- Email: [email protected]
Extracorporeal Membrane Oxygenation for Newborn Respiratory Failure
Extracorporeal Membrane Oxygenation for Neibom Respiratory Failze Thomas R. Weber, M.D., D. Glenn Pennington, M.D., Robert Connors, M.D., William Kennan, M.D. , Suresh Kotagal, M.D., Paul Braun, C.C.P., and Victor Martychenko, C.C.P. ABSTRACT Jugular vein-carotid artery extracorporeal membrane oxygenation (ECMO) was utilized in 22 newborns (16 male and 6 female) 1 to 12 days old with respiratory failure due to meconium aspiration (12 patients), diaphragmatic hernia (41, persistent fetal circulation (3), hyaline membrane disease (21, and Rh incompatibility (1). Prior to ECMO, all patients had alveolar-arterial O2 pressure gradients greater than 580 mm Hg (predictedmortality greater than 90%), weighed more than 1,800 gm, had a gestation period of longer than 35 weeks, and had no cerebral hemorrhage. The duration of ECMO was 41 to 310 hours (mean, 134.5 hours). Nineteen (86%) of the 22 patients survived ECMO. Death was caused by lung disease (2) and cerebral hemorrhage (1). Four other patients died 6 to 40 days after ECMO of pulmonary hypoplasia (11, pneumonia (11, cerebral edema (11,and hepatorenal failure (1). Complications during ECMO were few and easily managed. Fifteen infants (68%)are alive 1 to 18 months after ECMO. Three have neurological deficit (2 severe, 1 mild). Bayley Developmental Examinations in 4 survivors now more than 12 months old are normal. Extracorporeal membrane oxygenation is an aggressive but effective technique of life support in newborns refractory to conventional respiratory management. Potential complications of ECMO mandate strict aseptic technique, constant monitoring, and multidisciplinary patient management.
Although the survival of infants with respiratory failure has improved markedly over the past ten to fifteen years, certain diseases continue to have a poor prognosis; among these are congenital diaphragmatic hernia, meconium aspiration syndrome, persistent pulmonary hypertension of the newborn (persistent fetal circulation), and hyaline membrane disease. Contemporary management of these infants with ventilators, drug therapy (inotropic and pulmonary vasodilators), and other From the Divisions of Pediatric Surgery and Pediahic Cardiothoracic Surgery, St. Louis University Medical Center and Cardinal Glennon Children's Hospital, St. Louis, MO. Presented at the Twenty-second Annual Meeting of The Society of Thoracic Surgeons, Washington, DC, Jan 27-29, 1986. Address reprint requests to Dr.Weber, Department of Pediatric Surgery, Cardinal Glennon Children's Hospital, 1465 S Grand Blvd, St. Louis, MO 64104.
529 Ann Thorac Surg 42529-535, Nov 1986
means of life support have resulted in 50% survival at best in most of these disorders. Extracorporeal membrane oxygenation (ECMO) has been utilized successfully in a few centers for treatment of term or near-term newborns with potentially reversible respiratory failure. Since much of the morbidity associated with conventional therapy is directly related to complications from the therapy (high-pressure ventilation, air leak syndromes), the provision of a period of lung "rest" with the use of ECMO seems an attractive means of support. In spite of a number of reports attesting to the usefulness of ECMO in newborn respiratory failure, a number of clinicians remain unconvinced as to its safety, efficacy, and absence of long-term sequelae. These concerns prompted a review of 22 newborns treated for respiratory failure with ECMO over an 18month period.
Material and Methods
Perfusion Circuit For the past five years, ECMO has been utilized at St. Louis University Medical Center for support in patients with cardiogenic shock, most of whom were recovering from a cardiac procedure [l].This experience included both adults and children, and in several instances, the technique acted as a "bridge" to cardiac transplantation. Encouraged by this experience, the medical center initiated ECMO for newborn respiratory failure in January, 1984. The perfusion circuit is very similar to that used by Bartlett and co-workers [2], and consists of a roller pump (Picker International, Ossining, NY), a membrane lung, and a heat exchanger (both SciMed, Minneapolis, MN), all of which are connected by polyethylene tubing. The roller pump is servo controlled by a pressure-sensitive switch that is in contact with the Silastic venous reservoir and interrupts the roller pump when venous filling of the reservoir is slowed. The membrane oxygenator is connected to gas sources (02and CO2) through a gas mixer, which allows precise, appropriate changes in gas supply. The heat exchanger is connected to a circulating liquid heater (Gaymar Industries, Orchard Park, NY), generally set at 38°C. The entire apparatus, including the activated clotting time (ACT) monitor (Hemochron, Metuchen, NJ), is mounted on a cart that is taken to the patient's bedside for cannulation and perfusion. The cannula for venous cannulation is either a 12F or 16F chest tube or a standard 14F cardiopulmonary cannula, and it is placed through the right jugular vein into the right atrium. The arterial cannula is either a 10F chest tube or 8F plastic pediatric feeding tube, placed through
530 The Annals of Thoracic Surgery Vol 42 No 5 November 1986
the right carotid artery into the aortic arch. The position of the cannulas is confirmed by roentgenography prior to initiation of perfusion if the clinical status of the patient allows such a delay. Each vessel (jugular vein and carotid artery) is ligated distally, and no attempt is made to reestablish continuity after cessation of ECMO. The cannulation procedure is performed with complete aseptic technique in the intensive care unit. The cannulas exit through the neck incision, which is closed loosely with skin sutures. An antibiotic-soaked gauze is applied to the incision and is changed daily in a sterile fashion. Removal of the cannulas is likewise performed at the patient's bedside. Just prior to cannulation, the baby is heparinized intravenously (150 units per kilogram of body weight), and the ACT is maintained at 180 to 200 seconds with a continuous heparin drip. The bypass flow rate must be sufficient to completely support and perfuse the infant. The normally contoured arterial pattern flattens as the flow rate is increased. When the arterial pressure tracing flattens to the point that there are no peaked contours, then approximately 70 to 80% of cardiac output is going into the ECMO circuit. This generally occurs at a flow rate of 140 to 160 ml/kg/min, but is dependent on the state of hydration at the time ECMO is initiated. Mainte-
Fig 1. Clinical profile of a typical newborn with meconium aspiration. Increasing ventilatory requirements (inspired Ozfraction [FiO,], peak ventilatory pressure, and ventilator rate) combined with failing arterial oxygenation (Pa02)prompted the use of extracorporeal membrane oxygenation (ECMO) at 48 hours of age. Lung "rest" allowed rapid pulmonary improvement and gradual weaning from ECMO, which was terminated at 120 hours of age. (CPAP = continuous positive airway pressure.) Birth Meconium
