Objective. To review demographic and procedural factors and their association with weaning rate and survival from extracorporeal membrane oxygenation (ECMO) in pediatric patients undergoing repair of cardiac malformations.
Methods. The hospital records of children requiring ECMO during cardiac operation due to failure to wean from cardio-pulmonary by pass (CPB) were retrospectively reviewed, and an analysis of variables affecting survival was performed.
Results. Thirty-five pediatric patients between January 1, 2000 and December 31, 2006 required ECMO for cardiopulmonary support during cardiac operations. ECMO survival was 54.3% and was comparable across all age groups. The lowest pH during ECMO treatment was the only predictor of mortality (P = 0.006). No other patient, surgical or anesthetic, factor was associated with either weaning from ECMO or hospital survival.
Conclusions. No clear risk factor could be identified for survival from ECMO in our pediatric patients who underwent cardiac surgery and failed weaning from cardiopulmonary bypass.

Keywords: congenital, cardiopulmonary bypass, infants, neonates, failure to wean for cardiopulmonary bypass, survival


Extensive repairs of complex cardiac anomalies utilizing cardiopulmonary bypass (CPB) may result in severe myocardial dysfunction leading to inability to wean pediatric patient from CPB. (1) Occasionally this complication may be refractory even to the most intensive pharmacologic therapy and the use of extracorporeal membrane oxygenator (ECMO) may be the only alternative. A significant body of literature exists regarding the use of ECMO in postoperative course after cardiac surgery in children, (2-5) and more limited information exists regarding the use of intraoperative ECMO. In present study we report outcome in 35 pediatric patients who received ECMO intraoperatively after failing weaning from CPB following repair of complex heart malformations. Incidence and survival analysis of this patient population was recently described. (6) In present report we expand the analysis to examine association between demographic, co-morbid, and procedural factors in these patients and failure to wean from surgical ECMO or hospital survival.


Study Design, Patients and Setting. After Mayo Clinic Institutional Review Board (Rochester, MN, USA) approval, we ascertained all patients 0-17 years of age who were treated with ECMO during cardiac operations requiring CPB at Mayo Clinic, Rochester, MN, between January 1, 2000 and December 31, 2006. All identified charts were authorized for review by the Institutional Review Board. Cases were identified from the Mayo Clinic ECMO registry. We reviewed only medical records of children who were treated with ECMO during their index cardiac surgery. Successful weaning from ECMO was considered if the child survived at least 48 hours after separation.
Data collection and analysis. The medical records of all index patients were abstracted. The primary objective was to examine the potential risk factors for failure to wean from ECMO in pediatric patients undergoing repair of cardiac malformations. We analyzed patient demographics (age, gender, American Society of Anesthesiologists Physical Status [ASA PS] score), patient-related factors: surgical urgency status, and dominant cardiac anatomy, univentricular physiology (encompasses a wide variety of heart defects that functionally and physiologically constitute a single ventricular chamber) (7,8) or biventricular heart physiology. Finally, we analyzed the following procedural factors: use of circulatory arrest, use of intraoperative ultrafiltration, continuous veno-venous hemofiltration (CVVHF), requirement for other means of mechanical circulatory support (intraaortic balloon pump [IAB], left ventricular assist device [LVAD]), duration of anesthesia before ECMO, duration of CPB, and completeness of surgical repair (complete vs. incomplete repair). The variables assessed while on-ECMO included: duration of ECMO therapy, and highest creatinine, and lowest pH during ECMO treatment.
Statistical Analysis. The bivariate association between each of the candidate risk factors and weaning success and hospital survival was first evaluated using the rank sum test for continuous variables and Fisher’s exact test for categorical variables. Duration of anesthesia before ECMO, duration of CPB, duration of ECMO therapy, highest creatinine level while on ECMO and lowest pH value while on ECMO were analyzed as continuous variables, while all other variables were analyzed as categorical variables. In all cases, two-tailed P-values ≤ 0.05 were considered statistically significant. Analyses were performed using SAS statistical software (Version 8.2; SAS Institute, Inc., Cary, NC, USA).


