J THORAC CARDIOVASC SURG 1989;968-71
Role of extracorporeal membrane oxygenation in selected pediatric respiratory problems Between 1984 and 1988, 89 infants and children with severe respiratory failure were supported by extracorporeal membrane oxygenation. Major clinical diagnoses included congenital diaphragmatic hernias (34), meconium aspiration syndrome (26), and sepsis (8). Extracorporeal membrane oxygenation was used for patients with a predicted mortality rate of at least 80 % based on an oxygenation index> 0.4. Venoarterial bypass was accomplished by way of right cervical cannulation of the common carotid artery and internal jugular vein. Overall survival was 71 % but varied widely by diagnosis and progressively improved over time. The average extracorporeal membrane. oxygenation run was 5.7 days. Intracranial hemorrhage was the most serious complication occurring in 16% of patients. Mechanical circuit complications were seen in 22 % but rarely related to significant morbidity. Extracorporeal membrane oxygenation appears to provide effective cardiopulmonary' support for selected pediatric respiratory problems. It affords those with potentially reversible pathophysiology the temporal opportunity for successful medical or surgical therapies.
Craig W. Lillehei, MD, P. Pearl O'Rourke, MD, Joseph P. Vacanti, MD, and Robert K. Crone, MD, Boston, Mass.
Considerable progress has been made during the past 3 decades in the understanding and application of cardiopulmonary bypass. It was designed for short-term support during intracardiac surgery. However, with the development of the membrane oxygenator, prolonged cardiopulmonary support became feasible. Initial attempts to apply this technology to respiratory failure were disappointing. 1 In 1978 multicenter, randomized, prospective study of severe adult acute respiratory failure concluded that the probability of long-term survival was not improved by extracorporeal membrane oxygenation (ECMO).2 Although ECMO was judged very effective supportive therapy, unless the underlying condition improved, mortality was not affected. Under the direction of Bartlett and colleagues;' the use of ECMO in newborn respiratory failure was investigated. Between 1973 and 1986, 100 neonates with respiratory failure were treated with remarkable success." Under their guidance and tutelage, ECMO has spread to more than 45 centers worldwide.' From the Departments of Surgery and Anesthesia, The Children's Hospital and Harvard MedicalSchool, Boston, Mass. Addressfor reprints:Craig W. Lillehei, MD, Department of Surgery, Fegan4, The Children's Hospital,300 Longwood Ave.Boston, MA 02115.
ECMO was first used in Boston for neonates in 1984. At the outset only infants with congenital diaphragmatic hernias (CDH) and progressive respiratory failure, in whom mortality was virtually certain, were treated. Since then ECMO has been used in a variety of circumstances. This article outlines that experience. Patients and methods After approval by the institutional review board, 89 infants and children were treated with ECMO between February 1984 and October 1988 at The Children's Hospital in Boston. All had received maximal medical therapy, which included pharmacologic paralysis, sedation, alkalinization, and hyperventilation. Initial criteria for ECMO use were based on 100% predicted mortality from a retrospective chart review of infants with CDH. These criteria included, preductal and postductal Pa02 < 50 and an inability to lower PaC02 < 40 despite maximal medical therapy. With the application to other diagnostic groups, oxygenation index (Fi0 2 X mean airway pressure/Pa02) > 0.4 was used instead. Patients < 35 weeks' gestation or with evidence of significant intracranial hemorrhage were excluded from consideration for ECMO. The technique of ECMO was identical to that described elsewhere by Bartlett and colleagues." After informed consent was obtained a right cervical cutdown allowed cannulation of the right atrium by way of the internal jugular vein. Blood was drained by gravity into a servo-regulated collapsible reservoir ("bladder box"). A roller pump was used at flows of approximately 120 ml/kg/rnin through a silicone membrane oxygenator and heat exchanger. Oxygenated blood was returned to the aortic arch by way of a cannula threaded retrograde through the
Volume 98 Number 5, Part 2 November 1989
ECMO in selected pediatric respiratory problems
Table I. ECMO survival: 1984 to 1988 Diagnosis
Congenital diaphragmatic hernia Meconium aspiration syndrome Sepsis Pneumonia Persistent fetal circulation Respiratory distress syndrome Pediatric Other
8 6 4 4 4 3
Survival (%) 80
41 96 100 83 100 100 25
60 40 ~ o
Table II. Mechanical complications of ECMO circuit Oxygenator Pump Heat exchanger Tubing Reservoir TOTAL
3 0 1 1 1 6/30
2 2 0 1 0 5/20
3 2 1 I 2 9/39
8 4 2 3 3 20/89 (22%)
common carotid artery. Occasionally a hemofilter was bridged into the circuit. While on ECMO muscle relaxation and sedation were weaned as possible. The lung was placed "at rest" using Fi0 2 0.21, ventilatory pressures of 20/5, and a rate of 4. Once successful oxygenation couid be maintained on conventional ventilation with an Fi0 2 0.3 to 0.4, ECMO was discontinued and the cannulas were removed. ECM0 was withdrawn earlier if severe bleeding developed, which precluded continued heparinization. Cranial ultrasound scans were obtained in all infants.
