Pediatric Annals

Extracorporeal Membrane Oxygenation in Neonatal Pulmonary Failure

Thomas M Krummel, MD; Lazar J Greenfield, MD; Barry V Kirkpatrick, MD; Dawn G Mueller, MD; Kathryn W Kerkering, MD; Arnold M Salzberg, MD

Abstract

Despite significant improvements in the care of infants with serious pulmonary dysfunction, current treatment is inadequate for the estimated 15,000 newborns who die yearly of respiratory failure. Respiratory distress syndrome (RDS) accounts for most of these deaths. The remainder die of meconium aspiration syndrome (MAS) or of persistent fetal circulation (PFC) which may complicate RDS or MAS, aggravate the pulmonary hypoplasia associated with congenital diaphragmatic hernia (CDH) or occur as a primary entity.

Current treatment of neonatal pulmonary failure sequentially includes supplemental oxygen, positive pressure mechanical ventilation, inotropic agents to augment the systemic circulation (notably dopamine), and pulmonary vasodilators to reduce pulmonary vascular resistance. These maneuvers have lowered the current mortality rate to approximately 15%. Unfortunately, the mortality of the infant with CDH repaired before eight hours of age is still 60%.

In addition to the acute therapeutic failures resulting in death, chronic failure occurs in a number of infants who survive the immediate neonatal period but are crippled by bronchopulmonary dysplasia (BPD) from barotrauma and oxygen toxicity.

Hill and co-workers first demonstrated the feasibility of long-term pulmonary support with extracorporeal circulation.1 Since then results with extracorporeal membrane oxygenation (ECMO) in adults and older children with respiratory failure have been discouraging because resolution of severe pulmonary disease proceeds to fibrosis rather than returning to normal architecture. However, neonatal pulmonary failure is followed by anatomic and functional lung healing in many instances if the integrity of the patient is maintained. Accordingly, the adult experience cannot be extrapolated to the newborn. ECMO thus might provide life support in moribund infants and purchase the critical time needed for return of lung function.

As a final resuscitative measure, Bartlett has used a modified heart lung apparatus to provide cardiopulmonary support for up to two weeks in neonates failing to respond to maximum medical therapy.2 In the ECMO system, venous blood is diverted externally, oxygen is added and carbon dioxide removed with a membrane oxygenator and the "arterialized" blood is pumped back into the ascending aorta. The pulmonary circulation is decompressed and the systemic circulation supported, alleviating the right to left shunting characteristic of PFC. Thus ECMO provides a direct approach to both components of respiratory failure in the newborn; achieving gas exchange and simultaneously decompressing pulmonary hypertension. This support permits ventilator adjustments to minimize barotrauma and oxygen toxicity. Following Bartlett's initial efforts, Hardesty has confirmed the value of ECMO in the management of certain moribund infants.3

ECMO creates the appropriate environment for pulmonary healing by providing gas exchange without damaging levels of inspired oxygen and by reducing barotrauma complications, including interstitial emphysema, p neu m o mediasti n um, and pneumothorax. When major bilateral air leaks are combined with poor compliance in the remaining lung, the additional hypoventilation and hypercapnea requires high inspiratory pressures which aggravate the alveolar fistulas. This lethal cycle is effectively treated with ECMO. Once the patient is on bypass and the ventilator pressure is decreased, air leaks seal promptly and lung recovery can be expected.

Under current criteria, ECMO will be considered in about 1 % to 2% of newborns with respiratory failure, who are moribund despite 100% oxygen, high pressure mechanical ventilation and vasoactive drug therapy. Final selection, however, must be based on objective criteria predictive of potential mortality. This may be provided by Bartlett's newborn pulmonary insufficiency index,5 or the alveolar-arterial oxygen grathent (A-a DO2) proposed by Raffely and Downes.6 The NPII is in essence a cumulative integration of hypoxia and acidosis during the first 24 hours of life. Using the NPII notnogram scoring system, a 90% probability of mortality can be determined for patients who have severe…

Despite significant improvements in the care of infants with serious pulmonary dysfunction, current treatment is inadequate for the estimated 15,000 newborns who die yearly of respiratory failure. Respiratory distress syndrome (RDS) accounts for most of these deaths. The remainder die of meconium aspiration syndrome (MAS) or of persistent fetal circulation (PFC) which may complicate RDS or MAS, aggravate the pulmonary hypoplasia associated with congenital diaphragmatic hernia (CDH) or occur as a primary entity.

