| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;76:1435-1441
© 2003 The Society of Thoracic Surgeons
a Department of Pediatrics and Communicable Diseases, Ann Arbor, MI, USA
b Department of Surgery, University of Michigan, C. S. Mott Children's Hospital, Ann Arbor, Michigan, USA
Accepted for publication May 14, 2003.
* Address reprint requests to Dr Bratton, F6884 Mott, 0243, University of Michigan Health System, 1500 East Medical Center Drive, Ann Arbor, MI 48109-0243, USA.
e-mail: brattons{at}med.umich.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentReferences |
|---|
METHODS: All pediatric patients who received ECLS support within 7 days after surgery between July 1995 and June 2001 were examined to describe clinical features associated with survival. We compared the results with our prior report to assess changes in practice and outcome.
RESULTS: Seventy-four patients were followed. Fifty percent survived to discharge. Hospital survival was not significantly related to patient age, cannulation site, or indication. Thirty-five percent of patients required hemofiltration while on ECLS and were significantly less likely to survive (23% vs 65%). A multivariate analysis combining all children from our prior report with the present cohort revealed that patients who received hemofiltration were five times more likely to die (odds ratio 5.01, 95% confidence interval 2.11–11.88). Children with an adequate two-ventricular repair had lower risk of death (odds ratio 0.42, 95% confidence interval 0.19–0.91) after adjusting for patient age, study period, and hours elapsed before initiation of ECLS after surgery.
CONCLUSIONS: Patients with an adequate two-ventricle repair have significantly higher hospital survival, whereas those with single ventricle physiology or need for dialysis have decreased survival.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
In 1996 we published the outcomes of pediatric patients treated with ECLS after cardiac surgery and described clinical features associated with death in those patients [1]. Only 33% of patients survived to hospital discharge. Patients with single ventricle physiology, those who were unable to be separated from cardiopulmonary bypass, and those with prolonged ECLS runs were significantly less likely to survive.
In this report we review our most recent experience among pediatric cardiac patients treated with ECLS after surgery to characterize changes in patient selection and clinical outcome. We then combined the present cohort with the patients reported in 1996 to increase statistical power in evaluation of factors associated with death for the entire University of Michigan experience.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Diagnostic and outcome definitions
In order to compare our present cohort with an earlier report from the University of Michigan, we used previously described surgical diagnostic groups, indications for ECLS and outcome categories [1]. Group 1 consisted of patients with two ventricles who underwent complete repair. Group 2 consisted of 6 children who were thought to have an inadequate two-ventricle repair. Group 3 patients had pulmonary blood flow provided by a surgically placed systemic-to-pulmonary artery shunt. Group 4 consisted of patients with pulmonary blood flow provided by a cavopulmonary connection (eg, a hemi-Fontan, bidirectional Glenn) or a Fontan operation. Patients in groups 3 and 4 had single ventricle physiology. Group 5 patients had cardiac transplantation.
The general indications for ECLS were inadequate organ perfusion, oxygenation, or ventilation. Specific indications for ECLS included ventricular dysfunction (systolic, diastolic, or both); pulmonary failure (oxygenation, ventilation, or a combination); increased pulmonary vascular resistance, or an occluded systemic-to-pulmonary artery shunt [1]. Multiple indications were recorded if a combination of two or more conditions existed.
Discontinuation of ECLS was based upon either sufficient cardiorespiratory stability to successfully wean support, or the determination that irreversible dysfunction in one or more organ systems had developed that precluded survival. The decision to wean a child from ECLS was determined by multiple factors including cardiac and pulmonary function, as well as the patient's extra vascular fluid status. Generally patients were treated with dopamine and low dose epinephrine infusions. Resuscitative fluids for volume loading were routinely available. Trials were generally instituted when the patient was near ideal body weight, had a stable cardiac rhythm, and no radiographic infiltrates. Patients were sedated and frequently muscle relaxed during trials off to minimize oxygen consumption. The ventilator was adjusted to provide full pulmonary support. Trials off support were generally less than 4 hours with frequent blood gas testing and clinical assessment of perfusion. Currently lactate levels are assessed with each blood gas determination.
Complications diagnosed while the patient was on ECLS were recorded and included: seizures, radiologic evidence of a stroke or intracranial hemorrhage, culture positive infection, and need for dialysis.
