Introduction

Extracorporeal membrane oxygenation (ECMO) is a potentially life-saving technology for newborns with refractory cardiopulmonary failure of reversible etiology. ECMO involves draining deoxygenated venous blood, extracorporeal gas exchange thereby oxygenating and removing carbon dioxide from the blood, and returning the blood via an infusion cannula either into a vein (venovenous ECMO) or artery (venoarterial ECMO; Fig. 82.1 ). Cannulation can be either peripheral, most commonly via the neck or femoral vessels, or central, usually via the right atrium and ascending aorta ( Fig. 82.2 ). In neonates, peripheral cannulation usually occurs via the internal jugular vein and carotid artery, because the femoral vessels are diminutive, technically challenging, and associated with greater complications. Dual lumen single-stage cannulas also exist, which have two separate lumens for infusion and drainage and are usually placed in the neck under fluoroscopic or echocardiographic guidance.

A Patient With a Life-Threatening Congenital Cardiac Anomaly Being Treated With Veno-Arterial Extracorporeal Membrane Oxygenation .

Chest X-Ray of a Neonate Being Treated With Extracorporeal Membrane Oxygenation .

Cannulas are seen in the right atrium and in the ascending aorta.

The use of ECMO in the neonatal population has expanded significantly since its inception in the early 1970s, due to a combination of improved technology, the expanding role of ECMO, and its encouraging outcomes. Indications for ECMO in the neonate can be broadly grouped into those that are intended to manage primary respiratory failure or cardiac failure or those following refractory cardiac arrest (extracorporeal cardiopulmonary resuscitation [ECPR]). This chapter uses the most recent literature to review current indications, applications, complications, and outcomes for ECMO in the neonate and discusses current controversies with respect to patient selection and ECMO management in this critically ill population.

ECMO—Respiratory

Cannulation for respiratory failure composes approximately 20% to 30% of all neonatal ECMO cannulations and usually occurs in the setting of pulmonary hypertension or persistent fetal circulation. ECMO for respiratory support of the neonate is associated with the highest survival percentage among all neonatal and pediatric ECMO support, with 65% to 70% survival to hospital discharge. In terms of specific disease processes, roughly one-third of all neonatal ECMO cannulations for respiratory failure are due to congenital diaphragmatic hernias, which have an average ECMO duration of 12 days and a 50% rate of survival to hospital discharge. Meconium aspiration syndrome represents the second-highest proportion (roughly 25%) of indications for neonatal ECMO, has an average ECMO run of 6 days, and is associated with greater than 90% survival to hospital discharge. Primary pulmonary hypertension represents roughly 20% of all neonatal ECMO cannulations for respiratory failure, has a 7-day average run duration, and is associated with roughly 75% survival to hospital discharge.

Physiologic indications for ECMO cannulation in neonatal respiratory failure include an oxygen index (defined as [mean airway pressure × fraction of inspired oxygen (Fi o ₂ )]/[postductal partial pressure of arterial blood oxygen (Pa o ₂ )] × 100) greater than 40 for more than 4 hours, failure to wean from 100% oxygen despite prolonged (greater than 48 hours) maximal medical therapy for persistent episodes of decompensation, severe refractory hypoxic respiratory failure with acute decompensation (Pa o ₂ less than 40), severe pulmonary hypertension with evidence of right ventricular and/or left ventricular dysfunction, and hypotension resistant to vasopressors or inotropic medications. Absolute contraindications include lethal chromosomal disorders or other lethal anomalies, irreversible brain damage, uncontrolled bleeding, and grade 3 or greater intraventricular hemorrhage. Relative contraindications include a gestational age of less than 34 weeks, birth weight loss of 2 kg, and mechanical ventilation for more than 10 to 14 days.

Complications suffered during ECMO are associated with greater mortality and can roughly be broken down into biomechanical, hemorrhagic, neurologic, cardiovascular, pulmonary, limb, and infectious complications. Among infants undergoing ECMO for respiratory support, 17% experience clots in the oxygenator and an additional 30% experience clots elsewhere in the circuit, 14% have separate cannulation site or surgical site bleeding, nearly 11% undergo significant hemolysis, disseminated intravascular coagulation occurs in 3%, and 1.7% of infants experience gastrointestinal hemorrhage. Fifteen percent of infants experience a neurologic injury, with 7% experiencing infarct and 7.5% experiencing hemorrhage. An additional 7% suffer cardiac tamponade, 4.5% suffer a pulmonary hemorrhage, and less than 1% suffer limb ischemia.

