Secondary Logo

Journal Logo

Clinical Outcomes–Devices

Postoperative Extracorporeal Life Support in Pediatric Cardiac Surgery: Recent Results

Ghez, Olivier*; Feier, Horea*; Ughetto, Fabrice; Fraisse, Alain; Kreitmann, Bernard*; Metras, Dominique*

Author Information
doi: 10.1097/01.mat.0000178039.53714.57
  • Free


Extracorporeal life support (ECLS) has been used in pediatrics for a long time. Major indications concern mainly patients with respiratory failure (persistent pulmonary hypertension of the neonate, diaphragmatic hernia, meconial inhalation), postcardiotomy patients, or patients awaiting transplantation. Recent articles advocate a greater use of this technique, considering improvements of the results in terms of survival.1–4 Also, many large centers have set a policy of permanent availability of an “ECMO team,” allowing early and rapid placement of ECLS, particularly in a resuscitation context, with very encouraging results.5–7 Over the last few years, in view of these recent publications and excellent results obtained on the first few patients presenting a refractory failure treated with ECLS in our center, we increased the number of indications of ECLS in our postcardiotomy patients. We now have an early ECLS placement policy in case of severe hemodynamic or respiratory failure. We review here the results of this policy.

Patients and Methods

We retrospectively reviewed the files of all patients who underwent heart surgery and were treated by ECLS in the postoperative period in our center between January 1, 2002, and January 1, 2005. Indications were hemodynamic, respiratory, or mixed failure, and are summarized in Table 1.

Table 1
Table 1:
Patients Undergoing Extracorporal Support after Surgical Repair of Congenital Cardiopathy

Different techniques of ECLS were applied depending on the context. Arteriovenous ECLS with oxygenator was used for hemodynamic failure or mixed failure, refractory to medical treatment. Biventricular assistance was used in isolated heart failure. The goal was to provide a systemic oxygenated blood flow of 2.4 l · min−1 · m−2 in full-support. Venovenous ECLS with oxygenator was used for isolated respiratory failure. The goal was then to provide an arterial oxygen saturation ≥90% with gentle artificial ventilation.

The circuit included a nonocclusive pump and pretreated tubing and cannulas. Cannulation site and specific material depended on the patient. Oxygenators were adapted to patients’ body surface area (Minimax or Affinity, Medtronic Inc., Minneapolis, MN). More recently, we used long-term oxygenators specifically designed for ECLS (Jostra Quadrox, Maquet cardiopulmonary AG, Germany, or Lilliput2 D902 ECMO, Dideco, Sorin Group, Italy). The pump used was nonocclusive in all cases (Biomedicus, Medtronic Inc., or more recently Revolution, Cobe, Sorin Group).

Central cannulation has been used in infants undergoing ECLS early after cardiotomy when the sternum was usually left open, establishing a circulation between the right atrium and the ascending aorta. Peripheral cannulation between the right jugular vein and the right carotid artery was preferred when the sternum was already closed at the time of ECLS institution in infants. Teenagers underwent peripheral cannulation between a femoral vein and a femoral artery. Ventricular assistance was established on the left side between the left atrium and the aorta, and on the right side between the right atrium and the pulmonary artery.

Peritoneal dialysis or hemodiafiltration connected on the ECLS circuit (Prisma, Gambro/Hospal AG, Germany) was instituted if needed.

Anticoagulation was obtained by continuous heparin infusion adapted to heparin level (between 0.3 and 0.5 IU/ml) and activated clotting time between 180 and 200 seconds. (Hemochron JR, Medtronic Inc.). Coagulation factors were maintained above 50%, anti–thrombin III above 70%, hemoglobin above 110 g/l, and platelets above 100,000/ml, with iterative transfusions to prevent hemorrhagic complications.

Inotropes were weaned in cases of arteriovenous ECLS or ventricular assistance. Peripheral resistances were adjusted with noradrenalin if necessary to maintain a perfusion pressure adapted to the patient age. Artificial ventilation was kept to a minimum for iFO2 and volume. The ventilation was increased for ECLS weaning.

ECLS Weaning

Monitoring of right and left atrial pressures, pulmonary arterial pressure, venous oxygen saturation, and arterial blood gases, associated to echocardiographic surveillance, helped in the decision of ECLS weaning and its modalities. ECLS weaning was done over a period of 2 or 3 days. The minimal flow considered was 150 ml/min.

