The evolution of patient selection criteria and indications for extracorporeal life support in pediatric cardiopulmonary failure

Abstract

Bill James, baseball statistician and author, tells the story of hungry cavemen sitting about a campfire, waiting for tomatoes to ripen. One has the inspiration to throw an ox on the fire, and the first barbecue ensued and was endured. After eating, the conversation goes something like this. “There were some good parts.” “Yeah, but there were some bad parts." And the smart one says, “This time, let's not eat the bones.” The evolution of patient selection criteria for the use of extracorporeal support (ECLS) is a bit like those cavemen and their first barbecued ox. Extracorporeal life support technology and application to patient care is the unique result of a long standing history of ambitious attempt, evaluation, debate, collaboration and extension.

Bill James, baseball statistician and author, tells the story of hungry cavemen sitting about a campfire, waiting for tomatoes to ripen. One has the inspiration to throw an ox on the fire, and the first barbecue ensued and was endured. After eating, the conversation goes something like this: “There were some good parts.” “Yeah, but there were some bad parts.” And the smart one says, “This time, let's not eat the bones.”1

The evolution of patient selection criteria for the use of extracorporeal support (ECLS) is a bit like those cavemen and their first barbecued ox. Extracorporeal life support technology and application to patient care is the unique result of a long standing history of ambitious attempt, evaluation, debate, collaboration and extension.

Rooted in cardiovascular surgery's holy grail of learning to repair hearts, ECLS today is the result of well over half a century of passionate commitment. The early feast of investigative opportunity, in a precedent that continues to foster innovation, was broadly shared. The disciplines of surgery, pediatrics and neonatology have collaborated with venture capitalists, basic scientists and industry to continuously improve the technology and pioneer clinical application.

These disciplines collaboratively shared experience, compared results, developed clinical feasibility trials, analytic methodology and applied this experience to broader applications. At this writing, over 42,000 infants, children and adults have been treated with this technology.2

Much of the technology and methodology in use today emerged in the 1950s. Gibbon reported the repair of a young woman with an ASD in Philadelphia in 1953. The roller pump used was invented by Gibbon and his wife. With very few modifications, the same pump technology remains prevalent 60 years later.3

C. Walton Lillehie and Clarence Dennis, among others, pioneered techniques and devices that would allow cardio pulmonary bypass in the operating theater. Each has written histories of the era.4,5 Goor's recent biography of Lillehie adds interesting details to the saga.6 Bartlett, in 1990, summarized the evolution of extracorporeal technology from the cardiac surgical patient in the operating theater to the bedside of the critically ill adult and neonate with cardiopulmonary failure in the intensive care unit.7

Two major contributions were fundamental to the evolution of long-term use of extracorporeal support for pulmonary and cardiac failure. The first is Bartlett and Drinker's recognition that heparin dosing could be titrated and its effect monitored in a timely manner. Early investigators were faced with the challenge of avoiding the Scylla of hemorrhage and the Charibdis of thrombus formation. Heparin was described early in the 20^th century, and remains in use today.8 The observation that the closed circuits developed for long-term support in the intensive care unit did not require infinite heparinization was fundamentally crucial in allowing long-term support. This was a somewhat serendipitous and subsequently overlooked discovery. Bartlett and Drinker, while performing experiments with various circuit and membrane designs (e.g., toroidal flow), recognized that the cardiac patient in the operating room required infinite heparinization due to the stagnant flow in lungs and field necessary during open surgical repair. However, long-term bypass support produced constant flow in both circuit and circulation. Thus significantly less heparin could be used, and early experiments demonstrated that circuits could be used for days without clot formation, and in animals, hemorrhage could be prevented.9–11

A remarkable element of these early studies was the development of a methodology and device to accurately (if not precisely) titrate near continuous adjustment in heparin requirement in experimental circuits and animal studies. Bartlett and Drinker described the bedside activated clotting time and a device for measurement based on what was known as the re-calcification time. This allowed the titration of small, controlled infusions of heparin. Fifty years later, the activated clotting time remains a cornerstone of anticoagulation management in extracorporeal support.12

The second cornerstone was the discovery that silicone rubber is a semipermeable and selective membrane.13 Kolobow, Kolf, Rashkind and others noted that the transfer of oxygen through silicone was an order of magnitude faster than in other materials, and developed the spiral coil membrane lung.14–16

