Introduction

What is extracorporeal life support and extracorporeal membrane oxygenation?

Extracorporeal life support (ECLS) describes prolonged cardiopulmonary bypass providing hemodynamic and respiratory support in critically ill patients who cannot be supported by conventional measures. ECLS and extracorporeal membrane oxygenation (ECMO) are often used interchangeably. ECMO was first reported in the 1970s and the term has remained [1]. Since support includes more than just oxygen delivery, ECLS is now the preferred terminology. Patients on ECLS receive improved O₂ delivery and CO₂ removal with augmentation of cardiac output and support of hemodynamics. Unlike a bypass pump used in cardiac surgery, ECLS provides prolonged support for days to weeks. In its simplest form, during ECLSdeoxygenated blood is removed from the patient, perfused through a porous membrane (the oxygenator) where gas exchange occurs, a heat exchanger which warms the blood, followed by return of warmed, arterialized blood to the patient.

simplified venoarterial ECLS circuit

Simplified venoarterial extracorporeal life support circuit showing basic components of pump, oxygenator, heat exchanger and cannulas taking flow both to and from the patient. (image courtesy of the Extracorporeal Life Support Organization)

simplified venovenous ECLS Circuit

Simplified venovenous extracorporeal life support circuit showing basic components of pump, oxygenator, heat exchanger and cannulas taking flow both to and from the patient. (image courtesy of the Extracorporeal Life Support Organization)

What is the goal of extracorporeal life support?

ECLS is a supportive therapy and not designed to cure the underlying disease [2]. While its use is expanding to include more complex patients, it should not be used in patients with irreversible pathology. The goal of ECLS is to support respiratory and cardiovascular function while resting the heart and lungs and providing improved oxygen delivery to end organs while the patient recovers. It allows for lung protective strategies, thus minimizing secondary injury from barotrauma, volutrauma or ongoing hypoperfusion. While neonates used to comprise the largest number of patients benefiting from ECLS in cases of severe respiratory failure, by far adult ECMO for respiratory failure forms the largest cohort each year. Typical neonatal problems which result in respiratory failure that requires ECMO include persistent pulmonary hypertension of the newborn, meconium aspiration, respiratory distress syndrome , sepsis, pneumonia, air leak syndrome, and congenital diaphragmatic hernia. Pediatric ECLS is generally used in the acute respiratory distress syndrome (ARDS) associated with viral or bacterial pneumonia and profound septic shock from other etiologies [3].

What is the Extracorporeal Life Support Organization?

Extracorporeal life support organization (ELSO) is a collaborative, voluntary registry established in 1989 to gather data on use of ECLS across the globe. As of 2015 there were over 250 ELSO members with chapters in Europe, Latin America, South, West and Pacific Asia. The registry provides semiannual reports on the use and complications of ECLS across the registry and individual centers [4]. Data can be queried to answer clinical questions addressing mechanical and procedural concerns. ELSO has established guidelines for equipment use, set up and maintenance, transport of patients on ECMO, infection control and anticoagulation and training of personnel in new ELSO centers. Quality improvement initiatives have expanded since its inception and ELSO now provides center of excellence awards to institutions with exceptional outcomes. Moreover, it delivers data to individual centers to help them benchmark their outcomes against those in the remainder of the registry. Finally, ELSO provides comparative data used for regulatory device approval such as ventricular assist devices and has contributed to the Current Procedural Terminology (CPT) and International Classification of Diseases (ICD) codes used for the surgical and medical care of ECLS patients.

see also Cannulation for ECLS

Content in this topic is referenced in SCORE Extracorporeal Life Support overview

Epidemiology

How has extracorporeal life support changed?

As of January 2016, 73,596 patients had been placed on extracorporeal life support(ECLS) internationally, of which 36,246 were neonates and 18,019 were older children. The number of registered ELSO centers and their yearly use of ECLS have increased precipitously in this last decade to 298 centers and 6,177 cases/year, respectively. This is particularly true for the last five years following the H1N1 epidemic of 2009 which saw a marked improvement in survival of patients treated with extracorporeal membrane oxygenation (ECMO) over those receiving conventional support [5]. While some of this trend is due to renewed enthusiasm about outcomes of extracorporeal support, the availability of simpler and safer equipment has truly revolutionized the way we think about ECMO. Nearly each component of the ECLS circuit has been re-engineered resulting in simpler and safer runs. Smaller and more compact circuits are now available as are highly efficient double lumen cannulas and oxygenators and both have simplified intra- and interfaciltiy transport on ECLS. We can now bring the technology to the patient, rather than transporting an unstable patient in hopes that they survive to cannulation. Magnetically levitated centrifugal pumps and pediatric specific circuit components (e.g. smaller, microporous, hollow fiber oxygenators) allow for small priming volumes, thus decreasing hemodilution and bleeding in small patients. Shorter tubing lengths and newer oxygenator designs have decreased rapid pressure changes - particularly in neonates. Improvement in pressure and flow monitoring technology has provided more acute, minute to minute information, making longer runs safer [2][6].

