| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1995;60:329-336
© 1995 The Society of Thoracic Surgeons
Departments of Cardiovascular Surgery and Pediatric General Surgery, Children's Hospital of Michigan, Wayne State University School of Medicine, Detroit, Michigan
 |
Abstract |
|---|
 TopFootnotes Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Methods. We retrospectively reviewed the records of 73 pediatric patients with congenital heart disease who were placed on ECMO support between August 1984 and February 1994. The patients were divided into groups defined by the timing of ECMO cannulation relative to the time of operation. Group 1 patients (n = 7, 9.6%) were placed on ECMO preoperatively. Group 2 patients (n = 66, 90.4%) were a heterogeneous population placed on ECMO at any interval after cardiac repair. Subgroup 2A consisted of patients (n = 17, 25.8%) who could not be weaned from cardiopulmonary bypass and were converted directly to ECMO support immediately after repair. Subgroup 2B patients (n = 49, 74.2%) were cannulated postoperatively after an initial period of clinical stability.
Results. Hospital survival for all study patients (42/73) and for group 2 patients (38/66) was 58%. Only 4 group 2A patients (23.5%) survived their hospitalization compared with 34 group 2B patients (69.4%) (p = 0.001). Multivariate analysis identified elevated right atrial pressure after ECMO decannulation (p = 0.049) and, possibly, membership in group 2A (p = 0.061) as independent risk factors for hospital death.
Conclusions. Extracorporeal membrane oxygenation is most effective in salvaging pediatric cardiac surgical patients who demonstrate medically refractory hemodynamic deterioration at some interval after being successfully weaned from cardiopulmonary bypass. The right atrial pressure after extracorporeal membrane oxygenation decannulation is an independent predictor of hospital death.
|
Introduction |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Extracorporeal membrane oxygenation (ECMO) provides prolonged mechanical cardiopulmonary support by combining venoarterial bypass with membrane oxygenation. Its use in patients with congenital heart disease dates back to 1970 when Baffes and colleagues [1] reported using ECMO to support infants with congenital heart defects and in unstable condition who were undergoing palliative cardiac surgical procedures. Seven years later, Bartlett and associates [2] reported promising results in 28 adult and pediatric patients treated with ECMO for various forms of cardiopulmonary disease and predicted a future expanded role for this technique in the treatment of pediatric patients with refractory cardiac failure. Since that time, a number of pediatric postcardiotomy ECMO series [3–12] with varying rates of hospital survival have been published.
Despite the availability of intraaortic balloon counterpulsation for children [13–16], ECMO has emerged as an effective technique for the mechanical support of many pediatric postcardiotomy patients with medically refractory cardiac failure. Pediatric applications of intraaortic balloon counterpulsation have been hampered by difficulties in balloon placement, constraints on balloon size imposed by the smaller pediatric vessel caliber, difficulties in synchronization of the balloon cycle to the faster pediatric heart rates, and the reduced diastolic augmentation achieved with the intraaortic balloon pump because of the very elastic pediatric aorta. Further, intraaortic balloon counterpulsation primarily supports the left ventricle and does little to treat the right heart failure that frequently accompanies pediatric low cardiac output (LCO) after cardiotomy.
The purpose of this retrospective review of our past 9
years' experience with pediatric ECMO in cardiac surgical patients was to analyze our results and to identify independent risk factors for hospital death.
|
Material and Methods |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
These 73 pediatric cardiac surgical ECMO patients were divided into groups defined by the timing of ECMO cannulation relative to the time of each patient's cardiac operation. Group 1 patients (n = 7, 10%) were placed on ECMO preoperatively. Group 2 patients (n = 66, 90%) were a heterogeneous population placed on ECMO at any interval after cardiac repair. Group 2 patients were arbitrarily divided into two subgroups. Subgroup 2A consisted of patients (n = 17, 26%) who could not be weaned from cardiopulmonary bypass (CPB) and were converted directly to ECMO support while still in the operating room. Subgroup 2B patients (n = 49, 74%) were successfully weaned from CPB after the surgical repair and were transferred to the pediatric intensive care unit. Medically refractory cardiac or pulmonary failure or both developed after 32.8 ± 4.2 hours (mean ± standard error of the mean) of initial clinical stability, and the patients were then placed on ECMO support.