Statistical Analysis

Our analysis is descriptive. Values are presented as their mean and standard deviation and their median if their distribution justifies it.


Between January 1, 2002, and January 1, 2005, we performed 19 ECLS installations in 15 patients. In 2004, of the 248 patients undergoing bypass surgery for congenital heart disease, 8 required ECLS in the postoperative period (3.2%). Mean age was 4.9 ± 7 years (median 5.9 months, range 11 days to 21 years). All patients were operated on for congenital heart disease, except for one patient who had a heart transplant (Table 1). Mean aortic cross-clamp time was 102 ± 65 minutes; mean bypass time was 196 ± 111 minutes.

Indications were hemodynamic failure in 12 cases, respiratory failure in 5 cases, and mixed failure in 2 cases. Four patients required cardiopulmonary resuscitation during ECLS placement (2–45 minutes of cardiopulmonary resuscitation, no deaths). Mean delay between surgery and ECLS placement was 3.2 ± 3.4 days (range 0–11 days, median 2 days). In four patients, ECMO was installed in the operating room because of inability to come off bypass (one graft failure, two arterial switch operations, and one triatrial heart, urgent repair). Mean ECLS duration was 145 ± 72 hours (6 ± 3 days, range 3–16 days). Two patients underwent surgery for residual lesions (Table 1).

Four patients underwent two successive periods of ECLS. Patient 4 presented a complex cyanotic cardiopathy consisting of an unbalanced atrioventricular canal with double outlet right ventricle. After a superior cavopulmonary anastomosis, he underwent complete correction of his malformation. ECLS was necessary in the postoperative period because of ventricular dysfunction. This arteriovenous ECLS was weaned after 6 days. However, refractory hypoxia led us to institute a venovenous ECLS the next day. This support allowed us to perform a catheterization, which revealed a residual pulmonary stenosis. This lesion could then be surgically addressed and the child could be weaned from his support. Unfortunately, this patient, who underwent a tracheotomy, died of respiratory complications 6 months after ECLS weaning, at a time when his cardiac status was considered successfully repaired.

Patient 8 presented a pulmonary atresia with ventricular septal defect and major aortopulmonary collateral arteries and had already undergone two Blalock-Taussig shunts. After complete correction (ventricular septal defect closure and transannular right ventricular patch), a low cardiac output refractory to medical treatment led us to place an arteriovenous ECLS, which could be weaned only after valvulation of the right ventricular outflow tract (3 days). However, development of a major tricuspid insufficiency due to a double-orifice tricuspid valve necessitated a new period of arteriovenous ECLS, which could be weaned only after tricuspid valve repair (7 days).

Patient 10 presented with acute respiratory distress syndrome 4 days after complete atrioventricular canal correction leading to refractory hypoxia necessitating a venovenous ECLS. This assistance could be weaned after 6 days. Unfortunately, during the next 24 hours, fluid accumulation and renal function impairment led us to recannulate for an arteriovenous ECLS for 6 days.

Patient 15 had an arterial switch operation, complicated by a cardiorespiratory failure and hemorrhagic syndrome necessitating arteriovenous ECLS (central cannulation). This ECLS was weaned after 5 days. However, a respiratory failure 24 hours after ECLS weaning necessitated a new period of arteriovenous ECLS (neck cannulation), which could finally be weaned after 6 days

Cannulation was done through the neck vessels in four cases (infants) and through the femoral vessels in four cases (teenagers). In 11 cases, a central cannulation was performed.

Mortality under ECLS

Two patients died while receiving ECLS. A 21-year-old boy (patient 9) operated on for complete correction of pulmonary atresia with ventricular septal defect presented a massive cerebral ischemic accident after 67 hours of ECLS instituted for hemodynamic failure. A 17-year-old girl (patient 7) operated on for total cavopulmonary connection presented with a ventricular dysfunction necessitating arteriovenous ECLS and died from multiorgan failure after 16 days of ECMO. No infant died while on ECLS in this series. Thirteen patients (86.7 %) survived to ECLS weaning.

