
ECMO duration resulted in higher mortality due to inadequate support. In the group
of neonates assisted for 4 to 12 days, the survival rate decreased gradually with
the longer duration, especially in children with assist time > 7 days.[13] Long-term assistance means poor recovery of heart function, especially in neonates
after cardiac surgery. If accompanied with serious complications, ECMO could not successfully
wean off in short- and long-term assist devices, or switch to left ventricular assist
device as bridge to heart transplantation should be considered.[16] In this group, there were eight patients with assist time of 1 to 3 days and five
survived, nine patients assisted for 3 to 5 days and seven patients discharged; however,
six patients assisted for more than 5 days and all died (p = 0.01). Our results were close to the ELSO statistics, indicating that ECMO duration
was one of the main factors affecting survival.
Bleeding
Complications in the ECMO support are important predictors of poor outcomes. The most
common complication in this study was bleeding. Sixteen patients experienced severe
surgical site bleeding requiring explorations and 11 patients died. Neonates have
small volume of circulating blood and immature blood coagulation mechanism, ECMO prefilling
dilute the anticoagulant components, and longer CPB time destroyed blood components.
Improving supplement coagulation components and try to correct coagulation defects
are very important.[13] Four patients experienced severe GI bleeding and only one survived, and two patients
with ECPR time more than 30 minutes before ECMO. GI bleeding resulted from infection,
ischemia and hypoxia, feeding, and blood flow to premature labor when patients were
under critical illness. Studies have shown that 90 to 95% necrotizing enterocolitis
(NEC) occurs in premature and low-birth-weight infants with a gestational age of less
than 36 weeks, which is the result of a combination of risk factors.[17] The literature reports that feeding factors (type, speed, and concentration) are
intrinsically linked to the occurrence of NEC because the neonatal digestive system
is immature and susceptible to ischemic injury, and digestive enzyme activity is low,
which will eventually increase the incidence of NEC.[18]
Residual Malformations
Residual cardiac malformations were found in six patients by transthoracic echocardiography.
One neonate's malformation was corrected during ECMO and discharged. Other patients
were weaned successfully, but died finally due to recurrent heart failure. There was
a statistically significant difference between the survival and nonsurvival (p = 0.04). Studies[19]
[20] have reported that one-quarter of children with ECMO support after open heart surgery
have residual anatomical problems. All residual lesions should be evaluated actively,
neonates who failed to wean from ECMO without a clear cause were investigated for
presence of residual lesions. Our center uses transthoracic echocardiography and cardiac
catheterization to evaluate residual lesions. A transthoracic echocardiography imaging
study was undertaken during the process of ECMO and was the initial imaging study
for evaluating residual lesions. A cardiac catheterization imaging study was undertaken
when the echocardiographic results were inconclusive and neonate was unable to be
weaned from ECMO support successfully. Early examination and interventional therapy
were related to better clinical outcomes. Therefore, residual malformations should
be discovered as soon as possible, and active intervention can improve cardiac function.
Summary
ECMO was an effective technology to support the neonates with cardiopulmonary failure
after open heart surgery to wait for heart function recovery. Control the lactate
acidosis and surgical site bleeding event may be helpful for patients' recovery.
