Prolonged Extracorporeal Support for Nonneonatal Respiratory Failure By Thomas
R. Weber,
Thomas
F. Tracy,
Jr, Robert
Connors,
Barbara
Kountzman,
and D. Glenn Pennington
St Louis, Missouri 0 Extracorporeal membrane oxygenation (ECMO) is effective for newborns with pulmonary failure unresponsive to conventional therapy. However, ECMO for the older child and adult has been controversial and not widely utilized. Over 4 years, 24 patients (aged 4 months to 16 years; 11 boys, 13 girls) underwent venoarterial ECMO (duration, 7 to 19 days) for respiratory failure. The diagnoses were: viral pneumonia (7). hydrocarbon aspiration (6). sepsis with adult respiratory distress syndrome (ARDS) (2). bacterial pneumonitis (2). tracheal stenosis (1), bilateral pulmonary contusion (1). diaphragmatic hernia with ARDS (I), ketoacidosis with ARDS (I), pulmonary artery injection of hydrocarbon (I), drowning (1). and epiglottis with barotrauma (1). Pre-ECMO blood gas ranges (and means) were PO* 16 to 65 (46). and PCOz 47 to 112 (65). Nineteen patients received dopamine, dobutamine, or other inotrope for associated cardiac and/or renal failure. Cannulation for ECMO was through neck or groin vessels in 17, and sternotomy in 7. ECMO flow rates were 150 to 250 mL/kg/min, to maintain PO2 >lOO and PCOl ~40. Nine patients (41%) survived ECMO, with eight long-term survivors, (4 hydrocarbon aspiration or injection, 1 pulmonary contusion, 1 viral pneumonia, 1 ARDS, 1 barotrauma), three of whom have mild neurological deficit. All patients with sternotomy, and 6 of 15 with neck and/or groin cannulation, required 1 to 5 explorations for hemorrhage while on ECMO. All survivors had primarily pulmonary failure; patients with combinations of pulmonary, cardiac, and renal failure did not survive. ECMO can be life-saving in the child with isolated pulmonary failure, but its efficacy in patients with multiorgan failure is uncertain. Copyright o 1992 by W.B. Saunders Company
for these diseases in the term newborn have undoubtedly improved over the past 8 to 10 years since ECMO has become available in a large number of neonatal centers. In contrast to these improvements in survival in neonatal respiratory failure, the outlook for the older child with respiratory failure, from a variety of causes, remains poor. Mortality rates of 60% or greater have been reportedlT* and have remained unchanged for several years. Thus, offering ECMO to this group of patients who fail conventional therapy seems worthwhile. We report herein our initial experience with ECMO for respiratory failure in the pediatric patient, unresponsive to conventional therapy. MATERIALS
AND METHODS
From the Divisions of Pediam’c Surgery and Cardiovascular Surgery, Depanment of Surgery, St Louis University School of Medicine and Cardinal Glennon Children’s Hospital, St Louis, MO. Presented at the 43rd Annual Meeting of the Surgical Section of the American Academy of Pediatrics, New Orleans, Louisiana, October 26-27, 1991. Address reprint requests to Thomas R. Weber, MD, Cardinal Glennon Children’s Hospital, 1465 S Grand Blvd, St Louis, MO 63104. Copyrtght o 1992 by WB. Saunders Company 0022-3468/92/2708-0033$03.00/O
Over a 4-year period, 32 infants and children (aged 4 weeks to 16 years) were evaluated for the use of ECMO for respiratory failure. Eight children were felt to be noncandidates for ECMO, primarily because of coexisting severe neurological deficit that was felt to be permanent, or non-reversible pulmonary disease. Twenty-four of the children were placed on ECMO, as described below. Specifically excluded from this review were patients in whom ECMO was used in the immediate postoperative period after operation for congenital heart disease. The primary diagnosis of these 24 patients are shown in Table 1. Infections (viral or bacterial pneumonia, sepsis with adult respiratory distress syndrome [ARDS]) accounted for respiratory failure in 11 of the 24 patients. Aspiration of various hydrocarbon substances occurred in 6 patients, whereas a variety of other causes of respiratory failure were present in the remaining patients. Of particular interest is a l&year-old girl who had a cleaning solution mistakenly injected into her pulmonary artery during cardiac catheterization, who developed immediate respiratory failure, prompting placement and air transport on cardiopulmonary bypass. All children were treated with conventional therapy including a ventilator (24 patients, 6 with high-frequency jet ventilation), with variable settings. Most patients required high end-expiratoty (6 to 25 cm H20) and peak inspiratory (20 to 82 cm HaO) pressures, and 16 had continuous air leaks through tube thoracostomy tubes, related to barotrauma. Dopamine, dobutamine, or epinephrine (19 patients), bronchodilators (20 patients), systemic (13 patients) or inhalation (4 patients) corticosteroids, and Ribaviron for respiratoIy syncytial virus (RSV) pneumonia (4 patients) were also used. As might be expected, chest radiographs just prior to the use of ECMO were uniformly significantly abnormal (Figs 1 and 2). It should be noted that 13 of the 24 patients were referred from remote states, and required fixed-wing air transport to our facility for ECMO. Blood gas determinations just prior to ECMO showed PO* mean 46 mm Hg (range, 18 to 65) and PC02 mean 65 mm Hg (range, 47 to 112). Cardiac arrest had occurred prior to ECMO in 9 patients. Venoarterial ECMO was initiated in this series using basic tech-
1100
JournalofPediatric Surgery, Vol27, No 8 (August),1992: pp 1100-l105
INDEX WORDS: Pulmonary brane oxygenation (ECMO).
