Extracorporeal Membrane Oxygenation in Pediatric Respiratory Failure
Rosa M. Ortiz, MD, Robert E. Cilley, MD, and Robert H. Bartlett, MD
Respiratory failure is one of the major medical problems of newborn infants. The leading causes are hyaline membrane disease, meconium aspiration, persistent fetal circulation syndrome, congenital diaphragmatic hernia, and sepsis. In full-term infants pulmonary hypertension with rightto-left shunting is usually the predominant pathophysiology, regardless of the primary diagnosis. The standard treatment for neonatal respiratory failure with pulmonary hypertension involves the use of mechanical ventilation, induced respiratory alkalosis, and vasodilating drugs. Most infants do well on this regimen, but a small minority-5 per cent to 10 per cent-fail to respond and die of respiratory failure. An additional 10 per cent develop chronic lung disease (bronchopulmonary dysplasia [BPD]) as a result of the pressure and oxygen used for treatment. This small group of nonresponders presents a difficult problem that has been addressed with systemic life support via extracorporeal membrane oxygenation (ECMO), which allows "lung rest" at low ventilator settings. Although ECMO has been under study for 15 years, it has been widely applied to neonatal respiratory failure only in the last 3 years. The results from several centers have been very encouraging, establishing ECM 0 as treatment of choice for specific patients who do not respond to ventilator treatment. TECHNIQUE ECMO is the process of prolonged extracorporeal circulation (cardiopulmonary bypass) achieved by extrathoracic vascular cannulation. A From the Departments of Surgery and Pediatrics, University of Michigan Medical Center, Ann Arbor, Michigan
Pediatric Clinics of North America-Vol. 34, No.1, February 1987
ORTIZ ET AL.
modified heart-lung machine is used, which consists of venous blood drainage, a servo-regulating blood pump, a membrane lung to exchange oxygen and CO2, and a heat exchanger to maintain temperature. When the decision to use ECMO has been made, and parental consent has been obtained, the heart-lung machine is prepared and primed, and the cannulation operation is begun. With the patient under local anesthesia, the right common carotid artery and the right internal jugular vein are exposed. After a loading dose of heparin (100-200 units per kg) is given, the largest possible catheters are advanced into the aortic arch and the right atrium. Because the flow resistance of the venous drainage catheter determines the maximal blood flow and therefore the ability to deliver oxygen, the venous catheter should be capable of delivering the total cardiac output (120-150 ml per kg per min) and should be in proper position to drain blood from the superior and inferior vena cava. The common carotid and the jugular vein are permanently ligated. The artery could be repaired, but even successful repair could lead to formation of small clots that could embolize to the brain. The collateral circulation to the right hemisphere is maintained by way of the external carotid and the vertebral arteries. The heart-lung machine fills by gravity drainage from the venous catheter. A servo-regulated pump is used to pump blood through the membrane lung without sucking directly on the right atrial catheter. The blood is pumped through a membrane lung. In the lung the blood is separated from the ventilating gas (02 or OiC0 2 mixture) by a thin silicone rubber membrane acros which O2 is added and CO2 and water vapor are removed. The blood leaving the membrane lung passes through a small heat exchanger to maintain temperature and returns to the patient. The priming volume is approximately 400 ml. The circuit is primed with fresh blood or packed cells and fresh frozen plasma, which has been adjusted to have physiologic acid base balance, electrolytes, and blood gases before instituting bypass. The apparatus is shown in Figure 1. MANAGEMENT OF ECMO
Establishing ECMO flow: Once cannulated, the patient is placed on bypass, gradually increasing blood flow to 100 to 120 ml per kg per min, which should support the patient. The ventilator is then set to low pressures, rate and F 10 2 (typically rate 10, pressure-20/4, F 1020.3). Patients generally remain intubated. Mild positive airway pressure is used to prevent alveolar collapse, which also serves as a safety measure if circuit components need to be changed and allows diagnostic trials off ECMO to be performed by adjusting the ventilator settings without intubating first. Pulmonary Management
The umbilical arterial and mixed venous blood gases are monitored frequently. The arterial P02 is adjusted by increasing or decreasing bypass flow through the membrane lung, and the pco2 is adjusted by the amount and type of gas flow through the membrane lung. The gas used is usually oxygen or a mixture of oxygen and carbon dioxide. In the initial period,
IN PEDIATRIC RESPIRATORY FAILURE
Figure 1. Diagram of a modified heart-lung machine.