nance volumes of total parenteral nutrition are added to the ECMO circuit, as the gastrointestinal tract is not functional while the baby is on ECMO. Hemolysis is not usually a problem with ECMO. Very high flow rates through a small arterial cannula or perfusion pressures exceeding 300 mm Hg have produced little or no clinically evident hemolysis. As lung function improves, the ECMO flow rate is decreased and the ventilator settings are increased to allow lung function to resume. The impending improvement is frequently heralded by a change in chest radiograph from opacification to aeration within lung fields. When the flow rate reaches 60 to 80 ml/min, a test period is attempted with the baby completely off ECMO. If the blood gases are adequate, the cannulas are removed and ECMO is terminated. A typical ECMO profile is shown in Figure 1. Clinical Material In the 18-month period, 36 inquiries were received concerning patients who might be candidates for ECMO. From this pool, 22 were found to satisfy the criteria set for ECMO at our institution and approved by the Institutional Review Board, St. Louis University Medical Center. The criteria are based on data published by other centers with established newborn ECMO programs [371, and identify a population with a 90% or greater risk for death related to respiratory failure. These criteria are as follows: Weight > 1,800 gm Gestational age > 35 weeks Cranial ultrasound negative for intracranial hemorrhage Cardiac evaluation (ultrasound) confirming cardiac lesion is not the cause of respiratory distress
Bilateral Dopaminel --Removed +From ECMO
Vent. Rate
CPAP Extubated
1 4
Peak Vent. Pressure
FiO,
150
ECMo 100 Flow cclKglmin 50 24
48
R
96
120
Hours
144
200 212
531
Weber, Pennington, Connors, et al: Extracorporeal Membrane Oxygenation for Newborns
Diagnosis Meconium aspiration Persistent fetal circulation Ventilator barotrauma (pneumothorax, pneumomediastinum, pulmonary interstitial emphysema) Diaphragmatic hernia Level of hypoxia (any one of the following with maximal ventilator and pharmacological therapy) Acute deterioration: arterial 0 2 tension < 40 mm Hg, pH < 7.15, cardiopulmonary resuscitation P(A-a)02 > 580 mm Hg for 4 hours, or rapidly increasing* P(A-a)02 > 620 mm Hg for 1 hour* Of the 22 newborns treated with ECMO, 16 were male and 6 were female. They ranged from 11 to 298 hours old (mean, 62.5 hours) and weighed from 1,990 to 4,100 gm (mean, 3,100 gm). The diagnoses were meconium aspiration (12 patients), diaphragmatic hernia (4), persistent fetal circulation (3), hyaline membrane disease (2), and Rh incompatibility (1). Ten of the 12 newborns with meconium aspiration and 1 of the 2 infants with hyaline membrane disease also had pulmonary hypertension and a fetal circulatory pattern that contributed to the hypoxia. Two infants had a cardiac defect in addition to meconium aspiration-tetralogy of Fallot in 1 and double-outlet right ventricle in the other. In neither patient did the cardiac defect complicate the ECMO therapy. All 22 newborns had a P(A - a)O2greater than 580 mm Hg, and 15 had a gradient greater than 600 mm Hg. All patients were treated with maximal therapy prior to the institution of ECMO. This consisted of rapid ventilation (60 to 120 breaths per minute), frequently with high peak inspiratory pressures (30 to 50 mm Hg), and inotropic (dopamine hydrochloride, Isuprel [isoproterenol hydrochloride]) and pulmonary vasodilator (tolazoline hydrochloride) medications. All patients were considered to be failures of medical therapy as judged by the attending neonatologist. Seven of the infants were transferred by air transport (fixedwing and helicopter), and the remainder were in the newborn intensive care unit at Cardinal Glennon Children's Hospital. Six of the 22 were undergoing cardiopulmonary resuscitation, including external cardiac massage, at the time of ECMO cannulation. After discharge from the hospital, the infants are followed by a multidisciplinaryteam of pediatric surgeons, neonatologists, and pediatric neurologists at frequent intervals. When each child is 1 year of age, the Bayley Developmental examination is administered to assess achievement of developmental milestones. Neurological deficits, developmental abnormalities, seizure disorders, and chronic lung diseases are followed closely and treated appropriately.
0)
3
In1
4:h 20
"
R DS
UA
PFC
DH
Diagnosis
Fig 2 . Age at which extracorporeal membrane oxygenation (ECMO) was initiated for each diagnosis. (PFC = persistent fetal circulation; MA = meconium aspiration; RDS = respiratory distress syndrome; DH = diaphragmatic hernia.)
Results The mean age at which ECMO was initiated for each diagnostic category is shown in Figure 2. In general, the oldest patients were treated early in the series, but as the experience grew, newborns were referred earlier for consideration for ECMO. The duration of ECMO therapy was 41 to 310 hours (mean, 134.5 hours). The longest runs were in infants with diaphragmatic hernia, and the shortest runs were in infants with persistent fetal circulation unrelated to meconium aspiration (Fig 3). When age at onset of ECMO is examined versus the duration of ECMO, no direct correlation is found. However, all of the survivors had ECMO durations of 160 hours or less, and the nonsurvivors had longer ECMO runs (Fig 4). Nineteen of the 22 patients survived the ECMO treatment. The 3 deaths during ECMO were due to massive cerebral hemorrhage (1 infant) and irreversible lung dis-
"OI 200
8 120
I
40
PFC
1 IDS
Diagnosis 'P(A-a)O, = 760 - (PaOz + PaCO, + 47) where P(A-a)Oz = alveolar-arterial O2pressure difference, Pa02 = arterial 0 2 tension, and PaCOz = arterial COz tension [7].
Fig 3 . Duration of extracorporeal membrane oxygenation (ECMO) for each diagnosis. (PFC = persistent fetal circulation; MA = meconium aspiration; RDS = respiratory distress syndrome; DH = diaphragmatic hernia.)
532 The Annals of Thoracic Surgery Vol 42 No 5 November 1986
0 Survivors
0
Non-survivors
2240 80]0
-99
40-
0
-
Fig 4. Duration of extracorporeal membrane oxygenation (ECMO) versus age when it was initiated. A distinct division (broken line) between survivors and nonsurvivors occurred at 140 to 160 hours of ECMO. There was no direct correlation between duration of ECMO and age at which it was initiated.
order (2 patients). These latter 2 newborns had ECMO durations of 220 and 230 hours, which indicated severe lung disease that was unresponsive to even the most aggressive therapy. Four other infants died 6 to 40 days after ECMO. Two of these infants had repair of diaphragmatic hernia and were placed on ECMO at 6 and 60 hours after repair. They died of pulmonary hypoplasia and Pseudomonas sp. pneumonia, respectively. In the third infant, massive cerebral edema developed and brain death occurred 48 hours after he was removed from ECMO. His pulmonary function had recovered fully. The fourth newborn died of pulmonary and hepatorenal failure related to the original diagnosis (immune hydrops from Rh incompatibility). Thus, the overall survival is 15 (68%) of 22. Patient complications included the following: seizures (8 patients, 6 survivors), renal failure (5 patients, 4 survivors), bleeding (4 patients, 3 survivors), and neurological deficit (2 severe, 1 mild). None proved to be significant determinants of survival. The seizures were successfully treated in each instance with intravenous administration of phenobarbital sodium. Bleeding from the neck wound was managed by exploration and cauterization of bleeding vessels. Renal failure (urine output less than 1 ml/kg/hr) was treated with an ultrafiltration unit (Amicon Corporation, Danvers, MA) inserted in parallel into the ECMO circuit. Flow through the ultrafiltration unit was generally 50 to 75 mumin, thereby allowing the production of ultrafiltrate at 10 to 15 mVhr.