The extracorporeal circuit continues to be improved. Refinements have included improvement in tubing flexibility in order to tolerate long use in a roller pump, the development of the centrifugal pump, the continued evolution of the hollow fiber oxygenator and efforts to coat circuits to avoid inflammatory change induced by artificial surfaces. Catheter design now includes the single dual lumen catheter. Catheters can be placed using traditional percutaneous techniques. The ability to employ veno-venous extracorporeal support, with percutaneous access and placement of a single, dual lumen catheter, avoids the problem of carotid arterial cannulation.17

The prevention of bleeding in the patient and prevention of thrombus formation in patient and circuit remains a major challenge in patient management. This subject continues to be actively investigated.18

Heparin remains the mainstay of anticoagulation management, and bleeding on ECLS is a continued cause of morbidity. Novel approaches to the control of bleeding and alternatives to heparinization have been described. Recent studies have explored the use of a direct thrombin inhibitor, Argatroban, and these suggest its use as an alternative to heparin in the presence of heparin- induced thrombocytopenia.19 Control of active hemorrhage in the anticoagulated patient requires the input of a skilled surgical team to assess and control any surgical bleeding.20 Medical management of the patient with extraordinary blood loss remains a challenge. The most experience in the medical control of bleeding beyond minimal heparin, surgical control and platelet supplementation is with 6-aminocaproic acid.21,22

Continued improvement in the understanding of the nature of the inflammatory response in both patient and circuit has led to novel attempts to “calm” the endothelium and decrease the effect of artificial membranes on blood components. Explorations into various surface coating methodologies to improve hemocompatibility may ultimately lead to a clinically useful artificial endothelium.23,24 Polymers that release nitric oxide are an extremely promising solution to the problem.25,26

Although Gibbons's pump and Kolobow's spiral coil silicone membrane oxygenator remain in use today, the centrifugal pump has emerged as an alternative. The centrifugal pump can be miniaturized and is the pump of choice for emergency deployment in resuscitations and for patient transport.27 The hollow fiber oxygenator is portable, lending an application in emergency and transport use. Plasma leakage with long-term use has diminished with refinement. The hollow fiber membrane can be coated to decrease bio-reactivity.28

Mechanical failure in circuit components is increasingly rare. Fleming has thoroughly reviewed complications due to circuit components in ECLS deployment in 28,000 courses of ECLS in neonates and children. Case exclusions were circuit runs of less than 25 hours, greater than 450 hours and cardiopulmonary resuscitation. Predictably, the authors noted an increase in component failure associated with the duration of use. The majority of failures occurred in neonatal use. Mechanical failure of a component was noted in nearly 15% of all cases. The frequency of failure was observed to decline in the past ten years.29

To the Bedside: Development of Clinical Applications

Several authors have reviewed the progression of ECLS technology from bench to clinical trial.30,31 Thorough coverage of the topic can be found in the excellent text published by the Extracorporeal Life Support Organization, now in its third edition.32 This review will primarily consider the problem posed by the pediatric patient with pulmonary failure. The unique challenges in this age group included the lack of description of the disease, intuitive but misguided concepts of disease severity and the rarity of respiratory failure in the child.

In the early 1970s, several case reports appeared as investigators applied ECLS technology in bedside care. It is notable that the humble case report was (and remains) an important contribution to the evolution of this technology. Dorson and White published their respective experience with infants supported with extracorporeal circuits.33,34 Baffes and colleagues reported the first infant supported with extracorporeal support post-palliative cardiac surgery in 1970.35

A historic symposium on extracorporeal technology was held in Copenhagen in 1975 and included a report of the first infant supported with ECLS technology. This precedent-setting meeting of experts presaged the evolution of interdisciplinary and interinstitutional cooperation that has characterized this endeavor.36

The publication of two early cases demonstrated the willingness of clinical investigators to come to the bedside of critically ill patients with new (some might have said controversial) technology. In 1976, Bartlett and colleagues, working at the University of California, Irvine, were confronted with the challenge of an abandoned infant in significant hypoxemic failure. It was recognized that the conditions leading to hypoxemia in the neonate, including high pulmonary vascular resistance were temporary and reversible. Bartlett's team placed the infant on an extracorporeal perfusion circuit. As fate would have it, the neonatal nurses named the child Esperanza (Hope). The infant survived, and this well chronicled episode was seminal in the spread of the technology to sick newborn infants in the then young specialty of neonatology.37

In 1972, Hill and a team of colleagues reported the case of a young trauma victim placed on long-term extracorporeal support. Hill and his team traveled with a circuit to Santa Barbara from San Francisco to attend to a young motorcycle accident victim with multiple injuries, including a ruptured aorta. The patient was supported for three days. This anticipated the still topical dilemma of transport of patient, device or patient on device, indication and eligibility, all of which remain an issue today.38 The willingness to offer this technology (and of families to trust and permit) when conventional wisdom predicted death characterized these reports.