failure;
extracorporeal
mem-
E
XTRACORPOREAL membrane oxygenation (ECMO) is an accepted means of life support for neonates with specific causes of respiratory failure. Persistent pulmonary hypertension, meconium aspiration, hyaline membrane disease, and congenital diaphragmatic hernia are examples of neonatal disorders that have responded well to this form of support, with excellent long-term survival. The mortality rates
PROLONGED ECMO FOR RESPIRATORY FAILURE
1101
Table 1. Diagnosis Leading to Pulmonary Failure No. of Diagnosis
Patients
Viral pneumonia Respiratory syncytial virus
4
Adenovirus
2
Herpes virus
1
Hydrocarbon aspiration Mineral oil
1
Furniture polish
1
Cleaning fluid
1
Xylocaineicharcoal
1
Lamp oil
1
Kerosene
1
SepsisiARDS
2
Bacterial pneumonia Klebsiella
1
Pseudomonas
1
Congenital tracheal stenosis
1
Congenital diaphragmatic hernia/ARDS Ketoacidosis/ARDS
1
Pulmonary artery injection of hydrocarbon Drowning
1
1 1
Epiglottisibarotrauma
1
Bilaferal pulmonary contusion
1
Prbbreviation:
ARDS, adult respiratory distress syndrome.
previously described.3,4 Intravenous and intraarterial cannulae were placed in either neck vessels (internal jugular vein, carotid artery), groin vessels (femoral artery and vein), or a combination of these in 17 patients, or via median sternotomy in 7 patients. Systemic heparinization was used to maintain activated clotting time at 180 to 220 seconds. ECMO flow rates were adjusted to 150 to 250 mL/kg/min, to achieve arterial PO2 levels of 80 to 100 mm Hg and PC02 of 30 to 40 mm Hg. After initiation of ECMO, niques
as
Fig 1. Chest radiograph of a 12-year-old patient who had sustained inadvertent injection of a cleaning solution into her pulmonary artery during cardiac catheterization. Note significant right lung injury and severe hyperinflation of the left lung.
Fig 2. Chest radiograph of a 2-year-old child with hydrocarbon aspiration, showing multiple tube thoracostomies. diffuse pulmonary injury, and the jugular vein ECMO cannula.
the ventilator pressure and inspired oxygen levels were reduced significantly to eliminate air leak and continuing barotrauma or oxygen toxicity. As pulmonary function improved, the ventilator settings were gradually increased and the ECMO flow rate weaned. Clinical improvement may or may not have correlated with radiologic improvement at this point (Figs 3 and 4). Blood, plasma, and platelet transfusions were used as needed to maintain these levels at or near normal range. ECMO was terminated when: (1) there was sufficient pulmonary recovery to allow management with conventional ventilator; (2)
Fig 3. Same patient as in Fig 1. Note marked improvement hours of ECMO.
after 72
WEBER ET AL
Fig 4. Same patient as in Fig 2. After 10 days of ECMO, followed by 14 more days of ventilator therapy, the chest radiograph is markedly improved.
after suffering respiratory arrest secondary to airway occlusion, 3 months after ECMO treatment. Among the long-term survivors, three have neurological deficits that were not focal and unrelated to intracerebral hemorrhage. These three children, as well as the remaining five survivors have no evidence of permanent pulmonary sequelae, although two continue to require periodic bronchodilator therapy. Associated organ failure was almost nonexistent in the survivors. The characteristics of the patients who died are shown in Table 3. Ten of the 15 patients had infections, either pneumonia or systemic as a contributing factor in their respiratory failure. In addition, 11 of the 15 patients had associated other organ failure that was either the direct cause, or contributed significantly to their death. Attempts to manage other organ failure, ie, dialysis for renal failure or inotropic agents for cardiac failure, were invariably unsuccessful in these patients. DISCUSSION
irreversible brain damage was determined to be present; (3) the pulmonary disorder appeared to be irreversible, usually after 2 to 3 weeks of ECMO without significant improvement: or (4) progressive multiple organ (cardiac, liver, renal) failure developed. RESULTS
Nine of the 24 treated patients (41%) recovered sufficient pulmonary function to allow weaning from ECMO. The survivors required ECMO for 64 to 323 hours (mean, 179.8 hours), whereas nonsurvivors received ECMO for 20 to 612 hours (mean, 208.8 hours). All patients with sternotomy, and 8 of 15 with neck or groin cannulas required 1 to 5 explorations for bleeding during ECMO. There was no correlation between the presence of hemorrhage during ECMO and ultimate survival. Four of the nine survivors underwent placement of cannulas through sternotomy approach. The characteristics of the survivors are shown in Table 2. Eight of the 9 patients successfully weaned from ECMO are long-term survivors, while one patient died of cerebral hypoxia and seizure disorder Table 2. Characteristics
ECMO is now an acceptable form of life support for the neonate with respiratory failure, and a number of series, including several prospective, randomized trials, have demonstrated its usefulness in infants who are unresponsive to conventional therapy. Longterm outlook, both pulmonary and neurological, also appears to be very satisfactory as well. However, the use of ECMO for support of the nonneonate with respiratory failure has not been as widely used. This lack of enthusiasm for ECMO in the older child is probably partly due to a randomized trial of ECMO in young adults,5 which showed no increase in survival over conventional therapy, and also early experience with ECMO in the postcardiac surgery patient, where survival rates were minimal. Because of the extremely limited experience with ECMO in older children thus far, questions persist regarding suitable candidates for the procedure, treatable diagnoses, duration of treatment, optimal methods of cannulation, and short- and long-term followup, just to mention a few. These and other questions
of Nine Survivors of Nonneonatal Pre-ECMO pHIPCO~IP0~
ECMO
AssociatedOrgan Failure
Case NO.