Despite these advances, the challenges of anticoagulation continue to plague many ECLS runs due to both thrombotic and bleeding events. The optimal anticoagulant and means of anticoagulation monitoring remain debated in the ELSO community [7].

2016 ELSO summary

Summary of International ELSO Registry data as of 2016. (from the ELSO International Registy report, January, 2016)

What have been the changes in the indications for extracorporeal life support over the last ten years?

With the advent of new technology and rising experience with ECLS, the indications for ECMO have expanded to include more complex patients. In the neonatal population, the number of respiratory runs has dramatically decreased from 1,516 in the at its peak in 1992 to 878 in 2014, likely due to the advent of iNO and other treatment modalities for neonatal respiratory failure.

neonatal ECLS runs for respiratory failure

These data demonstrate the number of neonatal ECLS runs for respiratory failure by year. Note the peak in 1992 such that the current number of cases/year are approximately fifty percent of the peak. (rom the ELSO International Registy report, January, 2016)

The overall survival has decreased as ECLS has been used in more complex patients.

neonatal respiratory runs by diagnosis

These ELSO data demonstrate the number of ECLS cases in neonatal respiratory failure patients as a function of diagnosis. (from the ELSO Registry Report, January, 2016)

The number of pediatric respiratory runs has increased steadily from 142 in 1992 to 486 in 2014 with a slow increase in the survival to 61% by 2014. A significant peak was seen in this population during the H1N1 epidemic as survival mirrored that of the adult population.

pediatric ECLS respiratory failure

These data demonstrate the yearly pediatric respiratory failure cases placed on ECLS per the ELSO Registry. Note the peak in 2009 at the time of the H1N1 epidemic. (from the ELSO International Registy report, January, 2016)

The majority of pediatric respiratory failure cases were due to the acute respiratory distress syndrome (ARDS) and non-ARDS respiratory failure or acute respiratory failure associated with viral or bacterial pneumonia.

pediatric respiratory runs by diagnosis

These ELSO data demonstrate the number of ECLS cases in pediatric respiratory failure patients as a function of diagnosis. (from the ELSO Registry Report, January, 2016)

The number of neonatal cardiac runs slowly rose in the early 2000s, but has remained relatively steady in the 400 to 450 per year range over the last few years.

cardiac ECLS By diagnoses 0-30 days

These data demonstrate the number of ECLS cases by diagnosis among patients age 0 to 30 days of age with cardiac failure. (from the ELSO International Registy report, January, 2016)

Pediatric cardiac runs remain steady at about 550 per year. The vast majority of these are in patients with congenital heart disease (over 5,000 to date) in children age thirty days to sixteen years of age.

cardiac ECLS by diagnosis 30 Days to < 16 Years

These data demonstrate the number of ECLS cases by diagnosis among patients age 30 days to less than 16 years with cardiac failure. (from the ELSO International Registy report, January, 2016)

Cystic fibrosis patients and those with severe ARDS have become more recent candidates for ECLS for bridge to transplant and the management of primary graft dysfunction following lung transplantation [8][9]. Pretransplant rehabilitation with early initiation of ambulatory venovenous ECMO has resulted in excellent post-transplant outcomes [10][11]. While this cohort of patients is relatively small, the strategies of awake and ambulatory care have become more common in the management of pediatric ARDS patients on ECMO leading to less deconditioning and better recovery or reduced disability at the time of lung transplantation.

Basic Science

What are the elements of oxygen dynamics on extracorporeal life support?

The basics of oxygen dynamics are reviewed in detail in the Respiratory Care and Surgical Infection topics. The venous oxygen saturation (SvO₂) on extracorporeal life support (ECLS) is an important measure of the relationship between oxygen delivery (DO₂) and consumption (VO₂) . A low SvO₂ indicates either inadequate O₂ delivery or excessive consumption. In taking care of a patient on ECLS it is important to optimize the components of delivery and minimize the components of consumption. On ECLS, maneuvers to maximize oxygen delivery include maintaining the hematocrit, using an appropriately sized membrane lung and increasing the flow rate. Minimizing consumption with paralytics, sedatives or antibiotics may help to maintain the SVO₂ toward a normal of 75% [12][13].