Fig. 1. ECMO survival: 1984-1988
-.-s VI l:: ell
Fig. 2. Intracranial hemorrhage: 1984-1988. ICH, Intracranial hemorrhage; V/S, ultrasound.
Results During the past 5 years, ECMO has been used in 89 infants and children with respiratory failure at The Children's Hospital. Overall survival was 71 %. The mortality rate progressively improved'over the study from an initial rate of 40% to 87% in 1988 (Fig. 1). Survival varies widely by diagnosis (Table I). ECMO has been most successful in neonates with primary pulmonary hypertension (including meconium aspiration syndrome and persistent fetal circulation). Survival in infants with CDH or older children with cardiopulmonary failure has been substantially lower. The length of ECMO runs ranged from 8 to 496 hours, including one infant who had cannulation twice. The average ECMO run lasted 138 hours (5.7 days). Runs were somewhat longer in CDH (150 hours = 6.25 days). Significantly longer runs were noted in nonsurvivors (178 hours) compared with survivors (108 hours) with this diagnosis. As many as four patients have been treated simultaneously. Cannulation was accomplished by way of the right,
common carotid artery and internal jugular vein in all except one patient. In that case the jugular vein was occluded, and direct atrial cannulation through a sternotomy was required. Volume repletion was often necessary initially to maintain adequate venous return and bypass flows.Attempts to augment venous return with additional cannulas in the femoral or proximal jugular veins were not useful. Cannulation site bleeding was seen in nine patients (10%) but was controlled with repletion ofclotting factors and local hemostatic measures. Pediatric ECMO was associated with intrathoracic bleeding complications in all four cases. Intracranial hemorrhage was identified by ultrasound scans in 16% of infants on ECMO. This complication clearly decreased over the study (Fig. 2). The incidence of mechanical circuit complications was 22% (Table II), but such problems were rarely related to any substantial morbidity. Recently continuous hemofiltration has been used in 19 patients to remove excess fluid. Three infants were brought to the operating room for
The Journal of
Thoracic and Cardiovascular
Lillehei et al.
major surgery while on bypass. Two underwent repair of diaphragmatic hernias, and one had coarctation of the aorta corrected. Discussion
The utility of ECMO for neonatal respiratory failure has been clearly demonstrated by Bartlett and coworkers" and others. 7-10 This study confirms those favorable results in a large series from a single institution. Presently the primary indication for ECMO in neonates is persistent pulmonary hypertension of the newborn. In infants with a predicted mortality of 80% based on an oxygenation index> 0.4, survival rates of 80% to 90% have been regularly achieved. 11 The results reported herein compare favorably. ECMO has also been applied to other causes of pediatric cardiopulmonary failure with varied success. The challenge is to identify those patients with diseases that are reversible during a limited period of cardiopulmonary support. This study also demonstrates the evolution of an ECMO program. Although selection criteria and case mix have certainly changed over time, the improved survival has been dramatic. There is certainly a "learning curve." The incidence of mechanical circuit complications has been acceptable, but bypass requires constant vigilance. More reliable circuitry and monitoring are still required. The long-term morbidity of carotid ligation is not known. However, the techniques of venovenous bypass are encouraging in this regard.'? The most serious complication of ECMO has been intracranial hemorrhage associated with the requisite systemic heparinization. Its reduction over time has been gratifying. This probably reflects multiple factors such as decreasing heparin dosages, maximizing venous drainage, and minimizing pre-ECMO trauma. Nonetheless, 10% of infants treated still suffer this complication. Application of ECMO technology to the respiratory problems of premature infants awaits the development of heparinless circuits. Infants with CDH remain a major challenge. The results reported herein reflect a strategy in which no infants, regardless of pre-ECMO gas exchange, are considered unsalvageable. It appears that although recovery is possible in some, others have insufficient pulmonary parenchyma to support gas exchange. Studies are currently in progress that are using early ECMO as a possible bridge to neonatal lung transplantation in those who cannot be successfully weaned from support. The role of ECMO is still evolving. However, it has clearly emerged from the laboratory and operating room into the modem intensive care unit.