Group 1 (Placed on ECMO Preoperatively)
Group 1 had 5 boys (71.4%) and 2 girls (28.6%). Their median age and weight were significantly less than those of the group 2 patients (0.1 month versus 7.9 months [p = 0.002] and 3.9 kg versus 6.2 kg [p = 0.01]) (Table 1
). Five group 1 patients (71.4%) were seen with profound LCO, hypoxia, or both (Table 2). Four of these 5 patients were moribund at presentation because of severe pulmonary venous obstruction related to total anomalous pulmonary venous connection (n = 3) and cor triatriatum (n = 1). The fifth patient was seen with profound hypoxia resulting from transposition of the great arteries with intact interventricular septum. Despite the presence of a patulous, nonrestrictive interatrial communication after balloon atrial septostomy and a large patent ductus arteriosus on prostaglandin E1 infusion, profound hypoxia with acidosis persisted. These 5 patients were placed on ECMO to stabilize their clinical condition prior to definitive cardiac repair.
|
|
Demographics for Total Population and Group 2
The demographic data for all study groups are summarized in Table 1
. The total ECMO population consisted of 44 boys (60.3%) and 29 girls (39.7%) with a median age at operation of 7.2 months and a median weight of 5.6 kg. Group 2 (cannulated at any interval after repair) comprised 39 boys (59.1%) and 27 girls (40.9%) with a median age of 7.9 months and a median weight of 6.2 kg. There were no significant differences in demographic variables between subgroup 2A and subgroup 2B (see Table 1).
Cardiac Diagnoses for Group 2
Of the 20 congenital cardiac diagnoses represented in the study population, complete atrioventricular canal defect (n = 13, 19.7%), single ventricle (n = 11, 16.7%), tetralogy of Fallot (n = 9, 13.6%), total anomalous pulmonary venous connection (n = 5, 7.6%), and double-outlet right ventricle (n = 5, 7.6%) were the most common in group 2. All of the patients with complete atrioventricular canal defect were in group 2, and all had associated trisomy 21.
Indications for ECMO in Total Population and Group 2
The indications for ECMO included the following medically refractory clinical conditions: (1) LCO, (2) pulmonary artery hypertensive crisis (PAHCs), (3) profound hypoxia, (4) cardiac arrest, (5) cardiac arrhythmia, and (6) lack of resumption of spontaneous electric activity after surgical repair (see Table 2). Our patients were placed on ECMO only after all reasonable alternative therapies had been exhausted, and it was thought that the patient's death was inevitable within the next 12 to 24 hours without mechanical support. Absolute contraindications to ECMO included the presence of multiple-organ failure, severe neurologic dysfunction, and uncontrollable hemorrhage. A relative contraindication was the presence of a systemic-to-pulmonary artery shunt and documented untreated sepsis. The adequacy of the cardiac repair was confirmed in all patients by two-dimensional echocardiography, cardiac catheterization with angiography, or both.
Eighteen patients (24.7%) in the total study population were seen with more than one indication for ECMO. The most common indications for ECMO in the total group were LCO (n = 64, 87.7%) and PAHC (n = 14, 19.2%). Low cardiac output (n = 61, 92.4%) and PAHC (n = 14, 21.2%) were also the most common indications for institution of ECMO support in the patients who were cannulated at any interval after cardiac repair (group 2) (see Table 2).
ECMO Protocol
CANNULATION.
Neck cannulation, if not anatomically contraindicated, was generally preferred to reduce the probability of mediastinal sepsis. However, as might be expected, a higher proportion of subgroup 2A patients were cannulated in the chest than in subgroup 2B (58.8% versus 16.3%; p = 0.001) because the cannulas were already in place, and the institution of ECMO involved only switching the cannula connections to the ECMO circuit. In subgroup 2B, chest cannulation was reserved for patients whose anatomy precluded the placement of neck cannulas or for patients in whom it was thought that adequate cardiac decompression was not achieved by neck cannulation. Hence, 83.7% of subgroup 2B patients were cannulated in the neck rather than the chest.