What are the elements of gas exchange on extracorporeal life support?

As in normal physiology, gas exchange on ECLS consists of both oxygen delivery and CO₂ removal. In ECLS, the membrane lung is responsible for gas exchange. Gas flows down a diffusion gradient dependent on the partial pressure of the gas in the blood and the membrane lung. Gas flow to the membrane lung is regulated by a flow meter and gas blender. Gas running through the oxygenator is the sweep gas [13][14].

CO₂ removal depends on the gas diffusion gradient, the sweep gas flow rate and the membrane surface area. The diffusion gradient is contingent upon the relative CO₂ concentration on either side of the membrane (i.e. the gradient between the venous blood and the ventilating gas). The gas flow rate or the sweep rate equates to the speed in which the air in the membrane is exchanged with new air. Consequently, a high sweep rate will clear the CO₂ faster due to the higher gradient. A lower sweep rate will lessen the gradient as CO₂ is removed less rapidly from the membrane lung. CO₂ removal also depends on the surface area of the membrane. A greater surface area gives greater C0₂ clearance by allowing more blood to interact with gas. Unlike oxygen, CO₂ removal is not dependent on flow rate of blood through the circuit. The diffusion of CO₂ through the membrane of the artificial lung is so rapid that equilibrium is almost instantaneous [13].

Oxygen delivery on ECLS is determined by blood oxygenation in the membrane lung, flow through the circuit, oxygen uptake through the native lung and cardiac output [15]. As with CO₂, the relative amounts of oxygen in the blood versus the gas phase in the membrane lung is one factor in determining the oxygen exchange from gas to blood. In addition, the oxygen solubility in plasma (i.e,. the ease with which oxygen diffuses through blood) is important. As blood enters the membrane lung, the first red blood cells will become saturated with oxygen. As it advances through the membrane, oxygen will diffuse through the plasma to saturate more blood [14]. Time must be allotted for this to happen. If blood passes too quickly through the gas exchange area there will not be enough time to fully saturate the blood. In fact, if the pump flow rate is increased in an attempt to increase oxygenation the SO2 may decrease (although oxygen delivery will likely stay the same). The pump flow at which maximal oxygen delivery is met is called the rated flow and is unique to each oxygenator. Specifically, the rated flow is defined as the flow at which blood enters with an SO₂ = 75% and leaves at an SO₂ = 95% [13][14].

As with CO₂ removal, the surface area of the oxygenator affects the amount of blood exposed to oxygen at any given time. For any given membrane, diffusion characteristics, surface area and blood path thickness are relatively fixed.

What are the physiologic principles of extracorporeal life support?

The goal of ECLS is to improve tissue oxygen delivery and remove carbon dioxide allowing normal aerobic metabolism to continue while resting the lung and, in venoaterial (VA) extracorporeal membrane oxygenation (ECMO), the heart. This is achieved by draining venous blood and providing gas exchange through an artificial lung, warming the blood via a heat exchanger, and returning the blood to the body through a vein or artery.Since gas exchange is performed by the membrane lung, the conventional ventilator pressures and oxygen can be reduced, adjusted and in certain circumstances the patient can even be extubated.

extracorporeal life support circuit

The image demonstrates an extracorporeal life support circuit with a roller pump, oxygenator and an integrated heat exchanger.

In VA bypass the function of the lung and heart are taken over by the machine. Blood is drained via a vein and passes through the circuit where oxygen is added, carbon dioxide is removed, and the blood is warmed. It is returned to the aorta and mixes with the remaining blood being pumped from the left ventricle. Nearly all of the venous return can be drained into the circuit. In fact, unless limited by the size of the venous cannula, the maximum flow is a good estimate of intravascular volume: if a patient is dehydrated, the maximum flow will not reach expected levels.

Total cardiac output is the native heart output plus the bypass flow rate. The hemoglobin saturation is determined by native lung function and the rate of bypass flow. The blood from the circuit should have a saturation of 100%. This may drop as a consequence of poorly saturated blood leaving the patient’s diseased lungs mixing with the circuit blood. Nevertheless, if most of the blood is passing through the circuit the saturation should still be 95 to 100% [12][13].