Age
1
16 mo
Mineral oil aspiration
7.51143133
None
Moderate neurological deficit
2
5mo
RSV pneumonia
7.44/50/43
None
Mild neurological deficit
Tracheal stenosis
7.10/91/55
None
Died of cerebral hypoxia at 6 mo
Furniture polish aspiration
7.39142134
None
Normal
Bilateral pulmonary contusion
7.26153147
None
Normal
3
3mo
4
16 mo
5 6
6 vr 8mo
7
13 mo
Diagnosis
outcome
CDHIARDS
7.04171 I67
None
Mild neurological deficit
Epiglottislbarotrauma
7.29156182
None
Normal
8
12yr
Pulmonary artery injection of disinfectant
7.24146126
Mild cardiac
Normal
9
28 mo
Xylocaine/charcoal
7.171801120
None
Normal
aspiration
Abbreviations: RSV, respiratory syncytial virus: CDH, congenital diaphragmatic hernia; ARDS, adult respiratory distress syndrome.
PROLONGED ECMO FOR RESPIRATORY FAILURE
1103
Table 3. Characteristics Case
of 15 Nonsurvivors of Nonneonatal
Pre-ECMO
ECMO
Organ
NO.
Age
1
5mo
SepsislARDS
7.31125139
Cardiac, renal
Multiorgan failure
2
2mo
RSV pneumonia
7.09166134
Cardiac
Cardiac failure
SepsislARDS
7.39/483/57
Cardiac, renal
Multiorgan failure
Klebsiella pneumonia
7.32149126
Renal
Irreversible pulmonary disease
Herpes pneumonia
7.13/54/85
None
Irreversible pulmonary disease
3 4 5
14yr 3 vr 1 mo
Diagnosis
Associated
pH/PCOZ/POz
Failure
Cause of Death
Pseudomonas pneumonia
7.20/47/21
Cardiac, liver, renal
Multiorgan failure
7
16yr
Adenovirus pneumonia
7.27167160
None
Irreversible pulmonary disease
8
10mo
Cleaning solution aspiration
7.02/71/37
Renal, cardiac, liver
Multiorgan failure
9
14yr
Ketoacidosis/ARDS
7.23/56/60
Cardiac, renal
Multiorgan failure
7.56137124
Cardiac
Cardiac failure
None
Irreversible pulmonary disease
6
2.5 mo
10
3mo
RSV pneumonia
11
9mo
Adenovirus pneumonia
12
2.5 y’
Lamp oil aspiration
731i58140
None
Irreversible pulmonary disease
13
4.5 yr
Kerosene aspiration
7.20129177
Liver, renal, cardiac, cerebral
Multiorgan failure
14 15
2 vr 2mo
Drowning
7.15/88/56
Cardiac, renal
Multiorgan failure
RSV pneumonia
7.28167133
Cardiac, cerebral
Multiorgan failure
6.91125154
Abbreviations: ARDS, adult respiratory distress syndrome; RSV, respiratory syncytial virus.
have led to several pleas for a randomized, controlled trial of nonneonatal ECMO. v In spite of these important questions, there are a number of survivors of ECMO for nonneonatal respiratory failure, most (if not all) of whom probably would have died without ECMO.%r-1” Analysis of the present series has shown several important points. Any program offering ECMO for older children will be presented with an extremely wide variety of patients to evaluate for this form of therapy. This series included patients with such diverse diagnoses as pulmonary infections, aspiration of various substances, ARDS associated with other disorders, pulmonary contusion, and barotrauma. Unfortunately, at the present time we have not accumulated enough patients in any category to make definitive statements regarding the prognosis of patients with specific diagnoses. However, it is quite clear from this series that patients with isolated pulmonary failure, in whom other organs (brain, kidney, heart, liver) are functioning normally, stand the best chance for survival. This would appear to exclude at present the patient with multiple organ failure associated with sepsis, patients who have suffered sufficient hypoxia failure associated with sepsis, and patients who have suffered sufficient hypoxia to cause infarct or severely compromised function in organs other than the lung. However, assessing such organ damage prior to ECMO is difficult, and at present we tend to offer ECMO to virtually all children with apparently reversible pulmonary disorder, but warn the families that additional organ failure, if it should become manifest or develop during ECMO may be a cause for discontinuing therapy. The length of time ECMO is required in the older
patient can be considerable. Six of the nine survivors required ECMO for over 7 days, with two of these requiring more than 10 days of ECMO. These are generally greater times than required by most neonates treated with ECMO, suggesting a more severe pulmonary parenchymal injury. Other reported pediatric survivors have also required prolonged ECMO treatments.7J0 Patients in the present series have been successfully treated with cannulas placed in the neck or groin vessels, or a combination of these sites, or directly into the right atrium and aorta through a stemotomy approach. Although sternotomy results in significant hemorrhage during ECMO, necessitating reexploration in each case, this approach allows the placement of large cannulas, which facilitates ECMO flow at lower pressures. This is especially useful in larger children, in whom flow rates of 2.5 to 4.0 L/min are frequently needed. We have had no episodes of sternal infection in the four survivors of stemotomy in this series. The short-term follow-up of the survivors has been gratifying, with no significant pulmonary disorders, and only mild neurological deficits in three patients, all of whom have improved with time. Thus, it would appear that the long-term outlook for these critically ill children is excellent, an observation that has been mentioned in other series as well.7J0 Although this series is somewhat preliminary, it would appear that ECMO is now a viable alternative therapy for the older child with respiratory failure, unresponsive to conventional therapy. Future multiinstitutional studies of ECMO will hopefully resolve important questions regarding patient selection, optimal techniques of ECMO management, and longterm sequelae.