Between 1972 and 1977, several single institution case series of experience with neonates and adult patients were published in reference 7. A notable feature of these case reports, interestingly enough, is the application of support in what would later often (for decades, in some cases) be considered by subsequent author-investigators as “controversial indications.” The continued publication of observational reports was and is fundamental in the evolution of rational, data-based indications and contraindications of this type of therapeutic intervention. Some of these conditions include the use of extracorporeal circulation in patients with opportunistic infections.39 Viral pneumonia was later presumed a contraindication, but was the subject of an early report.40 Hanson reported successful respiratory support in a woman following pulmonary hemorrhage; nearly two decades would elapse before it was demonstrated that this was application is not a contraindication.41 An adult with pulmonary emboli was successfully supported with a membrane oxygenator, and an embolectomy was performed while on circuit by Cooper and colleagues.42

In 1982, Bartlett and colleagues, having moved to the University of Michigan in Ann Arbor, had accumulated enough experience to publish a study of forty-five infants with meconium aspiration, diaphragmatic hernia and pulmonary hypertension. At the time of publication, these were uniformly morbid conditions. Each of these conditions were reversible, thus extracorporeal support was not a rescue modality but a temporary support of organ failure.43

Although critics and supporters would plea for randomized trials, the publication of case experience proved invaluable to the dissemination of the technology. History would show that randomized clinical trials, with consensus agreement on definition of outcome and eligibility, would be difficult to execute and controversial in analysis. The path to legitimization and standard of care is a tribute to persistent investigators confronting problems in defining efficacy, trial design and the willingness of patients and families to take risks in desperate situations. The publication of the proceedings of the Copenhagen meeting stimulated the development of a clinical trial in the United States.44 In 1974, the NIH proposed a clinical trial in adults. This NIH collaborative study was designed to evaluate the use of extracorporeal membrane oxygenation (ECMO) in adults with hypoxemic respiratory failure.45 The results of this trial were published in the New England Journal of Medicine in 1979.46 The results were interpreted as negative, with no benefit to patients treated with extracorporeal support. In depth analysis revealed ignorance of the natural history and epidemiology of disease, imprecise definitions of organ failure, disease severity and, pointedly, a lack of consensus of definitions of conventional treatment. The study was terminated at 92 patients (anticipated enrollment was 300) as death rates were equally high (90%) in each arm. A nationwide epidemic of influenza pneumonia coincided with the trial and skewed results. Although investigators intended to provide lung rest while on extracorporeal support, many patients remained on what would be considered today as remarkably high inflation and distending pressure. A notable observation was that the presence of multiple organ failure increased the risk of death in a patient with lung injury. A subsequent report in 1986 suggested precise definition, characterization of mortality rates, estimation of disease severity and scoring systems for comparison and inclusion.47,48 A benefit of the NIH adult trial was insight into the pathology of respiratory failure in the adult.49

The enthusiasm for the investigation of extracorporeal support for respiratory failure in adults was dampened in the United States following the NIH trial. However, groups in Stockholm, Japan, Germany and, notably, Luciano Gattinoni continued to investigate the possibility of improving outcome in adults with respiratory failure. Gattinoni and colleagues, recognizing that mechanical ventilation was destructive to the injured lung, developed an animal model for extracorporeal carbon dioxide removal, using the native lungs for oxygenation, with minimal inflation pressures.50 A landmark paper published by the Milan group demonstrated a 49% survival rate in adult patients with selection criteria similar to those used in the NIH trial (in which the survival rate was 10%).51 Similar successes were published by groups in Marburg, Dusseldorf, Toronto and elsewhere. Much of this work was presented at a conference in Marburg, Germany in 1988.7,52

The adult NIH trial of 1979 dismayed investigators, particularly in the United States. In order to respond to critics of the technology, proponents were faced with the dilemma of randomizing morbidly ill patients to a novel, invasive therapy, with death as an outcome. An innovative approach to study design led to publication of a successful, statistically valid trial in critically ill newborn infants. The methodology stimulated an important discussion and eventual reconsideration of trial design.53–56 The difficulty of achieving equipoise clinical trials is illustrated by the experience of the Boston Group. A prospective attempt was stopped early for ethical reasons, with similar, favorable results to those published by the Ann Arbor group in 1985.57 [A decade later, using a model possible in the United Kingdom's Health Care Delivery System, a prospective randomized trial of the use of ECMO in neonates was published noting a substantial better survival with ECMO (32% vs. 59%) when compared to conventional management].58