VA ECMO also provides hemodynamic support. With mixed SvO2 monitoring, the oxygen delivery to consumption ratio can be continuously adjusted. A low SvO₂ means either inadequate delivery or excessive consumption. The components of oxygen delivery need to be optimized and components of consumption need to be minimized. This is accomplished by transfusion to an adequate hematocrit, delivering more blood through an appropriate size oxygenator and increasing the flow rate to enhance cardiac output. Paralytics, sedatives, antibiotics or hypothermia can be used to minimize oxygen consumption and to achieve an SvO₂ of 75% [1]. (see Respiratory Care Pathophysiology)

In venovenous (VV) bypass, the patient’s heart is used to pump the returned blood from the circuit through the body. The oxygenated blood returning from the circuit to the right atrium mixes with the deoxygenated blood already in the right atrium. This mixed blood may either be returned to the circuit or may be pumped into the systemic circulation by the patient’s heart. The infused ECMO circuit blood that is returned to the ECMO circuit is termed recirculation. When a high proportion of blood that is infused from the ECMO circuit is recirculated, the pre-oxygenator blood may have a very high oxygen saturation (e.g. 80 to 90%). The problem under those circumstances is that the amount of oxygen which can be added to the blood by the gas exchange device is low and SaO₂ may be adversely affected. The recirculation can often be reduced and the cannula function optimized by adjusting the position of the cannula. Because of recirculation, ECLS gas exchange is not as efficient during VV ECMO. As a result, SaO₂ is often lower and it is acceptable to maintain arterial saturation at 80 to 90%. As long as cardiac output is adequate, oxygen delivery will be more than adequate to prevent acidosis and oxygenate the peripheral tissues [12][13].

In many centers, as the lung recovers, the flow in the ECMO circuit is decreased to let more blood pass through the native lungs while oxygenation is maintained. Otherwise, a blender is used to reduce the FiO₂ of the sweep flow so that the PaO₂ is maintained at reasonable levels. CO₂ removal is much more efficient than oxygenation. Higher blood flow rates are not necessary to extract CO₂ in the membrane lung if CO₂ removal is the main objective. The sweep flow is adjusted to keep the PaCO₂ in a physiologic range. On occasion, the patient will be hypocarbic despite the sweep flow being reduced to minimum levels. Under those circumstances, carbogen (5% CO₂ and 95% O₂) may be added to the sweep flow gas in small amounts [14].

How does the circuit interact with the native cardiovascular system?

The ECMO circuit is primed initially with CO₂, which more easily dissolves in fluid than air (nitrogen), and then with saline. In newborns, this saline prime is then displaced by one to two units of packed red blood cells. The hematocrit of the pump prime should be 35 to 40%. A saline prime can be used to start ECLS in older children and adults with subsequent transfusions used to reach the desired hematocrit. Maintaining a normal hematocrit is important. If the patient is anemic, higher pump flows will be necessary for adequate oxygen delivery. Giving fluid as opposed to blood exacerbates anemia and contributes to more edema [1][14].

ECMO blood flow is nonpulsatile. As the pump flow increases, the arterial wave form will become flatter and eventually become completely flat as one approaches total bypass. As long as the flow is adequate there is no evidence that there is a physiologic difference between pulsatile and nonpulsatile perfusion. The left ventricle may fill with blood from small amounts of native cardiopulmonary blood flow as well as drainage from bronchial and thebesian veins. Although the heart may attempt to contract, the ECMO flow presents substantial afterload to the left ventricle. As a result, the heart may be unable to open the aortic valve and eject blood. As a result, the left ventricle will distend and need to be decompressed. This can be done with a trans-septal atriotomy performed in the cath lab, sometimes with a cannula placed through the atriotomy. In some circumstances, a thoracotomy with direct cannulation of the left atrium or ventricle may be performed [13].

Oxygen delivery on ECLS begins with the oxygenator. The oxygenator variables are the geometry, thickness of the blood phase, membrane material and thickness and blood flow rate. The rated flow of each particular oxygenator takes into consideration all the above variables. As long as the ECMO blood flow is less than the rated flow of the oxygenator (rated flow = flow at which blood with a Hgb of 12 mg% enters with an SO₂ of 75% and leaves at an SO₂ of 95%) and the venous saturation is greater than 75%, the blood leaving the circuit will be near fully saturated. Oxygen delivery is then controlled by hemoglobin and blood flow [13][14].

What limits flow in the circuit?

Poiseulle’s Law states that with the movement of a gas or fluid through a tube, at a constant pressure the flow of the gas or fluid will vary directly with the fourth power of the radius of the tube.