Between January 1, 2000 and December 31, 2006 35 pediatric patients failed weaning from CPB and received ECMO intraoperatively. Table 1 shows demographics, primary diagnosis, surgical details and outcome of our ECMO patients. Table 2 shows association between demographic, patient and procedural characteristics (all categorical variables) and ability to be weaned from ECMO (n=24) as well as hospital survival (P values in table 2 compare those who were “able to wean” vs. “unable to wean”, and “survived vs. died”). None of the tested characteristics were significantly associated with weaning success or survival. The frequency of successful weaning from ECMO was 68.6% (24/35), however subsequently 5 patients died from other postoperative complications (cardiac arrest, malignant dysrhythmias, etc), and the overall survival rate was 54.3%. All 5 deaths occurred greater than 48 hours after separation from ECMO.
From exploratory analyses, no patient or procedural characteristics were found to differ significantly between patients who were successfully weaned from ECMO vs. not. Furthermore, none of the tested characteristics were different between the survivors and non-survivors. However, there was a trend toward better survival for patients who received either ultrafiltration or hemofiltration during surgery (P = 0.063). Table 3 shows association between patient and procedural (continuous) variables and weaning success and hospital survival. None of the tested variables was associated with weaning success, and only one, the lowest pH while on ECMO, was associated with worse hospital survival (P=0.006). While the duration of ECMO in those who were successfully weaned tended to be shorter (61 vs. 106 hours) this difference did not reach statistical significance (tested using the rank sum to account for outliers). Among our ECMO patients, 9 required surgical reexploration for bleeding, 4 had intracranial hemorrhage, 1 intracranial thromboembolic event, while 3 received orthotropic heart allotransplant (1 as a primary procedure, and 2 following failed index surgery).


Inability to regain cardiopulmonary function to sustain life, regardless of maximal pharmacologic support, is one of the most significant complications of surgery for congenital heart malformations. A mechanical cardiopulmonary support via ECMO has been increasingly used in our institution for pediatric patients with post-CPB refractory heart failure. At Mayo Clinic the decision to institute the ECMO is always made by cardiovascular surgeon and typically when inotropic support of epinephrine exceeds 0.15 mcg/kg/min, when one or more following conditions exist: systemic ventricular dysfunction with elevated filling pressures (right atrial pressure (RAP) >20 mmHg, left atrial pressure (LAP) >15 mmHg), difficulty with oxygenation and/or ventilation, pulmonary hypertension with right ventricular dysfunction.
We recently reported survival of 35 children who received ECMO for pharmacologically refractory heart failure during repair of congenital heart malformations. (6) Over half of our children (54.3%) requiring ECMO were neonates followed by infants (22.3%). The weaning rate from ECMO was 68.7%, and the hospital survival was 54.3% which is the range reported by others. (5,9,10) The small number of children in present study does not allow for comprehensive analysis of risk factors. Therefore, we conducted only an exploratory analysis which did not identify any demographic or procedural characteristics associated with outcome including age, gender, ASA PS, primary cardiac physiology (univentricular vs. biventricular), type of repair (complete vs. incomplete), use of circulatory arrest, intraoperative use of ultrafiltration or need for preoperative ventricular assist (left ventricular assist device or intra-aortic balloon pump), and shorter anesthetic course before initiation of ECMO. Only one factor, the lowest pH during ECMO treatment, was associated with worse hospital survival.
In other ECMO studies several predictors of mortality have been identified, albeit majority of predictors were associated with complications that occurred in later postoperative course. For example, Morris et al. (11) identified kidney or hepatic failure, and Montgomery et al. (12) the presence of multiple organ dysfunction and infections during ECMO treatment as predictors of survival. However, these are not consistent findings, as another study failed to establish association between the degree of renal dysfunction or infection and survival. (13) The only predictor of mortality in our study was the lowest pH value while on ECMO, which is consistent with a study (13) that showed that the highest lactate level within 48 hours on ECMO predicted death.
Intuitively, the type of heart defect or completeness of surgical repair could affect the survival. For example, Black et al. (14) found better ECMO survival in children who had complete repair than in those with residual heart defect (incomplete repair). However, the completeness of repair was not a significant factor for either ECMO weaning or survival in our patients. Kolovos et al. (13) reported that children who received ECMO after an adequate two-ventricular repair (biventricular physiology) had lower risk of death compared to those with single ventricle physiology. In contrast, another study reported 61% survival to discharge in infants with univentricular physiology, compared with 43% in those with biventricular physiology. (9) Morris et al. (11) did not confirm the association of univentricular physiology with increased mortality, which agrees with our findings (table 2).