ECLS technique
The ECLS circuit consists of a servoregulated pump, membrane lung, and heat exchanger [5]. Cannulation was performed either in the operating room or in the intensive care unit. Children who required ECLS in the immediate postoperative period, or who experienced a precipitous circulatory decompensation or cardiac arrest, underwent transthoracic cannulation. The neck was the preferred site of cannulation among children who received ECLS support later in the postoperative period. Cannulation techniques were modified to suit specific anatomic details, particularly in patients with prior cavopulmonary shunts.
All patients were anticoagulated with heparin sulfate to achieve an activated clotting time of 180 to 200 seconds, unless patients were bleeding and the goal was lowered to 160 to 180 seconds. Platelets were transfused to keep the platelet count greater than 100,000/mm3 and red blood cells were generally transfused to maintain a hematocrit greater than 35% while on ECLS, and between 40% and 45% before stopping ECLS among patients with cyanotic heart disease.
Prophylactic antibiotics were routinely administered in all children with an open chest. ECLS flow was generally maintained between 80 and 120 mL/kg per minute. Vasoactive infusions were discontinued once adequate blood pressure was achieved using extracorporeal flow. Diuretics were administered in cases of fluid overload and a hemofilter was used to augment fluid removal at the discretion of the attending physician. Hemofiltration was performed using a Renaflo II Hemofilter (Minnitech Inc, Minneapolis, MN) in line with the ECLS circuit. Blood flow through the shunt is estimated using a Transonic HT 110 bypass flow meter (Transonic Systems Inc, Ithaca, NY). Patient fluid removal was largely dictated by clinical stability; generally, rate of removal did not exceed 2 to 3 mL/kg per hour.
The decision to decompress the systemic ventricle was made on the basis of elevated atrial pressure or echocardiographic evidence of ventricular distension and poor ejection. The ventricle was vented using a venous return cannula placed in the left atrium or by performing an atrial septostomy in the catheterization laboratory. Regarding intentional shunt occlusion, we extended our review to include our cardiac ECLS experience from January 1989 to June 2001 because relatively few patients with a systemic-to-pulmonary artery shunt were treated with ECLS. This information is presented separately in the Results and Comment sections.
Analysis
The demographic and outcome information were analyzed with a standard statistical package (SPSS 10.0 for Windows; SPSS Incorporated, Chicago, IL). Data are presented as medians with 25th and 75th quartiles. Children who had more than one course of ECLS during a hospitalization were counted once and information from the initial surgery reported. Total duration of ECLS was the sum of all time on ECLS for a hospital admission.
Children who survived to hospital discharge are compared with those who died in the present cohort. Categoric data are compared with the 
2 test, while continuous data are compared using the Mann-Whitney U test. In order to better determine features associated with death for the entire University of Michigan experience, we combined our prior cohort of children [1] and the present study group. The relationship between clinical features that were significantly associated with death in the univariate analysis was examined by calculating the adjusted odds ratio using multiple logistic regression. All variables that were significantly associated with death in the univariate analysis were added simultaneously to the multivariate model.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
|
|
The demographic and clinical characteristics of surviving patients and those who died before hospital discharge are presented in Table 3. Sixty-nine percent of patients were placed on ECLS for cardiac failure. Sixty-five percent of patients had two ventricles, whereas 28% had a single ventricle and a systemic-to-pulmonary artery shunt. Thirty-five percent were cannulated through the neck and 99% received venoarterial support. Twenty-four percent of patients sustained a cardiac arrest before or at the time of ECLS cannulation.
|
Hospital survival was not significantly related to patient age, mode of ECLS support, site of cannulation, or occurrence of a treated cardiac arrest before cannulation. Twenty-eight children were placed on ECLS in the operating suite; of these, 64% survived to hospital discharge compared with 41% survival among the 46 children in whom ECLS was initiated in the cardiothoracic intensive care unit (p = 0.06). Eight of 28 children who started ECLS in the operating suite required dialysis (29%) compared with 39% who had ECLS initiated later (p = 0.4). Patients undergoing cardiopulmonary resuscitation (CPR) at the time of cannulation were significantly less likely to survive than those who did not require ongoing CPR (Table 3).