The malfunctions that developed with the ECMO apparatus were easily recognized and managed. None proved to be life threatening. Positive blood cultures developed in 4 patients while they were on ECMO. The organisms involved were Candida albicans (2 patients), Staphylococcus aureus (1 patient), and Pseudomonas sp. (1 patient). None of these 4 infants survived, although 1 of them lived 6 weeks before dying of Pseudomonas sp. pneumonia. In 3 patients, tubing rupture, apparatus leak, or both occurred; 2 of them survived. In 2 patients, both of whom lived, the oxygenator failed. The follow-up of the 15 long-term survivors emphasized neurological and developmental examinations. No chronic pulmonary disorders have been identified. One infant has mild neurological deficit (upper and lower extremity weakness) associated with seizures. However, developmentally the child seems normal. Two infants have severe, generalized neurological deficits that include seizures that are difficult to control. In addition, each child lacks cognitive function and is totally dependent for day-to-day care. The other 12 infants, who are 3 to 18 months old, have normal neurological examinations, and 4 more than 1 year of age have normal Bayley Developmental Examinations.
Comment The past fifteen to twenty years have seen major advances in neonatal intensive care. The improved survival in newborns, especially premature babies with pulmonary immaturity, is a direct result of improved technology. However, there continue to be several pulmonary disorders in newborns that have a high rate of death. Among these are meconium aspiration, persistent fetal circulation, and congenital diaphragmatic hernia. The high mortality has persisted despite the use of high-rate ventilation, jet ventilators, pulmonary
533 Weber, Pennington, Connors, et al: Extracorporeal Membrane Oxygenation for Newborns
vasodilators, and other measures. In fact, the high mortality in these disorders is due in part to barotrauma secondary to high-pressure ventilation, combined with excessive inspired 0 2 concentrations needed to adequately oxygenate these infants. These disorders must be treated with more aggressive maneuvers if survival is to be improved. In 1972, Hill and associates [8] treated an adult with ECMO; the patient survived. The use of ECMO in adults has been abandoned in many centers, however, because of an excessive number of complications and because of the results of a multicenter prospective randomized trial [9] from 1974 to 1979 that demonstrated no significant improvement in survival in this age group for a variety of diseases treated with ECMO. In contrast, the newborn seems to be peculiarly well suited to this form of therapy, as suggested by a growing body of data. In 1976, Bartlett and co-workers [lo] utilized ECMO for support of a newborn with meconium aspiration syndrome in whom all medical therapy was ineffective. The child survived, 1 of the first newborn survivors of ECMO. Since then, reports from a number of centers [37, 11, 141 have confirmed the usefulness of this technique. However, several controversial issues remain regarding its use. The present method of arteriovenous ECMO utilizes carotid artery ligation. Although this has not resulted in major neurological deficits, there are concerns about this invasive measure. The neurological complications that have developed in this series and in others center around intracerebral bleeding, brain edema, and seizures. The first two complications are frequently fatal, while the last (seizures) has been relatively common and does not seem to be an important determinant of survival. The short-term and long-term follow-up of infants treated with ECMO in several different centers [6,12,13] shows results similar to those presented here. The criteria used for the selection of infants who might benefit from ECMO have been unclear and vary from center to center. Most centers performing ECMO in newborns will accept patients with specific age (less than 10 to 14 days old), weight (greater than 1,800 to 2,000 gm), and neurological (normal head ultrasound) criteria, but the pulmonary severity that predicts greater than 90% mortality has been less straightforward. Attempts to use objective measures to assess candidates for ECMO, such as pulmonary insufficiency index [3] or P(A - a)O2 [7], have not been universally recognized. Our policy is to accept newborns who have received maximal medical therapy, who are referred by the attending neonatologist because of persistent or worsening hypoxia, and who satisfy the criteria listed earlier in this study. We believe this policy identifies a group of infants who have a high mortality risk (probably greater than 90%).This concept is reinforced by the fact that 6 of our infants were receiving chest massage and cardiopulmonary resuscitation at the time of cannulation. An analysis of the infants who died in this series reveals several important facts. All 4 infants in whom posi-
tive blood cultures developed died. Broad-spectrum antibiotics (gentamicin sulfate and ampicillin) are utilized during ECMO and continued for several days after termination of the procedure. However, development of resistant organisms and the emergence of organisms not covered by antibiotics (Cundidu sp.) resulted in infections that ultimately proved fatal. Strict aseptic technique during insertion of cannulas and each time the ECMO circuit is entered for blood sampling, heparin infusion, and intravenous administration of fluid or parenteral nutrition are critical. The 4 infants in whom infection developed had among the longest ECMO runs (greater than 7 days), although we and others have had numerous patients on ECMO longer than 7 days without the development of bacteremia. Longer ECMO durations mandate even closer vigilance with regard to aseptic technique. Because of the complex condition of the patients involved, the care of the infant before, during, and after ECMO is best performed by a multidisciplinaryteam. In our institution, the pediatric surgeons and pediatric cardiovascular surgeons are in charge of the ECMO programs and direct the care of the child. Neonatologists and cardiologistsare helpful in the assessment and management of respiratory and cardiac systems, and the pediatric neurologists manage seizure disorders, neurological deficits, and long-term developmental followup after the patient is discharged from the hospital. Other consultants are used as needed. The present series confirms the data presented from other centers concerning the efficacy of ECMO for respiratory failure, with satisfactory long-term neurological results. Further refinements in patient selection, aseptic management, and the multidisciplinary team approach should result in greater survival among these critically ill neonates. We thank the many referring physicians and neonatologists, the ECMO perfusion team, and the superb nursing staff in the intensive care unit of Cardinal Glennon Children’sHospital for their help in the management of these patients.