the patient's lungs "worsen" as the inflating pressures have been decreased, so that full support is needed for the first day. Improving lung function is evidenced by improvement in blood gases during stable flow and ventilator conditions, or lung function can be assessed during a trial without ECMO, during which the ventilator settings are increased and the cannulae are clamped. The third means is to evaluate the patient daily for breath sounds, for respiratory effort and drive, and for clearing of the "white-out" on the chest x-ray. The usual time for lung recovery (off ECMO on low ventilator settings) is 4 to 7 days. Anticoagulation It is necessary to give heparin to prevent clotting in the extracorporeal circuit. Because heparin is continuously infused in low doses (20-100 units per kg per hr), it is very important to monitor the coagulation status of the patient. This is done hourly by measuring whole blood activated clotting time (ACT) rather than plasma partial thromboplastin time because of the interaction among platelets, white cells, and heparin. The heparin dose is regulated to maintain the ACT between two and three times normal. In
ORTIZ ET AL.
addition, the platelet count must also be monitored carefully. Platelet aggregates form continuously during ECMO. These platelet aggregates are then removed from circulation by the reticuloendothelial cells, resulting in thrombocytopenia. Platelet transfusions are necessary when the level drops below 50,000 per mm 3 . Because of the sensitivity of platelets to contact and temperature, it is best to transfuse a whole unit of adult platelets suspended in plasma rather than platelets that have been concentrated. Fluids and Nutrition Besides heparin, the only routine medications given are antibiotics. Once stable on bypass, the patients are started on parenteral nutrition via the circuit. Because of the significant amount of water vapor lost via the membrane lung, daily weights and electrolytes must be followed. The hematocrit is kept above 45% to maximize the oxygen-carrying capacity of blood. Technical Support Care of the patient on ECMO requires attendance of a trained ECMO technician. These technicians are nurses, respiratory therapists, and perfusionists who are trained in the classroom and in the laboratory. They work in conjunction with the patient's nurse to carry out all the patient's needs. Teamwork from neonatologists, surgeons, nurses, and technicians, as well as consultants (cardiologists and ultrasonographers) is essential. Continuous concentration is required to run without errors. Practice in the laboratory keeps the physicians' and technicians' skills current and serves to evaluate new equipment. It is the effectiveness and the commitment of the manpower involved in ECMO that make it a successful therapeutic tool.
CASE SELECTION/INDICATIONS The indication for ECMO support is acute reversible respiratory (or cardiac) failure that is unresponsive to optimal ventilator and pharmacologic management, in which the recovery of the failing cardiopulmonary system can be expected within 1 week of extracorporeal support. It is indicated in infants older than 34 weeks' gestation. Because bronchopulmonary dysplasia is an iatrogenic disorder related to barotrauma from high airway pressure and toxicity from high oxygen concentration in the newborn infant, an additional indication is the requirement for high oxygen or high airway pressure to maintain stability. We use the oxygenation index (01) to predict which patients are at risk of dying or developing BPD. The oxygenation index is calculated by dividing the product of the F 10 2 (times 100) and the mean airway pressure by the postductal Pao 2 • 7 01
(Fr02(Pawl(lOO) Pa0 2
In our institution, an 01 greater than or equal to 40 correlates with a
IN PEDIATRIC RESPIRATORY FAILURE
Table 1. Per Cent Survival by Diagnosis of All Centers MAS RDS PFC Sepsis CDR Other
33/54 40/44 11/16 19/36 3/5
83% 61% 91% 69% 53% 60%
Key: MAS = Meconium aspiration syndrome; RDS = Respiratory distress syndrome; PFC = Persistent fetal circulation; and CDR = Congenital diaphragmatic hernia.
predicted mortality risk of 80 to 90 per cent, and an 01 greater than or equal to 25 but less than 40 with a 50 to 80 per cent mortality. Each neonatal center must derive its own mortality and BPD indicators by reviewing previous experience. Contraindications to ECMO support include an intracranial hemorrhage or other brain damage. The neurologic function of infants who have been paralyzed and have been transferred from other centers is the most difficult to assess. Any intracranial bleeding or other major hemorrhage (e.g., pulmonary or GI) could potentially be exacerbated when ,heparin therapy for bypass is begun. Infants who already have severe BPD indicated by prolonged ventilation (greater than 10 days) with high pressures and oxygen concentrations should not be considered for ECMO because BPD is not reversible within 1 week of lung rest. In congenital diaphragmatic hernias, bilateral hypoplasia incompatible with normal growth and development is a contraindication. This condition is indicated by a lack of a "honeymoon period." We would consider ECMO in a congenital diaphragmatic hernia patient only if there was some evidence of adequate parenchyma (i.e., Pao2 greater than 60 mm Hg postductal at some time during management). Prematurity (less than 35 weeks gestation) or birth weight less than 2 kg is not a contraindication per se, but the incidence of intracranial hemorrhage is very high in these infants.5 Of the sixteen infants weighing less than 2 kilograms treated by our group, all but three died of intracranial hemorrhage or cerebral edema. Of the three survivors, two infants were small for dates, 35 weeks' gestation. The third has severe neurologic deficits presumed to be due to an intracranial hemorrhage. We consider the current indications for ECMO as therapy for failure to respond to conventional management as outlined above. As research, we are investigating the use of ECMO earlier in the course of the infant's illness as an alternative to maximal ventilation therapy. RESULTS Two hundred ninety-three newborn infants have been treated with ECMO since 1975 as listed with the national ECMO registry. The overall survival is 75 per cent, and this includes all the early cases at these institutions (Table 1). Our own experience now includes 100 newborn cases
ORTIZ ET AL.