It can fairly be said that extracorporeal membrane oxygenation became a standard of care in the neonatal patient in the early 1990s. The NIH convened a consensus conference in May of 1990 and published the recommendation that diffusion of ECMO technology should continue.59 In response to a question at a symposium regarding the question of whether or not ECMO was a standard of care, a proponent posited the axiomatic response: “You can bill for it, you can be sued for doing it, sued for not doing it and sued for not doing it well.”60

The Extracorporeal Life Support Organization (ELSO) was established in 1989. This volunteer consortium sponsors a central database, known as the ELSO Registry, documenting experience, complications and survival in more than 42,000 cases world wide (as of January 2009) for analysis and evaluation of efficacy and categorization of complication and application. The registry is an international group with the participation of 100 centers. ELSO promotes coordination of prospective studies, develops practice guidelines and standards and, with annual symposia, provides a forum for the dissemination of technology and serves as a unique interface for dialogue between entrepreneurs, bench research, clinicians, technician and survivors.2 ELSO has published three editions of a core textbook of extracorporeal life support.

More than 42,000 total cases have been reported, including International, European and American centers. Approximate annual numbers of patients by age groups in US centers as of 2009 can be seen in Table 1.

Survival rates are reported in the ELSO registry by age and organ system, cardiac or respiratory failure and the use in cardiopulmonary resuscitation (emergency deployment during resuscitation).2

Selection Criteria and Clinical Indications for the Use of Extracorporeal Support: How Did We Get Here?

A legacy of the introduction of ECLS technology is a template for the analysis of transfer of technology from bench to bedside. A major clinical problem confronting clinicians and families was and remains the definition of the timing of intervention. Investigators studying respiratory failure in the non-neonatal, pediatric and the adult patient began a rigorous analysis of incidence, prevalence, definition, severity and predictors of survival. These collective studies, an exercise in cumulative thought, provided the foundation for the development of indication and comparison of results between management strategies. It became apparent that data to support a prediction of death or morbidity with conventional medical management of respiratory failure was inadequate. (This begs the question of what constitutes “conventional,” which remains an elusive definition.) Many investigators were ethically concerned with the use of an invasive technology without a randomized clinical trial to prove efficacy. An obstacle in the attempt to demonstrate efficacy was the lack of data characterizing the natural history of cardio pulmonary failure in various age groups. A transparent contribution of the emergence of ECLS, and challenges to its efficacy, may have been that it posed as a “straw target” for epidemiologists interested in characterizing disease severity and predictors of mortality in respiratory failure.

Northway and Rosan's seminal 1967 description of broncho-pulmonary injury sustained by infants surviving mechanical ventilation suggested the concept of ventilator toxicity, evident in survivors of neonatal mechanical ventilation.61 Phillips further characterized this phenomenon as attributable to a combination of oxygen, pressure and time on mechanical ventilation.62 Ashbaugh and Petty described hypoxemic respiratory failure in adults in a series of observations beginning in 1967. Noting a similarity to infant respiratory distress, the concept of an adult, acute respiratory distress syndrome (ARDS) was introduced into clinical practice.63,64 Murray and colleagues refined definitions and further characterized the syndrome.65

A consensus conference was held in 1994 that focused on the adult patient and produced guidelines for future investigation, definitions of outcome and disease mechanism.66 The understanding of the pathophysiology of the non-neonatal pediatric and adult patient with respiratory failure needed to be described. The challenge of promoting a meaningful dialogue between proponents and critics of extracorporeal support was supported by the investigative efforts of several groups, which now give us a (pre-contemporary) database of the natural history of the syndrome in children. Disease severity, prevalence and outcome required definition and description in order to provide a basis for meaningful comparison of treatment modalities. Bohn has summarized many of these and later studies in an exhaustive review.67

Several pediatric investigators made important observations linking disease severity, time of intervention with outcome. It is historically interesting that all confirmed Phillips 1975 observation that a combination of hypoxemia, duration of illness and exposure to high inflation pressures early in the development of respiratory failure predicted mortality. Each study noted that a combination of widened alveolar arterial gradients for oxygen, in excess of 450 mm Hg, duration of positive pressure ventilation for 12 hours to 4 days, high inflation pressure and/or requirement for end expiratory pressure predicted mortality rates of 55 to 80 percent with reasonable sensitivity and specificity.68–70