Hospital Discharge Survival

Twelve patients survived to hospital discharge (80%). Patient 6 died of intractable multiorgan failure 1 week after ECLS weaning. He presented with a shone syndrome with aortic and mitral stenosis. He went through two mitral repairs and eventually had a mitral valve replacement.

Late Mortality

As previously described, patient 4 died of respiratory complications 6 months after ECLS weaning.


No survivor presented obvious neurologic damage. Thorax was left open in five patients for 10 ± 6 days (0–21 days). Re-entry for bleeding occurred a mean of 2 ± 2 times per patient and occurred only in patients with delayed sternal closure and central cannulation, mainly during the first 2 days of ECLS. Multiple transfusions were necessary and blood unit consumption could be established as follows: red blood cells 13 ± 18 (4–76), plasma 4 ± 3, platelets 5 ± 3. Mediastinitis occurred in two patients (delayed sternal closure patients, Pseudomonas aeruginosa) and evolved favorably under irrigation and antibiotherapy. Three patients (the two with mediastinitis and another infant) who had a delayed sternal closure presented difficulties of wound healing after ECLS weaning and sternal closure. Satisfactory healing was obtained by using a vacuum-assisted closure system (Kinetics Systems Inc., San Antonio, TX), which allowed for sternal healing. A skin graft was subsequently applied on the sternum with the help of plastic surgeons. Hemodialysis was used during six ECLS periods, and peritoneal dialysis during nine ECLS periods. Renal function returned to normal in all survivors to hospital discharge.


Excellent results with survival rates superior to 80% can be achieved with ECLS in respiratory indications.2,8 Postcardiotomy situations yield poorer results, with survival rates around 40% and a certain neurologic risk.9 However, recent publications confirm interest in ECLS after pediatric cardiac surgery.1,4,10–13 According to European and American recommendations, centers of congenital heart surgery should have ECLS available.14,15 These variations in results could be explained by differences in ECLS installation criteria. It seems difficult to apply ECLS systematically, for instance after a Norwood operation as some advocate.13 However, this report presents evidence that good results can be achieved with an early ECLS installation, before irreversible organ damage occurs. Our recent experience has yielded very encouraging results, with a survival to hospital discharge of 80%.

The proportion of children benefiting from ECLS in our series (3.2%) is consistent with the literature.9 Postoperative indications can be either cardiopulmonary, secondary to bypass, aortic cross-clamping, or myocardial ischemia. In this case, recuperation of organ function can be expected in a 5-day delay.11 The latter needs to be addressed in cases of failure secondary to a residual lesion. ECLS allows one to go safely to the catheterization laboratory to improve diagnosis of such residual lesion that can then be addressed either surgically or by catheter intervention.16 Two of our 15 patients presented a residual lesion, but correction by surgery and catheterization allowed for ECLS weaning. Our good results could be explained, as hypothesized, by differences in indications.9,11 Even though we did not find obvious differences in ECLS indication criteria as compared to the literature, we recognize that the structure of our unit in itself may have led to a slightly earlier ECLS installation in cases of heart-lung failure. Our unit depends on a polyvalent pediatric intensive care unit. We therefore need to have a very low threshold for detection of potential problems. For instance, we systematically place a pulmonary artery catheter with continuous venous oxygen saturation monitoring in every neonate and in all complex cases, regardless of pulmonary hypertension risk. This allows early detection of low cardiac output. Our technique of ECLS is not different from those of other teams. However, we use nonocclusive pumps and long-term oxygenators, which are easier to manage in the intensive care unit. We also delay very easily sternal closure, which avoids some brutal accidents.17 Concerning ECLS installation during cardiac arrest, our results were surprising in that there were no deaths among four patients. However, cardiac arrest was anticipated in two cases; the ECLS circuit was being primed during these arrests, and the surgical team was ready and next to the child in both cases. In the two other cases, an adapted resuscitation, sometimes very long (45 minutes), allowed to keep the children alive while ECLS was installed. As noted in the literature, good results can be achieved in these situations by centers where ECLS is available.6,7 It is, however, difficult to obtain a permanent ”ECMO team” in most centers owing to a lack of human, technical, and financial resources. This is the case for our institution, which is a medium-sized center with two attending surgeons and one fellow operating 400 cases a year, including 250 pump cases. However, for the reasons described earlier, we have a low threshold to call in-hospital all the members of the ECMO team if needed. These good recent results are also explained by the preliminary experience of our center. Recent results in other centers certainly also have benefited from recent improvements in ECLS techniques.

