Extracorporeal membrane oxygenation for nonneonatal respiratory failure

Extracorporeal membrane oxygenation for nonneonatal respiratory failure

Extracorporeal Membrane Oxygenation for Nonneonatal Respiratory Failure ByVincent Adolph, John Heaton, New Rodney Steiner, l Extracorporeal memb...

839KB Sizes 0 Downloads 0 Views

Extracorporeal Membrane Oxygenation for Nonneonatal Respiratory Failure ByVincent

Adolph,

John

Heaton, New

Rodney

Steiner,

l Extracorporeal membrane oxygenation (ECMO) has been used for 20 years in neonates and children with cardiac and respiratory failure. The number of neonates treated with ECMO has increased exponentially, but the number of older children treated is small. The selection and exclusion criteria for pediatric ECMO are poorly defined, and the results vary because of variable selection criteria and institutional experience with the technique. In order to help define the role of pediatric ECMO, we reviewed our experience in nonneonatal pediatric respiratory failure. We have treated 22 patients ranging in age from 1 to 105 months and ranging in weight from 3 to 35 kg. Eighteen patients met the criteria for adult respiratory distress syndrome, two had respiratory scyncytial virus pneumonia, and one had severe barotrauma complicating the management of reactive airway disease. All patients were considered by the referring institutions and by us to be failing conventional management as evidenced by hypoxia, hypercarbia, excessive ventilatory pressures, or progressive barotrauma. All were considered likely to die with continued conventional management. Sixteen of the 22 patients had complications (73%), but half of the last 10 patients had no complications. Hemorrhagic complications occurred in 12 patients. Mechanical complications included membrane failure, raceway rupture, pump malfunction, and improper cannula positioning. Other complications included culture-proven infection and renal failure. Eleven of the 22 patients survived (50%); nine of the last 12 survived (75%). These results suggest that ECMO may be a useful technique in selected pediatric patients with respiratory failure. Survival and complication rates improve as experience with the technique increases. Copyright o 1991 by W.B. Saunders Company INDEX WORDS: (ECMO).

Extracorporeal

membrane

oxygenation

T

HE FIRST successful application of extracorporeal membrane oxygenation (ECMO) was reported in 1972 by Hill et al.’ They described the use of venoarterial ECMO in a 24-year old trauma patient with adult respiratory distress syndrome (ARDS). By 1976, a review of the world literature reported a total of 233 patients with respiratory failure treated with

From the Section of Pediatric Surgery, Ochner Clinic, New Orleans, LA, and the Section of Pediatric Surgery, University of Chicago, Chicago, IL. Presented at the 21st Annual Meeting of the American Pediatric Surgical Association, Vancouver, British Columbia, May 19-22, 1990. Address reprint requests to Robert M. Arensman, MD, Section of Pediatric Surgery, Box 163, The University of Chicago, 5841 S Maryland Ave, Chicago, IL 60637. Copyright o 1991 by W.B. Saunders Company 0022-3468/91/2603-0016$03.00/0

326

Stan Bonis,

Kenneth

Falterman,

and Robert

Arensman

Orleans, Louisiana and Chicago, Illinois

ECMO by 90 teams in seven countries.’ These patients were treated predominantly using venovenous circulation, and the survival rate was 15%. In 1979, Zapol et al reported the results of a nine-center collaborative study of venoarterial ECMO in adults with ARDS3 In this prospective randomized study involving 90 patients over 2.5 years, ECMO did not improve survival. Following these reports, interest in extracorporeal respiratory support diminished considerably. As the interest in adult ECMO was decreasing, Bartlett et al were beginning to demonstrate an apparent increase in survival for neonatal patients treated with ECMO for respiratory failure.4 The number of neonates treated with ECMO has increased exponentially in the last few years,’ with a dramatic improvement in survival compared with historical contro1s.6X7There are now two randomized prospective studies that demonstrate improved survival in neonates.8,9 There are 64 centers providing neonatal ECMO that report to the Extracorporeal Life Support Organization (ELSO) Registry; in 3,593 patients the survival rate has been 83%.’ ECMO has been used in a much smaller number of nonneonatal pediatric patients with respiratory failure. As of March 1990, there were 102 cases of pediatric respiratory failure reported to the ELSO Registry.’ Several reports from ECMO centers have included a small number of pediatric patients.1”4 The selection and exclusion criteria for pediatric ECMO are poorly defined because predicting mortality in these patients has been even more difficult than in neonates. The results with ECMO support also vary as a result of the variable selection criteria and institutional experience with the technique. To help define the role of ECMO in pediatric respiratory failure, we reviewed our experience using ECMO in nonneonatal respiratory failure. MATERIALS