Criticism of the use of extracorporeal technology in children continued into the 1990s. These were rooted in the lack of a rigorous, randomized, controlled trial. Attempts at controlled trials were controversial. There was a lack of equipoise amongst investigators over two main issues: (1) the safety and efficacy of ECLS therapies in children and (2) lack of robust predictive data with conventional therapy.71

Clinicians at the University of Michigan utilized extracorporeal support for non-neonatal respiratory failure in children beginning in the early 1980s. ECLS was instituted when clinicians determined that death was likely if the patient remained on conventional mechanical ventilation. A series of publications between 1992 and 1994 documented survival rates of 60 to 80 percent. The data was analyzed for elements that might predict survival, with the intention of providing a foundation for more robust eligibility criteria for the use of ECLS in respiratory failure. Reported survival rates were 60–80%, while estimates of mortality with (albeit undefined and non-standardized) conventional management of the period were over 60%, with a prevalence of roughly six patients per one thousand PICU admissions.72–74

The value of the ELSO registry database became apparent with the publication of a review of 220 children with morbid respiratory failure treated with ECLS prior to 1993. Several historical and clinical variables were statistically demonstrated to predict survival to discharge. These were age, days on mechanical ventilation, peak inflation pressures, the alveolar-arterial gradient, end expiratory and mean airway pressures.75 Green summarized a robust cohort analysis of 331 pediatric patients from 32 centers, and concluded that ECMO treatment of respiratory failure in children was associated with improved survival.76

The ability to accurately characterize historical and clinical variables predicting survival or disease severity has continued to inspire analysis and remains an active area of investigation.77,78 These and other investigations have supported relatively standardized indications for the use of extracorporeal support in childhood respiratory failure. Of particular clinical value is the oxygenation index as a useful tool in following patients and in the prediction of the need for ECMO.30,79,80

The history of the evolution of ECLS from bench to bedside is an interesting study in innovation in patient care. The principal challenge to analysis of new technologies, and confirmation of efficacy, is the necessary comparison to an ever shifting definition of conventional management. The care of the patient with severe respiratory failure continues to evolve and perhaps, prayerfully, improve. The sentinel revelation of the past two decades of respiratory intensive care is confirmation that stretch injury and biotrauma injure the lung. However, even this near axiom of modern conventional care has not generated the dramatic improvement in survival many would intuitively expect and the casual observer might assume. A recent observation casts doubt on a genuine reduction in ARDS mortality in the adult patient. A decrease in ARDS mortality was only seen in observational studies prior to 1993. Mortality has not decreased between 1994 and 2006 and is lower in randomized controlled trials than in observational studies.81

ECLS continues to be instituted with a combination of predictive data and clinical judgment. Dalton, noting that data suggests that the characterization “failure to respond to conventional management” is the most common indication for support reported to the ELSO registry, has suggested patient selection criteria for the pediatric patient.30

1. Oxygen requirement greater than FiO₂ > 0.6.

2. Mean airway pressure more than 20 cm H₂O on conventional ventilation or 25–35 cm H₂O with high frequency ventilation.

3. Presence of airleak or evidence of barotraumas

4. Oxygenation Index (= 100 × mean airway pressure × FiO₂/PaO₂)

5. Duration of mechanical ventilation of 7–14 days.

6. Evidence for reversibility of disease

7. Reasonable medical certainty of a reasonable quality of life

8. No condition associated with bleeding diathesis

9. Inability to remove carbon dioxide.

Defining a contraindication is difficult. The major contraindication is the medical certainty of a poor quality of life. An important confounder in patient selection has been the somewhat intuitive exclusion of patients deemed to be of high risk. This is often based on the art of judgment and is continually revised with experience. The latter point emphasizes the value of publication and analysis of anecdotal and single institution experience.82

Conditions historically perceived as being contraindications to extracorporeal therapy include immune deficiency, trauma, burn patients, cancer patients, pulmonary hemorrhage and bone marrow transplant recipients. Experience teaches that many groups of high risk patients can be successfully supported with extracorporeal life support by skilled teams of intensivists and surgeons. Trauma patients with respiratory failure have been successfully supported.83 Patients with immune suppression and post transplantation have been reported with acceptable results.84–86 Burn patients can be successfully treated while on extracorporeal support.87 ECMO has successfully been utilized in patients with acute pulmonary hemorrhage secondary to Goodpasture's syndrome or Wegener's Granulomatosis. Early presumptions that heparinization would worsen the bleeding in these patients have proven incorrect.88,89