Poiseuille's Law

Therefore, a smaller diameter tube requires a greater pressure to move a given volume: in other words, the resistance to flow is higher. This is the same with a longer tube although not to the same degree. Accordingly, blood flow is limited by the size of the venous drainage cannula. The best cannula is the shortest and largest diameter that can fit into the vein. There is also resistance across the oxygenator and in the reinfusion cannula. The pump delivers the desired flow and post pump pressures are monitored. Pressures as high as 300 mm Hg can be tolerated, although the higher the pressure the greater the risk of circuit disruption.

The French size of the catheter doesn’t completely reflect its flow characteristics as catheters have irregular internal diameters and side holes. Therefore, an M number is assigned to each specific catheter. This gives the pressure and flow relationship of each cannula [16]. Once the M number is known, a nomogram may be used to determine whether a given cannula will provide sufficient flows for a patient given the anticipated pressure gradient across the cannula.

M number nomogram

If one knows the M number of a cannula and the expected pressure drop across the cannula, one can predict the maximum flow that the cannula will allow. This is most important for the venous cannula which is often the limiting factor for ECMO flow. (from Bartlett RH. Physiology of ECLS. In ECMO: Extracorporeal Cardiopulmonary Support in Critical Care. ELSO Publisher. Van Meurs, KV, et al., editors. 3rd additon, 2005)

M numbers of ECLS cannulas

These are the M numbers of commonly used cannulas for venous cannulation. (from Pranikoff T and Hines MH. Vascular Access for Extracorporeal Support. In ECMO: Extracorporeal Cardiopulmonary Support in Critical Care. ELSO Publisher. Van Meurs, KV, et al, editors. 3rd additon, 2005)

General guidelines to VA and VV cannulation are provided:

guidelines to venoarterial cannula

These are suggestions for venoarterial cannula size as a function of patient weight. (from Michael McMullen, M.D. Seattle Children’s Hospital)

guidelines to venovenous cannula

These are suggestions for venoveous cannula size as a function of patient weight. (from Michael McMullen, M.D. Seattle Children’s Hospital)

Origen venovenous cannulas

Patients require pump flows of 80 to 120 mL/kg/min for full ECMO support with neonates requiring the higher and late adolescents the lower rates. If the cannulae are too small, the circuit cannot provide adequate flow [13][14].

see Cannulation for ECLS Surgical Decision Making

Anatomy

see Cannulation for ECLS Anatomy

Medical Treatment

Indications

What are the indications for extracorporeal life support?

Extracorporeal life support (ECLS) is used in patients with pulmonary and/or cardiac failure. Often chosen once the patient has been unresponsive to other modalities, ECLS should only be used for support of a reversible cause of cardiorespiratory failure or when there is a reasonable exit strategy from ECLS. Furthermore, the longer the patient remains on elevated ventilator settings or cardiac support, the less reversible the condition becomes. Therefore, deciding early in the course of an illness if the patient requires ECLS increases the likelihood of being weaned off ECLS[17][18].

Several objective criteria are followed for determining whether a patient needs ECLS including the oxygenation index (OI), alveolar-arterial oxygen difference and ventilation index in patients with respiratory failure[19]. It should be recognized that these criteria may need to be altered as enhancements in care and institutional capabilities develop over time.

The oxygenation index (OI) is calculated as follows:

OI = (MAP x FiO₂ x 100)/PaO₂

where FiO₂ is the fraction of inspired oxygen, MAP is the mean airway pressure and the PaO₂ is the arterial oxygenation

OI is measured in three postductal arterial gases, each drawn one half to one hour apart. A value > 40 is suggestive of 80% mortality and served as the indication in the early years of ECLS. It has been suggested that initiating ECLS with OI > 25 may have more favorable neurological and functional outcomes [19][20]. More recently, an approach has been adopted by some centers wherein initiation of ECLS is considered in patients with OI > 25 and initiated when the OI is > 40 .

The alveolar-arterial oxygen difference [(A-a) DO₂] is the difference in the level of oxygen between the alveoli and the capillaries. A level greater than or equal to 610 Torr despite eight hours of maximal management is associated with a mortality of 79% [20][21][22].

A ventilation index (respiratory rate x PaCO₂ x peak inspiratory pressure/1000) >40 along with an oxygenation index > 40 is associated with a 77% chance of mortality in children with respiratory failure.[19]

Special consideration must be given to those patients that are on modes of ventilation other than conventional ventilator support (e.g. high frequency jet ventilation, high frequency oscillator ventilation) [23]. With these modes, there are alterations in the relationship between mean airway pressure, gas exchange and survival [24]. However, the principle is still that of criteria which allow early recognition of a deteriorating condition.