AND

METHODS

From September 1983 to June 1989, we treated 162 pediatric patients, 130 of whom were neonates. In the neonatal group survival to extubation was achieved in 104 patients (80%). Ten older patients had cardiac failure: nine after cardiotomy and one after cardiac transplant with rejection. Five of the 10 patients with cardiac failure were successfully decannulated, but only two are long-term survivors. Twenty-two patients had nonneonatal respiratory failure of various etiologies, and they constitute the study group.

Journalof Pediatric Surgery, Vol26,

No 3 (March),

1991:

pp326-332

NONNEONATAL

327

ECMO

ampicillin and an aminoglycoside when there is no proven infection. Culture-proven infections are treated as indicated by the antibiotic sensitivities of the cultured organism. Patients are given sedation and/or pain medications as required, but are not generally paralyzed during the ECMO run. Fentanyl drips have been used occasionally when sedation was required frequently. Most patients require only periodic sedation and are awake and alert for most of the ECMO run. If movement interferes with venous return and is not controlled with sedation, paralysis is maintained. Parenteral nutrition is provided through the ECMO circuit throughout the run. Transgastric tube placement for enteral feeding during the ECMO run has not been attempted. Standard parenteral formulas using high concentrations of glucose are used to minimize free water administration. Respiratory treatments are given throughout the ECMO run. Static and dynamic compliance as well as inspiratory and expiratory resistance are measured daily or every other day, and treatments are changed to improve resistance if pulmonary function testing indicates increased resistance. Chest physiotherapy is started immediately and continues throughout the ECMO run.

Circuit Design The circuit used is very similar to the standard neonatal ECMO circuit. The modifications to the neonatal circuit include changes in membrane size, heat exchanger, tubing size, raceway tubing, and, in larger patients, a bridge around the bladder. A4embrane. Kolobow (Washington, DC)/Sci-Med (Minneapolis, MN) membranes were used in all cases. A membrane that is rated for twice the expected flow is used as shown in Table 1. The larger membrane is often required to provide adequate ventilation through the circuit and provides the added benefit of “respiratory” reserve in long ECMO runs. This decreases the need to change oxygenators while on bypass. Heat exchanger. For patients weighing less than 11 kg, the Sci-Med P-714 heat exchanger was used as in neonates. For larger patients who require the 2,500 oxygenators or larger, the membrane has an integrated heat exchanger. Tubing size. For patients who require the 800 or 1500 oxygenator, M-in tubing is used in the entire circuit. The larger membranes have %-in blood ports, and %-in tubing is used throughout the circuit. Raceway tubing. Super-Tygon tubing (Bioces Inc, Bolingbrook, IL) is used in the circuit that occupies the raceway. This tubing is advanced every 100 to 110 hours or earlier if it shows signs of wear. Bladder bridge. A bladder box is used in neonatal and pediatric circuits to servoregulate the roller pump. In patients weighing more than 20 kg (membranes 2500 or larger with %-in tubing), a bridge is placed around the bladder box. This allows some blood to bypass the bladder box, which can hold only 50 mL of fluid.