WEBER ET AL
1104
REFERENCES 1. Royal1 JA, Levitt DL: Adult respiratory distress syndrome in children. I. Clinical aspects, pathophysiology, pathology, and mechanisms of lung injury. J Pediatr 112:169-180,1988 2. Vernon DD, Dean JM, McGough EC: Pediatric extracorporeal membrane oxygenation. The time for anecdotes is over. Am J Dis Child 144:855-856,199O (editorial) 3. Weber TR, Pennington DG, Connors R, et al: Extracorporeal membrane oxygenation for newborn respiratory failure. Ann Thorat Surg 42:529-535,1986 4. Scalzo AJ, Weber TR, Jaeger RW, et al: Extracorporeal membrane oxygenation for hydrocarbon aspiration. Am J Dis Child 144:867-871,199O 5. Zapol W, Snider M, Hill JD, et al: Extracorporeal membrane oxygenation in severe acute respiratory failure. JAMA 242:21932196,1979
6. O’Rourke P, Crone RK: Pediatric applications of extracorporeal membrane oxygenation. J Pediatr 116:393-3951990 (editorial) 7. Steinhorn RH, Green T: Use of extracorporeal membrane oxygenation in the treatment of respiratory syncytial virus bronchiolitis: The national experience, 1983-1988. J Pediatr 116:338-342, 1990 8. Anderson HL, Attorri R, Custer J, et al: Extracorporeal membrane oxygenation for pediatric cardiopulmonary failure. J Thorac Cardiovasc Surg 99:1011-1021,199O 9. Redmond CR, Graves ED, Falterman KW, et al: Extracorporeal membrane oxygenation for respiratory and cardiac failure in infants and children. J Thorac Cardiovasc Surg 93:199-204,1987 10. Steiner RB, Adolph VR, Bonis SL, et al: Pediatric extracorporeal membrane oxygenation in posttraumatic respiratory failure. J Pediatr Surg 26:1011-1015,199l
Discussion Jay Wilson (Boston, MA): Dr Weber and his colleagues have reported to you today a 24-patient series concerning ECMO for pediatric failure encompassing such diverse diagnoses as sepsis, hydrocarbon aspiration, pulmonary contusion, etc. The overall survival was 9 of 24 patients (41%), which is comparable to the collective national experience maintained in the ELSO registry. This first slide, which I obtained from the ELSO registry, shows the national survival statistics from the ELSO registry. It shows that diagnoses are diverse and that survival is variable, but the overall survival is about 48%. In the current report, which is a large series of pediatric ECMO, the authors note that because of the diverse nature of the diagnoses, no definitive conclusions can be reached regarding prognosis of individual patients or even who is a suitable ECMO candidate. They are not alone in this quandary because, unfortunately, this is the state of pediatric ECMO today. The second slide, stratified by year, shows that although our experience with pediatric ECMO was growing immensely, our ability to determine who will survive the course of pediatric ECMO is not. This speaks to the need for a prospective randomized national trial. The unique contribution of this report, however, appears to be the finding that patients with isolated pulmonary failure, irrespective of diagnosis, have a higher survival, 75% in this series, than those with multiorgan system failure. Since I think this is the take-home message of this report, and because I believe patient selection is critical, I have a few questions for the authors. Eight patients were excluded from ECMO in the manuscript because their disease was felt to be
nonreversible. The question I have is how do you determine this? And have you used lung biopsies to help you in this decision? Second, you noted that survival was best when only pulmonary disease was present. However, 19 of 24 patients required pressure agents prior to ECMO, yet only one survivor in the manuscript was noted to have mild cardiac failure. What were your criteria for defining cardiac failure in this group of patients? Among the long-term survivors, three patients had neurological deficits deemed to be nonfocal. In several other patients, ECMO was terminated secondary to irreversible brain damage. What method was used to determine irreversible brain damage? Was CT scan used routinely to follow up with these patients? What methods of cannulation were used in the patients who had neurological deficits? And were any of those deficits attributable to the cannulations? Finally, on the basis of your report, would you be prepared at this time to exclude patients in your institution with multiorgan system failure from ECMO? I want to congratulate the authors on this report. I believe their findings about isolated pulmonary disease being the best prognosis is real, and, given our current lack of outcome predictors, important. Furthermore, until such time as a prospective, randomized trial can be organized, a large, carefully analyzed, single institution study such as this contributes significantly to the aggregate knowledge of this very complex topic. Arnold Coran (Ann Arbor, MI): When I left Ann Arbor Friday, I got the latest update on the ELSO registry, both for the University of Michigan and for the international registry, and as of that date we have treated 62 pediatric patients with ECMO, 28 for
PROLONGED
ECMO
FOR RESPIRATORY
FAILURE
respiratory failure and the others for cardiac failure, with a survival rate in the respiratory failure group of 57%, and a survival rate in the cardiac failure group of 38%. At this point, with the international registry, we have 285 patients with respiratory failure and 494 with cardiac failure with very similar survival rates. Our experience with the pediatric patient with respiratory failure on ECMO, is that you get to a point where you are not sure when to get them off the ECMO, and that occurs in 4 to 6 weeks. That is, they just can’t be weaned off. I would like to know what criteria you are using for continuing or stopping ECMO. One of the factors that appears to be important in weaning from the ECMO is the development of pulmonary fibrosis, especially in the hydrocarboningestion patients. Are you giving them anything during the course of ECMO to try to stop that fibrotic process? Robert Foglia (St Louis, MO): Your patients were on a ventilator 3 to 14 days. Could you stratify those patients in terms of survival and how long they were on a ventilator? Was survival improved in those patients who were on a ventilator for a shorter period of time pre-ECMO? T.R. Weber (response): In terms of how long pa-
1105
tients were on a ventilator prior to ECMO, we have several patients who were ventilated longer than 10 days. It’s not reasonable in a small group of patients to try to draw too many conclusions in that regard, however. I think both Dr Wilson and Dr Coran bring up some very important points; that is, determining which patients have reversible disease and how long should they remain on ECMO? Those are very difficult questions, and obviously, with a small series like this, we’re not going to have the answers. Our policy has been to leave these children on for 2 weeks and then to attempt to reassess their status. Occasionally, this must be done with lung biopsy. Sometimes the parents help to make the decision for you. We obviously try to choose patients preoperatively who we think have reversible disease. We have not noticed any difference in terms of site of cannulation and ultimate outcome, particularly with regard to neurological deficit. I think that Dr Wilson was questioning the possibility of carotid ligation in the older child causing neurological deficit. We have used carotid ligation in this group with survival, and without significant or noticeable neurological defect. We have used our neurologist to determine when a child has developed neurological deficits, rather than relying on CAT scan in this setting.