References 1. Pennington DG, Merjavy JP, Codd JE, et al: Extracorporeal membrane oxygenation for patients with cardiogenic shock. Circulation 7O:Suppl 1:130, 1984 2. Bartlett RH, Andrews AF, Toomasian JM, et al: Extracorporeal membrane oxygenation for newborn respiratory failure: forty-five cases. Surgery 92:425, 1982 3. Wetmore N, McEwen D, OConnor M, et al: Defining indications for artificial organ support in respiratory failure. Trans Am SOCArtif Intern Organs 25:459, 1979 4. Krummel TM, Greenfield LJ, Kirkpatrick BU, et al: Extracorporeal membrane oxygenation in neonatal pulmonary failure. Pediatr Ann 11:905, 1982 5. Bartlett RH, Roloff DW, Cornell RG, et al: Extracorporeal circulation in neonatal respiratory failure: a prospective randomized study. Pediatrics 76:479, 1985 6. Loe WA, Graves ED, Ochsner JL, et al: Extracorporeal
534 The Annals of Thoracic Surgery Vol 42 No 5 November 1986
membrane oxygenation for newborn respiratory failure. J Pediatr Surg 20:684, 1985 7. Krummel TM, Greenfield LJ, Kirkpatrick BU, et al: Alveolar-arterial oxygen gradients versus the neonatal pulmonary insufficiency index for prediction of mortality in ECMO candidates. J Pediatr Surg 19:380, 1984 8. Hill JD, O'Brien TG, Murray JJ, et al: Prolonged extracorporeal oxygenation for acute post-traumatic respiratory failure (shock-lung syndrome). N Engl J Med 286:629, 1972 9. Zapol WM, Snider MT, Hill JD, et al: Extracorporeal membrane oxygenation in severe acute respiratory failure: a randomized prospective study. JAMA 2422193, 1979 10. Bartlett RH, Gazzaniga AB, Jefferies R, et al: Extracorporeal membrane oxygenation (ECMO) cardiopulmonary support in infancy. Trans Am SOCArtif Intern Organs 22:80, 1976 11. Hardesty RL, Griffith BP, Debski RF, et al: Extracorporeal membrane oxygenation: successful treatment of persistent fetal circulation following repair of diaphragmatic hernia. J Thorac Cardiovasc Surg 81:556, 1981 12. Worcester CC, Huxtable RF, Rucker RW, et al: Meconium aspiration: management with extracorporeal membrane oxygenation (ECMO). Am SOCArtif Intern Organs 7:139, 1984 13. Krummel TM, Greenfield LJ, Kirkpatrick BU, et al: The early evaluation of survivors after extracorporeal membrane oxygenation for neonatal pulmonary failure. J Pediatr Surg 19:585, 1984 14. Towne BH, Lott IT, Hicks DA, et al: Long-term follow-up of infants and children treated with extracorporeal membrane oxygenation (ECMO): a preliminary report. J Pediatr Surg 20:410, 1985
Discussion (San Francisco, CA): I congratulate Dr. Weber on these extraordinarily good results. ECMO has come a long way since it was first begun in the early 1970s, and many of the technological problems from those times have largely been corrected, or at least carefully managed. This is a clinical paper that I was lucky enough to be able to review beforehand and look carefully at the data. There is no hypothesis in it to prove or disprove. The value lies in its contribution to the cumulative knowledge of clinical management and in the selection of patients for ECMO that has allowed this group to have such great success. It also adds to the work of Bartlett in Ann Arbor and the Pittsburgh group in showing that ECMO can be a very successful therapy in a selected group of neonates. I have a few comments and several questions. When one examines ECMO, there are three things that are important to patient survival. They are (1) to carefully select patients who have a reversible pulmonary lesion, (2) to always be able to control the circulation and oxygenation, and (3) to avoid complications. An important selection criterion for this neonate group was that the babies were full term and did not have low weights. The mean weight was 3,100 gm, and some weights were as high as the normal weight. Another characteristic of this group was that the vast majority of patients had single-organ disease. One cannot expect ECMO or a related therapy to resurrect patients who are dying of multiple-organ failure. This therapy must be directed at a specific problem to be successful. One difficulty that has plagued patient selection and the success of ECMO concerns the kind of respiratory characteristics DR. J. DONALD HILL
we select when trying to determine the reversibility of a pulmonary lesion. This has been a particular problem in neonates when one is attempting to achieve a 10% survival without ECMO versus a 70 to 80% survival with ECMO. Those of you who deal with pediatricians know that they are reluctant to refer these children to a high-technology therapy that they don't understand. I have a few comments on the technological problems experienced in this study. I was surprised to see that in 5 patients there were technological problems that I consider largely preventable; I am referring to leaks, tubing breaks, and oxygenator failure. This is not a criticism of this excellent group from St. Louis but rather a criticism of the technology. There has been a failure at the industrial research level to fabricate oxygenators for use in this neonate group that have an acceptable standard of reliability. To have technological problems in almost 25% of the patients is unacceptable. However, as far as I could understand from the manuscript, there were no deaths or complications related to these problems. What about complications? The complications largely involved problems of the central nervous system and failure of the lung to recover. The central nervous system problems were related to the disease itself and the tendency for these children to have cerebral hemorrhages. The incidence of hemorrhage in this group is related to the need to heparinize. This is another area where research has to be done to develop antithrombogenic surfaces to minimize this problem. I ask the authors whether they have considered using hypothermia to manage the problem of damage to the central nervous system that is possibly related to the carotid ligation and to the existing hypoxia . In management, I was surprised and pleased to see that the authors were able to reduce the inspired O2 fraction and the barrel pressure. It was not exactly clear to me whether they put the lungs completely or only partially to rest. Dr. Weber, would you give us a detailed account as to what was done with the lungs during ECMO? My experience has been that when the lungs are brought to complete rest, they turn into liver, which precipitates a hopeless situation. Thus, we always maintain some pulmonary function in terms of ventilation and keep the pulmonary artery pressures up. Dr. Weber, do you monitor these pressures during this period? What do you do if you do not get a good flow? Do you readjust your cannula or go with what you have? The results of Dr. Weber and his co-authors are excellent, and it would be very hard to improve on them with current technology. Was there a relationship between survival and the weights of the infants? Do you have a favored lesion among the four common syndromes that you presented? I again congratulate Dr. Weber on these excellent results, and thank him for the opportunity to review this manuscript in advance. DR. WEBER: Dr. Hill's pioneering efforts in this field are well known, and I consider it an honor to comment on his remarks. One aspect of ECMO that has been a bit puzzling and varies from center to center is the selection of the babies. As Dr. Hill mentioned, the neonatologists you are working with have a great deal to do with this. Many neonatologists are reluctant to refer patients for this invasive technique without first exhausting all possibilities of medical therapy. Because of our success with ECMO, many neonatologists are becoming more aggressive in terms of referring the patients. For obvious reasons, we must wait until the neonatologists specifically refer the babies for ECMO before we can institute the therapy. Dr. Bob Bartlett
535
Weber, Pennington, Connors, et al: Extracorporeal Membrane Oxygenation for Newborns
has addressed the problem of selection of patients in several prospective randomized studies, and has tried to determine more precisely when the babies should be placed on the apparatus and which babies are the best candidates for the procedure. He and others are continuing these studies at present. In answer to your question concerning hypothermia, we have not utilized that approach as yet, although it does seem attractive. I am also unaware of any other ECMO program for newborns that uses hypothermia at this time. As for the ventilator settings, we decreased the rate to 10 per minute, the peak ventilator pressure to 15 cm H 2 0 , and the inspired O2 fraction to 0.3. We don’t completely eliminate the ventilator, but rather utilize very low settings until the lungs have recovered. This generally results in opacification of the lung fields on roentgenograms for several days, but when the
roentgenogram begins improving, that is usually a signal that the baby is ready for weaning from ECMO. As for the age and weight of the patient, we have not noticed any difference in survival in regard to the patient’s weight. Bob Bartlett has noticed that the much smaller babies, with weights in the range of 1,200 to 1,400 gm, don’t seem to tolerate the procedure well and have a high incidence of intracerebral bleeding. However, to date, we have not treated any babies that small. Finally, the lesions that seem to respond best to ECMO include meconium aspiration and persistent fetal circulation. The highest mortality to date seems to be in babies with diaphragmatic hernia and severe pulmonary hypoplasia. We will continue to look for improvements in the technique that might offer survival to an increasingly large group of babies.