The patients with immune compromise were perceived to be poor candidates for support due to the suspicion that super infection would be fatal. Recent reviews have documented successful use of ECLS in significantly immuno-suppressed patients. A 2008 review of the ELSO registry experience in children with ICD-9 codes or Current Procedural Terminology coding associated with immune compromise reports a 31% survival rate.90 Patients with malignancy and cardio pulmonary failure were originally not placed on extracorpeal support because of uncertainties of ultimate prognosis for quality of life. Gow and colleagues reviewed the ELSO registry experience of 107 children less than 21 years of age with both hematogenous (73 patients) and solid tumor (34 patients). Forty-five of 107 (42%) patients survived ECLS. Eight patients died in hospital after ECLS support for an overall survival to discharge of 35%.91

Sepsis as an indication has been controversial. Skeptics reasoned that coagulopathy would be difficult to control during ECMO and made the unfounded presumption that the circuit would be colonized by the infecting organism.92,93 ECMO is now listed in the American College of Critical Care Medicine Guidelines as a therapy in refractory shock in infants and children.94

The Melbourne group has published the largest single institution experience with 45 children and observed a 47% survival to discharge. (This series includes 12 patients with meningococcemia).95

A recent development in the use of ECLS as part of septic shock management is the publication of experience with rarely encountered infections. Meningococcemia remains a difficult problem, and results with extracorpeal support are mixed.96 Community-associated, methicillin-resistant Staphylococcus aureus sepsis is increasing in both prevalence and case severity in children.97 A retrospective study examined treatment of pediatric patients with ECLS at a single institution and reviewed the ELSO Registry data base. Survival to discharge rates varied with age. Children less than 4 years old had a survival rate of 65%, with no survivors in children aged 5–9 years and a 31% survival in adolescents.98

Infections in children with Bordetella pertussis are associated with poor outcome if both pneumonia and septicemia are present.99 A very high mortality rate has been noted in pediatric patients with Pertussis infections placed on extracorporeal support. Some have questioned the use of ECLS in the pertussis victim presenting with shock and pneumonia. It remains to be seen if these poor results (and those observed in Meningococcemia) are due to a lack of understanding of the mechanism of the infection, inadequate predictive data that might provide a basis for early intervention or a failure to understand the inflammatory nature of interaction between circuit component and patient.100,101

Criteria for Use of ECLS in Children with Sepsis

Clinical judgement remains the basis for the decision to utilize ECLS in the septicemic child. Clinically useful guidelines include shock refractory to goal-directed fluid resuscitation, failure to respond to two or more vaso active agents and presence of multi-organ failure. Extracorporeal membrane oxygenation is included in recent consensus guidelines for the treatment of septic shock.94

Cardiac ECLS

In the past decade this is the most rapidly growing area of ECLS usage. Diagnostic groups include post-operative congenital heart surgery, resuscitation, cardiac failure and bridge to transplant. Baffes is credited with first report of extracorporeal, post-operative support.35 Soeter and colleagues reported the 48 hour support of a four-year old following repair of Fallot's tetrology.102

Early reports of modest successes led to increased usage in myocardial failure as well as the post operative patient.103–107

The first review of the ELSO registry for pediatric cardiac patients was published in 1991. The authors reviewed the case experience in 189 patients and a 43% survival was observed. Most deaths were attributed to irreversible cardiac or central nervous system injury, and the authors suggested that earlier intervention might improve results.108

As data accumulated, a significant number of patients treated with ECLS could be rigorously analyzed for predictors of survival and establishment of indications. The Michigan group observed that non-survival was associated with the initiation of support in the operating theater or institution more than 50 hours following surgery. Patients with two ventricle repair had a 42% survival, while those requiring a cavopulmonary connection had a 83% survival rate.109

A subsequent report updated experience and published a multivariate analysis of variables associated with survival. Variables analyzed included time after surgery, ongoing CPR at institution of ECMO, need for renal replacement therapy and surgical diagnosis. Children with an adequate two ventricle repair were observed to have lower risk of death, and need for renal replacement therapy was associated with a five-fold increased risk of death. An improvement in overall survival in this series from 33% to over 50% compared to historical reports was attributed to earlier initiation of therapy and the introduction of the systemic to pulmonary shunt as opposed to a cavo-pulmonary connection. These authors noted that, in contrast to previously published data, the initiation of ECMO in the operating theater was now associated with improved survival and a reduction in need for renal replacement therapy. They further suggested that the timings of earlier intervention should be explored in future investigation.110