Selection Criteria All patients were considered to be failing conventional management as evidenced by hypoxia, hypercarbia, excessive ventilatory pressures, or progressive barotrauma. Ail were considered likely to die with continued conventional management. ECMO was offered to all patients as an attempt at rescue. The indications for ECMO in nonneonatal ECMO are not yet defined, but all of these patients met at least one of the standard neonatal criteria for ECMO. Twenty patients met the barotrauma criteria; eight had an oxygenation index greater than 0.40 for 2 hours; seven had an alveolararterial gradient greater than 600 for 12 hours; and five had acute deterioration (PaO, < 40 mm Hg for 2 hours). Nine patients had peak airway pressures greater than 50 cm H,O and eight patients met the fast entry criteria for the NIH ECMO study (PaO, < 50 mm Hg for more than 2 hours with an F,O, of 1.0 and positive end-expiratory pressure of at least 5).

Cannulation Techniques All patients were supported using venoarterial perfusion. All patients were cannulated via the right common carotid artery and internal jugular vein. The vein was ligated in all patients that smvived to decannulation. In four of the last five survivors, the carotid artery was repaired at the time of decannulation. Colorflow Doppler scans of the carotid in all four showed antegrade flow in the common carotid artery. There was no ultrasound evidence of stenosis. In addition, there were no clinical or computed tomographic (CT) signs of infarction or embolic complications.

RESULTS

Patient Management on ECMO

Twenty-two patients with pulmonary failure were treated with ECMO (Table 2). They ranged in age from 1 to 105 months (mean, 31 months); weight ranged from 3 to 35 kg (mean, 14 kg). Eighteen patients had ARDS, as defined in Table 3, and eight survived (44%). Two had respiratory failure due to respiratory syncytial virus (RSV) infection and both survived. One patient had severe barotrauma complicating the management of reactive airway disease

The management of older patients on ECMO is in most ways similar to neonates on ECMO. Systemic heparin is administered and the activated clotting time (ACT) is maintained at 200 to 250 seconds. The ACT is kept at 160 to 180 seconds in patients who have active bleeding. Earlier in the series, ACTS were kept between 250 and 350 seconds. The platelet count is maintained at 75,000/mm3 in pediatric patients in whom there is no significant bleeding. If bleeding develops platelet counts are maintained above 100,OOO/mm’.Patients are generally given prophylactic antibiotics consisting of

Table 1. Membrane Characteristics and Uses Patient Weight (kg) <6

Membrane

Area (m’)

BloodFlow’ (Urnin)

Gas Flow’ (Urnin)

Volumet (mL)

Ports Iin) ‘/n

800

0.8

1.2

2.4

100

6-11

1,500

1.5

1.8

4.5

175

%

11-20

2,500

2.5

4.5

7.5

330

%

20-30

3,500

3.5

5.5

10.5

4,500

4.5

6.5

13.5

450 540

3/s

>30 *Maximum

rated flow according to manufacturer’s recommendations.

tPriming volume.

3/s

328

ADOLPH ET AL

Table 2. Pediatric Pulmonary Patients

Case NO.

Barotrauma Criteria Met

Respiratory Severity Index

Time on Ventilator (d)

Hours on ECMO

Hoursto Extubation After ECMO

Age (mol

Weight [email protected])

1

60

20

ARDS; valproic acid ingestion/aspiration

Y

1.80

28

32

N

-

2

1

10

ARDS; aspiration

N

0.72

2

112

Y

36

N

Diagnosis

Survival

3

76

18

DIP

Y

2.00

10

197

4

16

11

ARDS; hydrocarbon ingestion

Y

1.53

6

312

N

5

15

20

ARDS; Reye’s syndrome

Y

4.56

42

140

N

6

17

10

ARDS; hydrocarbon aspiration

N

1.51

2

450

N

7

11

9

ARDS: near drowning

Y

1.57

4

68

Y

8

18

10

SJS

Y

2.83

15

61

N

9

32

14

ARDS; near drowning

Y

4.92

9

261

N

10

13

20

ARDS; hydrocarbon aspiration

Y

4.25

14

595

N

Pneumonia; RSV

N

2.08

6

450

Y

*

ARDS; trauma

Y

1 .OJ

9

176

Y

72

-

*

11

8

5

12

45

15

13

20

9

ARDS; near drowning

Y

2.23

5

249

Y

14

55

19

ARDS; near drowning

Y

3.00

5

377

N

55 -

0.99

64

8

138

Y

14

516

N

4 *

89

Y

72

3.08

112

Y

J2t

Y

2.22

14

107

N

15

10

8

ARDS; near drowning

Y

16

100

35

ARDS; trauma

Y

I?