REFERENCES 1. White JJ, Andrews HG, Risenberg H, Mazur 0, Haller JA. Prolonged respiratory support in newborn infants with a membrane oxygenator. Surgery 1971;70:288-96. 2. Zapol WM, Snider MT, Hill JD, et al. Extracorporeal membrane oxygenation in severe acute respiratory failure: a randomized prospective study. JAMA 1979;242:2193-6. 3. Bartlett RH, Gazzaniga AB, Jeffries MR, Huxtable R, Haiduc J, Fong S. Extracorporeal membrane oxygenation (ECMO) cardiopulmonary support in infancy. Trans Am Soc Artif Intern Organs 1976;22:80-93. 4. Bartlett RH, Toomasian JM, Roloff OW, Gazzaniga AB, Corwin AG, Rucker R. Extracorporeal membrane oxygenation (ECMO) in. neonatal respiratory failure: 100 cases. Ann Surg 1986;204:236-44. 5. Toomasian JM, Snedecar SM, Correll RG, Cilley RE, Bartlett RH. The national experience with extracorporeal membrane oxygenation for newborn respiratory failure. Data from 715 cases. Trans Am Soc Intern Organs 1988;34:140-7. 6. Bartlett RH, Andrews AF, Toomasian JM, Haiduc N, Gazzaniga AB. Extracorporeal membrane oxygenation for newborn respiratory failure: 45 cases. Surgery 1982;92:42533. 7. Short BL, Lotze A. Extracorporeal membrane oxygenation therapy. Ann Pediatr 1988;17:516-25. 8. Krummell TM, Greenfield LJ, Kirkpatrick BY, et al. Clinical use of an extracorporeal membrane oxygenator in neonatal pulmonary failure. J Pediatr Surg 1982;17:525-31. 9. Redmond CR, Graves ED, Falterman KW, Ochsner JL, Arensman R M. Extracorporeal mem brane oxygena tion for respiratory and cardiac failure in infants and children. J THORAC CARDIOVASC SURG 1987;93:199-204. 10. Weber TR, Pennington DG, Connors R, et al. Extracorporeal membrane oxygenation for newborn respiratory failure. Ann Thorac Surg 1986;42:529-35. 11. Ortiz RM, Cilley RE, Bartlett RH. Extracorporeal membrane oxygenation in pediatric respiratory failure. Pediatr Clin North Am 1987;34:39-46. 12. Klein MD, Andrews AF, Wesley JR, et al. Yenovenous perfusion in extracorporeal membrane oxygenation for newborn respiratory insufficiency: a clinical comparison with venoarterial perfusion. Ann Surg 1985;201 :5206.