Cannulation was performed by placing the tip of the venous cannula in the right atrium through the right internal jugular vein (ECMO cannula; Bio-Medicus, Inc, Eden Prairie, MN) or by placing it directly through the open sternum (Pacifico venous cannula; DLP, Inc, Grand Rapids, MI). When chest cannulation was employed, a left atrial cannula was placed, if necessary, to separately decompress the left side of the heart. The arterial cannula was placed in the aortic arch through the right common carotid artery (Bio-Medicus ECMO cannula) or into the ascending aorta (Argyle angled aortic perfusion cannula; Sherwood Medical, St. Louis, MO) directly through the open sternum.
ECMO CIRCUIT.
Blood drained by gravity through the venous cannula into a servoregulating venous return monitor (Seabrook Medical Systems, Inc, Cincinnati, OH) to a roller pump (Cobe Laboratories, Inc, Lakewood, CO). The venous blood was then pumped through an appropriately sized membrane oxygenator and heat exchanger (Avecor Cardiovascular, Inc, Plymouth, MN) to be returned to the patient through the arterial cannula. Temperature (Seabrook, Medical Systems, Inc), pressure (Escort patient monitor; Medical Data Electronics, Inc, Arleta, CA), and oxygen saturation (7820 oxygen monitor; Puritan-Bennett, Inc, Carlsbad, CA) monitors as well as an air detector (Minnilert RS-3275; Minnitech Corporation, Inc, Minneapolis, MN) are standard components of our circuit. A gas blender with two flowmeters (for oxygen and carbon dioxide) to the membrane oxygenator facilitate control of the arterial oxygen tension and arterial carbon dioxide tension without changing the pump flow.
ANTICOAGULATION.
Anticoagulation was achieved by a constant heparin sodium infusion maintaining an activated clotting time of 180 to 220 seconds. Platelet transfusions were given to maintain a platelet count of greater than 100,000/µL [17]. The prothrombin time was kept as near normal as possible with the infusion of 10 mL/kg of fresh frozen plasma every 6 hours and with vitamin K administration. The hemoglobin was maintained greater than 10 mg/dL with transfusions of packed red blood cells as indicated. In patients with an open fontanel, daily cranial ultrasound studies were obtained to detect intracranial hemorrhage.
FLOWS.
Whereas the pump flows are adjusted to maintain satisfactory oxygenation in a neonate with persistent pulmonary hypertension, the primary goal of postcardiotomy ECMO therapy usually is to restore and maintain optimal tissue perfusion while cardiac recovery takes place. The partial or nearly total bypass produced by ECMO in these patients decompresses the heart and reduces the work it must perform, thereby decreasing its energy requirements. The improvement in arterial oxygen saturations produced by the membrane oxygenator also lowers pulmonary vascular resistance. This theoretically maximizes the potential for myocardial recovery, resolution of the PAHC, or both. Pump flows (100 to 150 mL • kg-1 • min-1) were targeted to maintain a urine output of greater than 1 mL • kg-1 • h-1, toe temperatures greater than 30°C, and a brisk capillary refill and to reduce left and right atrial pressures maximally. Immediately after the institution of ECMO, all inotropic medications, except renal dose dopamine hydrochloride, were discontinued. A peripheral vasodilator, such as sodium nitroprusside, was frequently used to treat systemic arterial hypertension [18] and to improve peripheral perfusion.
VENTILATOR.
Shortly after the institution of ECMO, the ventilator inspired oxygen fraction was typically reduced to 0.25 if the patient was totally bypassed without any major cardiac contribution to the total systemic flow. The positive end-expiratory pressure was set at 6 to 10 cm H2O, and the respiratory rate was reduced to 15 to 25 breaths/min. During periods of partial bypass, when the patient's own ejection contributed to the total systemic flow, the inspired oxygen fraction was elevated to at least 0.40 to ensure adequate arterial oxygen saturations at the level of the coronary ostia. At all times during the ECMO run, the arterial oxygen saturations were maintained at 98% to 100%, and the acid-base status was adjusted to within the normal range. The patient's temperature was maintained as close to normothermia as possible.
FLUID MANAGEMENT.
Prior to the institution of ECMO, the intravascular volume was invariably expanded to optimize atrial filling pressures and cardiac output. After the beginning of ECMO, ultrafiltration (Minifilter Plus; Amicon Corp, Danvers, MA) and diuresis were routinely employed to decrease interstitial edema and to lower filling pressures by reducing the intravascular volume. Ultrafiltration was used in 56 (76.7%) of the 73 ECMO patients in this series. Hemodialysis was used in 3 patients (4.1%) to palliate acute renal insufficiency.