17

12

RAD

Y

2.03

18

14

8

Pneumonia; RSV

N

19

2

3

ARDS; ? viral pneumonia

20

2

4

ARDS; s/p lobectomy

Y

3.16

7

475

Y

*

21

47

17

ARDS; trauma

Y

0.57

5

91

Y

28

22

105

35

ARDS; ? aspiration/barotrauma

Y

3.19

2

68

Y

Abbreviations: ARDS, adult respiratory distress syndrome; DIP, desquamative

interstitial pneumonitis; SJS, Stevens-Johnson

72 syndrome; RSV,

respiratory syncytial virus; RAD, reactive airway disease; S/P, status post. *Prolonged ventilator support required after ECMO but survived to extubation. tReturned to prior ventilator settings.

(survived) and one patient with desquamative interstitial pneumonitis died. Eleven of 22 patients survived (50%). Only two of the first 10 patients survived, but nine of the last 12 patients survived (75%). Table 4 shows the trend of improving success in patients treated with ECMO as well as the complication rates by year. Table 5 shows the comparison data for blood gas data and ventilator settings of survivors and nonsurvivors. There was not a significant difference for any variable between survivors and nonsurvivors. The average time of mechanical ventilation prior to ECMO was 10 days (range, 2 to 42 days). For survivors, the average was 5 days (range, 2 to 9) and for nonsurvivors the average was 14 days (range, 2 to 42). No patient on mechanical ventilation for 10 days or longer prior to ECMO survived. One patient with RSV was 14 months old and was still hospitalized requiring intermittent ventilator support prior to the viral infection. He was placed on ECMO 5 days after the onset of the illness. Table 3. Criteria for Diagnosis of ARDS

The average time on ECMO was 230 hours (range, 32 to 595 hours). Survivors were on ECMO for an average of 184 hours (range, 68 to 475 hours) compared with an average of 277 hours for patients who died (range, 32 to 595 hours). Seven of the 11 survivors were extubated shortly after decannulation, with the average time to extubation being 49.8 hours (range, 28 to 72 hours). Three patients required mechanical ventilation for more than 2 weeks after ECMO but survived to extubation. The patient with RSV pneumonia who required ventilator support prior to the acute illness returned to his previous ventilator settings within 72 hours of decannulation. Complications Six of the 22 patients (27%) had uncomplicated courses on ECMO that ranged from 64 to 249 hours,

Table 4. Survival and Complication Rates by Year

YeFir

No. of Patients

NO.

Survived(W)

No. of Complications1%)

1984

0 (0)

1 (100)

1985

1 (50)

2 (100)

PaO, < 75 on F,O, > 50% and PaO,/F,O, < 200

1986

0 (0)

1 (100)

Diffuse bilateral radiographic changes

1987

1 (20)

4 (80)

Noncardiogenic

1988

3 (60)

4 (80)

Initiating event or ventilator support > 7 days

1989

6 (75)

4 (50)

NONNEONATAL

329

ECMO

Table 5. Blood Gas Data and Ventilator Parameter

Settings Before Bypass

Survivors

PaO, (mm Hg)

74.3 + 33.5

54.8 -+ 18.3

PaCO, (mm Hg)

60.5 ? 22.6

56.3 f 25.2

Peak airway pressure (mm Hg)

43.5 lr 10.8

50.7 f 10.6

Mean airway pressure (mm Hg)

20.4 2 6.3

23.6 ? 4.6

Compliance (mUcm H,O)

