Regardless of extracorporeal life support mode, following cannulation, the venous cannula is connected to the tubing of the circuit allowing the drainage of blood from the body and into the extracorporeal life support pump (either roller pumps or centrifugal pumps can be used). The pump circulates blood forward to a semi-permeable membrane lung. In the membrane lung, gradient-related diffusion of gases, namely oxygen and carbon dioxide, occurs. Oxygen diffuses from the sweep gas being applied to the membrane into the blood; carbon dioxide diffuses from the blood into the sweep gas. The now-oxygen rich blood is warmed during its flow through a heater and infused into the aorta via the carotid artery cannula placed to provide venoarterial extracorporeal life support. In the case of venovenous extracorporeal life support, this oxygen-rich blood is infused into the right atrium via the infusion lumen of the double lumen venous cannula. During venoarterial extracorporeal life support, the central venous oxygen saturation can be continuously monitored by a fiber-optic catheter placed in the venous side of the circuit.
The venous drainage limb and the arterial infusion limb of the circuit are connected by a short stretch of tubing called the “bridge,” which is kept clamped during extracorporeal life support. In an emergency, if the patient has to come off extracorporeal life support, the arterial and venous cannulae are clamped proximal to the bridge, and the bridge is unclamped allowing blood to continue to flow through the circuit via the bridge (but without reaching the patient), preventing stasis and clotting.
Obstruction of the venous drainage can create increased negative pressure in the venous drainage limb of the circuit, allowing air to be drawn into the circuit. To avoid this, negative pressure in the venous limb of the circuit is sensed by a pressure transducer applied across a compliant piece of the circuit called a “bladder.” This mechanism can either slow down or shut off the pump if there is excessive negative pressure on the venous side of the circuit, thereby decreasing the risk of air entry into the circuit and air embolism.
Following cannulation, blood flows are initiated and slowly advanced to desired rates. During neonatal venoarterial extracorporeal life support, the initial flow is set to approximately 120 mL/kg/min, providing approximately 80% of neonatal cardiac output. In venovenous extracorporeal life support, because recirculation of the blood can occur in the right atrium due to the double lumen catheter providing drainage and forward flow, the flow requirement for adequate oxygen delivery is higher, 140 to 150 mL/kg/min. Once the desired flow is achieved, ventilator settings are decreased to “rest settings” limiting baro- and volutrauma to the lungs and allowing rest and time for recovery from the underlying disease. In venovenous extracorporeal life support, the ventilator settings are decreased gradually over the 24 to 36 hours following cannulation rather than rapidly (as in venoarterial extracorporeal life support).
To monitor the patient’s progress on extracorporeal life support, the adequacy of the patient’s oxygenation is determined by serially measuring partial pressure of arterial oxygen (pa02), pulse oxygen saturation, and, in venoarterial extracorporeal life support, the mixed venous oxygen saturations. Anticoagulation is required given blood’s propensity to clot on interfacing with the nonbiological material of the extracorporeal life support circuit, and continuous heparin infusion is used in most centers. Close monitoring of multiple anticoagulation laboratory markers is required. Head ultrasounds are performed daily for the first days of extracorporeal life support in neonates and then as clinically indicated to detect intracranial hemorrhage. Although nutrition is usually exclusively provided by parenteral nutrition, reports indicate that nutrition may be provided at least partially by enteral route for patients receiving extracorporeal life support (18). Continuous renal replacement therapy (CRRT) can be used to prevent or manage fluid overload though the ideal time for CRRT initiation is unknown.
Indications
Extracorporeal life support is standard of care for term and late preterm infants with hypoxic respiratory failure who have failed to improve with other standard medical interventions (31). Unfortunately, exact criteria for neonatal respiratory extracorporeal life support are difficult to establish and failure to respond to other medical therapies, as judged by the neonatal provider, is commonly the indication. With the advent of high-frequency ventilation, exogenous surfactant, and inhaled nitric oxide, the “other medical therapy” options are significantly increased from the initial days of neonatal extracorporeal life support, and compared to historical cohorts, extracorporeal life support is less commonly applied for respiratory disease in neonates presently.