WEANING.
The patient was generally maintained on full ECMO support for a minimum of 72 hours. At that time, the flows were gradually reduced by small increments at hourly intervals over a 12- to 24-hour period. During this period of reduction in flow, the ventilator settings were gradually adjusted upward to maintain 100% arterial saturations, normocarbia, and a normal acid-base status. The left and right atrial pressures were allowed to rise to normal levels, augmented by volume boluses as indicated. As the reduction in flow continued, inotropic agents, systemic vasodilators, and pulmonary vasodilators were added, as necessary, to optimize the cardiac output and to treat mild pulmonary artery hypertension. As long as the clinical assessment of cardiac output remained satisfactory, the weaning from the ECMO flow was continued. Arbitrarily, the ECMO flow was not allowed to drop to less than 100 to 200 mL/min to avoid thrombosis with embolization. Assuming a satisfactory cardiac output during the entire wean, the patient was given a 10-minute trial off ECMO and was then decannulated if the trial was well tolerated. The cervical vessels were not reconstructed after ECMO decannulation. Delayed sternal closure [19] was performed at the time of decannulation in patients who were cannulated by the transthoracic route. In patients who were cannulated cervically and who also had an open sternum, delayed sternal closure was usually performed the day after ECMO decannulation.
Data Collection and Analysis
The data collected for each patient consisted of 189 variables that were sorted on Microsoft Excel for Windows program version 4.0 (Microsoft Corp, Redmond, WA). These data were downloaded into SPSS for Windows release 5.0 (SPSS Inc, Chicago, IL) for statistical analysis. Quantitative variables that approximated a normal distribution were reported as the mean ± the standard error of the mean and were analyzed by Student's unpaired t test. Quantitative variables that did not approximate a normal distribution were reported as the median and were analyzed by the Mann-Whitney U test (Wilcoxon rank-sum test). Nominal variables were analyzed nonparametrically by Fisher's exact test or 
2 test.
Fifty-one clinical variables were subjected to an initial univariate analysis to screen for possible predictors of the outcome event hospital death. To detect independent predictors of hospital death, the following nine variables (identified in the univariate analysis with a p value of less than 0.1) were subjected to multivariate analysis: (1) duration of CPB, (2) group 2A membership, (3) serum creatinine and (4) blood urea nitrogen values 48 hours after ECMO cannulation, (5) red blood cells and (6) fresh frozen plasma transfused (mL • m-2 • h-1) while on ECMO, (7) serum creatinine value after ECMO decannulation, (8) right atrial pressure averaged over the first 8 hours after ECMO decannulation, and (9) PAHC as the indication for ECMO. The multivariate analysis employed the stepwise forward Wald multiple logistic regression method to construct a model to predict the probability of hospital death in group 2.
|
Results |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Group 1
The duration of ECMO in the 7 group 1 patients (placed on ECMO preoperatively) was 100.0 ± 17.6 hours. Five (71.4%) of these patients were weaned from ECMO support, and 4 of these 5 survived hospitalization (57.1% group 1 hospital survival). Both patients whose cardiac diagnosis had not been correctly established preoperatively survived and are doing well. The 3 patients who died had (1) severe pulmonary venous obstruction associated with total anomalous pulmonary venous connection plus complete atrioventricular canal defect and severe preoperative atrioventricular valve insufficiency, (2) cor triatriatum status post cardiac arrest in the cardiac catheterization laboratory, and (3) recurrent pulmonary venous stenosis 10 weeks after repair of the mixed form of total anomalous pulmonary venous connection. One (25.0%) of the 4 group 1 survivors experienced two clinically significant complications: a left-hemispheric stroke and gram-positive sepsis (Table 3). This patient survived and recovered full neurologic function.
|
2 test]) than that in subgroup 2A (n = 4, 23.5%).