5.5 f 2.3

6.3 + 2.9

562 f 89

592 2 45

Oxygenation index

0.32 2 0.14

0.48 f 0.20

Respiratory severity index

1.88 f 0.98

2.86 f 1.29

AaDO, (mm Hg)

NOTE. All data given as mean 2 SD.

and all survived to extubation. Half of the last 10 ECMO runs were uncomplicated. Sixteen of the 22 patients had complications on ECMO (73%). The most significant complications have been hemorrhagic. Earlier in the series, platelets were given when thrombocytopenia was associated with bleeding and ACTS were kept between 250 and 350 seconds. Using the range of 200 to 250 seconds for systemic anticoagulation and maintaining a higher platelet count, incidence and severity of hemorrhagic complications is lower. Seven of the first 10 patients had significant hemorrhagic complications, whereas only five of the last 12 had significant bleeding using the lower ACTS. In addition, in only one of the last 12 patients had a hemorrhagic complication that was the direct cause of death. By contrast, in four of the first 10 patients, bleeding either directly lead to the patient’s death or was significant and required discontinuation of ECMO, shortly after which the patient died. There were no thrombotic complications, and fibrin was not observed in the bridge around the oxygenator or in the bladder. Intracranial hemorrhage (ICH) confirmed by CT scan, or acute deterioration in neurological status attributed to ICH, occurred in three patients. Two patients developed severe intrapulmonary hemorrhage that prompted decannulation. One patient died after ECMO was discontinued due to a rapidly expanding retroperitoneal hematoma that deveIoped at the same time as an acute neurological change. Seven other patients had significant blood loss from the gastrointestinal tract or from surgical sites, but the bleeding was not severe enough to prompt decannulation. Twelve of the 22 patients (55%) had a hemorrhagic complication on ECMO. Mechanical complications were secondary only to hemorrhagic complications in frequency and severity. In the first 14 circuits Tygon tubing was used in the circuit and raceway. In four cases raceway rupture complicated the ECMO run; in two cases, raceway rupture occurred twice despite periodic advancing of the raceway tubing. Super-Tygon tubing was used in the raceway in the last eight cases, and was advanced

every 100 to 110 hours. No raceway ruptures have occurred in runs as long as 516 hours. Oxygenator deterioration requiring replacement or addition of a second oxygenator occurred in seven patients. In one patient the membrane failed within minutes of initiating bypass, and bypass was stopped because of the suspicion of clots in the circuit. A new membrane functioned well and the patient had no evidence of embolic complications; the membrane was found to contain several large clots occupying almost the entire surface area. The pump malfunctioned in one patient and the patient tolerated clamping out from the circuit while it was changed. Two patients required repositioning of a cannula while on ECMO due to migration after the initial placement. In one patient placement of a second venous catheter in the femoral vein was attempted because of inadequate ffow through the venous catheter and inability to provide full support. In one patient air was found entering the venous cannula, but no source could be found. This was aspirated via the bladder without incident and resolved. Mechanical complications occurred in 10 of the 22 patients (45%). Three patients developed renal insufficiency or failure and required hemofiltration or dialysis while on ECMO; all three died. Mild renal insufficiency (elevated creatinine; no dialysis or hemofiltration) occurred in two patients. Five patients had cultureproven infections while on ECMO, and two survived. DISCUSSION

Severe respiratory failure affects only a small percentage of admissions to pediatric intensive care units; in one series, only 1%.15Affected patients have a high mortality, ranging from 30% to 74%.16-18 Survival in patients with ARDS is not improving significantly, despite advances in ventilatory support.” Long-term pulmonary function is generally good in patients who survive despite the severity of their respiratory failure.“’ Extracorporeal support of gas exchange is intuitively an ideal means of treating severe respiratory failure if it is reversible, but enthusiasm for ECMO waned after the National Institutes of Health (NIH) ECMO study.3 With the wide acceptance of ECMO as therapy for otherwise fatal neonatal respiratory failure, enthusiasm for extracorporeal support in older patients is increasing. Selecting patients with a high predicted mortality in whom the lung disease is still reversible is critical to the intehigent use of ECMO as a rescue therapy or in randomized trials. Until recently, there have been few predictive data in pediatric respiratory failure. Zobel et al have proposed a respiratory severity index (RSI),