Calculation of the oxygenation index (oxygenation index = [(mean airway pressure x Fi02 x 100)/postductal pa02] can aid in quantifying disease severity. In the past, the extracorporeal life support qualifying criteria for neonates included those with predicted mortality greater than 80% if conventional medical management is continued without extracorporeal life support. In the modern era, infants with an oxygenation index of more than 25 should be cared for in a center capable of providing neonatal extracorporeal life support and if not at such a center, should be transferred expediently. An oxygenation index greater than 40 to 45 is a commonly applied indication of neonatal respiratory extracorporeal life support (31).
Relative eligibility and qualifying criteria for neonatal extracorporeal life support secondary to respiratory failure are found in Tables 1a and 1b.
Table 1a. Neonatal Extracorporeal Membrane Oxygenation Eligibility Criteria
Eligibility criteria: |
| • Birth weight greater than 2 kilograms
• Gestational age older than 34 weeks
• Reversible cardiopulmonary disease
• Absence of intracranial hemorrhage greater than Papile grade III
• Absence of irreversible organ damage
• Absence of lethal anomalies
• Pre-extracorporeal life support ventilation 14 days or longer |
Adapted from (31).
|
Table 1b. Neonatal Extracorporeal Membrane Oxygenation Qualifying Criteria
Qualifying criteria: |
| Respiratory failure and cardiovascular shock with at least one of the following: |
| • Oxygenation index consistently greater than 40 for 0.5 to 6 hours or more, greater than 20 with lack of improvement despite maximal medical therapy
• Partial pressure of arterial oxygen (Pa02) less than 40 mm Hg for longer than 2 to 12 hours, despite intervention
• Metabolic acidosis with shock (pH < 7.25) for 2 hours or more with hypotension
• Progressive, refractory pulmonary hypertension with either right ventricular dysfunction or high inotropic requirements |
Adapted from (31).
|
For neonates with cardiac disease, extracorporeal life support indications are broadly categorized as per-operative stabilization extracorporeal life support, perioperative extracorporeal life support, and extracorporeal life support remote from surgical intervention but specific criteria are lacking (31). Preoperatively, extracorporeal life support can be used to optimize hemodynamics and oxygen delivery and expedite surgical repair. Perioperatively, extracorporeal life support is used for infants who are unable to successfully separate from cardiopulmonary bypass in the immediate postoperative period. In patients remote from surgical intervention, those with nonsurgical cardiac diseases, or patients awaiting cardiac transplant, extracorporeal life support is frequently utilized.
Some of the newer applications of extracorporeal life support include the following: (1) extracorporeal cardiopulmonary resuscitation; (2) initiating extracorporeal life support after cesarean section delivery and prior to clamping the umbilical cord (ie, ex utero intrapartum treatment procedure) in neonates diagnosed prenatally with life-threatening airway anomalies or severe congenital diaphragmatic hernia (21); and (3) for the administration of systemic hypothermia therapy using the extracorporeal life support circuit in a neonate with hypoxic ischemic encephalopathy who is requiring extracorporeal life support for hypoxic respiratory failure (25).
Extracorporeal cardiopulmonary resuscitation, or extracorporeal support cannulation to aid in CPR, is an increasingly common indication for extracorporeal life support, particularly in infants with cardiac disease. In the cases of extracorporeal cardiopulmonary resuscitation detailed in the Pediatric ELSO Registry International Report, nearly all cardiac arrests were witnessed with more than 80% occurring in the intensive care unit, operating room, or emergency department (04).
Contraindications
Extracorporeal life support remains a lifesaving but “high risk and resource intense intervention” and, thus, should be utilized in infants with a high likelihood of meaningful survival and quality of life after extracorporeal life support (31). Infants with complicated pathologies, regardless of the degree of respiratory failure or shock, should not be considered for extracorporeal life support. Strict contraindications to neonatal extracorporeal life support include (but are not limited to): significant hemorrhaging (including intracranial hemorrhages > Papile III), irreversible organ damage, severe preexisting brain damage, the presence of lethal anomalies, and the presence of irreversible extra pulmonary organ damage. Relative contraindications include weight less than 2 kg, gestational age less than 34 weeks, and prolonged duration of mechanical ventilation prior to extracorporeal life support.