The indications and contraindications for ECLS have evolved over time. Generally, patients are placed on ECLS for failure of ventilation, oxygenation or perfusion and broadly categorized into respiratory failure, cardiac failure or shock [25]. In general, pulmonary support is provided via veno-venous (VV) ECLS; cardiac or cardiopulmonary support via veno-arterial (VA) ECLS; and emergent support of a patient in shock via ECPR[26]. The disease states for which ECLS is instituted for neonatal and pediatric respiratory failure is demonstrated.

neonatal respiratory runs by diagnosis

These ELSO data demonstrate the number of ECLS cases in neonatal respiratory failure patients as a function of diagnosis. (from the ELSO Registry Report, January, 2016)

pediatric respiratory runs by diagnosis

These ELSO data demonstrate the number of ECLS cases in pediatric respiratory failure patients as a function of diagnosis. (from the ELSO Registry Report, January, 2016)

Previous contraindications to ECLS have been re-evaluated and the criteria for inclusion broadened. Originally, an estimated gestational age less than 34 weeks, intraventricular hemorrhage (IVH), mechanical ventilation pre-ECLS greater than seven days and cardiac arrest were considered contraindications to ECLS. However, these criteria have been relaxed.

• in the preterm newborn, estimated gestational age greater than thirty weeks and birth weight greater than 1.5 kg demonstrate reasonable outcomes, although cannulation is still a limitation
• intraventricular hemorrhage (IVH) with grade II IVH or less are candidates for ECLS with close surveillance of the bleed
• mechanical ventilation pre-ECLS less than fourteen days are candidates for ECLS
• cardiac arrest is not considered a contraindication in many centers that practice ECPR

Relative exclusion criteria are the presence of irreversible multiorgan system failure, congenital diaphragmatic hernia with severe pulmonary hypoplasia, major burns, severe immunodeficiency, active bleeding, congenital cardiac anomalies that are not amenable to surgical repair and the presence of an incurable disease process.

In those pediatric patients with cardiac failure, initiation of ECLS is based on clinical criteria including compromise of perfusion as indicated by oliguria (urine output less than 0.5 mL/kg), metabolic acidosis and hypotension despite aggressive treatment. Pediatric cardiac patients are placed on extracorporeal membrane oxygenation (ECMO) for cardiac arrest (20%), cardiogenic shock (20%), acute deterioration (10%) and an inability to be weaned from operating room bypass (20%).

What are the ethical issues surrounding extracorporeal life support?

Questions remain as to whether ECLS is effective in enhancing outcome in patients with cardiorespiratory failure. One reason is that many of the trials that have been undertaken evaluating effectiveness of ECLS have used adaptive approaches because of the ethics of withholding a lifesaving management approach. For example, Bartlett et al utilized a play the winner randomization strategy in evaluating ECLS for newborn respiratory failure [27]. Standard randomization was initially applied. However, randomization was weighted toward the successful and away from the unsuccessful experimental or control. Unfortunately, the first patient randomized to ECLS lived, the second randomized to control and died and the third randomized to ECLS and lived. From that point forward, because of the weighting, all patients ended up randomizing to ECLS and lived. Thus, in this study there was one control who died and eleven ECLS patients that survived with a statistical difference suggesting effectiveness of ECLS.

O’Rourke et al conducted a trial of ECLS with an atypical randomization scheme utilizing a statistical design by Zelen. Randomization would place the patients in the conventional or ECLS group, but a maximum of four deaths would be allowed at which point the randomization would stop and all patients would be placed in the other group [28]. In phase I, four of ten babies in the conventional group died and nine of nine babies in the ECMO group survived [29]. Randomization was halted as planned and twenty subsequent newborns were managed with ECMO . Overall, the survival of ECMO treated infants was 97% (28 of 29) compared with 60% (6 of 10) in the conventional group which was significantly different (P < 0.05).

Finally, a study of newbornECLS in the United Kingdom used a standard randomization approach [30]. Newborns (n=185) with respiratory failure at neonatal intensive care unit centers were randomized to remain at the center or to be transferred to one of five ECMO centers. In the ECLS group, 30 of 93 infants died compared with 54 of 92 allocated conventional care (p = 0.0005). A study of similar design was performed in adult patients with severe ARDS [5]. The data suggested that mortality or severe disability at six months was lower in the ECMO group (36.7%) when compared to controls (52.9%)

Attempts have been made to perform a randomized, controlled trial of ECMO in pediatric patients. However, difficulties with the ethics around refusing ECLS in patients with severe respiratory failure and low enrollment made the study impossible to accomplish. Green et al demonstrated that among 53 diagnosis and risk matched pairs, there was a significantly lower mortality rate in the ECMO treated patients (26.4% versus 47.2%, p < .01) [31].