Important Notice for Contributors to The Annals of Thoracic Surgery Beginning with this issue, The Annals is instituting several changes in the processing of manuscripts. The most important of these are: (1) One original and two duplicate manuscripts will be required, including three separated and clearly labeled sets of illustrations. Formerly, only the original and one duplicate manuscript and two sets of illustrations were required. (2) Because of escalating postage costs, manuscripts and
illustrations routinely will not be returned to the authors. The Editorial Office will be responsible for their proper disposal so that confidentiality is preserved. Manuscripts and illustrations will be returned only if requested and a self-addressed stamped (US postage) envelope provided. There are also a number of other changes. We urge prospective contributors to read the Information for Authors sheet carefully before submitting manuscripts.
Although the survival of infants with respiratory failure has improved markedly over the past ten to fifteen years, certain diseases continue to have a poor prognosis; among these are congenital diaphragmatic hernia, meconium aspiration syndrome, persistent pulmonary hypertension of the newborn (persistent fetal circulation), and hyaline membrane disease. Contemporary management of these infants with ventilators, drug therapy (inotropic and pulmonary vasodilators), and other From the Divisions of Pediatric Surgery and Pediahic Cardiothoracic Surgery, St. Louis University Medical Center and Cardinal Glennon Children's Hospital, St. Louis, MO. Presented at the Twenty-second Annual Meeting of The Society of Thoracic Surgeons, Washington, DC, Jan 27-29, 1986. Address reprint requests to Dr.Weber, Department of Pediatric Surgery, Cardinal Glennon Children's Hospital, 1465 S Grand Blvd, St. Louis, MO 64104.
529 Ann Thorac Surg 42529-535, Nov 1986
means of life support have resulted in 50% survival at best in most of these disorders. Extracorporeal membrane oxygenation (ECMO) has been utilized successfully in a few centers for treatment of term or near-term newborns with potentially reversible respiratory failure. Since much of the morbidity associated with conventional therapy is directly related to complications from the therapy (high-pressure ventilation, air leak syndromes), the provision of a period of lung "rest" with the use of ECMO seems an attractive means of support. In spite of a number of reports attesting to the usefulness of ECMO in newborn respiratory failure, a number of clinicians remain unconvinced as to its safety, efficacy, and absence of long-term sequelae. These concerns prompted a review of 22 newborns treated for respiratory failure with ECMO over an 18month period.
Material and Methods
Perfusion Circuit For the past five years, ECMO has been utilized at St. Louis University Medical Center for support in patients with cardiogenic shock, most of whom were recovering from a cardiac procedure [l].This experience included both adults and children, and in several instances, the technique acted as a "bridge" to cardiac transplantation. Encouraged by this experience, the medical center initiated ECMO for newborn respiratory failure in January, 1984. The perfusion circuit is very similar to that used by Bartlett and co-workers [2], and consists of a roller pump (Picker International, Ossining, NY), a membrane lung, and a heat exchanger (both SciMed, Minneapolis, MN), all of which are connected by polyethylene tubing. The roller pump is servo controlled by a pressure-sensitive switch that is in contact with the Silastic venous reservoir and interrupts the roller pump when venous filling of the reservoir is slowed. The membrane oxygenator is connected to gas sources (02and CO2) through a gas mixer, which allows precise, appropriate changes in gas supply. The heat exchanger is connected to a circulating liquid heater (Gaymar Industries, Orchard Park, NY), generally set at 38°C. The entire apparatus, including the activated clotting time (ACT) monitor (Hemochron, Metuchen, NJ), is mounted on a cart that is taken to the patient's bedside for cannulation and perfusion. The cannula for venous cannulation is either a 12F or 16F chest tube or a standard 14F cardiopulmonary cannula, and it is placed through the right jugular vein into the right atrium. The arterial cannula is either a 10F chest tube or 8F plastic pediatric feeding tube, placed through
530 The Annals of Thoracic Surgery Vol 42 No 5 November 1986
the right carotid artery into the aortic arch. The position of the cannulas is confirmed by roentgenography prior to initiation of perfusion if the clinical status of the patient allows such a delay. Each vessel (jugular vein and carotid artery) is ligated distally, and no attempt is made to reestablish continuity after cessation of ECMO. The cannulation procedure is performed with complete aseptic technique in the intensive care unit. The cannulas exit through the neck incision, which is closed loosely with skin sutures. An antibiotic-soaked gauze is applied to the incision and is changed daily in a sterile fashion. Removal of the cannulas is likewise performed at the patient's bedside. Just prior to cannulation, the baby is heparinized intravenously (150 units per kilogram of body weight), and the ACT is maintained at 180 to 200 seconds with a continuous heparin drip. The bypass flow rate must be sufficient to completely support and perfuse the infant. The normally contoured arterial pattern flattens as the flow rate is increased. When the arterial pressure tracing flattens to the point that there are no peaked contours, then approximately 70 to 80% of cardiac output is going into the ECMO circuit. This generally occurs at a flow rate of 140 to 160 ml/kg/min, but is dependent on the state of hydration at the time ECMO is initiated. Mainte-
Fig 1. Clinical profile of a typical newborn with meconium aspiration. Increasing ventilatory requirements (inspired Ozfraction [FiO,], peak ventilatory pressure, and ventilator rate) combined with failing arterial oxygenation (Pa02)prompted the use of extracorporeal membrane oxygenation (ECMO) at 48 hours of age. Lung "rest" allowed rapid pulmonary improvement and gradual weaning from ECMO, which was terminated at 120 hours of age. (CPAP = continuous positive airway pressure.) Birth Meconium
nance volumes of total parenteral nutrition are added to the ECMO circuit, as the gastrointestinal tract is not functional while the baby is on ECMO. Hemolysis is not usually a problem with ECMO. Very high flow rates through a small arterial cannula or perfusion pressures exceeding 300 mm Hg have produced little or no clinically evident hemolysis. As lung function improves, the ECMO flow rate is decreased and the ventilator settings are increased to allow lung function to resume. The impending improvement is frequently heralded by a change in chest radiograph from opacification to aeration within lung fields. When the flow rate reaches 60 to 80 ml/min, a test period is attempted with the baby completely off ECMO. If the blood gases are adequate, the cannulas are removed and ECMO is terminated. A typical ECMO profile is shown in Figure 1. Clinical Material In the 18-month period, 36 inquiries were received concerning patients who might be candidates for ECMO. From this pool, 22 were found to satisfy the criteria set for ECMO at our institution and approved by the Institutional Review Board, St. Louis University Medical Center. The criteria are based on data published by other centers with established newborn ECMO programs [371, and identify a population with a 90% or greater risk for death related to respiratory failure. These criteria are as follows: Weight > 1,800 gm Gestational age > 35 weeks Cranial ultrasound negative for intracranial hemorrhage Cardiac evaluation (ultrasound) confirming cardiac lesion is not the cause of respiratory distress