R. Weber,
Thomas
F. Tracy,
Jr, Robert
Connors,
Barbara
Kountzman,
and D. Glenn Pennington
St Louis, Missouri 0 Extracorporeal membrane oxygenation (ECMO) is effective for newborns with pulmonary failure unresponsive to conventional therapy. However, ECMO for the older child and adult has been controversial and not widely utilized. Over 4 years, 24 patients (aged 4 months to 16 years; 11 boys, 13 girls) underwent venoarterial ECMO (duration, 7 to 19 days) for respiratory failure. The diagnoses were: viral pneumonia (7). hydrocarbon aspiration (6). sepsis with adult respiratory distress syndrome (ARDS) (2). bacterial pneumonitis (2). tracheal stenosis (1), bilateral pulmonary contusion (1). diaphragmatic hernia with ARDS (I), ketoacidosis with ARDS (I), pulmonary artery injection of hydrocarbon (I), drowning (1). and epiglottis with barotrauma (1). Pre-ECMO blood gas ranges (and means) were PO* 16 to 65 (46). and PCOz 47 to 112 (65). Nineteen patients received dopamine, dobutamine, or other inotrope for associated cardiac and/or renal failure. Cannulation for ECMO was through neck or groin vessels in 17, and sternotomy in 7. ECMO flow rates were 150 to 250 mL/kg/min, to maintain PO2 >lOO and PCOl ~40. Nine patients (41%) survived ECMO, with eight long-term survivors, (4 hydrocarbon aspiration or injection, 1 pulmonary contusion, 1 viral pneumonia, 1 ARDS, 1 barotrauma), three of whom have mild neurological deficit. All patients with sternotomy, and 6 of 15 with neck and/or groin cannulation, required 1 to 5 explorations for hemorrhage while on ECMO. All survivors had primarily pulmonary failure; patients with combinations of pulmonary, cardiac, and renal failure did not survive. ECMO can be life-saving in the child with isolated pulmonary failure, but its efficacy in patients with multiorgan failure is uncertain. Copyright o 1992 by W.B. Saunders Company
for these diseases in the term newborn have undoubtedly improved over the past 8 to 10 years since ECMO has become available in a large number of neonatal centers. In contrast to these improvements in survival in neonatal respiratory failure, the outlook for the older child with respiratory failure, from a variety of causes, remains poor. Mortality rates of 60% or greater have been reportedlT* and have remained unchanged for several years. Thus, offering ECMO to this group of patients who fail conventional therapy seems worthwhile. We report herein our initial experience with ECMO for respiratory failure in the pediatric patient, unresponsive to conventional therapy. MATERIALS
AND METHODS
From the Divisions of Pediam’c Surgery and Cardiovascular Surgery, Depanment of Surgery, St Louis University School of Medicine and Cardinal Glennon Children’s Hospital, St Louis, MO. Presented at the 43rd Annual Meeting of the Surgical Section of the American Academy of Pediatrics, New Orleans, Louisiana, October 26-27, 1991. Address reprint requests to Thomas R. Weber, MD, Cardinal Glennon Children’s Hospital, 1465 S Grand Blvd, St Louis, MO 63104. Copyrtght o 1992 by WB. Saunders Company 0022-3468/92/2708-0033$03.00/O
Over a 4-year period, 32 infants and children (aged 4 weeks to 16 years) were evaluated for the use of ECMO for respiratory failure. Eight children were felt to be noncandidates for ECMO, primarily because of coexisting severe neurological deficit that was felt to be permanent, or non-reversible pulmonary disease. Twenty-four of the children were placed on ECMO, as described below. Specifically excluded from this review were patients in whom ECMO was used in the immediate postoperative period after operation for congenital heart disease. The primary diagnosis of these 24 patients are shown in Table 1. Infections (viral or bacterial pneumonia, sepsis with adult respiratory distress syndrome [ARDS]) accounted for respiratory failure in 11 of the 24 patients. Aspiration of various hydrocarbon substances occurred in 6 patients, whereas a variety of other causes of respiratory failure were present in the remaining patients. Of particular interest is a l&year-old girl who had a cleaning solution mistakenly injected into her pulmonary artery during cardiac catheterization, who developed immediate respiratory failure, prompting placement and air transport on cardiopulmonary bypass. All children were treated with conventional therapy including a ventilator (24 patients, 6 with high-frequency jet ventilation), with variable settings. Most patients required high end-expiratoty (6 to 25 cm H20) and peak inspiratory (20 to 82 cm HaO) pressures, and 16 had continuous air leaks through tube thoracostomy tubes, related to barotrauma. Dopamine, dobutamine, or epinephrine (19 patients), bronchodilators (20 patients), systemic (13 patients) or inhalation (4 patients) corticosteroids, and Ribaviron for respiratoIy syncytial virus (RSV) pneumonia (4 patients) were also used. As might be expected, chest radiographs just prior to the use of ECMO were uniformly significantly abnormal (Figs 1 and 2). It should be noted that 13 of the 24 patients were referred from remote states, and required fixed-wing air transport to our facility for ECMO. Blood gas determinations just prior to ECMO showed PO* mean 46 mm Hg (range, 18 to 65) and PC02 mean 65 mm Hg (range, 47 to 112). Cardiac arrest had occurred prior to ECMO in 9 patients. Venoarterial ECMO was initiated in this series using basic tech-
1100
JournalofPediatric Surgery, Vol27, No 8 (August),1992: pp 1100-l105
INDEX WORDS: Pulmonary brane oxygenation (ECMO).