Patients with two ventricles tended to have better survival compared with those with one ventricle (58% vs 34%). Only 20% of patients with a cavopulmonary connection survived to hospital discharge, whereas 38% of patients with an aortopulmonary shunt and a single ventricle survived to hospital discharge (Table 3). The physiologic indication for initiation of ECLS was not significantly related to survival; however, all patients (n = 4) with increased pulmonary vascular resistance survived.
Hospital survival was not significantly related to serum creatinine measured immediately before ECLS or to the highest level while on ECLS (Table 4). Thirty-five percent of patients underwent hemofiltration while receiving ECLS; those patients were significantly less likely to survive than those who did not (p = 0.001). Serum lactate levels were available for 44 patients of the most recently treated patients. The highest serum lactate within 48 hours of cannulation was inversely related to survival. However, the lowest lactate within 24 hours following initiation of ECLS was not significantly related to survival. Fourteen percent of patients had seizures and 7% had radiologic evidence of central nervous system injury. Fifteen percent of patients had a documented infection while on ECLS despite prophylactic antibiotic therapy. However, the only complication that was significantly related to death was need for hemofiltration during ECLS.
|
Systemic-to-pulmonary artery shunt management
Between January 1989 and June 2001, 28 patients underwent ECLS within 7 days of a Norwood or Damus-Kaye-Stansel palliation. The systemic-to-pulmonary shunt was left patent intentionally in 18 children (although 3 patients were found to have thrombus partially occluding their shunt by angiography performed during their periods of support). In 5 patients (4 patients in the current study period), the shunt was intentionally completely occluded at the time of cannulation and for at least 24 hours thereafter. One patient had the shunt partially occluded by a Heifletz clip. Whether the shunt had been occluded in 4 patients could not be ascertained from review of the medical records. Three of 5 patients whose shunt was intentionally occluded developed one or more of the following complications: severe pulmonary edema noted on chest radiography, pulmonary hemorrhage, and noncompliant lungs. These 3 patients all died, at least in part due to respiratory failure. Only 1 patient with a nonintentionally occluded systemic-to-pulmonary shunt died of respiratory failure, an infant with RSV pneumonia.
Factors associated with survival
A multivariate analysis of factors associated with death was performed, which included hours after surgery before ECLS, presence of ongoing CPR, dialysis during ECLS, and surgical diagnosis. Because we were limited by relatively small sample size, surgical diagnosis was simplified to adequate two ventricle repair or not. Children with an adequate two ventricle repair tended to have lower risk of death (odds ratio 0.40, 95% confidence interval 0.1–1.2), whereas need for dialysis was associated with more than fivefold increased risk of death (odds ratio 5.6, 95% confidence interval 1.8–17.9), but neither ongoing CPR nor hours elapsed before ECLS were significantly associated with death.
In an effort to determine why survival was higher in our present population compared with our previous experience, we compared the 64 patients from our previous report with our present study; the data are presented in Table 5. Overall survival significantly improved from 33% in the past cohort to 50% in the current group. Patients in the present cohort were significantly younger and had initiation of ECLS sooner in the postoperative period. Among patients with a single ventricle, significantly more recent patients had pulmonary blood flow supplied by a systemic-to-pulmonary shunt than by a cavopulmonary connection.
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Initiation of ECLS in the operating suite tended to be associated with improved survival and decreased need for dialysis in the most recent cohort, although this difference was not statistically significant. These results differ from our earlier report, and several others [2–3, 6, 8], which suggest that likelihood of survival is decreased among patients in whom ECLS was initiated in the operating room. The decision to place a child on ECLS in the operating suite is determined by the surgeon's judgment and is affected by many factors including ventricular function, length of cardiopulmonary bypass, and pulmonary function.
Our center does not currently offer rapid resuscitation ECLS for patients with ongoing CPR. The experience with rapid resuscitation ECLS is included in the current cohort and was associated with 80% mortality. We can provide ECLS support within 15 minutes using a crystalloid primed circuit and currently attempt to initiate ECLS before cardiac arrest. However, if a patient is actively receiving chest compressions, we generally do not start cannulation for ECLS.