Complications in Group 2 Survivors
Sixteen (42.1%) of the 38 group 2 survivors experienced one or more clinically significant complications, with 7 (18.4%) having multiple complications (see Table 3). Neurologic complications (n = 9, 23.7%) were the most common, with 2 survivors (5.3%) experiencing multiple neurologic events. (Nine (32%) of the 28 group 2 nonsurvivors also sustained neurologic complications, which were the probable cause of death in 4; the remaining 5 patients had seizures.) Intracranial hemorrhage (n = 4, 10.5%), seizures (n = 3, 7.9%), and stroke (n = 2, 5.3%) were the most common neurologic events. Five (13.2%) of the group 2 survivors had to undergo mediastinal reexploration one or more times for bleeding while on ECMO support. (Four (14.3%) of the 28 group 2 nonsurvivors underwent mediastinal exploration for bleeding, and in 1 patient, the mediastinal hemorrhage was probably the cause of death.) Infectious complications (n = 4, 10.5%) included 2 patients with gram-negative sepsis, 1 with gram-positive sepsis, and 1 with a superficial infection at the neck cannulation site. In none of the patients in the total study population did mediastinitis develop. Although ultrafiltration was used frequently (n = 50, 75.8%) along with diuretics to treat anasarca, critical renal failure requiring hemodialysis occurred in only 3 (7.9%) of the group 2 survivors. Hepatic failure and upper gastrointestinal bleeding requiring transfusion each occurred in 2 group 2 survivors (5.3%) (see Table 3).
Univariate Analysis for Group 2 Survivors Versus Nonsurvivors
PRE–ECMO VARIABLES.
The basic demographic variables (age at operation, weight, and body surface area) of group 2 survivors versus nonsurvivors were similar, although the difference in the male to female ratio between the two groups did approach significance (P = 0.053). Preoperative indices of cardiac function, including the echocardiographically determined shortening fraction and the systolic and diastolic left ventricular dimensions, were also similar when group 2 survivors were compared with nonsurvivors. The duration of CPB was the only intraoperative variable that was significantly different (P = 0.045) between survivors (120.7 ± 8.4 minutes) and nonsurvivors (147.0 ± 9.7 minutes) (Table 4). Although the aortic cross-clamp and total circulatory arrest times were longer in the nonsurvivors, these differences were not significant. A higher proportion of nonsurvivors (46.4%) than survivors (10.5%) were cannulated in the operating room (subgroup 2A membership) (P = 0.001). The proportion of group 2 survivors (89.5%) versus nonsurvivors (96.4%) whose indications for ECMO included LCO was not significantly different (P = 0.385), but the difference in the proportion of survivors (28.9%) versus nonsurvivors (10.7%) whose indication for ECMO was PAHC did approach significance (P = 0.073) (see Table 4).
|
).
POST–ECMO VARIABLES.
The serum creatinine level (P = 0.014) determined 48 hours after ECMO decannulation and the right atrial pressure (P = 0.003) averaged over the initial 8 hours after ECMO decannulation were significantly higher in group 2 nonsurvivors (see Table 4).
Multivariate Analysis for Group 2 Outcome Event-Hospital Death
The multivariate analysis identified higher right atrial pressure (averaged over the initial 8 hours after ECMO decannulation) as an independent predictor of hospital death (p = 0.049). The only other variable included in the model by the multivariate analysis as a possible independent risk factor for hospital death was membership in subgroup 2A (p = 0.061).
|
Comment |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Frequency of ECMO Use
The literature contains little information concerning the frequency of the use of ECMO support in pediatric cardiac surgical patients. Over the last 9 years, we have used this technique in 3.0% of our patients who underwent open cardiotomy. During the last year of this series (1994), we used ECMO support in 1.4% of our patients who underwent open procedures. Dalton and colleagues [3] and Ziomek and co-workers [12] reported a frequency of usage in open procedures of 1.5% and 6.5%, respectively.
Many confounding factors can influence the pediatric postcardiotomy ECMO usage rate from one institution to another. Its use in institutions with limited access to this technique is, of necessity, very low. In institutions with efficient and well-established ECMO services, the ready availability of this valuable tool may be reflected in higher rates of use. The surgeon's bias and case selection also undoubtedly affect the statistics substantially. Surgeons who believe that ECMO has little to offer pediatric cardiac surgical patients rarely use it. On the other hand, surgeons who are very familiar and comfortable with the technique may tend to overuse it. For example, some surgeons may place on ECMO neonates who have refractory cardiopulmonary failure after a first-stage Norwood procedure, whereas others consider the presence of a systemic–pulmonary artery shunt a relative contraindication to the use of this technique. These and other confounding factors make meaningful comparisons of the frequency of usage of pediatric postcardiotomy ECMO between institutions difficult to interpret.