Determining futility when deciding whether to offer ECMO as a management option is a source of debate for the ECLS community [32][33]. In patients with severe burns, major head injury or disseminated malignancy the prognosis is in general poor. Moler et al published a retrospective study of 2,879 pediatric patients supported with ECLS of which 183 were immunocompromised. The subgroups analyzed were those patients with immune deficiency, leukemia, lymphoma, cancer, opportunistic infections, solid organ transplant and bone marrow transplant. Of these patients, there was reduced hospital survival ranging from 0 to 37% [34]. However, multiple centers are exploring the use of ECMO in this population. Often attempts are made to identify the overall prognosis of the illness with a multidisciplinary discussion including the oncology team and family to determine candidacy for ECLS.

Since futility may sometimes be difficult to ascertain at the time of intervention, some children are placed on ECLS allowing time for a discussion to occur between all the involved parties including the family, the palliative care team, and a member of the institutional ethics committee. Patients with profound neurologic impairment, multiple congenital anomalies or other conditions not compatible with meaningful life should be excluded as candidates for ECLS.

When should one stop extracorporeal life support in the patient that is not making progress?

Another ethical dilemma is around the point where ECLS is futile and should be discontinued. The initial discussion with a patient’s family regarding ECLS should include potential outcomes and complications. Having clear goals and an expected time line mitigates extensive periods of ECLS in patients who may not have meaningful outcomes. That being said, the time on ECLS which indicates that recovery is futile is rapidly changing and is a significant controversial issue. Although ELSO guidelines suggest a cutoff period of two weeks without resumption of pulmonary function in a patient that is not a transplant candidate and three days in patients that are not ventricular assist device or cardiac transplant candidates, at pediatric and adult centers patients with respiratory failure have been on ECLS for many months [35]. One child was reported to be on ECLS for 600 days and was successfully weaned. The indication of futility is also a controversy in newborns with CDH where it has been suggested that longer periods on ECLS should be undertaken before discontinuation of ECLS is considered since at four weeks, 43% of patients still on ECMO survived to discharge [36].

Clearly, there are many ethical considerations that need to be taken into account when making the decision to take a patient off ECLS. No strict guidelines are in place with regards to the duration of ECLS and it is not possible to determine what constitutes an ECLS run that is too long. Ultimately the decision as to when a patient who is not recovering with ECLS should come off support lies with the care provider and the family.

Medical Decision Making

When is venoarterial versus venovenous extracorporeal life support used?

Venoarterial (VA) bypass is indicated if cardiac support is required or in any patient where access for venovenous (VV) support cannot be obtained. For VA access, the preferred site for venous drainage is the right atrium via the right internal jugular vein with reinfusion into the aortic arch via the right common carotid artery.

Venovenous (VV) extracorporeal life support (ECLS) is used in patients requiring only respiratory support. Therefore, the cardiac function has to be adequate. This may be difficult to determine when the patient is severely hypoxemic since high pressure ventilation may depress cardiac function and increase the need for inotropic support. After ECLS is initiated, airway pressures are decreased, SaO₂ increased, cardiac output is enhanced and inotropic support may be successfully weaned. VV bypass has the advantage of avoiding arterial cannulation which eliminates the potential for arterial embolization and ischemia. Arterial ligation is unnecessary, pulmonary blood flow is preserved and oxygenation of the pulmonary circulation is augmented. In addition, there are no hemodynamic effects such as an increase in left ventricular afterload. Finally, during VA bypass the blood ejected by the left ventricle perfuses the coronary arteries/myocardium and is relatively hypoxic. During VV support, the myocardium is instead perfused by blood that is relatively well oxygenated. For VV access, the preferred site of access is the right internal jugular vein for a double lumen cannula and the right internal jugular vein and femoral vein for multiple site cannulation.

How do you evaluate the patient on extracorporeal life support?

Patients on extracorporeal life support (ECLS) are evaluated daily. Weight is assessed in order to determine whether fluid overload exists. Diuresis is initiated with a goal body weight just above that of dry weight in order to optimize pulmonary and cardiac function. Daily chest radiographs are obtained to assess for aeration. Once the lung starts demonstrating early aeration, the conventional ventilator may be placed on recruitment settings. (see Medical Treatment)

The amount of ECLS a patient requires can be assessed by the required level of ECLS flow and sweep gas. Once these start to derease and the patient maintains normal hemodynamics and gas exchange, it is an indication that the cardiopulmonary system is regaining function. With VV ECLS, the flow and sweep gas may be decreased with eventual discontinuation of sweep flow and capping of the gas phase of the oxygenator so that blood continues to perfuse the oxygenator, but there is no contribution to gas exchange. Assessment of serial blood gases and SvO₂ levels during this trial allows evaluation of the patient’s lung function.