Congenital Diaphragmatic Hernia

ECMO has long been used for neonates with congenital diaphragmatic hernias (CDHs) refractory to maximal conventional therapy. Although criteria for patient selection can be institution-specific, in general, ECMO is used for patients with CDHs who otherwise would not survive due to severe pulmonary hypertension. Criteria that ECMO centers may use to decide whether an infant with a CDH is a candidate for ECMO include an inability to maintain preductal oxygen saturations above 80%, refractory hypotension, persistent metabolic acidosis or lactatemia, peak inspiratory pressures greater than 28 to 30 cm H ₂ O, or mean airway pressures greater than 15 cm H ₂ O.

As with other indications for ECMO due to respiratory failure, relative exclusions may include patients with a gestational age less than 34 weeks or weight less than 2 kg, and absolute exclusions may include patients with legal chromosomal abnormalities and those with irreversible multiorgan failure and severe intracranial hemorrhage. Because improvements in pulmonary hypertension and CDH can take several weeks and may be independent of surgical repair, patients may have prolonged, complicated courses of ECMO and are at increased risk of complications as the duration of the ECMO run increases. Routine neurologic monitoring is indicated, and discussions regarding the futility of care and potential withdrawal of support are indicated in the setting of significant intracranial hemorrhage, worsening clinical status, multiorgan failure, or failure of clinical improvement despite adequate surgical repair and otherwise optimal medical therapy.

Persistent Pulmonary Hypertension of the Newborn

Persistent pulmonary hypertension of the newborn (PPHN) is a disease process characterized by abnormally increased pulmonary vascular resistance soon after birth, with subsequent right-to-left shunting of blood through the fetal circulatory system resulting in severe hypoxemia that is commonly refractory to conventional respiratory support. In severe cases of PPHN (up to 40%) that fail to respond to medical therapy including the use of vasodilatory agents, ECMO may be used to provide adequate gas exchange and prevent or mitigate irreversible lung injury until pulmonary vascular resistance decreases and pulmonary hypertension resolves. Criteria for ECMO in this setting generally consist of an oxygenation index greater than or equal to 40. Patients who are cannulated on ECMO for PPHN have an excellent prognosis, with survival rates exceeding 80%. Shorter duration of ECMO support is associated with improved survival. Prematurity, pre-ECMO pH less than 7.2, pre-ECMO arterial blood oxygen saturation (Sa o ₂ ) less than 65%, and ECMO duration greater than or equal to 7 days were independently associated with mortality in this patient population.

Meconium Aspiration Syndrome

For patients with meconium aspiration syndrome refractory to conventional medical therapy, ECMO may be an otherwise lifesaving therapy because it provides cardiopulmonary support while the underlying pulmonary disease process resolves. ^, ECMO is indicated in approximately 1.5% of all patients who are admitted to the neonatal intensive care unit with meconium aspiration syndrome. The use of ECMO in this setting is associated with a favorable prognosis (greater than 90% survival to discharge), because patients tend to do well once the reversible sequela of the aspiration syndrome resolve, the lungs improve, and the child is able to successfully wean from the ECMO circuit and survive to discharge.

ECMO—Cardiac

Congenital heart disease is the most common reason neonates are cannulated on ECMO for cardiac failure. Hypoplastic left heart syndrome, cyanotic defects with decreased or shunted pulmonary flow, left or right ventricular outflow tract obstruction, and septal defects can all result in refractory cardiac arrest in which ECMO support prior to definitive repair is associated with increased survival. The average ECMO run duration for congenital heart disease is 6 days, and 40% to 50% of all patients supported on ECMO in the setting of congenital heart disease survive to hospital discharge. Additional indications for cardiac ECMO include myocarditis, cardiomyopathy, cardiogenic shock, and cardiac arrest, which together compose less than 10% of all neonatal cardiac ECMO runs and are associated with a 40% to 50% rate of survival to hospital discharge.

As with ECMO in the setting of respiratory failure, physiologic indications exist for cardiac ECMO and include refractory cardiogenic shock (generally seen in the setting of cardiomyopathy, acute myocarditis, or cardiac dysfunction and severe sepsis), postoperative refractory cardiac failure (such as postcardiotomy shock when neonates fail to wean from cardiopulmonary bypass, postoperative low cardiac output syndrome, refractory cardiac arrhythmias, and pulmonary retention), cardiac arrest refractory to conventional CPR (ECPR; see next section), procedural support (during heart or lung transplantation), or as a bridge to heart or lung transplantation or durable mechanical circulatory support such as a ventricular assist device. No absolute contraindications exist specific to cardiac ECMO, but relative contraindications, which significantly overlap with those for ECMO and respiratory failure, include end-stage primary disease or incurable malignancy with poor prognosis, multiorgan failure, severe neurologic injury or intracranial hemorrhage, uncontrolled visceral bleeding, prematurity (less than 34 weeks’ gestational age), small size (less than 2 kg), chromosomal abnormalities, and patient care directives specifying against or limiting ECMO use. ECMO is also generally not recommended in this population if the neonate has or will have significant residual lesions after cardiac surgery or is unlikely to be a transplant candidate.