Table 1. Patients who failed separation from CPB received intraoperative ECMO and survived to leave the operating room.

Event Age/ Diagnosis Type of Operation ECMO survival Hospital Survival
(year) Gender/
2000 5d/F/3 IAA Type A; VSD, ASD Complete repair, closure of ASD and VSD, and end-to-end aortoplasty between descending aorta and aortic arch. No No
2000 17m/M/4 Ebstein’s anomaly, PFO, severe TR, atrial flutter TV replaced, PFO closed, cryoablation right atrial isthmus Yes Yes
2000 20m/M/3 TGA, tricuspidal atresia, subaortic stenosis Relief subaortic stenosis No No.
2000 3m/F/4 Anomalous left CA from PA; severe LV dysfunction Repair of anomalous CA, CABG No No
2001 10d/F/3 Complete TGA ASO Yes No
2001 17y/M/4 TV atresia, failing Fontain circulation Intraatrial conduit, Right Maze procedure, left AV valve anuloplasty; heart transplant (second operation) No No
2001 6y/M/3 Pulmonary atresia, VSD, previous Blalock-Taussig shunt Complete repair Yes Yes
2001 6d/M/4 Unbalanced complete AVSD; hypoplastic LV, IAA Type B Modified Norwood; repair VSD and IAA Yes Yes
2001 2m/M/4 Severe congenital AS, PFO Ross-Konno procedure (aortoventriculoplasty) Yes Yes
2002 5m/M/5 TOF, complete ASD, severe regurgitation common AV valve Complete repair DORV/AV canal (AVSD) defect; residual LVOT-repaired in second operation No No
2002 3d/M/4 TGA, IVS ASO Yes Yes
2002 12y/M/4E Prosthetic MV and AV endocarditis MV & AV replacement, CABG x1 Yes Yes
2002 3d/M/3 TGA, PDA, IVS, ASD ASO No
2002 48d/F/4E Hypoplastic left heart Heart transplant; aortic arch reconstruction Yes Yes
2003 14d/F/4 Double outlet R ventricle, VSD, AS Modified Norwood procedure Yes Yes
2003 5d/M/3 IAA Type A, CoA, ASD, VSD,PI Arch reconstruction, ASD+VSD closure No
2003 2d/M/4 HLHS, ASD Norwood procedure, stage I Yes Yes
2003 3m/F/3 Complete AVSD; Down syndrome Complete repair Yes No
2003 7d/M/4 Truncus arteriosus, severe truncal valve insufficiency Complete repair Yes Yes
2003 3d/M/4 TGA, VSD, DORV ASO, VSD repair, PDA ligation No No
2003 1d/M/5E Obstructive TAPVC, PDA Complete repair, ligation PDA Yes Yes
2004 6d/M/4 Obstructive TAPVC (infracardiac). PDA, pulmonary edema Complete repair TAPVC, ligation of PDA No No
2004 40d/M/3 Pulmonary atresia, VSD, PDA Compete repair, ligation PDA Yes Yes
2004 5y/F/4E Ebstein’s anomaly, ASD, severe RV dysfunction TVR, ASD closure, bidirectional Glenn shunt; Converted to fenestrated Fontain No No.
2004 6d/M/4 HLHS, AVSD Norwood procedure, pulmonary homograft, reconstruction RV, PDA closure No No
2005 7y/F/4 Cleft MV, MVR, TVR, CHF MV, TV repair Yes Yes
2005 4d/F/4 CoA, hypoplastic arch, large PDA, PFO Repair of CoA, ligation PDA Yes Yes
2005 7d/M/3 TGA, VSD, ASD, bicuspidal pulmonary valve ASO, VSD+ASD closure Yes Yes
2005 11y/M/3 Congenital AS AVR Yes Yes
2005 4d/M/4 Congenital AS Ross-Konno aortoventriculoplasty Yes No
2006 21d/F/3 Obstructive TAPVC Compete repair Yes Yes
2006 21d/F/3 AVSD, 2 AV valve orifices, large ostium primum defect, VSD, hypoplastic L ventricle Biventricular repair Yes Yes
2006 19d/F/3 Unbalanced AVSD; HLHS and aortic arch; coarctation aorta. Modified Norwood, stage 1 No No
2006 25d/M/4 Obstructive TAPVS; PHTN Complete repair Yes Yes
2006 3m/F/4 Interrupted aortic arch; severe subaortic stenosis, severe valvular stenosis, VSD Complete repair; Heart transplantation (second operation);acute allograft rejection No No