Bilateral Dopaminel --Removed +From ECMO
Vent. Rate
CPAP Extubated
1 4
Peak Vent. Pressure
FiO,
150
ECMo 100 Flow cclKglmin 50 24
48
R
96
120
Hours
144
200 212
531
Weber, Pennington, Connors, et al: Extracorporeal Membrane Oxygenation for Newborns
Diagnosis Meconium aspiration Persistent fetal circulation Ventilator barotrauma (pneumothorax, pneumomediastinum, pulmonary interstitial emphysema) Diaphragmatic hernia Level of hypoxia (any one of the following with maximal ventilator and pharmacological therapy) Acute deterioration: arterial 0 2 tension < 40 mm Hg, pH < 7.15, cardiopulmonary resuscitation P(A-a)02 > 580 mm Hg for 4 hours, or rapidly increasing* P(A-a)02 > 620 mm Hg for 1 hour* Of the 22 newborns treated with ECMO, 16 were male and 6 were female. They ranged from 11 to 298 hours old (mean, 62.5 hours) and weighed from 1,990 to 4,100 gm (mean, 3,100 gm). The diagnoses were meconium aspiration (12 patients), diaphragmatic hernia (4), persistent fetal circulation (3), hyaline membrane disease (2), and Rh incompatibility (1). Ten of the 12 newborns with meconium aspiration and 1 of the 2 infants with hyaline membrane disease also had pulmonary hypertension and a fetal circulatory pattern that contributed to the hypoxia. Two infants had a cardiac defect in addition to meconium aspiration-tetralogy of Fallot in 1 and double-outlet right ventricle in the other. In neither patient did the cardiac defect complicate the ECMO therapy. All 22 newborns had a P(A - a)O2greater than 580 mm Hg, and 15 had a gradient greater than 600 mm Hg. All patients were treated with maximal therapy prior to the institution of ECMO. This consisted of rapid ventilation (60 to 120 breaths per minute), frequently with high peak inspiratory pressures (30 to 50 mm Hg), and inotropic (dopamine hydrochloride, Isuprel [isoproterenol hydrochloride]) and pulmonary vasodilator (tolazoline hydrochloride) medications. All patients were considered to be failures of medical therapy as judged by the attending neonatologist. Seven of the infants were transferred by air transport (fixedwing and helicopter), and the remainder were in the newborn intensive care unit at Cardinal Glennon Children's Hospital. Six of the 22 were undergoing cardiopulmonary resuscitation, including external cardiac massage, at the time of ECMO cannulation. After discharge from the hospital, the infants are followed by a multidisciplinaryteam of pediatric surgeons, neonatologists, and pediatric neurologists at frequent intervals. When each child is 1 year of age, the Bayley Developmental examination is administered to assess achievement of developmental milestones. Neurological deficits, developmental abnormalities, seizure disorders, and chronic lung diseases are followed closely and treated appropriately.
0)
3
In1
4:h 20
"
R DS
UA
PFC
DH
Diagnosis
Fig 2 . Age at which extracorporeal membrane oxygenation (ECMO) was initiated for each diagnosis. (PFC = persistent fetal circulation; MA = meconium aspiration; RDS = respiratory distress syndrome; DH = diaphragmatic hernia.)
Results The mean age at which ECMO was initiated for each diagnostic category is shown in Figure 2. In general, the oldest patients were treated early in the series, but as the experience grew, newborns were referred earlier for consideration for ECMO. The duration of ECMO therapy was 41 to 310 hours (mean, 134.5 hours). The longest runs were in infants with diaphragmatic hernia, and the shortest runs were in infants with persistent fetal circulation unrelated to meconium aspiration (Fig 3). When age at onset of ECMO is examined versus the duration of ECMO, no direct correlation is found. However, all of the survivors had ECMO durations of 160 hours or less, and the nonsurvivors had longer ECMO runs (Fig 4). Nineteen of the 22 patients survived the ECMO treatment. The 3 deaths during ECMO were due to massive cerebral hemorrhage (1 infant) and irreversible lung dis-
"OI 200
8 120
I
40
PFC
1 IDS
Diagnosis 'P(A-a)O, = 760 - (PaOz + PaCO, + 47) where P(A-a)Oz = alveolar-arterial O2pressure difference, Pa02 = arterial 0 2 tension, and PaCOz = arterial COz tension [7].
Fig 3 . Duration of extracorporeal membrane oxygenation (ECMO) for each diagnosis. (PFC = persistent fetal circulation; MA = meconium aspiration; RDS = respiratory distress syndrome; DH = diaphragmatic hernia.)
532 The Annals of Thoracic Surgery Vol 42 No 5 November 1986
0 Survivors
0
Non-survivors
2240 80]0
-99
40-
0
-
Fig 4. Duration of extracorporeal membrane oxygenation (ECMO) versus age when it was initiated. A distinct division (broken line) between survivors and nonsurvivors occurred at 140 to 160 hours of ECMO. There was no direct correlation between duration of ECMO and age at which it was initiated.
order (2 patients). These latter 2 newborns had ECMO durations of 220 and 230 hours, which indicated severe lung disease that was unresponsive to even the most aggressive therapy. Four other infants died 6 to 40 days after ECMO. Two of these infants had repair of diaphragmatic hernia and were placed on ECMO at 6 and 60 hours after repair. They died of pulmonary hypoplasia and Pseudomonas sp. pneumonia, respectively. In the third infant, massive cerebral edema developed and brain death occurred 48 hours after he was removed from ECMO. His pulmonary function had recovered fully. The fourth newborn died of pulmonary and hepatorenal failure related to the original diagnosis (immune hydrops from Rh incompatibility). Thus, the overall survival is 15 (68%) of 22. Patient complications included the following: seizures (8 patients, 6 survivors), renal failure (5 patients, 4 survivors), bleeding (4 patients, 3 survivors), and neurological deficit (2 severe, 1 mild). None proved to be significant determinants of survival. The seizures were successfully treated in each instance with intravenous administration of phenobarbital sodium. Bleeding from the neck wound was managed by exploration and cauterization of bleeding vessels. Renal failure (urine output less than 1 ml/kg/hr) was treated with an ultrafiltration unit (Amicon Corporation, Danvers, MA) inserted in parallel into the ECMO circuit. Flow through the ultrafiltration unit was generally 50 to 75 mumin, thereby allowing the production of ultrafiltrate at 10 to 15 mVhr.
The malfunctions that developed with the ECMO apparatus were easily recognized and managed. None proved to be life threatening. Positive blood cultures developed in 4 patients while they were on ECMO. The organisms involved were Candida albicans (2 patients), Staphylococcus aureus (1 patient), and Pseudomonas sp. (1 patient). None of these 4 infants survived, although 1 of them lived 6 weeks before dying of Pseudomonas sp. pneumonia. In 3 patients, tubing rupture, apparatus leak, or both occurred; 2 of them survived. In 2 patients, both of whom lived, the oxygenator failed. The follow-up of the 15 long-term survivors emphasized neurological and developmental examinations. No chronic pulmonary disorders have been identified. One infant has mild neurological deficit (upper and lower extremity weakness) associated with seizures. However, developmentally the child seems normal. Two infants have severe, generalized neurological deficits that include seizures that are difficult to control. In addition, each child lacks cognitive function and is totally dependent for day-to-day care. The other 12 infants, who are 3 to 18 months old, have normal neurological examinations, and 4 more than 1 year of age have normal Bayley Developmental Examinations.