Results
Initially in respiratory diseases process such as persistent pulmonary hypertension or meconium aspiration syndrome, the lungs are typically noncompliant and appear opaque on chest radiographs. Typically, 24 to 36 hours after initiating extracorporeal life support as the infant has stabilized, lung conditioning including suctioning of the endotracheal tube, and administration of deep breaths every 4 to 6 hours, is begun. If the neonate is receiving extracorporeal life support for cardiac disease, lung conditioning is begun once definitive surgical treatment is complete. In neonates who receive extracorporeal life support for congenital diaphragmatic hernia, some improve dramatically with extracorporeal life support, receive lung conditioning as they improve, and move toward decannulation before surgical intervention is pursued; other infants require congenital diaphragmatic hernia surgical repair on extracorporeal life support, and lung conditioning is not begun until stabilization after surgery. Regardless of disease process, recovery from underlying lung disease is associated with improvement in aeration of lungs, improved lung compliance as well as increased pa02, pulse oxygen saturation, and venous oxygen saturation.
Extracorporeal life support flow can then be decreased gradually as respiratory support via the ventilator is increased. Eventually the patient undergoes decannulation and if successful, is supported by mechanical ventilation. CT scans or MRI of the brain, and in some centers, an EEG, are done prior to discharge.
Extracorporeal life support neonatal outcomes. The short- and long-term outcomes in survivors of neonatal extracorporeal life support are related to pre-extracorporeal life support and extracorporeal life support variables. Pre-extracorporeal life support conditions, namely primary disease, pre-extracorporeal life support cardiopulmonary resuscitation, hypoxia, acidosis, hypotension, and management strategies such as hyperventilation and alkalosis will affect the outcome. With the exception of one randomized trial done in the United Kingdom in which the control cases were critically ill neonates not randomized to extracorporeal life support, follow-up studies have compared extracorporeal life support outcomes with either near-miss extracorporeal life support cases or noncritical neonates. There have been no randomized controlled trials comparing short- and long-term outcomes of venoarterial versus venovenous extracorporeal life support. However, survival is decreased in infants who require conversion from venovenous to venoarterial compared to infants undergoing venoarterial or venovenous extracorporeal life support alone without conversion (20).
In the most recent Pediatric ELSO Registry International Report, neonatal respiratory and cardiac extracorporeal life support cases from 2009 to 2015 are detailed by primary diagnosis, survival, and average run length (Table 2) (04).
Table 2. Neonatal Respiratory and Cardiac Extracorporeal Life Support Cases From 2009-2015
ECLS disease category and primary diagnoses | Proportion of cases, n (%) | Survival to discharge | Average run duration, days |
Neonatal respiratory ECLS | CDH | 1851 (32) | 50 | 12 |
MAS | 1393 (24) | 93 | 6 |
PPHN | 1233 (21) | 74 | 7 |
RDS | 71 (1) | 79 | 6 |
Sepsis | 256 (4) | 55 | 7 |
Other | 1035 (18) | 59 | 9 |
Total respiratory cases | 5839 | 68 | 9 |
Neonatal Cardiac ECLS | CHD | 2301 (81) | 44 | 6 |
Hypoplastic left heart syndrome | 644 (23) | 40 | 6 |
Left ventricular outflow tract obstruction | 178 (6) | 41 | 6 |
Right ventricular outflow tract obstruction | 95 (3) | 39 | 6 |
Septal defects | 172 (6) | 44 | 6 |
Cyanotic with decreased pulmonary blood flow | 348 (12) | 48 | 7 |
Cardiac arrest | 41 (1) | 41 | 7 |
Cardiogenic shock | 57 (2) | 39 | 5 |
Cardiomyopathy | 44 (2) | 59 | 9 |
Myocarditis | 38 (1) | 50 | 11 |
Other | 368 (13) | 50 | 7 |
Total cardiac cases | 2849 | 45 | 6 |
Adapted from (04)
|
Survival. ELSO provides an annual, international report detailing ELSO-center extracorporeal life support cases worldwide. Table 3, adapted from this report, details types of extracorporeal life support and survival rates (13).