330

ADOLPH ET AL

which predicted 100% mortality if greater than 0.75 at 48 hours.2l Butt and McDougall found that a peak inspiratory pressure greater than 40 mm Hg and an alveolar-arterial gradient of greater than 580 predicted an 80% mortality.22 Tamburro et al reported that an alveolar-arterial gradient of greater than 450 mm Hg for 16 hours predicted mortality, with a 90% sensitivity and a 100% specificity.23 They also reported a 90% sensitivity and 92% specificity with a gradient of 450 mm Hg for 12 hours. The NIH ECMO study criteria for adults with ARDS predicted a 90% mortality.3 Table 6 shows the results for each group of entry criteria. None of the criteria for high mortality or “nonresponders” to select out “responders” ECMO. Furthermore, 15 of the present patients had PaCO, greater than 45 and would have been excluded

Table 6. Results by “Entry” Criteria Criteria (predictedmortality)

No. of Patients

No. of Survivors(%)

20

9 (45)

Oxygenation index (?80%)

8

3 (37)

AaDO, > 600 12 h (?80%)

7

2 (29)

Acute deterioration (80%)

5

3 (80)

21 20

10 (48) 10 (50)

PIP > 40 and AaDO, > 580 (80%)

9

3 (33)

Fast NIH criteria (90%)

8

3 (38)

PIP > 50 (7)

9

3 (33)

Barotrauma (?80%)

RSI (100%) AaDO, > 450 12 h (90%)

from the NIH study. Eight of these 15 (53%) survived. The results shown in Table 6 suggest that ECMO was effective in rescuing 40% to 50% of the treated patients.

REFERENCES 1. Hill J, O’Brien T, Murray J: Prolonged extracorporeal membrane oxygenation for acute post-traumatic respiratory failure (shock-lung syndrome). N Engl J Med 286:629-634,1972 2. Gille J, Bagniewski A: Ten years of use of extracorporeal membrane oxygenation (ECMO) in the treatment of acute respiratory insufficiency (ARI). Trans ASAIO 22:102-1081976 3. Zapol W, Snider M, Hill J, et al: Extracorporeal membrane oxygenation in severe acute respiratory failure. JAMA 242:21932196,1979 4. Bartlett R, Gazzaniga A, Jeffries M, et al: Extracorporeal membrane oxygenation (ECMO): Cardio-pulmonary support in infancy. Trans ASAIO 22:80X$1976 5. Extracorporeal Life Support Organization Registry. Ann Arbor, MI, University of Michigan, March 1990 6. Bartlett R, Andrews A, Toomasian J, et al: Extracorporeal membrane oxygenation (ECMO) for newborn respiratory failure: Forty-five cases. Surgery 92:425-433,1982 7. Short B, Miller M, Anderson K: Extracorporeal membrane oxygenation in the management of respiratory failure in the newborn. Clin Perinatol 14:737-748,1987 8. Bartlett R, Roloff D, Cornell R, et al: Extracorporeal circulation in neonatal respiratoty failure: A prospective randomized study. Pediatrics 76:479-487,198s 9. O’Rourke P, Crone R, Vacanti J, et al: Extracorporeal membrane oxygenation and conventional medical therapy in neonates with persistent pulmonary hypertension of the newborn: A prospective randomized study. Pediatrics 84:957-963,1989 10. Bartlett R, Gazzaniga A: Extracorporeal circulation for cardiopulmonary insufficiency. Curr Probl Surg lS:l-96,1978 11. Gattinoni L, Pesenti A, Mascheroni D, et al: Low-frequency positive-pressure ventilation with extracorporeal CO, removal in severe acute respiratory failure. JAMA 256:881-886,1986 12. Egan T, Duffin J, Glynn M, et al: Ten-year experience with extracorporeal membrane oxygenation for severe respiratory failure. Chest 94:681-687,1988