failure;
extracorporeal
mem-
E
XTRACORPOREAL membrane oxygenation (ECMO) is an accepted means of life support for neonates with specific causes of respiratory failure. Persistent pulmonary hypertension, meconium aspiration, hyaline membrane disease, and congenital diaphragmatic hernia are examples of neonatal disorders that have responded well to this form of support, with excellent long-term survival. The mortality rates
PROLONGED ECMO FOR RESPIRATORY FAILURE
1101
Table 1. Diagnosis Leading to Pulmonary Failure No. of Diagnosis
Patients
Viral pneumonia Respiratory syncytial virus
4
Adenovirus
2
Herpes virus
1
Hydrocarbon aspiration Mineral oil
1
Furniture polish
1
Cleaning fluid
1
Xylocaineicharcoal
1
Lamp oil
1
Kerosene
1
SepsisiARDS
2
Bacterial pneumonia Klebsiella
1
Pseudomonas
1
Congenital tracheal stenosis
1
Congenital diaphragmatic hernia/ARDS Ketoacidosis/ARDS
1
Pulmonary artery injection of hydrocarbon Drowning
1
1 1
Epiglottisibarotrauma
1
Bilaferal pulmonary contusion
1
Prbbreviation:
ARDS, adult respiratory distress syndrome.
previously described.3,4 Intravenous and intraarterial cannulae were placed in either neck vessels (internal jugular vein, carotid artery), groin vessels (femoral artery and vein), or a combination of these in 17 patients, or via median sternotomy in 7 patients. Systemic heparinization was used to maintain activated clotting time at 180 to 220 seconds. ECMO flow rates were adjusted to 150 to 250 mL/kg/min, to achieve arterial PO2 levels of 80 to 100 mm Hg and PC02 of 30 to 40 mm Hg. After initiation of ECMO, niques
as
Fig 1. Chest radiograph of a 12-year-old patient who had sustained inadvertent injection of a cleaning solution into her pulmonary artery during cardiac catheterization. Note significant right lung injury and severe hyperinflation of the left lung.
Fig 2. Chest radiograph of a 2-year-old child with hydrocarbon aspiration, showing multiple tube thoracostomies. diffuse pulmonary injury, and the jugular vein ECMO cannula.
the ventilator pressure and inspired oxygen levels were reduced significantly to eliminate air leak and continuing barotrauma or oxygen toxicity. As pulmonary function improved, the ventilator settings were gradually increased and the ECMO flow rate weaned. Clinical improvement may or may not have correlated with radiologic improvement at this point (Figs 3 and 4). Blood, plasma, and platelet transfusions were used as needed to maintain these levels at or near normal range. ECMO was terminated when: (1) there was sufficient pulmonary recovery to allow management with conventional ventilator; (2)
Fig 3. Same patient as in Fig 1. Note marked improvement hours of ECMO.
after 72
WEBER ET AL
Fig 4. Same patient as in Fig 2. After 10 days of ECMO, followed by 14 more days of ventilator therapy, the chest radiograph is markedly improved.
after suffering respiratory arrest secondary to airway occlusion, 3 months after ECMO treatment. Among the long-term survivors, three have neurological deficits that were not focal and unrelated to intracerebral hemorrhage. These three children, as well as the remaining five survivors have no evidence of permanent pulmonary sequelae, although two continue to require periodic bronchodilator therapy. Associated organ failure was almost nonexistent in the survivors. The characteristics of the patients who died are shown in Table 3. Ten of the 15 patients had infections, either pneumonia or systemic as a contributing factor in their respiratory failure. In addition, 11 of the 15 patients had associated other organ failure that was either the direct cause, or contributed significantly to their death. Attempts to manage other organ failure, ie, dialysis for renal failure or inotropic agents for cardiac failure, were invariably unsuccessful in these patients. DISCUSSION
irreversible brain damage was determined to be present; (3) the pulmonary disorder appeared to be irreversible, usually after 2 to 3 weeks of ECMO without significant improvement: or (4) progressive multiple organ (cardiac, liver, renal) failure developed. RESULTS
Nine of the 24 treated patients (41%) recovered sufficient pulmonary function to allow weaning from ECMO. The survivors required ECMO for 64 to 323 hours (mean, 179.8 hours), whereas nonsurvivors received ECMO for 20 to 612 hours (mean, 208.8 hours). All patients with sternotomy, and 8 of 15 with neck or groin cannulas required 1 to 5 explorations for bleeding during ECMO. There was no correlation between the presence of hemorrhage during ECMO and ultimate survival. Four of the nine survivors underwent placement of cannulas through sternotomy approach. The characteristics of the survivors are shown in Table 2. Eight of the 9 patients successfully weaned from ECMO are long-term survivors, while one patient died of cerebral hypoxia and seizure disorder Table 2. Characteristics
ECMO is now an acceptable form of life support for the neonate with respiratory failure, and a number of series, including several prospective, randomized trials, have demonstrated its usefulness in infants who are unresponsive to conventional therapy. Longterm outlook, both pulmonary and neurological, also appears to be very satisfactory as well. However, the use of ECMO for support of the nonneonate with respiratory failure has not been as widely used. This lack of enthusiasm for ECMO in the older child is probably partly due to a randomized trial of ECMO in young adults,5 which showed no increase in survival over conventional therapy, and also early experience with ECMO in the postcardiac surgery patient, where survival rates were minimal. Because of the extremely limited experience with ECMO in older children thus far, questions persist regarding suitable candidates for the procedure, treatable diagnoses, duration of treatment, optimal methods of cannulation, and short- and long-term followup, just to mention a few. These and other questions
of Nine Survivors of Nonneonatal Pre-ECMO pHIPCO~IP0~
ECMO
AssociatedOrgan Failure
Case NO.
Age
1
16 mo
Mineral oil aspiration
7.51143133
None
Moderate neurological deficit
2
5mo
RSV pneumonia
7.44/50/43
None
Mild neurological deficit
Tracheal stenosis
7.10/91/55
None
Died of cerebral hypoxia at 6 mo
Furniture polish aspiration
7.39142134
None
Normal
Bilateral pulmonary contusion
7.26153147
None
Normal
3
3mo
4
16 mo
5 6
6 vr 8mo
7
13 mo
Diagnosis
outcome
CDHIARDS
7.04171 I67
None
Mild neurological deficit
Epiglottislbarotrauma
7.29156182
None
Normal
8
12yr
Pulmonary artery injection of disinfectant
7.24146126
Mild cardiac
Normal
9
28 mo
Xylocaine/charcoal
7.171801120
None
Normal
aspiration
Abbreviations: RSV, respiratory syncytial virus: CDH, congenital diaphragmatic hernia; ARDS, adult respiratory distress syndrome.