Comment The past fifteen to twenty years have seen major advances in neonatal intensive care. The improved survival in newborns, especially premature babies with pulmonary immaturity, is a direct result of improved technology. However, there continue to be several pulmonary disorders in newborns that have a high rate of death. Among these are meconium aspiration, persistent fetal circulation, and congenital diaphragmatic hernia. The high mortality has persisted despite the use of high-rate ventilation, jet ventilators, pulmonary
533 Weber, Pennington, Connors, et al: Extracorporeal Membrane Oxygenation for Newborns
vasodilators, and other measures. In fact, the high mortality in these disorders is due in part to barotrauma secondary to high-pressure ventilation, combined with excessive inspired 0 2 concentrations needed to adequately oxygenate these infants. These disorders must be treated with more aggressive maneuvers if survival is to be improved. In 1972, Hill and associates [8] treated an adult with ECMO; the patient survived. The use of ECMO in adults has been abandoned in many centers, however, because of an excessive number of complications and because of the results of a multicenter prospective randomized trial [9] from 1974 to 1979 that demonstrated no significant improvement in survival in this age group for a variety of diseases treated with ECMO. In contrast, the newborn seems to be peculiarly well suited to this form of therapy, as suggested by a growing body of data. In 1976, Bartlett and co-workers [lo] utilized ECMO for support of a newborn with meconium aspiration syndrome in whom all medical therapy was ineffective. The child survived, 1 of the first newborn survivors of ECMO. Since then, reports from a number of centers [37, 11, 141 have confirmed the usefulness of this technique. However, several controversial issues remain regarding its use. The present method of arteriovenous ECMO utilizes carotid artery ligation. Although this has not resulted in major neurological deficits, there are concerns about this invasive measure. The neurological complications that have developed in this series and in others center around intracerebral bleeding, brain edema, and seizures. The first two complications are frequently fatal, while the last (seizures) has been relatively common and does not seem to be an important determinant of survival. The short-term and long-term follow-up of infants treated with ECMO in several different centers [6,12,13] shows results similar to those presented here. The criteria used for the selection of infants who might benefit from ECMO have been unclear and vary from center to center. Most centers performing ECMO in newborns will accept patients with specific age (less than 10 to 14 days old), weight (greater than 1,800 to 2,000 gm), and neurological (normal head ultrasound) criteria, but the pulmonary severity that predicts greater than 90% mortality has been less straightforward. Attempts to use objective measures to assess candidates for ECMO, such as pulmonary insufficiency index [3] or P(A - a)O2 [7], have not been universally recognized. Our policy is to accept newborns who have received maximal medical therapy, who are referred by the attending neonatologist because of persistent or worsening hypoxia, and who satisfy the criteria listed earlier in this study. We believe this policy identifies a group of infants who have a high mortality risk (probably greater than 90%).This concept is reinforced by the fact that 6 of our infants were receiving chest massage and cardiopulmonary resuscitation at the time of cannulation. An analysis of the infants who died in this series reveals several important facts. All 4 infants in whom posi-
tive blood cultures developed died. Broad-spectrum antibiotics (gentamicin sulfate and ampicillin) are utilized during ECMO and continued for several days after termination of the procedure. However, development of resistant organisms and the emergence of organisms not covered by antibiotics (Cundidu sp.) resulted in infections that ultimately proved fatal. Strict aseptic technique during insertion of cannulas and each time the ECMO circuit is entered for blood sampling, heparin infusion, and intravenous administration of fluid or parenteral nutrition are critical. The 4 infants in whom infection developed had among the longest ECMO runs (greater than 7 days), although we and others have had numerous patients on ECMO longer than 7 days without the development of bacteremia. Longer ECMO durations mandate even closer vigilance with regard to aseptic technique. Because of the complex condition of the patients involved, the care of the infant before, during, and after ECMO is best performed by a multidisciplinaryteam. In our institution, the pediatric surgeons and pediatric cardiovascular surgeons are in charge of the ECMO programs and direct the care of the child. Neonatologists and cardiologistsare helpful in the assessment and management of respiratory and cardiac systems, and the pediatric neurologists manage seizure disorders, neurological deficits, and long-term developmental followup after the patient is discharged from the hospital. Other consultants are used as needed. The present series confirms the data presented from other centers concerning the efficacy of ECMO for respiratory failure, with satisfactory long-term neurological results. Further refinements in patient selection, aseptic management, and the multidisciplinary team approach should result in greater survival among these critically ill neonates. We thank the many referring physicians and neonatologists, the ECMO perfusion team, and the superb nursing staff in the intensive care unit of Cardinal Glennon Children’sHospital for their help in the management of these patients.