Neurologic sequelae will be finely researched in survivors, none of whom has any obvious neurologic damage.

Morbidity can be high, consisting essentially of re-entry for bleeding, multiple transfusion, and infections. This can lead to high management costs, but the good results obtained justify our approach.


These results support early placement of ECLS after cardiac surgery in children whenever a severe postoperative hemodynamic or respiratory failure, refractory to medical treatment, is present.


1.Aharon AS, Drinkwater Jr. DC, Churchwell KB, et al: Extracorporeal membrane oxygenation in children after repair of congenital cardiac lesions. Ann Thorac Surg 72: 2095–2101; discussion 2101–2092, 2001.
2.Bennett CC, Johnson A, Field DJ, Elbourne D: UK collaborative randomised trial of neonatal extracorporeal membrane oxygenation: follow-up to age 4 years. Lancet 357: 1094–1096, 2001.
3.Duncan BW, Hraska V, Jonas RA, et al: Mechanical circulatory support in children with cardiac disease. J Thorac Cardiovasc Surg 117: 529–542, 1999.
4.Morris MC, Ittenbach RF, Godinez RI, et al: Risk factors for mortality in 137 pediatric cardiac intensive care unit patients managed with extracorporeal membrane oxygenation. Crit Care Med 32: 1061–1069, 2004.
5.Duncan BW, Ibrahim AE, Hraska V, et al: Use of rapid-deployment extracorporeal membrane oxygenation for the resuscitation of pediatric patients with heart disease after cardiac arrest. J Thorac Cardiovasc Surg 116: 305–311, 1998.
6.Morris MC, Wernovsky G, Nadkarni VM: Survival outcomes after extracorporeal cardiopulmonary resuscitation instituted during active chest compressions following refractory in-hospital pediatric cardiac arrest. Pediatr Crit Care Med 5: 440–446, 2004.
7.Zaritsky A: Extracorporeal cardiopulmonary resuscitation: ready for prime time? Pediatr Crit Care Med 5: 495–496, 2004.
8.UK Collaborative ECMO Trail Group: UK collaborative randomised trial of neonatal extracorporeal membrane oxygenation. Lancet 348:75–82, 1996.
9.Hintz SR, Benitz WE, Colby CE, et al: Utilization and outcomes of neonatal cardiac extracorporeal life support: 1996-2000. Pediatr Crit Care Med 6: 33–38, 2005.
10.Kolovos NS, Bratton SL, Moler FW, et al: Outcome of pediatric patients treated with extracorporeal life support after cardiac surgery. Ann Thorac Surg 76:1435–1441; discussion 1441–1432, 2003.
11.Thiagarajan RR, Nelson DP: Should we be satisfied with current outcomes for cardiac extracorporeal life support? Pediatr Crit Care Med 6: 89–90, 2005.
12.Undar A, McKenzie ED, McGarry MC, et al: Outcomes of congenital heart surgery patients after extracorporeal life support at Texas Children’s Hospital. Artif Organs 28: 963–966, 2004.
13.Ungerleider RM, Shen I, Yeh T, et al: Routine mechanical ventricular assist following the Norwood procedure–improved neurologic outcome and excellent hospital survival. Ann Thorac Surg 77: 18–22, 2004.
14.Daenen W, Lacour-Gayet F, Aberg T, et al: Optimal structure of a congenital heart surgery department in Europe. Eur J Cardiothorac Surg 24: 343–351, 2003.
15.Pediatrics AAo: Guidelines for pediatric cardiovascular centers. Pediatrics 109: 544–549, 2002.
16.Booth KL, Roth SJ, Perry SB, et al: Cardiac catheterization of patients supported by extracorporeal membrane oxygenation. J Am Coll Cardiol 40: 1681–1686, 2002.
17.Samir K, Riberi A, Ghez O, et al: Delayed sternal closure: A life-saving measure in neonatal open heart surgery; could it be predictable? Eur J Cardiothorac Surg 21: 787–793, 2002.
Copyright © 2005 by the American Society for Artificial Internal Organs