PROLONGED ECMO FOR RESPIRATORY FAILURE
1103
Table 3. Characteristics Case
of 15 Nonsurvivors of Nonneonatal
Pre-ECMO
ECMO
Organ
NO.
Age
1
5mo
SepsislARDS
7.31125139
Cardiac, renal
Multiorgan failure
2
2mo
RSV pneumonia
7.09166134
Cardiac
Cardiac failure
SepsislARDS
7.39/483/57
Cardiac, renal
Multiorgan failure
Klebsiella pneumonia
7.32149126
Renal
Irreversible pulmonary disease
Herpes pneumonia
7.13/54/85
None
Irreversible pulmonary disease
3 4 5
14yr 3 vr 1 mo
Diagnosis
Associated
pH/PCOZ/POz
Failure
Cause of Death
Pseudomonas pneumonia
7.20/47/21
Cardiac, liver, renal
Multiorgan failure
7
16yr
Adenovirus pneumonia
7.27167160
None
Irreversible pulmonary disease
8
10mo
Cleaning solution aspiration
7.02/71/37
Renal, cardiac, liver
Multiorgan failure
9
14yr
Ketoacidosis/ARDS
7.23/56/60
Cardiac, renal
Multiorgan failure
7.56137124
Cardiac
Cardiac failure
None
Irreversible pulmonary disease
6
2.5 mo
10
3mo
RSV pneumonia
11
9mo
Adenovirus pneumonia
12
2.5 y’
Lamp oil aspiration
731i58140
None
Irreversible pulmonary disease
13
4.5 yr
Kerosene aspiration
7.20129177
Liver, renal, cardiac, cerebral
Multiorgan failure
14 15
2 vr 2mo
Drowning
7.15/88/56
Cardiac, renal
Multiorgan failure
RSV pneumonia
7.28167133
Cardiac, cerebral
Multiorgan failure
6.91125154
Abbreviations: ARDS, adult respiratory distress syndrome; RSV, respiratory syncytial virus.
have led to several pleas for a randomized, controlled trial of nonneonatal ECMO. v In spite of these important questions, there are a number of survivors of ECMO for nonneonatal respiratory failure, most (if not all) of whom probably would have died without ECMO.%r-1” Analysis of the present series has shown several important points. Any program offering ECMO for older children will be presented with an extremely wide variety of patients to evaluate for this form of therapy. This series included patients with such diverse diagnoses as pulmonary infections, aspiration of various substances, ARDS associated with other disorders, pulmonary contusion, and barotrauma. Unfortunately, at the present time we have not accumulated enough patients in any category to make definitive statements regarding the prognosis of patients with specific diagnoses. However, it is quite clear from this series that patients with isolated pulmonary failure, in whom other organs (brain, kidney, heart, liver) are functioning normally, stand the best chance for survival. This would appear to exclude at present the patient with multiple organ failure associated with sepsis, patients who have suffered sufficient hypoxia failure associated with sepsis, and patients who have suffered sufficient hypoxia to cause infarct or severely compromised function in organs other than the lung. However, assessing such organ damage prior to ECMO is difficult, and at present we tend to offer ECMO to virtually all children with apparently reversible pulmonary disorder, but warn the families that additional organ failure, if it should become manifest or develop during ECMO may be a cause for discontinuing therapy. The length of time ECMO is required in the older
patient can be considerable. Six of the nine survivors required ECMO for over 7 days, with two of these requiring more than 10 days of ECMO. These are generally greater times than required by most neonates treated with ECMO, suggesting a more severe pulmonary parenchymal injury. Other reported pediatric survivors have also required prolonged ECMO treatments.7J0 Patients in the present series have been successfully treated with cannulas placed in the neck or groin vessels, or a combination of these sites, or directly into the right atrium and aorta through a stemotomy approach. Although sternotomy results in significant hemorrhage during ECMO, necessitating reexploration in each case, this approach allows the placement of large cannulas, which facilitates ECMO flow at lower pressures. This is especially useful in larger children, in whom flow rates of 2.5 to 4.0 L/min are frequently needed. We have had no episodes of sternal infection in the four survivors of stemotomy in this series. The short-term follow-up of the survivors has been gratifying, with no significant pulmonary disorders, and only mild neurological deficits in three patients, all of whom have improved with time. Thus, it would appear that the long-term outlook for these critically ill children is excellent, an observation that has been mentioned in other series as well.7J0 Although this series is somewhat preliminary, it would appear that ECMO is now a viable alternative therapy for the older child with respiratory failure, unresponsive to conventional therapy. Future multiinstitutional studies of ECMO will hopefully resolve important questions regarding patient selection, optimal techniques of ECMO management, and longterm sequelae.