Current treatment of neonatal pulmonary failure sequentially includes supplemental oxygen, positive pressure mechanical ventilation, inotropic agents to augment the systemic circulation (notably dopamine), and pulmonary vasodilators to reduce pulmonary vascular resistance. These maneuvers have lowered the current mortality rate to approximately 15%. Unfortunately, the mortality of the infant with CDH repaired before eight hours of age is still 60%.

In addition to the acute therapeutic failures resulting in death, chronic failure occurs in a number of infants who survive the immediate neonatal period but are crippled by bronchopulmonary dysplasia (BPD) from barotrauma and oxygen toxicity.

Hill and co-workers first demonstrated the feasibility of long-term pulmonary support with extracorporeal circulation.1 Since then results with extracorporeal membrane oxygenation (ECMO) in adults and older children with respiratory failure have been discouraging because resolution of severe pulmonary disease proceeds to fibrosis rather than returning to normal architecture. However, neonatal pulmonary failure is followed by anatomic and functional lung healing in many instances if the integrity of the patient is maintained. Accordingly, the adult experience cannot be extrapolated to the newborn. ECMO thus might provide life support in moribund infants and purchase the critical time needed for return of lung function.

As a final resuscitative measure, Bartlett has used a modified heart lung apparatus to provide cardiopulmonary support for up to two weeks in neonates failing to respond to maximum medical therapy.2 In the ECMO system, venous blood is diverted externally, oxygen is added and carbon dioxide removed with a membrane oxygenator and the "arterialized" blood is pumped back into the ascending aorta. The pulmonary circulation is decompressed and the systemic circulation supported, alleviating the right to left shunting characteristic of PFC. Thus ECMO provides a direct approach to both components of respiratory failure in the newborn; achieving gas exchange and simultaneously decompressing pulmonary hypertension. This support permits ventilator adjustments to minimize barotrauma and oxygen toxicity. Following Bartlett's initial efforts, Hardesty has confirmed the value of ECMO in the management of certain moribund infants.3

ECMO creates the appropriate environment for pulmonary healing by providing gas exchange without damaging levels of inspired oxygen and by reducing barotrauma complications, including interstitial emphysema, p neu m o mediasti n um, and pneumothorax. When major bilateral air leaks are combined with poor compliance in the remaining lung, the additional hypoventilation and hypercapnea requires high inspiratory pressures which aggravate the alveolar fistulas. This lethal cycle is effectively treated with ECMO. Once the patient is on bypass and the ventilator pressure is decreased, air leaks seal promptly and lung recovery can be expected.

Under current criteria, ECMO will be considered in about 1 % to 2% of newborns with respiratory failure, who are moribund despite 100% oxygen, high pressure mechanical ventilation and vasoactive drug therapy. Final selection, however, must be based on objective criteria predictive of potential mortality. This may be provided by Bartlett's newborn pulmonary insufficiency index,5 or the alveolar-arterial oxygen grathent (A-a DO2) proposed by Raffely and Downes.6 The NPII is in essence a cumulative integration of hypoxia and acidosis during the first 24 hours of life. Using the NPII notnogram scoring system, a 90% probability of mortality can be determined for patients who have severe respiratory failure.

Early and accurate prediction of 100% mortality despite best conventional therapy can be furnished by A-a DO2. In order to evaluate the limits of current medical therapy, the charts of the last 18 patients treated for pulmonary failure with 100% oxygen and high inspiratory pressures in the Newborn ICU at the Medical College of Virginia were reviewed.7 Patients were restricted to those with hypoxemia and pulmonary hypertension. No patient has ever survived with a post-ductal alveolararterial oxygen grathent greater than 620 torr for 12 hours despite maximum ventilatory support, tolazoline and dopamine.

Since October of 1980, nine newborns with predictable fatal pulmonary failure have been managed with ECMO at the Medical College of Virginia.8 Using the above criteria, seven patients were moribund and accepted for ECMO. Two additional newborns with ?-a DOz markedly above 620 torr for seven and eight hours developed refractory hemodynamic instability and were also placed on ECMO. Six survived. Most complications were manageable and early followup verifies normal growth and development in all but one of the survivors.

Partial veno-arterial cardiopulmonary bypass was achieved in all cases, using the right neck for access to the internal jugular vein and common carotid artery. Under local 1% xylocaine anesthesia, following systemic heparinization, a #14-16 F venous cannula was directed through the internal jugular vein into the right atrium. A fflO-12 F perfusion catheter was then advanced through the common carotid artery to the proximal aortic arch, assuring adequate admixture of blood traversing the lung with blood externally oxygenated. Both vessels were ligated; no attempt was made to restore patency at decan nutation. AP and lateral chest x-rays were obtained to verify correct location of the catheters.