AR, aortic regurgitation; ARF, acute renal failure; AS, aortic stenosis; ASA, American Society of Anesthesiologists; ASD, atrial septa defect; ASO, arterial switch operation; AV, atrio-ventricular; AVR, aortic valve replacement; AVSD, atrioventricular septal defect (AV canal); BV, biventricular; CA coronary artery; CABG, coronary artery bypass grafting; CoA, coarctation of the aorta; CHF, congestive heart failure; d, day; DORV, double outlet right ventricle; E, emergency; ECMO, extracorporeal membrane oxygenation; F, female; HLHS, hypoplastic left heart syndrome; IAA, interrupted aortic arch; IVS, intact ventricular septum; LV, left ventricular; LVOT, left venticular outflow tract; m, month; M, male; MS, mitral stenosis; MV, mitral valve; MVR, mitral valve regurgitation; PA, pulmonary artery; PDA, patent ductus arteriosus; PFO, patent foramen ovale; PHTN, pulmonary hypertension; PI, pulmonary insufficiency;  RV, right ventricle; S/P, status post; TAPVC, total anomalous pulmonary venous connection; TGA, transposition of great arteries; TOF, tetralogy of Fallot; TR, tricuspidal regurgitation; TV,  tricuspidal valve; TVR, tricuspidal valve regurgitation; VSD, ventricular septal defect. RACHS-1 risk for in-hospital death, see classification in Jenkins et al. (16)

Table 2. Bivariate association between patient and procedural characteristics and ability to be weaned from extracorporeal membrane oxygenation (ECMO) or hospital survival.

Variable N* Able to Wean
N (%)
P-value† Hospital Survival
N (%)
Overall 35 24 (68.6) 19 (54.3)
Gender 0.708 0.503
Male 22 16 (72.7) 13 (59.1)
Female 13 8 (61.5) 6 (46.2)
ASA PS 0.733 0.562
3 12 9 (75.0) 5 (41.7)
4 21 14 (66.7) 13 (61.9)
5 2 1 (50.0) 1 (50)
Age groups
0-30 days
31 days-1-17 years
14 (73.7)
5 (62.5)
0.701 10 (52.6)
4 (50.0)
5 (62.5)
Heart physiology 0.226 1.000
Univentricular 8 4 (50) 4 (50)
Biventricular 27 20 (74.1) 15 (55.6)
Surgical repair 0.226 1.000
Incomplete 8 4 (50) 4 (50)
Complete 27 20 (74.1) 15 (55.6)
Emergency 0.640 0.187
No 29 19 (65.5) 14 (48.3)
Yes 6 5 (83.3) 5 (83.3)
Stable at arrival to OR‡ 1.000 0.503
No 13 9 (69.2) 6 (46.2)
Yes 22 15 (68.2) 13 (59.1)
Circulatory Arrest 1.000 0.723
No 10 7 (70.0) 6 (60)
Yes 25 17 (68.0) 13 (52)
Intraoperative use of UF/HF (any) 0.063 0.379
No 6 2 (33.3) 2 (33.3)
Yes 29 22 (75.9) 17 (58.6)
LVAD or IABP on arrival to OR 0.314 0.457
No 34 24 (70.6) 19 (55.9)
Yes 1 0 (0.0) 0 (0)