WEBER ET AL
1104
REFERENCES 1. Royal1 JA, Levitt DL: Adult respiratory distress syndrome in children. I. Clinical aspects, pathophysiology, pathology, and mechanisms of lung injury. J Pediatr 112:169-180,1988 2. Vernon DD, Dean JM, McGough EC: Pediatric extracorporeal membrane oxygenation. The time for anecdotes is over. Am J Dis Child 144:855-856,199O (editorial) 3. Weber TR, Pennington DG, Connors R, et al: Extracorporeal membrane oxygenation for newborn respiratory failure. Ann Thorat Surg 42:529-535,1986 4. Scalzo AJ, Weber TR, Jaeger RW, et al: Extracorporeal membrane oxygenation for hydrocarbon aspiration. Am J Dis Child 144:867-871,199O 5. Zapol W, Snider M, Hill JD, et al: Extracorporeal membrane oxygenation in severe acute respiratory failure. JAMA 242:21932196,1979
6. O’Rourke P, Crone RK: Pediatric applications of extracorporeal membrane oxygenation. J Pediatr 116:393-3951990 (editorial) 7. Steinhorn RH, Green T: Use of extracorporeal membrane oxygenation in the treatment of respiratory syncytial virus bronchiolitis: The national experience, 1983-1988. J Pediatr 116:338-342, 1990 8. Anderson HL, Attorri R, Custer J, et al: Extracorporeal membrane oxygenation for pediatric cardiopulmonary failure. J Thorac Cardiovasc Surg 99:1011-1021,199O 9. Redmond CR, Graves ED, Falterman KW, et al: Extracorporeal membrane oxygenation for respiratory and cardiac failure in infants and children. J Thorac Cardiovasc Surg 93:199-204,1987 10. Steiner RB, Adolph VR, Bonis SL, et al: Pediatric extracorporeal membrane oxygenation in posttraumatic respiratory failure. J Pediatr Surg 26:1011-1015,199l
Discussion Jay Wilson (Boston, MA): Dr Weber and his colleagues have reported to you today a 24-patient series concerning ECMO for pediatric failure encompassing such diverse diagnoses as sepsis, hydrocarbon aspiration, pulmonary contusion, etc. The overall survival was 9 of 24 patients (41%), which is comparable to the collective national experience maintained in the ELSO registry. This first slide, which I obtained from the ELSO registry, shows the national survival statistics from the ELSO registry. It shows that diagnoses are diverse and that survival is variable, but the overall survival is about 48%. In the current report, which is a large series of pediatric ECMO, the authors note that because of the diverse nature of the diagnoses, no definitive conclusions can be reached regarding prognosis of individual patients or even who is a suitable ECMO candidate. They are not alone in this quandary because, unfortunately, this is the state of pediatric ECMO today. The second slide, stratified by year, shows that although our experience with pediatric ECMO was growing immensely, our ability to determine who will survive the course of pediatric ECMO is not. This speaks to the need for a prospective randomized national trial. The unique contribution of this report, however, appears to be the finding that patients with isolated pulmonary failure, irrespective of diagnosis, have a higher survival, 75% in this series, than those with multiorgan system failure. Since I think this is the take-home message of this report, and because I believe patient selection is critical, I have a few questions for the authors. Eight patients were excluded from ECMO in the manuscript because their disease was felt to be
nonreversible. The question I have is how do you determine this? And have you used lung biopsies to help you in this decision? Second, you noted that survival was best when only pulmonary disease was present. However, 19 of 24 patients required pressure agents prior to ECMO, yet only one survivor in the manuscript was noted to have mild cardiac failure. What were your criteria for defining cardiac failure in this group of patients? Among the long-term survivors, three patients had neurological deficits deemed to be nonfocal. In several other patients, ECMO was terminated secondary to irreversible brain damage. What method was used to determine irreversible brain damage? Was CT scan used routinely to follow up with these patients? What methods of cannulation were used in the patients who had neurological deficits? And were any of those deficits attributable to the cannulations? Finally, on the basis of your report, would you be prepared at this time to exclude patients in your institution with multiorgan system failure from ECMO? I want to congratulate the authors on this report. I believe their findings about isolated pulmonary disease being the best prognosis is real, and, given our current lack of outcome predictors, important. Furthermore, until such time as a prospective, randomized trial can be organized, a large, carefully analyzed, single institution study such as this contributes significantly to the aggregate knowledge of this very complex topic. Arnold Coran (Ann Arbor, MI): When I left Ann Arbor Friday, I got the latest update on the ELSO registry, both for the University of Michigan and for the international registry, and as of that date we have treated 62 pediatric patients with ECMO, 28 for
PROLONGED
ECMO
FOR RESPIRATORY
FAILURE
respiratory failure and the others for cardiac failure, with a survival rate in the respiratory failure group of 57%, and a survival rate in the cardiac failure group of 38%. At this point, with the international registry, we have 285 patients with respiratory failure and 494 with cardiac failure with very similar survival rates. Our experience with the pediatric patient with respiratory failure on ECMO, is that you get to a point where you are not sure when to get them off the ECMO, and that occurs in 4 to 6 weeks. That is, they just can’t be weaned off. I would like to know what criteria you are using for continuing or stopping ECMO. One of the factors that appears to be important in weaning from the ECMO is the development of pulmonary fibrosis, especially in the hydrocarboningestion patients. Are you giving them anything during the course of ECMO to try to stop that fibrotic process? Robert Foglia (St Louis, MO): Your patients were on a ventilator 3 to 14 days. Could you stratify those patients in terms of survival and how long they were on a ventilator? Was survival improved in those patients who were on a ventilator for a shorter period of time pre-ECMO? T.R. Weber (response): In terms of how long pa-
1105
tients were on a ventilator prior to ECMO, we have several patients who were ventilated longer than 10 days. It’s not reasonable in a small group of patients to try to draw too many conclusions in that regard, however. I think both Dr Wilson and Dr Coran bring up some very important points; that is, determining which patients have reversible disease and how long should they remain on ECMO? Those are very difficult questions, and obviously, with a small series like this, we’re not going to have the answers. Our policy has been to leave these children on for 2 weeks and then to attempt to reassess their status. Occasionally, this must be done with lung biopsy. Sometimes the parents help to make the decision for you. We obviously try to choose patients preoperatively who we think have reversible disease. We have not noticed any difference in terms of site of cannulation and ultimate outcome, particularly with regard to neurological deficit. I think that Dr Wilson was questioning the possibility of carotid ligation in the older child causing neurological deficit. We have used carotid ligation in this group with survival, and without significant or noticeable neurological defect. We have used our neurologist to determine when a child has developed neurological deficits, rather than relying on CAT scan in this setting.