The ECMO circuit employs gravity venous drainage to a servo-regulated roller pump, a heat exchanger and oxygenator primed with heparinized blood ventilated with a warm mixture of oxygen and carbon dioxide to achieve a pCO± of 30 to 35 torr, a pO: of 400 torr at a temperature of 37° (Figure I).

Following can nula t ion, bypass is gradually instituted over the next 30 to 60 minutes until approximately 80% of the predicted cardiac output is externally diverted through the ECMO circuit, usually requiring flow rates between 100 cc to 120 cc per kilogram per minute.

Once hemodynamics are stabilized and acidosis, hypoxia and hypercapnea corrected, the peak inspiratory pressure on the ventilator is reduced to 25 cm of water, PEEP (positive end expiratory pressure) adjusted to 4 cm of water and the ventilator rate set at 10 to 12 breaths per minute. The FiO2 is maintained at 40% and the pO2 maintained at 80 to 100 torr by adjusting the extrac orporeal flow. Vasoactive drugs are discontinued.

Whole blood activated clotting time is maintained at two to three times normal (240 to 360 seconds) with a constant heparin infusion of 30 to 60 units per kilogram per hour. Excessive bleeding, imminent operation or low serum fibrinogen is an indication for fresh frozen plasma administration. Platelets are given if thrombocytopenia of 50,000 or less is noted.

Figure 1. The ECMO circuit employs gravity venous drainage toa servo-regulated roller pump, a heat exchanger and oxygenator primed with a warm mixture of oxygen and carbon dioxide.

Figure 1. The ECMO circuit employs gravity venous drainage toa servo-regulated roller pump, a heat exchanger and oxygenator primed with a warm mixture of oxygen and carbon dioxide.

Maintenance intravenous fluids are continued and hyperalimentation is begun on the fourth or fifth day of life. Penicillin and gentamycin are routinely administered along with cimetidine. Blood cultures and chest x-rays are performed daily.

ECMO is discontinued if irreversible brain damage or other lethal organ failure is documented or when lung function recovers. If the pulmonary process resolves, oxygenation in that fraction of the cardiac output traversing the lungs improves and is observed as a rise in the peripheral pOi with fixed ECMO flow. As lung function improves ECMO flow is proportionally decreased, diverting less blood to the extracorporeal circuit while maintaining low ventilator settings. ECMO flows are further reduced over a period of days and subsequently discontinued. The cannulas are removed and the vessels ligated. Survivors are weaned from the ventilator in the usual fashion and extubated. Chest x-ray resolution lags behind clinical improvement but invariably occurs in survivors.

ECMO, then, is used when the patient who would otherwise die is identified, remains moribund despite best current methods of therapy and has a reasonable possibility, that time provided, will allow for resolution of pulmonary pathology and subsequent survival. Contraindications include weight below 1300 grams, clear documentation of intracranial bleed or neurologic deficit, multiple congenital anomalies, and renal failure.

In our experience, mean arterial pressure at the time of ECMO institution was 30 mm mercury with hemodynamic instability. Average arterial blood gases prior to ECMO in nine patients wereapH of 7.1, pO2 26, pCOi of 46. One hour after ECMO, the pH averaged 7.35, pO2 76, pCO2 34 with improved hemodynamics. After six hours of ECMO the average pH was 7.45, pOi 96 and pCOi of 33 with a hemodynamically stable patient (Table 1). A-a DO: grathents were improved by 16% within one hour after ECMO in eight of nine patients with a similar change in the ninth patient by the third hour.

Of nine newborns with predictably fatal pulmonary failure who were supported with ECMO at MCV, five had persistent fetal circulation presumably secondary to birth asphyxia. ECMO was maintained from 77 to 313 hours in these nine infants, with six survivors (Table 2). Five survivors have normal neurologic, cardiopulmonary and growth and development parameters. These statistics are gently supported prospectively by four infants who died after meeting ECMO requirements but denied perfusion because the apparatus was in use.

In a much larger group of patients Bart Ie tt has treated 45 with ECMO and 25 have survived.4 Of his 25 survivors 20 are well, with normal growth and development. These results are also strengthened prospectively by a 68% survival in 22 patients in the MAS group with ECMO. It was not used in 11 identical patients and only one survived.

Complications associated with ECMO have been manageable. Bleeding was most commonly seen in those infants requiring operation or chest tubes. Careful maintenance of activated clotting time and the judicious administration of platelets and fresh frozen plasma made serious bleeding an infrequent problem. A single critical complication happened during venous decannulation when an air embolus occurred; this patient has growth and development failure.