Group 1 (Cannulated Preoperatively)
Pediatric cardiac surgical ECMO is most commonly used in patients at some interval after cardiac repair, but there are valid indications for its use in patients prior to the surgical procedure (group 1) [3, 21, 22]. We elected to analyze our experience with preoperative ECMO (group 1) separately from postoperative use (group 2) for two reasons. First, group 1, as opposed to group 2, was generally a neonatal population. Second, the poor clinical condition that prompted the institution of ECMO in group 1 patients was related solely to the underlying congenital cardiac disorder. In group 2 patients, on the other hand, the detrimental effects of the operation itself (CPB, global myocardial ischemia, and total circulatory arrest) contributed to the patients' poor condition prior to ECMO. Our group 1 hospital survival of 57% compares favorably with the 62% overall survival in the report by Hunkeler and associates [21]. If the patient is thought to have a diagnosis of persistent pulmonary hypertension of the newborn before ECMO and the response to ECMO does not appear to be consistent with that diagnosis, it is important to reconsider the possibility of an undiagnosed congenital cardiac disorder and to perform a repeat echocardiogram.
Group 2 (Cannulated at Any Interval After Cardiac Repair)
INDICATIONS.
The decision to proceed with postcardiotomy ECMO is purely a clinical judgment; no reliable predictors of imminent postoperative death have been rigorously established. The decision to cannulate must be made in time for the ECMO team to mobilize and before irreversible end-organ damage occurs. Undoubtedly some patients are placed on ECMO prematurely, when perhaps medical therapy alone would suffice. This subjective aspect of postcardiotomy ECMO therapy hinders retrospective attempts to determine the true degree of efficacy of this technique in salvaging medically ``unsalvageable'' patients. Finally, some pediatric patients with medically refractory postcardiotomy LCO respond dramatically to delayed sternal reopening in the intensive care unit. We recommend this maneuver, especially in the neonatal population where the response rate is high and the morbidity is low [19]. Delayed sternal reopening does not preclude later cervical or transthoracic ECMO cannulation.
CANNULATION SITE.
The majority of patients cannulated in the intensive care unit (subgroup 2B) underwent neck cannulation. This often obviated the need to violate the mediastinum. In our opinion, this accounted, in large part, for the absence of mediastinitis in this series and probably reduced the incidence of major hemorrhage requiring mediastinal reexploration. In subgroup 2B patients, the adequacy of decompression by the neck cannulas was documented by a significant drop in the mean left and right atrial pressures (p < 0.001 for both) after the start of ECMO.
SURVIVAL.
Our group 2 hospital survival of 57.6% has been constant since our previous study (1990) [6] involving the first 36 patients in this series. At that time, we reported a hospital survival of 58.0%. The pediatric ECMO survival rate for cardiac support in the 1994 report of the Extracorporeal Life Support Organization [23] was 44%.
UNIVARIATE ANALYSIS (PRE–ECMO VARIABLES).
By univariate analysis, three pre–ECMO variables potentially related to hospital death were identified: (1) longer CPB time, (2) membership in subgroup 2A, and (3) possibly, absence of PAHC as an indication for instituting ECMO (see Table 4). A longer CPB time is associated with increased activation of complement with its deleterious consequences, increased trauma to the formed elements of the blood, and increased capillary permeability. Membership in subgroup 2A is undoubtedly associated with more profound postoperative ventricular dysfunction, as the patient is not able to be weaned from CPB at the end of the repair. Finally, because PAHC is a reversible process, its presence as an indication for ECMO should be positively associated with survival.
UNIVARIATE ANALYSIS (ON–ECMO VARIABLES).
Elevated blood urea nitrogen and serum creatinine levels 48 hours after ECMO cannulation were associated with the nonsurvivors in the univariate analysis. This could represent progression of end-organ damage that began prior to ECMO cannulation or a lack of adequate tissue perfusion during the ECMO run itself. The red blood cell and fresh frozen plasma transfusions were indexed to the patient's body surface area and also to the duration of the ECMO run. The increased need of red blood cell and fresh frozen plasma transfusions probably correlates with increased blood loss on ECMO. These transfusion variables, therefore, may be related to hospital death because of the detrimental effects of persistent hemorrhage and the adverse consequences of large transfusion volumes.