*Overall number in each category used as the denominator when calculating the percentage of patients able to wean from ECMO.
†P-values are from Fisher’s exact.
‡Patients were considered unstable if they required either mechanical ventilation or vasopressor support on arrival to the operating room (OR).
Abbreviations:  UF/HF-ultrafiltration or hemofiltration; American Society of Anesthesiologists Physical Status (ASA PS); LVAD, left ventricular assist device; IABP, intraaortic balloon pump.

Table 3. Bivariate association between procedural characteristics (continuous variables) and ability to be weaned from extracorporeal membrane oxygenation (ECMO) and hospital survival.

Variable ECMO
Unable to Wean
Able to Wean
P† Hospital
Anesthesia time before   ECMO (hr) 7.5 (4.3, 9.4) 6.5 (4.0, 10.5) 0.099 7.5 (4.1, 9.4) 6.6 (4.0, 10.5) 0.191
Duration of CPB (min) 256 (87, 433) 189.5 (99, 401) 0.140 233 (87, 433) 191 (105, 401) 0.778
Time of ECMO (hr) 106 (4, 1261) 60.5 (19, 171) 0.126 100.5 (4, 1261) 57 (20, 171) 0.132
Highest creatinine on ECMO 0.8 (0.5, 2.7) 0.75 (0.3, 2.4) 0.421 0.85 (0.3, 2.7) 0.7 (0.4, 2.4) 0.309
Lowest pH on ECMO 7.32 (7.14, 7.36) 7.35 (7.14, 7.49) 0.105 7.32 (7.14, 7.36) 7.36 (7.15, 7.49) 0.006

†P-values are from the rank-sum test.
Values are median (min-max).

Generally speaking, longer organ ischemia and/or time on bypass, use of circulatory arrest, are all surrogate markers of increased complexity of repair, and therefore could be associated with mortality. However, none of these factors played a significant role in weaning from ECMO or hospital survival in our patients. Another surrogate index of less reversible organ dysfunction is extended time spent on ECMO (14), and its use for more than 72 hours has been related with poor outcome. (9) Others reported no survival after use of ECMO more than 208 (15) or 256 hours. (12) While our non-survivors spent longer time on ECMO (ECMO times were 60.5 vs. 106 hours, for survivors vs. non-survivors, respectively) the difference was not statistically significant. In contrast to some reports (12,14,15), Morris et al. (11) found that ECMO duration over 72 hours is not necessarily the predictor of poor prognosis. Therefore, protracted use of ECMO appears not to be a clear indication for its discontinuation.
Study Limitations. The limitations of this study are the small number of ECMO cases which precludes the performance of meaningful statistical analyses of risk factors associated with weaning and mortality. The retrospective nature of our review carries the inherent limitation of the potentially inconsistent nature of information in the medical record. We cannot exclude that relevant information was omitted from the description of some events.
In conclusion, in pediatric patient who require intraoperative ECMO following CPB hospital survival was 54.3%. We identified association between lowest pH during ECMO treatment and survival and no other demographic or procedural factors were associated with either weaning from ECMO or hospital survival. Therefore, it appears that the complexity of repair and surgical factors related to complex heart anomalies determined the outcome in our patient cohort.


We are grateful to: Shonie L. Buenvenida, R.N., for help in organizing data collection for this study; Mr. Carl Walin (Mayo Clinic Anesthesia Information Services) and Mrs. Judy Lenoch (Data Management Specialists, Division of Cardiovascular Surgery Research, Mayo Clinic) for assistance in data retrieval.