implemented under a variety of protocols ranging from patients with cardiac arrest to those selected for a controlled randomized prospective study. Our overall survival is 72 per cent. These patients presented with meconium aspiration (44), RDS (26), diaphragmatic hernia (9), persistent fetal circulation (10), and others, including sepsis (11). In infants with birth weight greater than 2 kg, the overall survival is 83 per cent (69 of 83 patients). Of the last 60 patients older than 34 weeks' gestation, 59 survived. In experienced hands, the technique is safe and provides effective life support in full-term critically ill infants. Of our 72 survivors, 45 (63 per cent) are experiencing normal growth and development. 3 Twelve (17 per cent) have major neurologic dysfunction and developmental delay. This brain damage may have resulted from hypoxia or hypotension suffered before ECMO was instituted. These results are similar to the experience reported by Kirkpatrick et al., who found normal neurologic function in five of six neonatal ECM survivors.9 Six of our patients had residual lung disease when discharged from the hospital. One patient with presumed pulmonary hypoplasia, one with total anomalous pulmonary venous return and two with classic bronchopulmonary dysplasia (BPD) ultimately died of these complications. This represents a lower incidence of BPD than would be expected with maximal ventilator support in this age group and supports the hypothesis that BPD is caused by high pressure or high oxygen concentrations that are avoided by ECMO, rather than by newborn respiratory failure per se. 14 Krummel et al. have found several minor and one major neurologic sequelae but no evidence of pulmonary dysfunction on follow up of their eight survivors.12 Towne et al. recently reported 59 per cent normal growth and development of the first cohort of neonates treated with ECMO. 16 We conclude that the quality of survivors justifies the continued use of ECMO. One of the most important aspects of the ECMO experience is the finding that the newborn lung recovers from apparently lethal injury or vasospasm in virtually every case. ECMO provides passive life support and the intervention that results in recovery is "lung rest"-avoiding high pressure and FP2' This suggests that ventilator management causes or exacerbates pulmonary vasospasm, and supports the argument of Kolobow10 and Wung18 that high airway pressure should be avoided.
OUTLOOK, EXTENDED APPLICATIONS, AND POSSIBLE REFINEMENTS ECMO is life saving in lethal newborn respiratory failure and has been adopted in 18 centers as treatment for infants who fail on conventional management, but new treatment modalities should be subjected to controlled trials. Our randomized study showed that ECMO was the preferable treatment in newborns weighing over 2 kg with severe respiratory failure 4 ; however, the novel randomization method used in that study has raised questions. 13, 17 Other prospective randomized trials should be done. We are now addressing the question, "Will ECMO be cost effective when employed earlier in the course of severe respiratory failure?" in another randomized
ECMO IN PEDIATRIC RESPIRATORY FAILURE
Table 2. Pediatric Cases: Survival UCIIU ofM Others* TOTAL
4/9 44% 6/3219%
3/1127% 3/9 33%
*Others: summarized from literature--see representative cases in refs. 2, 7, 8, 10, 14.