Table 3. Neonatal and Pediatric Extracorporeal Life Support Cases and Survival through 2021
ECLS population and underlying process | Total runs | Survived ECLS | Survival to discharge or transfer |
Neonatal ECLS, n (%) | Pulmonary | 34,239 | 29,972 (87) | 25,005 (73) |
Cardiac | 10,233 | 7085 (69) | 4546 (44) |
Extracorporeal cardiopulmonary resuscitation | 2439 | 1700 (69) | 1042 (42) |
Pediatric ECLS, n (%) | Pulmonary | 11,935 | 8685 (72) | 7295 (61) |
Cardiac | 15,230 | 11,074 (72) | 8293 (54) |
Extracorporeal cardiopulmonary resuscitation | 6182 | 3647 (58) | 2619 (42) |
Adapted from (ELSO Registry Report: International Summary – October 2022)
|
Extracorporeal cardiopulmonary resuscitation outcomes. A study from the American Heart Association’s Get With The Guidelines - Resuscitation registry compared outcomes between patients who received pediatric extracorporeal cardiopulmonary resuscitation and those who received conventional CPR and found extracorporeal cardiopulmonary resuscitation was associated with improved survival to hospital discharge (OR 2.80; 95% CI 2.13-3.69, p < 0.011) and survival with favorable neurologic outcome (OR 2.64; 95% CI 1.91-3.64, p < 0.001) (22).
Lung functional outcome. Despite respiratory failure being the most common indication for neonatal extracorporeal life support, long-term lung function is typically minimally impacted. In infants enrolled in the only neonatal randomized controlled trial of extracorporeal life support (the UK trial), extracorporeal life support receivers had slightly better lung function than conventionally treated children at 1-year follow-up (08). However, maximal exercise tolerance and endurance at school age in extracorporeal life support survivors deteriorated overtime regardless of underlying extracorporeal life support diagnosis (34).
Renal outcomes. Acute kidney injury is unfortunately common among neonates treated with extracorporeal life support (16). The Kidney Intervention during Extracorporeal Membrane Oxygenation (KIDMO) study group, a multidisciplinary group of international investigators, studies acute renal complications and renal replacement therapy utilization during neonatal and pediatric extracorporeal life support and has worked to quantify the renal effects of extracorporeal life support. In a follow-up study of 169 neonatal extracorporeal life support survivors with a median age of 8.2 years, Zwiers and colleagues found that 54 patients (32%) had at least one sign of chronic kidney disease or hypertension (36). Though the chronic kidney disease values were marginally abnormal, they raised a concern of risk of more rapid decline in kidney function with increasing age and, hence, a need for screening for chronic kidney disease in adulthood.
Neurologic outcomes. Neurologic abnormalities are unfortunately common among extracorporeal life support survivors; depending on the neuroimaging technique, the hemorrhagic or nonhemorrhagic abnormalities have been reported to occur in 28% to 52% of neonates who have undergone extracorporeal life support. The severity of these abnormalities is predictive of the degree of impairment of neuropsychological status at 5 years of age. However, due to brain plasticity and compensation, even children with moderate to severe neuroimaging abnormality may not be disabled (11).
In a follow-up study of infants with extracorporeal life support-related left-sided seizures, Campbell and colleagues found no evidence of lateralization of neurologic disability (12). However, seizures associated with extracorporeal life support placed these infants at higher risk for cerebral palsy and developmental delays at 2 and 5 years of age than those without seizures. At 5 years of age, they were also noted to have lower IQ and speech language delays (30).
The United Kingdom collaborative extracorporeal life support trial group reported neurodevelopmental follow-up at 1 year of age in surviving extracorporeal life support patients (n=62) and control infants managed with conventional therapy (n=37) (09). Two infants (1 in each group) were severely disabled (01). Fifteen had evidence of tone changes in the extremities (16% extracorporeal life support, 13.5% conventional). There was no evidence of lateralization of signs. Neurodevelopmental examination at 4 years of age in surviving infants available for follow-up showed that there were more survivors without disability in the extracorporeal life support group versus the conventional therapy group (50% extracorporeal life support vs. 37% conventional). Sensorineural hearing loss, distributed equally in both groups, occurred in 11% of the children (09). Clinical variables that were predictive of poorer outcome amongst survivors (extracorporeal life support and controls) were seizures or supplemental oxygen at discharge. In infants randomized to extracorporeal life support, episodes of sepsis, inability to establish full oral feedings by 14 days of age, or hospital stay greater than 30 days were associated with increased risk of disability (10). Neurodevelopmental outcome at 7 years of age of children enrolled in the United Kingdom trial showed that the majority (76%) had normal cognitive function. Learning problems were similar in extracorporeal life support and control groups with notable difficulties with spatial and processing tasks (26).