References 1. Pennington DG, Merjavy JP, Codd JE, et al: Extracorporeal membrane oxygenation for patients with cardiogenic shock. Circulation 7O:Suppl 1:130, 1984 2. Bartlett RH, Andrews AF, Toomasian JM, et al: Extracorporeal membrane oxygenation for newborn respiratory failure: forty-five cases. Surgery 92:425, 1982 3. Wetmore N, McEwen D, OConnor M, et al: Defining indications for artificial organ support in respiratory failure. Trans Am SOCArtif Intern Organs 25:459, 1979 4. Krummel TM, Greenfield LJ, Kirkpatrick BU, et al: Extracorporeal membrane oxygenation in neonatal pulmonary failure. Pediatr Ann 11:905, 1982 5. Bartlett RH, Roloff DW, Cornell RG, et al: Extracorporeal circulation in neonatal respiratory failure: a prospective randomized study. Pediatrics 76:479, 1985 6. Loe WA, Graves ED, Ochsner JL, et al: Extracorporeal
534 The Annals of Thoracic Surgery Vol 42 No 5 November 1986
membrane oxygenation for newborn respiratory failure. J Pediatr Surg 20:684, 1985 7. Krummel TM, Greenfield LJ, Kirkpatrick BU, et al: Alveolar-arterial oxygen gradients versus the neonatal pulmonary insufficiency index for prediction of mortality in ECMO candidates. J Pediatr Surg 19:380, 1984 8. Hill JD, O'Brien TG, Murray JJ, et al: Prolonged extracorporeal oxygenation for acute post-traumatic respiratory failure (shock-lung syndrome). N Engl J Med 286:629, 1972 9. Zapol WM, Snider MT, Hill JD, et al: Extracorporeal membrane oxygenation in severe acute respiratory failure: a randomized prospective study. JAMA 2422193, 1979 10. Bartlett RH, Gazzaniga AB, Jefferies R, et al: Extracorporeal membrane oxygenation (ECMO) cardiopulmonary support in infancy. Trans Am SOCArtif Intern Organs 22:80, 1976 11. Hardesty RL, Griffith BP, Debski RF, et al: Extracorporeal membrane oxygenation: successful treatment of persistent fetal circulation following repair of diaphragmatic hernia. J Thorac Cardiovasc Surg 81:556, 1981 12. Worcester CC, Huxtable RF, Rucker RW, et al: Meconium aspiration: management with extracorporeal membrane oxygenation (ECMO). Am SOCArtif Intern Organs 7:139, 1984 13. Krummel TM, Greenfield LJ, Kirkpatrick BU, et al: The early evaluation of survivors after extracorporeal membrane oxygenation for neonatal pulmonary failure. J Pediatr Surg 19:585, 1984 14. Towne BH, Lott IT, Hicks DA, et al: Long-term follow-up of infants and children treated with extracorporeal membrane oxygenation (ECMO): a preliminary report. J Pediatr Surg 20:410, 1985
Discussion (San Francisco, CA): I congratulate Dr. Weber on these extraordinarily good results. ECMO has come a long way since it was first begun in the early 1970s, and many of the technological problems from those times have largely been corrected, or at least carefully managed. This is a clinical paper that I was lucky enough to be able to review beforehand and look carefully at the data. There is no hypothesis in it to prove or disprove. The value lies in its contribution to the cumulative knowledge of clinical management and in the selection of patients for ECMO that has allowed this group to have such great success. It also adds to the work of Bartlett in Ann Arbor and the Pittsburgh group in showing that ECMO can be a very successful therapy in a selected group of neonates. I have a few comments and several questions. When one examines ECMO, there are three things that are important to patient survival. They are (1) to carefully select patients who have a reversible pulmonary lesion, (2) to always be able to control the circulation and oxygenation, and (3) to avoid complications. An important selection criterion for this neonate group was that the babies were full term and did not have low weights. The mean weight was 3,100 gm, and some weights were as high as the normal weight. Another characteristic of this group was that the vast majority of patients had single-organ disease. One cannot expect ECMO or a related therapy to resurrect patients who are dying of multiple-organ failure. This therapy must be directed at a specific problem to be successful. One difficulty that has plagued patient selection and the success of ECMO concerns the kind of respiratory characteristics DR. J. DONALD HILL
we select when trying to determine the reversibility of a pulmonary lesion. This has been a particular problem in neonates when one is attempting to achieve a 10% survival without ECMO versus a 70 to 80% survival with ECMO. Those of you who deal with pediatricians know that they are reluctant to refer these children to a high-technology therapy that they don't understand. I have a few comments on the technological problems experienced in this study. I was surprised to see that in 5 patients there were technological problems that I consider largely preventable; I am referring to leaks, tubing breaks, and oxygenator failure. This is not a criticism of this excellent group from St. Louis but rather a criticism of the technology. There has been a failure at the industrial research level to fabricate oxygenators for use in this neonate group that have an acceptable standard of reliability. To have technological problems in almost 25% of the patients is unacceptable. However, as far as I could understand from the manuscript, there were no deaths or complications related to these problems. What about complications? The complications largely involved problems of the central nervous system and failure of the lung to recover. The central nervous system problems were related to the disease itself and the tendency for these children to have cerebral hemorrhages. The incidence of hemorrhage in this group is related to the need to heparinize. This is another area where research has to be done to develop antithrombogenic surfaces to minimize this problem. I ask the authors whether they have considered using hypothermia to manage the problem of damage to the central nervous system that is possibly related to the carotid ligation and to the existing hypoxia . In management, I was surprised and pleased to see that the authors were able to reduce the inspired O2 fraction and the barrel pressure. It was not exactly clear to me whether they put the lungs completely or only partially to rest. Dr. Weber, would you give us a detailed account as to what was done with the lungs during ECMO? My experience has been that when the lungs are brought to complete rest, they turn into liver, which precipitates a hopeless situation. Thus, we always maintain some pulmonary function in terms of ventilation and keep the pulmonary artery pressures up. Dr. Weber, do you monitor these pressures during this period? What do you do if you do not get a good flow? Do you readjust your cannula or go with what you have? The results of Dr. Weber and his co-authors are excellent, and it would be very hard to improve on them with current technology. Was there a relationship between survival and the weights of the infants? Do you have a favored lesion among the four common syndromes that you presented? I again congratulate Dr. Weber on these excellent results, and thank him for the opportunity to review this manuscript in advance. DR. WEBER: Dr. Hill's pioneering efforts in this field are well known, and I consider it an honor to comment on his remarks. One aspect of ECMO that has been a bit puzzling and varies from center to center is the selection of the babies. As Dr. Hill mentioned, the neonatologists you are working with have a great deal to do with this. Many neonatologists are reluctant to refer patients for this invasive technique without first exhausting all possibilities of medical therapy. Because of our success with ECMO, many neonatologists are becoming more aggressive in terms of referring the patients. For obvious reasons, we must wait until the neonatologists specifically refer the babies for ECMO before we can institute the therapy. Dr. Bob Bartlett
535
Weber, Pennington, Connors, et al: Extracorporeal Membrane Oxygenation for Newborns
has addressed the problem of selection of patients in several prospective randomized studies, and has tried to determine more precisely when the babies should be placed on the apparatus and which babies are the best candidates for the procedure. He and others are continuing these studies at present. In answer to your question concerning hypothermia, we have not utilized that approach as yet, although it does seem attractive. I am also unaware of any other ECMO program for newborns that uses hypothermia at this time. As for the ventilator settings, we decreased the rate to 10 per minute, the peak ventilator pressure to 15 cm H 2 0 , and the inspired O2 fraction to 0.3. We don’t completely eliminate the ventilator, but rather utilize very low settings until the lungs have recovered. This generally results in opacification of the lung fields on roentgenograms for several days, but when the
roentgenogram begins improving, that is usually a signal that the baby is ready for weaning from ECMO. As for the age and weight of the patient, we have not noticed any difference in survival in regard to the patient’s weight. Bob Bartlett has noticed that the much smaller babies, with weights in the range of 1,200 to 1,400 gm, don’t seem to tolerate the procedure well and have a high incidence of intracerebral bleeding. However, to date, we have not treated any babies that small. Finally, the lesions that seem to respond best to ECMO include meconium aspiration and persistent fetal circulation. The highest mortality to date seems to be in babies with diaphragmatic hernia and severe pulmonary hypoplasia. We will continue to look for improvements in the technique that might offer survival to an increasingly large group of babies.
Important Notice for Contributors to The Annals of Thoracic Surgery Beginning with this issue, The Annals is instituting several changes in the processing of manuscripts. The most important of these are: (1) One original and two duplicate manuscripts will be required, including three separated and clearly labeled sets of illustrations. Formerly, only the original and one duplicate manuscript and two sets of illustrations were required. (2) Because of escalating postage costs, manuscripts and
illustrations routinely will not be returned to the authors. The Editorial Office will be responsible for their proper disposal so that confidentiality is preserved. Manuscripts and illustrations will be returned only if requested and a self-addressed stamped (US postage) envelope provided. There are also a number of other changes. We urge prospective contributors to read the Information for Authors sheet carefully before submitting manuscripts.