UNIVARIATE ANALYSIS (POST–ECMO VARIABLES).
The association between hospital death and an elevated serum creatinine level determined 48 hours after ECMO decannulation probably signifies persistent end-organ damage precipitated by LCO prior to ECMO, poor tissue perfusion during ECMO, or other potentially nephrotoxic events. The underlying adverse factor, such as LCO, that produces the renal insufficiency is probably the variable that, of the two, is most closely related to hospital death. An elevated right atrial pressure averaged over the first 8 hours after ECMO decannulation is probably related to poor cardiac function caused by incomplete myocardial recovery during the ECMO run. This integrates well with renal insufficiency as a possible risk factor for hospital death in the univariate analysis.
MULTIVARIATE ANALYSIS.
It is not surprising that an elevated right atrial pressure after ECMO decannulation was identified in the multivariate analysis as an independent predictor of hospital death. Compared with all of the variables related to hospital death that were isolated by the univariate analysis, an elevated right atrial pressure after ECMO reflects a failure of myocardial recovery during the ECMO run with the most certainty. Similarly, a failure to wean from CPB after cardiac repair, thereby necessitating ECMO cannulation in the operating room (group 2A membership), probably denotes profound myocardial dysfunction from which recovery is unlikely.
INFERENCES.
(1) Extracorporeal membrane oxygenation is most effective in salvaging pediatric cardiac surgical patients who have been weaned from CPB and have experienced an interval of relative clinical stability before requiring ECMO support. (2) Although survival varies widely depending on the timing of ECMO cannulation relative to the time of the cardiac repair, the results in all of the groups defined in this report are encouraging enough to justify the continued use of ECMO in appropriate pediatric cardiac surgical candidates who have no contraindications to this therapy. (3) Higher right atrial pressure averaged over the initial 8 hours after weaning from pediatric postcardiotomy ECMO is an independent predictor of hospital death. (4) Cannulation for ECMO in the operating room (subgroup 2A membership) after unsuccessful attempts to wean from CPB may possibly be an independent predictor of hospital death. A multivariate analysis with larger numbers of patients is required to clarify this assertion and to identify other possible independent predictors of hospital death. (5) Finally, neurologic complications remain the greatest source of morbidity in pediatric postcardiotomy ECMO survivors.
|
Acknowledgments |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
Footnotes |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Address reprint requests to Dr Walters, Department of Cardiovascular Surgery, Children's Hospital of Michigan, 3901 Beaubien Blvd, Detroit, MI 48201.
|
References |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
|
|
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]
|
|
|
|
 |
|
 J. T. Baldwin, H. S. Borovetz, B. W. Duncan, M. J. Gartner, R. K. Jarvik, W. J. Weiss, and T. R. Hoke The National Heart, Lung, and Blood Institute Pediatric Circulatory Support Program Circulation, January 3, 2006; 113(1): 147 - 155. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Alsoufi, I. Shen, T. Karamlou, C. Giacomuzzi, G. Burch, M. Silberbach, and R. Ungerleider Extracorporeal Life Support in Neonates, Infants, and Children After Repair of Congenital Heart Disease: Modern Era Results in a Single Institution Ann. Thorac. Surg., July 1, 2005; 80(1): 15 - 21. [Abstract] [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]
|
|
|
|
 |
|