A case for earlier intervention was strengthened with the observation that the need for dialysis, presumably due to multi-organ failure prior to ECMO rescue, was associated with a poor prognosis.111 Several authors noted that early intervention, timed prior to the onset of multi-organ failure, is associated with improved survival.112,113 The presence of lactic acidosis and a strong ion gap as indices of inadequate perfusion, predict mortality in children following bypass and are useful variables for the clinician.114,115

Consensus guidelines for the deployment of ECLS in the post operative patient do not exist, remain institutional and require experience in the management of these patients. Clinicians must be experienced in the recognition of the low cardiac output syndrome (LCOS). Notably, hypotension is an unreliable and a very late manifestation of inadequate cardiac output.116

Suggested Guidelines for use of ECLS in the Cardiac Patient

1. Failure to come off intraoperative bypass

2. Progressive multi-organ failure

   1. Fluid overload

   2. Need for three or more vasoactive agents and failing to achieve reasonable goals for perfusion

   3. Metabolic acidosis

   4. Need for more than 60% ambient oxygen and/or oxygenation index greater than 40.

3. Low cardiac output syndrome

4. Cardiac arrest

Nearly 7,000 cardiac patients have been reported to the ELSO Registry. Survival rates by age (excluding cardiopulmonary resuscitation patient date) are seen in Table 2. ECMO has been successfully utilized to bridge cardiac patients to transplantation with very reasonable survival rates in several centers.117–119 Recently, the emergence of ventricular assist devices (VAD) sized for neonates and children has improved survival in bridging a patient to transplant. A 2006 review of outcomes in children bridged to heart transplantation reports an 86% survival rate in pediatric patients bridged to transplant with a VAD compared to a historical 47 to 57% survival noted with ECMO as a bridge to transplantation. Outcomes in these children were comparable to children medically bridged to transplantation. Description of the various devices under active investigation, with NIH support, was reviewed by Blume and co authors in 2006. Davies and colleagues have reviewed the United Network for Organ Sharing (UNOS) database. Ventricular assist device support was used in 471 of 2,532 transplantations and ECMO in 171. The investigators concluded that pediatric patients requiring a pre-transplantation ventricular assist device had similar outcomes when compared to patients with medical support. A generally worse outcome was noted in patients placed on ECMO.120,121 A large single institution long-term follow up analysis observed more complications in a group of children supported by ECMO when compared to a group supported with ventricular assist devices. Ventricular assist devices (VAD) have superceded ECMO in most clinical situations.122 There are a few indications in which ECMO remains superior to the VAD in the care of the post operative cardiac surgical patient. ECMO is superior to VAD in cases of pulmonary hypertension, in newborns with biventricular failure and when cardiac failure is complicated by respiratory failure.123

ECMO and Cardiopulmonary Resuscitation (ECPR)

A recent development is the use of ECMO in cardiac resuscitation (ECPR). Survival to discharge rates approach 40% in neonates and children but only 24% in adults. Registry data for 2009 report the use of ECPR in 386 neonates, 639 children and 150 adults. Chen et al. and Thourani and colleagues have reported higher survival rates, above 50%.124,125 Larger series, including reports from the ELSO registry, report lower survival rates of 38%. Morris has observed that the conditions existing prior to the arrest are an important variable in survival. Higher survival rates are observed in children with isolated cardiac disease. These data are not comparable to children who arrest with multiple organ failure. The latter group of patients has a poor prognosis after cardiac arrest. In this series, no patient survived if more than 30 minutes of unsuccessful CPR elapsed prior to extracorporeal support.126

The endeavor is labor- and cost-intensive. It is imperative that a complete, skilled multi-disciplinary team be close at hand in order to cannulate the patient expeditiously. Thiagarajan has pointed out that the procedure is neither universally available nor accepted by consensus in clinical practice. In the largest series reported to date, an attempt was made to identify variables associated with survival. Cardiac disease, neonatal respiratory disease and a pre-arrest pH of >7.17 were associated with decreased odds of mortality. Renal failure, metabolic acidosis, pulmonary bleeding and neurologic injury occurring after institution of ECLS were predictive of death.127 A cost analysis provides insight into the expense of maintaining a team and an inventory of equipment.128 ECPR requires a substantive institutional and personnel commitment. Data to define criteria, with the possible exception of the post-op congenital heart experience, do not yet exist. Variables such as duration of arrest prior to ECLS, predictors of survival and contraindications await more robust investigation.

Severe neurologic complications are quite frequent in this population. Barret and colleagues have reviewed 682 cases of the use of ECPR in the pediatric age group.