In addition to characteristics that are relative contraindications to cannulation, several patient- and disease-specific characteristics are associated with mortality in neonatal patients supported by ECMO for cardiac failure. Prior to cannulation, patients with single-ventricle physiology, myocarditis, or cardiomyopathy, patients requiring high levels of inotropes or with profound acidosis (pH less than 7.2) or high lactate, patients requiring CPR, and neonates with evidence of renal failure or fluid overload are likely to suffer higher mortality once cannulated onto ECMO. Once on ECMO, patients who fail to clear lactate within 24 hours, remain fluid overloaded, experience bleeding complications, suffer extracardiac complications including stroke, intracranial hemorrhage, or renal failure, or remain on ECMO for more than 7 days, experience significantly high mortality.

Conversely, several characteristics prior to ECMO cannulation exist that are associated with lower mortality. Neonates with larger body weights (greater than 3.3 kg), without chromosomal abnormalities, with two-ventricle physiology, with low inotropic requirements, with a duration of ventilation less than 14 days prior to ECMO, and without significant acidosis or elevated lactates as well as who do not require CPR and do not have renal failure or extracardiac involvement prior to ECMO experience lower mortality. Similarly, neonates who clear lactate within 24 hours of being cannulated onto ECMO, those who do not suffer the above-mentioned bleeding or extracardiac systemic complications, and those who have a duration of ECMO support less than 5 days are more likely to survive until hospital discharge.

Among all neonates supported by ECMO for cardiac failure, approximately 1% experience oxygenator clots, with an additional 25% to 30% experiencing clots elsewhere in the ECMO circuit. Compared with patients supported for respiratory failure, a much higher proportion experience surgical site bleeding (30% versus 6%) and cannulation site bleeding (11% versus 8%), 11% experience significant hemolysis, and 4% suffer disseminated intravascular coagulation; gastrointestinal hemorrhage occurs in 1%. Additionally, infants supported on ECMO for cardiac failure experience high rates of intracranial hemorrhage (11% versus 7.5%) and lower rates of intracranial infarction (3% versus 7%) compared with neonates supported for respiratory failure; 6% experience additional cardiac complications, 5% experience pulmonary hemorrhage, and 0.2% develop limb ischemia.

ECMO—ECPR

Extracorporeal cardiopulmonary resuscitation, or ECPR, is defined as the rapid deployment of venoarterial ECMO to provide circulatory support in patients in whom conventional CPR is unsuccessful in achieving sustained return of spontaneous circulation, which is deemed to have occurred when chest compressions are not required for 20 consecutive minutes and signs of circulation persist. ECPR is different than venoarterial ECMO in that with ECPR, ECMO is in the setting of conventional CPR, whereas venoarterial ECMO is initiated for low cardiac output either without cardiac arrest or following sustained return of spontaneous circulation. Neonatal ECPR is associated with a 35% to 45% rate of survival to hospital discharge.

Survival among neonates who suffer out-of-hospital cardiac arrest is more favorable compared with those who suffer in-hospital cardiac arrest. According to American Heart Association guidelines, there currently is insufficient evidence for or against the use of ECPR for infants with noncardiac diagnoses who suffer in-hospital cardiac arrest. ^, However, ECPR may be considered for infants with cardiac diagnoses who have in-hospital cardiac arrest and for institutions in settings with existing ECMO protocols, expertise, and equipment (class IIb). ^, In a study looking at outcomes of 593 cases of pediatric ECPR in the setting of in-hospital cardiac arrest at 32 hospitals between 2010 and 2014, mortality remained relatively constant, with overall mortality of 59% (60% survival to ECMO decannulation and 41% survival to hospital discharge). Neonates composed more than one-third of the study population.