The major complication and the most common cause of death in other series is intracranial hemorrhage. Intracranial bleeding is frequently the terminal event in any severely stressed newborn and review does not suggest an increased likelihood of intracranial bleeding in babies on ECMO when compared with other sick newborns.4 OUguric renal failure is reasonably frequent and totally morbid, in that no patient requiring dialysis has survived. Seizures are not unusual and the mortality will be about 50% with ECMO. Finally, mechanical complications from malfunctioning heat exchanger, membrane lung, or the cannulas are unavoidable in large numbers of patients with prolonged ECMO runs, yet to date no death directly related to failure of ECMO technology has occurred.

Three of our patients supported with ECMO died, two with meconium aspiration and one with persistent fetal circulation not associated with Bochdalek hernia. All three perfusions were complicated by oliguric renal failure. Prior to ECMO several other patients have been oliguric. In the patients who survived, the hemodynamic support afforded by ECMO led to the re-establishment of adequate urine flow, but this did not occur in the deaths despite diuretics and osmotic loading. Hemodialysis, which is simplified by ECMO, was not helpful.

Table

TABLE 1ARTERIAL BLOOD GASES: DETERMINATIONS PRIOR TO AND IMMEDIATELY AFTER ECMO

TABLE 1

ARTERIAL BLOOD GASES: DETERMINATIONS PRIOR TO AND IMMEDIATELY AFTER ECMO

Table

TABLE 2RESULTS IN NEONATES SUPPORTED WITH ECMO

TABLE 2

RESULTS IN NEONATES SUPPORTED WITH ECMO

Followup of ECMO patients for 24 months has consisted of a full evaluation of neurologic and pulmonary status, progress of growth and development, non-invasive vascular studies and 2Dechocardiogram. In all of the six survivors chest x-ray and arterial blood gas studies are within normal limits. In six of six patients the Doppler vascular studies show adequate and normal cerebral flow on the right side, either in a prograde or retrograde manner. The 2D echocardiograms do not demonstrate any abnormality suggestive of pulmonary hypertension and are uniformly unremarkable.

Neurologic evaluation and growth and development in five patients show right-handed preference, with brisker reflexes on the left. EEG and CT scans demonstrate no obvious abnormality although there are nonspecific changes noted on both. Obviously the substrate for cortical insult has been created by the hypoxia caused by the basic disease, and carotid ligation is cause for additional concern; therefore further followup is required.

In our sixth survivor neurologic and somatic growth and development have been retarded by cerebral hypoxia and ischemia presumably secondary to air embolization. In conclusion, ECMO has provided life-saving cardiopulmonary support in some neonates with predictably fatal pulmonary failure after current standard therapy has been exhausted. Complications are manageable and normal development appears possible. This partial venoarterial bypass has now been used successfully by three teams in four separate institutions on a very limited patient population. More widespread application and documentation is urgently needed.

REFERENCES

1. Hill JD, O'Bricn TC, Murray JJ, et al: Prolonged e x tracorpo real oxygenation for acute post-traumatic respiratory failure (Shock-Lung Syndrome). N Engl J Med l972;286-629.

2. Bartlett RH, Gazza ni ga AB, Fang SW, et al: Extracorpo real circulation (ECMO) in neonatal respiratory failure. J Thorac Cardiovasc Surg 1977; 74:826.

3. Hardesty RL. Griffith BP, Debski RF, et al: Extracorporeal membrane oxygenation: Successful treatment of persistent fetal circulation. Surgery 1981; 81:556.

4. Bartlett RH, Andrews AF, Toomasian JM, et al: Extracorporeal membrane oxygenation (ECMO) for newborn respiratory failure - 45 cases. Surgery 1982; 92:425.

5. We t m ore N. Mc Ewe n D. O'Conner M, et al: Defining indications for artificial organ support in respiratory failure. Am Soc Artif Inter Organs Trims 1979; 25:459.

6. Raphaely RC. Dowries JJ: Congenital diaphragmatic hernia: Prediction of survival. J Pediair Surg 1973; 8:815.

7. Ormazabal M. K ir k pa t rie k B, Mutiler D: Alteration of ?-a DO: in response to tolazoline as a predictor of outcome in neonates with persistent pulmonary hypertension. Pediair Res 1980; 14:607.

8. Krummd TM, Greenfield LJ, Kirkpatrick BV, et al: Clinical use of an extracorporeal membrane oxygenator in neonatal pulmonary failure. J Pediatr Surg, to be published.

TABLE 1

ARTERIAL BLOOD GASES: DETERMINATIONS PRIOR TO AND IMMEDIATELY AFTER ECMO

TABLE 2

RESULTS IN NEONATES SUPPORTED WITH ECMO

10.3928/0090-4481-19821101-09

Sign up to receive

Journal E-contents