Need for dialysis was associated with a markedly worse prognosis. This association has been reported by others [1–3, 9] and probably reflects secondary organ injury due to cardiovascular collapse. The fivefold increase in risk was of slightly higher magnitude than the unadjusted estimates (odds ratio 2 to 4) reported by Duncan and colleagues [2], and by Montgomery and associates [3]. When to initiate hemofiltration during ECLS remains unclear. Early initiation of continuous venovenous hemofiltration among critically ill patients can decrease volume overload, which has been associated with increased mortality [9, 10]. However, aggressive fluid removal can exacerbate renal insufficiency [11]. Nonetheless, recent studies suggest that initiation of intensive dialysis once a patient has developed renal failure is associated with improved survival [12, 13]. We generally initiate hemofiltration if a patient has sustained low urine output despite adequate blood pressure and aggressive diuretic therapy. By far the most common indication for hemofiltration is volume overload, rather then uremia or electrolyte imbalances.
We cannot fully explain why survival was better in the present cohort compared with the prior one. We continue to reserve ECLS for patients that we think are at very high risk of death without additional cardiorespiratory support. ECLS is used in a small minority of the most unstable patients (2.2% of surgical patients who required cardiopulmonary bypass). ECLS technology has not changed in substantive ways over the past 6 years at the University of Michigan. There were differences in patient surgical diagnoses with a larger proportion of infants having a single ventricle with an systemic-to-pulmonary shunt who tended to be more likely to die compared with patients with two ventricle repairs, but fewer patients with single ventricles and cavopulmonary connections. There were also fewer patients with residual surgical lesions in the most recent cohort. Indications for ECLS also differed with no patients supported with ECLS for sepsis in the recent group and fewer patients with multiple indications for ECLS. However, because of the diversity of diagnoses and indications these differences were not statistically significant. The overall proportion of children who developed need for dialysis was similar in the two groups whereas the time after surgery to initiation of ECLS was decreased in the recent cohort. Although time elapsed to initiation of ECLS was not statistically significant in the multivariate analysis, our current practice is to rapidly identify patients with severe cardiorespiratory failure after cardiac surgery and to provide ECLS if other measures do not facilitate rapid improvement.
As previously reported, myocardial function among postcardiotomy patients who survive returns within 3 to 5 days, and is very unlikely to improve sufficiently after 8 to 10 days of support [1, 3, 14]. Return of adequate function sufficient to survive off ECLS occurred within a median of 4.5 days in this cohort. However, children with primarily respiratory failure after cardiac surgery may benefit from longer ECLS [6, 15].
For patients with hypoplastic left heart syndrome and a systemic-to-pulmonary artery shunt we do not recommend intentional occlusion of the shunt because of the association of pulmonary injury. Jaggers and coworkers [16] have reported similar complications of pulmonary ischemic injury, and likewise recommend leaving the shunt open. Maneuvers to increase pulmonary vascular resistance, cardiac output, and ECLS support should be implemented if these infants have hypotension while on ECLS. We have used subambient FiO2 and decreased lung ventilation to increase pulmonary vascular resistance, as well as vasopressors and increased ECLS flow to increase systemic blood pressure. Darling and associates [17] describe the use of ECLS without an oxygenator to optimize systemic output. Our institution has yet to use this approach; however, this is an option in the management of low cardiac output.
Our study has limitations. The patients are from a single center with a large congenital heart surgery program; therefore, our findings may not be applicable in other clinical settings. Patients were identified prospectively; however, some of the clinical data were collected by chart review. Some laboratory data, such as lactate levels, were not available for the entire cohort. Clearly, surgical decisions to initiate ECLS in the operating suite may differ between surgeons and across institutions. However, we have analyzed a relatively large cohort of patients, and individual patient indications, complications, and reasons to discontinue ECLS were collected prospectively. These factors are frequently difficult to retrospectively determine by an audit of the medical record and have strengthened the analysis and conclusions.