In a population-based study of 5-year-old neonatal venoarterial extracorporeal life support survivors born between 1993 and 2000 in the Netherlands, Nijhuis-van der Sanden and colleagues tested 86% of the extracorporeal life support survivors for motor skills, intelligence, and child behavior (28). They reported that 49% were normal in all domains whereas severe disabilities were present in 13%; an additional 9% had impaired motor development combined with cognitive or behavioral problems. Interestingly, the follow-up of extracorporeal life support survivors born between 1996 and 2004 in the Netherlands showed that the motor performance declined at 12 years of age. The motor performance at 5, 8, and 12 years of age was normal in 73.7%, 74.8%, and 40.5%, respectively, highlighting the need for long-term follow-up as new developmental milestones are acquired (35). A 7-year follow-up assessed 90 of 100 children and recorded that 66 patients had a cognitive level that was within normal limits. The distribution of learning difficulties was similar between groups. A higher level of behavioral problems was noted in conventionally treated children, and there was some evidence of reduced morbidity in the extracorporeal life support group (26).
The reported incidence of sensorineural hearing loss in different studies has varied from 3% to 21% in extracorporeal life support survivors as compared to 0% to 37% reported in infants who were treated with conventional therapy without extracorporeal life support. Fligor and colleagues reported that the diagnosis of congenital diaphragmatic hernia, length of extracorporeal life support run greater than 160 hours, and duration of aminoglycoside therapy longer than 14 days were associated with increased risk of sensorineural hearing loss (17). In a population of post- extracorporeal life support infants followed at 9 to 13 years of age, Murray and colleagues found that out of eight infants with sensorineural hearing loss, five were identified by hearing screening prior to discharge from the neonatal intensive care unit, one was identified at 1 year of age, and two were diagnosed between 9 and 13 years of age (27). Also, the occurrence of pre-extracorporeal life support clinical seizures and the duration of extracorporeal life support were independently associated with hearing loss. Thus, the sensorineural hearing loss can be late onset and progressive, underlying the importance of continued vigilance for its detection during follow-up.
Long-term neuropsychological outcomes. A survey of young adult survivors of neonatal extracorporeal life support demonstrated that most are satisfied with their lives, working or in college, in good health, and have families (14). However, neuropsychologic assessment at between 17 and 18 years of age indicates that verbal, visual-spatial, and working memory problems persist into late adolescence for neonatal extracorporeal life support survivors (24).
Adverse effects
Complications occurring during extracorporeal life support. Table 4 lists some of the more frequent mechanical and patient complications reported by the ELSO Registry in neonates placed on extracorporeal life support between 2009 and 2015 (04).
Table 4. Neonatal Extracorporeal Life Support Complications
ECLS complications | Neonatal respiratory ECLS | Neonatal cardiac ECLS | Neonatal eCPR |
Mechanical complications, n (%) | Oxygenator failure | 280 (5) | 123 (4) | 26 (3) |
Pump malfunction | 84 (1) | 37 (1) | 12 (1) |
Cannula problem | 696 (12) | 156 (5) | 61 (7) |
Air in circuit | 209 (4) | 101 (3) | 58 (6) |
Patient complications, n (%) | Seizures (by EEG) | 158 (3) | 100 (4) | 54 (6) |
Cerebral infarct | 180 (3) | 93 (3) | 46 (5) |
Intracranial hemorrhage | 643 (11) | 326 (11) | 128 (14) |
Brain death | 23 (0.4) | 21 (1) | 14 (2) |
Cardiac tamponade | 13 (0.2) | 148 (5) | 20 (2) |
Surgical site bleeding | 386 (7) | 739 (26) | 150 (17) |
Gastrointestinal hemorrhage | 89 (2) | 35 (1) | 12 (1) |
Amputation | 0 (0) | 3 (0.1) | 0 (0) |
Adapted from (04).
|
Cardiac stun. Cardiac stun, severely diminished to absent myocardial function with decreased or absent arterial pressure wave form, can occur transiently during the first 24 to 48 hours of venoarterial extracorporeal life support.