 R R Chaturvedi, D Macrae, K L Brown, M Schindler, E C Smith, K B Davis, G Cohen, V Tsang, M Elliott, M de Leval, et al. Cardiac ECMO for biventricular hearts after paediatric open heart surgery Heart, May 1, 2004; 90(5): 545 - 551. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. L. Booth, S. J. Roth, R. R. Thiagarajan, M. C. Almodovar, P. J. del Nido, and P. C. Laussen Extracorporeal membrane oxygenation support of the Fontan and bidirectional Glenn circulations Ann. Thorac. Surg., April 1, 2004; 77(4): 1341 - 1348. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. M. Ungerleider, I. Shen, T. Yeh Jr, J. Schultz, R. Butler, M. Silberbach, C. Giacomuzzi, E. Heller, L. Studenberg, B. Mejak, et al. Routine mechanical ventricular assist following the Norwood procedure--improved neurologic outcome and excellent hospital survival Ann. Thorac. Surg., January 1, 2004; 77(1): 18 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. R. Johnson, M. S. Schamberger, J. C. Hart, M. W. Turrentine, and J. W. Brown After repair, atrioventricular valve regurgitation during cardiac extracorporeal membrane oxygenation predicts survival Ann. Thorac. Surg., September 1, 2003; 76(3): 848 - 852. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. G. Hamrick, D. B. Gremmels, C. A. Keet, C. H. Leonard, J. K. Connell, S. Hawgood, and R. E. Piecuch Neurodevelopmental Outcome of Infants Supported With Extracorporeal Membrane Oxygenation After Cardiac Surgery Pediatrics, June 1, 2003; 111(6): e671 - 675. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-S. Chen, A. Chao, H.-Y. Yu, W.-J. Ko, I.-H. Wu, R. J.-C. Chen, S.-C. Huang, F.-Y. Lin, and S.-S. Wang Analysis and results of prolonged resuscitation in cardiac arrest patients rescued by extracorporeal membrane oxygenation J. Am. Coll. Cardiol., January 15, 2003; 41(2): 197 - 203. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. W. Duncan Mechanical circulatory support for infants and children with cardiac disease Ann. Thorac. Surg., May 1, 2002; 73(5): 1670 - 1677. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. Drinkwater Jr, A. S. Aharon, S. V. Quisling, D. Dodd, V. S. Reddy, A. Kavanaugh-McHugh, T. Doyle, N. R. Patel, F. E. Barr, J. K. Kambam, et al. Modified Norwood operation for hypoplastic left heart syndrome Ann. Thorac. Surg., December 1, 2001; 72(6): 2081 - 2087. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. C. Patel, M. Jothi, D. B. Trivedi, G. Sabino, P. Daly, P. D. Booker, and M. Pozzi Post-cardiotomy mechanical circulatory support using a conventional bypass circuit in children Eur. J. Cardiothorac. Surg., October 1, 2001; 20(4): 811 - 815. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Y. Schure, P. C. Laussen, and F. X. McGowan JR Mechanical Cardiopulmonary Support of Infants and Children With Congenital Heart Disease Seminars in Cardiothoracic and Vascular Anesthesia, March 1, 2001; 5(1): 46 - 54. [Abstract] [PDF]
|
|
|
|
|
|
C. Pizarro, D.A. Davis, R.M. Healy, P.J. Kerins, and W.I. Norwood Is there a role for extracorporeal life support after stage I Norwood? Eur. J. Cardiothorac. Surg., March 1, 2001; 19(3): 294 - 301. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Jacobs, J. W. Ojito, T. W. McConaghey, B. D. Boden, A. C. Chang, A. Aldousany, E. M. Zahn, and R. P. Burke Rapid cardiopulmonary support for children with complex congenital heart disease Ann. Thorac. Surg., September 1, 2000; 70(3): 742 - 750. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. Jaggers, J. M. Forbess, A. S. Shah, J. N. Meliones, P. M. Kirshbom, C. E. Miller, and R. M. Ungerleider Extracorporeal membrane oxygenation for infant postcardiotomy support: significance of shunt management Ann. Thorac. Surg., May 1, 2000; 69(5): 1476 - 1483. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. E. Ibrahim, B. W. Duncan, E. D. Blume, and R. A. Jonas Long-term follow-up of pediatric cardiac patients requiring mechanical circulatory support Ann. Thorac. Surg., January 1, 2000; 69(1): 186 - 192. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y.-S. Chen, M.-J. Wang, N.-K. Chou, Y.-Y. Han, I.-S. Chiu, F.-Y. Lin, S.-H. Chu, and W.-J. Ko Rescue for acute myocarditis with shock by extracorporeal membrane oxygenation Ann. Thorac. Surg., December 1, 1999; 68(6): 2220 - 2224. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. I. Cohen, J. W. Gaynor, V. Ramesh, T. R. Karl, J. M. Steven, J. Posner, B. J. Clark III, and L. A. Rhodes EXTRACORPOREAL MEMBRANE OXYGENATION FOR PATIENTS WITH REFRACTORY VENTRICULAR ARRHYTHMIAS J. Thorac. Cardiovasc. Surg., November 1, 1999; 118(5): 961 - 963. [Full Text] [PDF]
|
|
| < |