Those patients with single ventricle physiology, continued elevation of lactate after institution of ECLS, and development of renal failure have increased risk of death. Patients with an adequate two-ventricle repair who do not develop other organ dysfunction have higher hospital survival rates. Advances in patient selection may improve outcome.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. J. E. Schuerer, N. S. Kolovos, K. V. Boyd, and C. M. Coopersmith Extracorporeal Membrane Oxygenation: Current Clinical Practice, Coding, and Reimbursement Chest, July 1, 2008; 134(1): 179 - 184. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. R. Davies, M. J. Russo, K. N. Hong, M. L. O'Byrne, D. P. Cork, A. J. Moskowitz, A. C. Gelijns, S. Mital, R. S. Mosca, and J. M. Chen The use of mechanical circulatory support as a bridge to transplantation in pediatric patients: an analysis of the United Network for Organ Sharing database. J. Thorac. Cardiovasc. Surg., February 1, 2008; 135(2): 421 - 427. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D Picarelli, C. Kreutzer, M Barboza, S Antunez, G Pose, G Touya, L Liguera, D Abdala, and G Echegaray Post-cardiotomy circulatory support with the Terumo Baby - Rx tm oxygenator in a newborn Perfusion, November 1, 2007; 22(6): 377 - 379. [Abstract] [PDF]
|
|
|
|
 |
|
 R. R. Thiagarajan, P. C. Laussen, P. T. Rycus, R. H. Bartlett, and S. L. Bratton Extracorporeal Membrane Oxygenation to Aid Cardiopulmonary Resuscitation in Infants and Children Circulation, October 9, 2007; 116(15): 1693 - 1700. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Wagner, I. Risnes, M. Abdelnoor, H. M. Karlsen, and J. L. Svennevig Is it possible to predict outcome in cardiac ECMO? Analysis of preoperative risk factors Perfusion, July 1, 2007; 22(4): 225 - 229. [Abstract] [PDF]
|
|
|
|
|
|
C. K. Allan, R. R. Thiagarajan, P. J. del Nido, S. J. Roth, M. C. Almodovar, and P. C. Laussen Indication for initiation of mechanical circulatory support impacts survival of infants with shunted single-ventricle circulation supported with extracorporeal membrane oxygenation J. Thorac. Cardiovasc. Surg., March 1, 2007; 133(3): 660 - 667. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. G. Ohye Invited commentary Ann. Thorac. Surg., November 1, 2006; 82(5): 1641 - 1642. [Full Text] [PDF]
|
|
|
|
|
|
V. H. Thourani, P. M. Kirshbom, K. R. Kanter, J. Simsic, B. E. Kogon, S. Wagoner, F. Dykes, J. Fortenberry, and J. M. Forbess Venoarterial extracorporeal membrane oxygenation (VA-ECMO) in pediatric cardiac support. Ann. Thorac. Surg., July 1, 2006; 82(1): 138 - 144. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-C. Huang, E.-T. Wu, W.-J. Ko, L.-P. Lai, J. Hsu, C.-I. Chang, I.-S. Chiu, S.-S. Wang, M.-H. Wu, F.-Y. Lin, et al. Clinical Implication of Blood Levels of B-Type Natriuretic Peptide in Pediatric Patients on Mechanical Circulatory Support Ann. Thorac. Surg., June 1, 2006; 81(6): 2267 - 2272. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Hoskote, D. Bohn, C. Gruenwald, D. Edgell, S. Cai, I. Adatia, and G. Van Arsdell Extracorporeal life support after staged palliation of a functional single ventricle: Subsequent morbidity and survival J. Thorac. Cardiovasc. Surg., May 1, 2006; 131(5): 1114 - 1121. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Hannan, M. A. Ybarra, J. A. White, J. W. Ojito, A. F. Rossi, and R. P. Burke Patterns of Lactate Values after Congenital Heart Surgery and Timing of Cardiopulmonary Support Ann. Thorac. Surg., October 1, 2005; 80(4): 1468 - 1474. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. T. Mahle, J. M. Forbess, P. M. Kirshbom, A. R. Cuadrado, J. M. Simsic, and K. R. Kanter Cost-utility analysis of salvage cardiac extracorporeal membrane oxygenation in children J. Thorac. Cardiovasc. Surg., May 1, 2005; 129(5): 1084 - 1090. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Rellensmann, T. Krasemann, and H.-G. Kehl Should All Stage-One Norwood Patients Receive a Prolonged Period of Postoperative Mechanical Circulatory Support? Ann. Thorac. Surg., March 1, 2005; 79(3): 1098 - 1099. [Full Text] [PDF]
|
|
|
|
 |
|
 L. Lequier Extracorporeal Life Support in Pediatric and Neonatal Critical Care: A Review J Intensive Care Med, September 1, 2004; 19(5): 243 - 258. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