prospective study. In this study, morbidity and cost will be investigated; we expect that all patients will survive. Our own preliminary experiences point toward possible indications in older infants and children both for support in respiratory failure (i.e., in bronchiolitis) and for cardiac support following open heart surgery or heart transplantation. Criteria for pediatric ECMO must take into account the irreversible acute fibrosis that results after about 5 days of mechanical ventilation. Fibrosis is suggested by high pulmonary vascular resistance unresponsive to local vasodilators and by extremely low lung compliance. 1 The overall survival of pediatric cases treated with ECMO reported in the literature is 26 per cent, but most of these patients were moribund when placed on ECMO. (Table 2).2,6,8,11.15 Current laboratory efforts include attempts to eliminate the need for systemic anticoagulation by coating those surfaces with heparin or other anticoagulants. Such coatings have been tested experimentally and the results demonstrate that the principle is sound. Another potential refinement is a single-cannula venovenous ECMO system using a double-lumen catheter or a tidal-flow system by way of the internal jugular vein. 19 In such a system, blood would be simultaneously withdrawn and restored to the venous circulation after oxygenation in the extracorporeal circuit. Such a system would not supply cardiac stIpport and therefore would require that the infant have normal cardiac function. SUMMARY ECMO is capable of safely supporting respiration and circulation in newborns with severe respiratory failure and a moribund clinical presentation. The results thus far suggest that term infants with respiratory failure are the best candidates for ECMO, with a survival rate of 83 per cent. Infants under 35 weeks' gestation have a very high incidence of intracranial hemorrhage. Consequently, we do not currently accept them for ECMO treatment. The outcome of the survivors is largely determined by the clinical condition before ECMO and by major complications. Research must be directed toward cost effectiveness, timing and earlier use, alternative vascular access, cannula and circuit design, and expanded indications. REFERENCES 1. Bartlett RH, Gazzaniga AB: Extracorporeal circulation for cardiopulmonary failure. Curr Prohl Surg 15: 1978
ORTIZ ET AL.
2. Bartlett RH, Cazzaniga AB et al: Extracorporeal membrane oxygenation (ECMO) in the treatment of cardiac and respiratory failure in children. Trans Am Soc Artif Intern Organs 24:578-579, 1980 3. Bartlett RH, Cazzaniga A et al: Extracorporeal membrane oxygenation (ECMO) in neonatal respiratory failure: 100 cases. Ann Surg (In Press) 4. Bartlett RH, Roloff DW, Cornell RC et al: Extracorporeal circulation in neonatal respiratory failure: A prospective randomized study. Pediatrics 76(4):479-487, 1985 5. Cilley RE, Zwischenberger JB, Andrews AF et al: Intracranial hemorrhage during extracorporeal membrane oxygenation in neonates. (Submitted for publication) 6. Cille JP: World census of long-term perfusions for respiratory support. Physiopathologic Respiratoire, 1975 7. Hallman M, Merritt A, Jarvenpaa A-L et al: Exogeneous human surfactant for treatment of severe respiratory distress syndrome: A randomized prospective clinical trial. J Pediatr 106:963-969, 1985 8. Heiden D, Mielke CH, Rodvien R et al: Platelets, hemostasis and thromboembolism during treatment of acute respiratory insufficiency with extracorporeal membrane oxygenation: Experience with 28 clinical perfusions. J Thorac Cardiovasc Surg 70(4):644-655, 1975 9. Kirkpatrick BV, Krummel TM, Mueller DC et al: Use of extracorporeal membrane oxygenation in respiratory failure in term infants. Pediatrics 72:872-876, 1983 10. Kolobow T, Morretti MP, Mascheroni D et al: Experimental meconium aspiration syndrome in the preterm fetal lamb: Successful treatment using the extracorporeal artificial lung. Trans Am Soc Artif Intern Organs 29:221-226, 1983 11. Kolobow T, Stool EW et al: Acute respiratory failure: Survival following ten days' support with a membrane lung. J Thorac Cardiovasc Surg 69(6):947-953, 1975 12. Krummel TM, Crennfield LJ, Kirkpatrick BV et al: Clinical use of an extracorporeal membrane oxygenator in neonatal pulmonary failure. J Pediatr Surg 17:525-531, 1982 13. Paneth N, Wallenstein S: Extracorporeal membrane oxygenation and the play-the-winner rule. Pediatrics 76(4):622-623, 1985 14. Phillips ACS: Oxygen plus pressure plus time: The etiology of bronchopulmonary dysplasia. Pediatrics 55:44, 1975 15. Splaingard ML, Fraizer OH et al: Extracorporeal membrane oxygenation: Its role in the survival of a child with adenoviral pneumonia and myocarditis. South Med J 76(9):1171-1173, 1983 16. Towne BH, Lott IT, Hicks DA et al: Long-term follow-up of infants and children treated with extracorporeal membrane oxygenation (ECMO). J Pediatr Surg 20:410-414, 1985 17. Ware JH, Epstein MP (Letter): Extracorporeal circulation in neonatal respiratory failure: A prospective randomized study. Pediatrics 76(5):849-851, 1985 18. Wung JT, James LS, Kilchevsky E et al: Management of infants with severe respiratory failure and persistence of the fetal circulation without hyperventilation. Pediatrics 76:488-494, 1985 19. Zwischenberger JB, Toomasian JM et al: Total respiratory support with single cannula venovenous ECMO: Double-lumen continuous How versus Single-lumen tidal How. Trans Am Soc Artif Intern Organs 31:610-615, 1985 Departments of Surgery and Pediatrics University of Michigan Medical Center Ann Arbor, Michigan 48109