Those who survive are more likely to have a primary medical or surgical etiology of cardiac arrest and have a shorter duration of arrest (42 minutes for survivors of hospital discharge versus 51 minutes for nonsurvivors). Additionally, odds of death increase with the presence of renal insufficiency and any adverse events or complications during ECMO and with an increasing number of adverse events during ECMO. Neurologic injury is the complication most associated with mortality; patients who suffer neurologic complications have nearly three times the odds of death compared with those who do not. The time of day (weekday versus weekend) of the arrest, the location of in-hospital arrest, whether the arrest was witnessed or not, and the first documented rhythm are not associated with survival. Among those who survive, the average duration is 4 days, and more than 80% of patients experience an adverse event (compared with 90% of nonsurvivors). These complications include cardiovascular (56%), hemorrhagic (33%), metabolic (26%), and renal (25%) complications, seizures (10%), intracranial hemorrhage (8%), cerebral infarction (3%), infections (6%), and pulmonary complications (4%). However, among those who do survive, 93% have a favorable neurologic outcome at hospital discharge. ^, In a study looking at up to 1-year survival and neurobehavioral outcomes among infants and children who were supported with ECPR after undergoing in-hospital cardiac arrest, 85% had a preexisting cardiac condition, 51% had postcardiac surgery, 57% were less than 1 year of age, 78% had a CPR duration greater than 30 minutes, more than 40% survived to 12 months after hospital discharge, and 30% survived to 12 months with a favorable neuropsychological outcome.

Neurologic Outcomes

As previously mentioned, neurologic complications among neonates supported by ECMO are common and associated with worse prognoses. Acute neurologic injury, which includes hypoxic-ischemic injury, intracranial hemorrhage, ischemic stroke, and clinical or electroencephalographic evidence of seizures, occurs in 10% to 35% of infants and increases mortality by roughly 100%. ^{, ,} Furthermore, infant and child ECMO survivors suffer lasting neurologic outcomes. A systematic review of neurologic outcomes after ECMO in children ages 0 to 18 years looked at studies consisting of patients with CDHs, cardiac disease, cardiac arrest, and next populations and found that 10% to 50% of children’s performed more than two standard deviations below the mean on cognitive testing. The brain is particularly susceptible to injury during ECMO due to the pre-ECMO hypoxic-ischemic injury suffered during cardiopulmonary failure cardiac arrest, combined with the inflammatory state of the ECMO circuit and the need for systemic anticoagulation. As such, accurate and timely neuromonitoring is particularly important. Neuromonitoring can be done via a variety of modalities, including cross-sectional imaging, Doppler ultrasound, electroencephalography, cerebral oximetry, and awake neuromonitoring, which is more common in neonates than in children or adults.

Controversies and Ethical Considerations

As previously discussed, gestational age <34 weeks and weight <2 kg have historically been contraindications to ECMO cannulation, due in part to technical challenges as well as to overall poor prognoses including increased risk of intraventricular hemorrhage and death. However, recent studies have challenged these criteria. A study by Church et al. looking at outcomes of patients with a gestational age of 34 weeks or less cannulated onto ECMO found no difference in intracranial hemorrhage between patients with a gestational age of 29 to 33 weeks compared with 34 weeks (21% versus 17%). However, there was a statistically significant higher incidence of cerebral infarct (22% versus 16%; P = .03) and overall lower survival among patients of 29 to 33 weeks’ gestational age compared with 34 weeks (48% versus 58%; P = .05). Intracranial hemorrhage and survival were not associated with gestational age during logistic regression analysis. The authors concluded that the modestly worse differences seen in the cohort of lower gestational age may be clinically acceptable, arguing against using a gestational age of <34 weeks as an absolute contraindication to ECMO.

In a retrospective study of the ELSO database looking at infants weighing less than or greater than 2 kg at the time of ECMO cannulation, Rozmiarek et al. found that although overall survival was lower in patients weighing less than 2 kg (53% versus 77%; P < .001), a survival rate of 40% or greater could be achieved in infants who weighed as little as 1.6 kg. Judicious use of anticoagulation in this population could also improve survival, because bleeding was associated with a significantly lower survival rate. In a separate study looking at extremely low birth weight infants with congenital diaphragmatic hernias supported with ECMO, Delaplain et al. found increased mortality among infants who weighed less than 2 kg but not among those less than 34 weeks’ gestational age.

There is no shortage of ethical considerations related to the deployment and use of ECMO, particularly in infants. Because ECMO is a highly resource-intensive therapy, deliberate and thoughtful consideration should be made for each potential infant that could potentially benefit from ECMO. We recommend a multidisciplinary group of clinicians and stakeholders for consideration and routine reviewing of ECMO candidates and institutional criteria for cannulation, with liberal use of ethics committees when deemed appropriate.

Once infants are cannulated onto ECMO, decisions related to withdrawal of support can be particularly challenging. Early and frequent discussions of the family’s specific goals and frequent evaluation and reevaluation of the abilities of ECMO to meet those goals throughout the patient’s ECMO course are particularly important. Additional ethical considerations include expanding ECMO to broader indications and expanded patient populations despite a lack of clear evidence regarding prognosis and outcomes; the role of ECMO in brain-dead potential organ donors; the role of regional ECMO centers; and decisions surrounding futility or continuation of ongoing care and/or ECMO support. ^,