01. Krian A, Kramer HH, Quaegebeur J, Ostermeyer J, Korbmacher B, Godehardt E, et al. The arterial switch-operation: early and midterm (6 years) results with particular reference to technical problems. Thorac Cardiovasc Surg 1991;39 Suppl 2:160-5.
02. Dalton HJ, Siewers RD, Fuhrman BP, Del Nido P, Thompson AE, Shaver MG, et al. Extracorporeal membrane oxygenation for cardiac rescue in children with severe myocardial dysfunction. Crit Care Med 1993;21:1020-8.
03. Langley SM, Sheppard SV, Tsang VT, Monro JL, Lamb RK. When is extracorporeal life support worthwhile following repair of congenital heart disease in children? Eur J Cardiothorac Surg 1998;13:520-5.
04. Walters HL, Hakimi M, Rice MD, Lyons JM, Whittlesey GC, Klein MD. Pediatric cardiac surgical ECMO: multivariate analysis of risk factors for hospital death. Ann Thorac Surg 1995;60:329-36.
05. Ziomek S, Harrell JE, Jr., Fasules JW, Faulkner SC, Chipman CW, Moss M, et al. Extracorporeal membrane oxygenation for cardiac failure after congenital heart operation. Ann Thorac Surg 1992;54:861-7.
06. Flick RP, Sprung J, Gleich SJ, Barnes RD, Warner DO, Dearani JA, et al. Intraoperative extracorporeal membrane oxygenation and survival of pediatric patients undergoing repair of congenital heart disease. Ped Anesth 2008;18:757-66.
07. Anderson RH, Cook AC. Morphology of the functionally univentricular heart. Cardiol Young 2004;14 Suppl 1:3-12.
08. Krishnan U. Univentricular heart: management options. Indian J Pediatr 2005;72:519-24.
09. Aharon AS, Drinkwater DC, Jr., Churchwell KB, Quisling SV, Reddy VS, Taylor M, et al. Extracorporeal membrane oxygenation in children after repair of congenital cardiac lesions. Ann Thorac Surg 2001;72:2095-101.
10. Raithel SC, Pennington DG, Boegner E, Fiore A, Weber TR. Extracorporeal membrane oxygenation in children after cardiac surgery. Circulation 1992;86:II305-10.
11. Morris MC, Ittenbach RF, Godinez RI, Portnoy JD, Tabbutt S, Hanna BD, et al. Risk factors for mortality in 137 pediatric cardiac intensive care unit patients managed with extracorporeal membrane oxygenation. Crit Care Med 2004;32:1061-9.
12. Montgomery VL, Strotman JM, Ross MP. Impact of multiple organ system dysfunction and nosocomial infections on survival of children treated with extracorporeal membrane oxygenation after heart surgery. Crit Care Med 2000;28:526-31.
13. Kolovos NS, Bratton SL, Moler FW, Bove EL, Ohye RG, Bartlett RH, et al. Outcome of pediatric patients treated with extracorporeal life support after cardiac surgery. Ann Thorac Surg 2003;76:1435-41.
14. Black MD, Coles JG, Williams WG, Rebeyka IM, Trusler GA, Bohn D, et al. Determinants of success in pediatric cardiac patients undergoing extracorporeal membrane oxygenation. Ann Thorac Surg 1995;60:133-8.
15. Kulik TJ, Moler FW, Palmisano JM, Custer JR, Mosca RS, Bove EL, et al. Outcome-associated factors in pediatric patients treated with extracorporeal membrane oxygenator after cardiac surgery. Circulation 1996;94:II63-8.
16. Jenkins KJ, Gauvreau K, Newburger JW, Spray TL, Moller JH, Iezzoni LI. Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg 2002;123:110-8.

Creative Commons LicenseThis work is licensed under a Creative Commons Attribution 4.0 International License