Acute injury was defined as brain death, infarction and/or intracranial hemorrhage. Nearly one quarter (22%) of the 682 cases experience a significant neurologic complication. Brain death occurred in 11%, infarction in 7% and hemorrhage was noted in 7%. The in-hospital mortality rate was 89%. Pre-existing conditions that were associated with decreased odds of a neurologic injury included a modest metabolic acidosis and the presence of cardiac disease. Pulmonary hemorrhage, use of renal replacement therapies and cardiac arrest occurring after institution of ECLS were associated with increased odds of a neurologic complication.129

Organ Preservation after Cardiac Death

Although several authors have used descriptors such as ECLS and ECMO, in this circumstance another term, such as “organ preservation after death” may be more appropriate and less confusing to the public. Extracorporeal technology has been used in some centers in donation after cardiac death protocols. The Michigan group reported a 10 year experience in 48 donors. Thirty-six procedures led to transplantation of 67 kidneys, 19 livers and two pancreases.130 Boucek and colleagues recently published a thought provoking experience in three infants undergoing heart transplantation with organs obtained from donors who died from cardiocirculatory causes.131 A spirited debate prompted by this publication is summarized in a series of editorials in the New England Journal of Medicine. Despite sanctions by the Institute of Medicine, the Joint Commission and a consensus conference, several areas of ethical interest and concern have been raised. These include the utility of the Dead Donor Rule in the area of organ transplantation, the beneficent intent of increasing the availability of transplantable organs, the definition of cardiac and brain death and the ethical boundaries in general of organ donation after circulatory death.132–134

Intravascular Devices

Intracorporeal CO₂ removal is an extension of extracorporeal support. Gattinnoni's original observation was that the native lung, even quite compromised, could support oxygenation if carbon dioxide could be removed with extracorporeal techniques.51,135 This concept has been extended to the development of several novel devices intended to provide oxygenation as well as carbon dioxide removal. The intravascular oxygenation or IVOX device was evaluated in a two-stage, international multicenter trial. Success was mixed, as about 30% of pulmonary function could be supported and survival was poor.136 The Hattler device, another innovation, consists of a small pulsating balloon in the midst of a hollow fiber bundle. This configuration enhances convective mixing. It has been trialed successfully in large animals.137

The next great achievement in the long saga of supporting patients with organ failure is the implantable artificial lung for the long-term support of the pulmonary failure patient. The topic and concepts and devices have been reviewed. This device will dramatically improve the quality of life for the pulmonary failure patient and be a practical bridge to transplantation.138–140

Summary

The continued evolution of patient selection and indications for extracorporeal support is a tribute to the foresight and ambition of early investigators. A notable feature of the extension of this technology has been, from its origins, a willingness to share experience, listen and respond to criticism and foster cooperative investigation. The use of extracorporeal support today is an example of collective, cumulative and collaborative thinking. This history is old, borrowed, new and blue. Extracorporeal support is a rather old idea; devices of fifty years ago remain familiar, technology has been borrowed from fields of basic science as well as industry, new indications and extensions of understanding of disease mechanism continue and over 40,000 blue patients have been treated with the technology.

The future is bright as investigations of the many facets of extracorporeal support continue to be refined. On the horizon is the next “Holy Grail,” the implantable artificial lung, and improved survival for the pulmonary failure patient. Improvement in the understanding and treatment of the inflammatory response of blood and patient to artificial circuitry will extend indications for and safety of extracorporeal support. Experience in resuscitation will likely continue to expand.

We can anticipate challenges to cost and re imbursement. The concept of regional transfer centers may well be developed in a time of high cost and the requirement of experience in the management of rare and increasingly complex disorders. If history is our teacher, the future for this area of critical care is challenging, compelling and promises to be rewarding. It all began, and continues, at the bedside.

Figures and Tables

Table 1 Extracorporeal life support results by age, organ system and survival to discharge in United States (January 2009)

Group	Cases	Percent (%) survival [to transfer of discharge]
Neonatal
Respiratory	19982	76
Cardiac	3089	39
CPR	386	38
Pediatric
Respiratory	3021	53
Cardiac	3442	44
CPR	639	39
Adult
Respiratory	1025	52
Cardiac	499	32
CPR	139	37
Total	32228	64

Table 2 Cardiac ECLS survival rates. United States summary, ELSO registry January 2009.2

Age group	Number	% Survived
Neonate	3089	39%
Pediatric	3442	44%
Adult	605	33%

Acknowledgements

Robert Hawes Bartlett, the keeper of the feast, who invited all to the table to share the good parts and begin the sorting-out of the bad.