13. Snider M, Campbell D, Kofke W, et al: Venovenous perfusion of adults and children with severe acute respiratory distress syndrome. Trans ASAIO 34:1014-1020,1988 14. Muller E, Knoch M, Lennartz H: Severe ARDS managed by extracorporeal CO, removal-A review of 78 consecutive adults and children. Proc Sixth Annual ECMO Symposium, Washington, DC, Children’s Hospital National Medical Center, 1990 15. Lyrene R, Truog W: Adult respiratory distress syndrome in a pediatric intensive care unit: Predisposing conditions, clinical course and outcome. Pediatrics 67:790-795, 1981 16. Holbrook P, Taylor G, Pollack M, et al: Adult respiratory distress syndrome in children. Pediatr Clin North Am 27:677-685, 1980 17. Katz R: Adult respiratory distress syndrome in children. Clin Chest Med 8:635-639, 1987 18. De Bruin W, Notterman D, Greenwald B: Mortality of ARDS in infants and children. Crit Care Med 17:Slll, 1989 (abstr) 19. Hudson L: Survival data in patients with acute and chronic lung disease requiring mechanical ventilation. Am Rev Respir Dis 140:X9-S24,1989 20. Elliott C, Morris A, Cengiz M: Pulmonary function and exercise gas exchange in survivors of adult respiratory distress syndrome. Am Rev Respir Dis 123:492-495,198l 21. Zobel G, Kuttnig M, Trop H: A respiratory severity index (RSI) for children with ARDS. Crit Care Med 17:SllO, 1989 (abstr) 22. Butt W, McDougall P: What is the role of pediatric ECMO in Australia? Aust Pediatr J 25:189-191, 1989 23. Tamburro R, Chyka D, Bugnitz M: The use of alveolararterial oxygen gradient to predict mortality from severe respiratory failure in pediatrics. Proc of the Sixth Annual ECMO Symposium, Washington, DC, Children’s Hospital National Medical Center, 1990

Discussion F. Ryckman (Cincinnati, OH): The need for extracorporeal circulatory support in the pediatric population is very rare, and it is going to be extremely

necessary to have informative and detailed reviews such as this one for all of us to spread the wealth of knowledge on this rare field. What I would like to

NONNEONATAL

ECMO

emphasize today is really the difficulty with defining patient-selection criteria for this patient population. Unlike neonatal ECMO, which has a common unifying diagnosis of persistent pulmonary hypertension, pediatric ECMO series have a tremendous variety of primary pathology. If we are to accurately assess the results of treatment such as ECMO in this group, it’s essential that we know the natural history of each of these diseases and their predicted mortality for the diagnostic specific groups. As has been pointed out here, there is a tremendous descrepancy in survival depending on the diagnosis for which the treatment is used. Selection criteria varied. Six different methods were used in this particular study although nine were mentioned on a slide as proposed methods of selection. Unlike neonatal ECMO for which we are attempting to predict hypoxic vasoconstrictive shunting damage and potential mortality, in pediatric ECMO, as emphasized by the need to institute bypass early, what we are hoping to address is a prediction of irreversible pulmonary fibrosis by direct lung damage. The question I would raise is can we hope to believe that any single criteria is going to be able to predict the rate of development of pulmonary fibrosis in this widely discrepent diagnostic group that contains entities as different from each other as hydrocarbon ingestion, viral pneumonia, and generalized ARDS? The second question is that of the venovenous technique. As many of us are aware, the present adult experience with extracorporeal circulatory support has really shown a tremendous advantage with the use of apneic oxygenation associated with extracorporeal CO, removal (ECC02R). Is this success something that should be brought into the pediatric ECMO realm at the present time, or do you feel with carotid reconstruction that the use of venoarterial bypass is as equally advantageous? My final question is both a query and a plea for the strong support of the national multicenter trial to attempt to address these questions. I think the first question that has to be answered is that of the definition of the natural history of each of these respiratory failure entities. An attempt at answering this question is ongoing right now through the ELSO Registry groups. It is very important that this be defined, and did you define this in your own patient population when you selected these patients? Second, I think we need a multicenter trial to really look at the fine results that you have here with venoarterial ECMO compared with venovenous ECCO,R compared with conventional medical management. I/. Adolph (response): In response to Dr Ryckman’s first question, our concern was selection criteria, which is clearly the most controversial aspect of pediatric ECMO. As he points out, unlike neonates,