Coagulation and bleeding. Thrombocytopenia occurs frequently during extracorporeal life support and is managed with the infusion of platelets. Clots in the circuit can adversely affect circuit function and pose risk of infection, systemic and pulmonary emboli, and coagulopathy. The risk of intracranial hemorrhage-infarction is related to a variety of underlying factors, such as systemic heparinization, ligation of carotid artery, ligation of jugular vein, and impaired autoregulation of cerebral blood flow.
Special populations
Congenital diaphragmatic hernias. Morbidity and mortality continue to remain high in infants with congenital diaphragmatic hernia requiring extracorporeal life support. These infants frequently require extracorporeal life support.
Hypoxic ischemic encephalopathy. Therapeutic hypothermia has been shown to provide neuroprotection in term infants with perinatal asphyxia. However, a multicenter randomized controlled trial undertaken to determine whether systemic hypothermia (cooling to 34ºC) would improve the neurologic outcome in neonatal extracorporeal life support cases (NEST trial) did not show improved outcomes up to 2 years of age (15).
COVID-19. The initial ELSO Registry reports describing extracorporeal life support use for COVID-19 reported in-hospital mortality rates (37.4%, 95% CI 34.4%-40.4%) comparable to outcomes after extracorporeal life support for acute respiratory distress syndrome unrelated to COVID-19, but the cohort was limited to patients treated prior to May 1, 2020 (02). As Barbaro and colleagues suggest, these findings imply providers could make similar clinical judgements about extracorporeal life support candidacy and use in patients with respiratory failure regardless of COVID-19 status, but comparisons to outcomes among a more modern cohort suggest extracorporeal life support receivers after May 1, 2020, experienced increased mortality (approximately 15%); most extracorporeal life support–treated patients with COVID-19 died (03). Ultimately, it remains unclear which patients are the best candidates for extracorporeal life support in the setting of COVID-19.
Summary
In summary, extracorporeal life support not only improves survival in a population of critically ill neonates, but the long-term outcome is also good in the majority of extracorporeal life support survivors. Although severe neuromotor abnormalities may be detected earlier, school aged follow-up is necessary to identify children at risk for scholastic problems and in need of appropriate intervention. New extracorporeal life support guidelines for follow-up after neonatal and pediatric extracorporeal life support detail recommendations for structured, neurodevelopmental follow-up of these patients, with early identification and prompt interventions for delays (19).
Clinical vignette
A 3.5 kg male infant was born by spontaneous vaginal delivery to a group B strep culture-positive female following rupture of membranes of 24 hours duration. The mother received two doses of ampicillin prior to delivery. The APGAR scores were 5 and 9 at 1 and 5 minutes, respectively. During the immediate neonatal course, the patient required positive pressure ventilation with 100% oxygen as well as ampicillin and gentamicin for suspected sepsis. Pulmonary hypertension was present requiring high frequency oscillatory ventilation and inhaled nitric oxide. The infant received exogenous surfactant and required chest tube placement for pneumothorax. Despite maximal medical therapy, the baby’s oxygenation worsened, and he was transferred to an extracorporeal life support center. Because the baby’s pa02 was less than 50 mm Hg, his oxygenation index was greater than 40, and echocardiogram confirmed the presence of pulmonary hypertension with a structurally normal heart, he was placed on venoarterial extracorporeal life support. Head ultrasound prior to extracorporeal life support did not demonstrate intracranial hemorrhage.
Following cannulation, the patient’s oxygenation improved rapidly, and the baby was supported on bypass flow of 300 mL/min (approximately 85 ml/kg/min) with mechanical ventilation at lung rest setting: oxygen requirement was 30%, peak inspiratory pressure/peak end-expiratory pressure was 20/10 cm H20, ventilator rate 10 per minute. Initial opacification “white out” of the lungs improved over time.