Discussion Dr. Gannon. Thank you very much, Craig. Are there any questions for Dr. Lillehei? I would like to ask two. First, could you go briefly through the weaning process? Secondly, could you also tell us how often you must change pump heads, whether you use a roller head, or a Bio-medicus, or Centrimed? Dr. Craig Lillehei. Typically we have used roller pumps with good success. We have not had to change the membrane oxygenator in most patients. The typical run as I have described is
Volume 98 Number 5, Part 2 November 1989
over in 7 days, and the membrane oxygenator functions quite well during that period. In the patient that went for 3 weeks on extracorporeal bypass, we electively changed the circuit after 7 days when there was some evidence of malfunction. As far as the weaning process is concerned, when pulmonary compliance seems reasonable and the x-ray film looks promising, these patients are cycled off extracorporeal membrane oxygenation (EMCO). We simply reduce our flows to about 3D mI/kg/min. The patient is taken off ECMO completely, and the arterial blood gases are measured. If good gas exchange is achieved over a 15- to 3D-minute period, then they are returned to bypass, and arrangements are made to decannulate these patients, again in the intensive care unit. Dr. Hunter. I would like to point out that Craig's middle initial is not W. He was Craig Owen Lillehei named after Owen Wangensteen. Last January I watched a lady with pulmonary emphysema practically die and there absolutely was nothing we coulddo for her. She was already on oxygen binasal. It occurred to me that there was a possibility that we could put a very small catheter in her inferior vena cava and microbubble oxygen into it and in that way make her her own oxygenator. When I came home from visiting this lady, I talked to Craig's father, and he said it isso simple, we never thought about it. Perhaps we should think of putting a small catheter in the inferior vena cava, microbubbling oxygen in there, and using that method of nonpulmonary oxygen transport. Dr. Craig Lillehei. Interesting idea. I have no experience. Dr. Walton Lillehei. You better correct the names. Dr. Cohn. Craig, that is a tremendous piece of work. Having had some experience with this in adults at the Brigham a few years before, it represents just an unbelievable amount of work. Youare to be congratulated. I guess you got the work ethic from your old man, genetically transferred. You mentioned survival. Is this long-term survival? Are these peopleon whom you have an appreciable follow-up? Assuming that this is the case, I would certainly appreciate knowing the length of your survivorship. Have you evaluated their lungs functionally at the late postoperative periods, and what type of residual changes, if any, have you noted? Finally, the other question that is always bugging us in "Taxachusetts," as you well know, is that this is a very costly clinical procedure. Is this a reimbursable clinical project now-a-days, or
ECMO in selected pediatric respiratory problems
is it something that is on an individual basis at your institution related to research funding because you are sort of in the forefront there? Dr. Craig Lillehei. Thank you, Dr. Cohn. Dr. Walton Lillehei. I rise to correct my good friend and former resident, Dr. Samuel Hunter. Sam, his name is Craig Walton Lillehei, but we do have a Kevin Owen Lillehei, who is also here today. Kevin is a younger brother of Craig's and is a neurosurgeon in the Department of Surgery at the University of Colorado in Denver. He was named after Owen Wangensteen. Dr. Hunter. I stand corrected. Dr. Craig Lillehei. Dr. Cohn, in response to your questions, first, the survivalship that we report in this series dates to October 1988. It therefore does not reflect long-term survival. We have had one late death in a patient with a diaphragmatic hernia. All of the remainder are still surviving and thriving at present. In neonates with persistent pulmonary hypertension, the critical period appears to be those first few weeks of life. Once we are able to get them over that hurdle, fortunately, late death has not been a major problem. However, late morbidity is a problem. These are patients who have often been banged away with the ventilator for several days, if not weeks, beforehand. The incidence of bronchopulmonary dysplasia or fibrosis in that residual lung is considerable, both in our ECMO population, as well as a control population. Those are the results that will be reported separately, but I think the message is that it does not make a lot of sense to beat these lungs up beforehand if, in fact, you're not gaining ground. That is where ECMO offers very promising support. In answer to your third question, yes, I live in "Taxachusetts." It has been very difficult financially for our EMCO program from the outset. To date we have received little financial reimbursement for this modality, although in most states it is accepted therapy. Most recently Blue Cross and Blue Shield agreed that it is a medically acceptable procedure, so we are hopeful that the reimbursement situation may improve. Dr. Castaneda. Sorry to rise again, but I just had to put on my hat as surgeon-in-chiefof Children's Hospital and would like to say that we are extremely happy to have this energetic, talented, and outstanding young man on our staff, who I might add, has the fortune of having combined in his genes the attributes of both his father and his mother.