331

in whom the final common pathway of pulmonary hypertension is exquisitively sensitive to ECMO support, the various etiologies in older pediatric patients complicates selection. I agree with his point that different selection criteria for different diagnoses may ultimately prove to be the correct way to select patients. There’s not enough information now on different diagnoses and their natural history to make that kind of selection at this point. He also mentioned the use of extracorporeal CO, removal. The difference in this technique is that it is a low-flow technique that uses venovenous support. The membrane only provides CO, removal and oxygenation still occurs through the natural lung as opposed to the membrane lung. This technique has several advantages. For one, it’s a low-flow system and it makes cannulation easier because it only enters the venous system. The other potential advantage is that it provides a more natural circulation to the lung. In the ECCO,R system, the well-oxygenated blood is returned to the lung rather than bypassing the lung. There are several experimental models that suggest venoarterial ECMO may actually damage the lung by providing well-oxygenated blood into the aorta. So there are several theoretical potential advantages to ECCO,R. Using criteria that were used in the NIH ECMO trial, several centers that use ECCO,R in adults have had about a 50% survival rate. The only difference in those series was that they had slightly earlier entry, and in every review of extracorporeal support in adults, neonates, or children, early entry equalled a better survival. R.P. Foglia (Los Angeles, CA): I enjoyed your paper very much for several reasons. One of which is there are only 85 children in the pediatric experience in the registry and instead of anecdotal experience, you have presented a large series. If we look at the pediatric experience, in many ways it’s like the adult experience 11 years ago that Zapol reported. In that experience, only 10% of the patients survived, primarily because they had been in severe respiratory failure for an average of 10 days. In the pediatric experience in the National Registry, the nonsurvivors had been on a ventilator for 8% days before going on ECMO, the survivors 4% days. So there is inherently a paradox because the risk of intracranial hemorrhage, the major morbidity in the neonate, should be low in the pediatric patient. V Adolph (response): We found the same in our series, the average duration of mechanical ventilation prior to ECMO in survivors was 5 days as opposed to 14 days in the patients who didn’t survive. As I said before, no patient who was put on ECMO 10 days after the institution of mechanical ventilation survived. This demonstrates that the critical factor is not

332

only patient selection but trying to identify a predictive criterion that exists at an early point after intubation. Some of the more recent predictive criteria predict mortality as early as 48 hours, and at least in one reported series had 100% sensitivity. Whether those predicted criteria will apply to other institutions is the crucial question. R. H. Connors (St Louis, MO): We have used ECMO in several children in St Louis as well. We have the same problem deciding who to put on, but I would also like your comments on deciding when to stop. Have you found any measures for patients on ECMO that predict when it’s going to work? How long have you had patients on ECMO; what were your longest successful runs? Do you have any feelings about when to stop? K Adolph (response): Several of the series that have

ADOLPH ET AL

reported the use of ECMO or ECCO,R in adults and pediatric patients have reported that they either observe improvement in 48 to 72 hours or the patient is not going to improve at all. In our experience, that wasn’t the case. We had several patients who were on ECMO as long as 10 to 14 days with no apparent improvement, and we were considering decannulation electively, expecting that the patient would die. One patient, was on ECMO for 19 days. We were planning to decannulate electively but waited until the next day. There was improvement in the chest x-ray and improvement in compliance so we waited and that patient was ultimately decannulated and survived. So, at this point, we generally terminate ECMO only when some complication forces us off ECMO.