With VA ECLS, ECLS flow may be decreased as native gas exchange improves. Eventually, a "trial off" is performed as the arterial and venous tubing between the circuit and the patient is clamped. Blood in the ECLS circuit is allowed to continue flowing through a "bridge" between the arterial and venous aspects of the circuit in order to maintain circuit patency and to avoid clot formation. Serial blood gases are followed to assess the oxygenation and ventilation while off ECLS and on acceptable ventilator settings. Echocardiography may be useful to assess cardiac function and right-sided heart pressures during the trial.

If the trial is successful, the cannulas may be left in place for 24 hours as the ECLS circuit is discontinued, especially if there is uncertainty as to whether the patient may or may not need to be placed on ECLS again. That being said, the cannulas should be removed at the earliest possible time [3]. Infusion of heparin through the cannulas prevents clot formation.

How does one trouble shoot when the flow is inadequate?

If pump flow is insufficient during ECLS, it is most often due to inadequate intravascular volume: cautious volume administration may address the issue. Other issues include cannula positioning, kinking or size; tension pneumo- or hemothorax which may decrease flow by impeding venous return;pericardial tamponade which will increase intrapericardial pressure and decrease venous return; patient Valsalva; reduced patient height when a servoregulated bladder is used (drainage pressure is dependent on height between the patient and the bladder); or pump malfunction [51]. When there is insufficient flow, check the ECLS circuit for kinks and raise the bed if using a bladder. If a centrifugal pump is being used without a bladder, high negative drainage pressures should be noted if the problem is with volume; cannula positioning, kinking or size; or in the drain tubing segment. A chest radiograph and an echocardiogram should be performed if a simple solution is not apparent. If flow and/or oxygenation is inadequate during VV ECLS, conversion to VA ECLS or an addition of another drainage cannula may be necessary.

How does one trouble shoot when the patient is hypoxic?

If the patient suffers refractory hypoxemia on ECLS that is not remedied by increasing oxygen or flow, consider oxygenator failure or a disconnected sweep O₂. Check that the blood leaving the oxygenator is bright red and that blood gases assessed at the outlet of the oxygenator demonstrate a PO₂ greater than 200 to 300 mm Hg. The sweep flow outlet of the oxygenator should be examined for evidence of blood which suggests a membrane leak [51][52].

The blood from the circuit should have a saturation of 100%. The SaO₂ may be lower than 100% as a consequence of poorly saturated blood leaving the patient’s diseased lungs mixing with the circuit blood. In fact, as intravascular volume increases, more blood passes through the native circulation relative to the ECLS circuit. Since the blood flowing through the native circulation is poorly saturated, the arterial mix of the ECLS and native cardiac output will result in a lower SaO₂ as blood volume increases as long as ECLS flow remains unchanged. Interestingly, despite the observed decrease in SaO₂, the total oxygen delivery from the ECLS circuit plus the native cardiac output remains constant. The optimal management of this arterial desaturation is not just to increase flow; rather, it is to diurese the patient.

Surgical Decision Making

see Cannulation for ECLS

Steps of the Procedure

see Cannulation for ECLS Steps of the Procedure

How is surgery performed while on extracorporeal life support?

When surgical procedures, such as thoracotomy, chest tube placement, paracentesis, vascular access, suprapubic bladder aspiration or diaphragmatic hernia repair are performed on extracorporeal life support (ECLS), the following protocol for patient management is recommended in an effort to minimize bleeding complications. Preoperatively, the platelet count should be greater than 100,000/mm3 and the ACT 180 to 200 seconds. Packed red blood cells should be available for immediate infusion. All tissues, including the skin, are incised with electrocautery. Dissection is performed slowly with meticulous hemostasis and ligation of all identifiable vessels. Topical hemostatic agents such as fibrin glue, Gelfoam® and thrombin are recommended. Complete hemostasis needs to be confirmed prior to closure [53].

In many institutions an aminocaproic acid infusion is used routinely in preparation for operations such as thoracotomy or CDH repair while on ECLS. A bolus of 100 mcg/kg is administered followed by an infusion of 30 mg/kg/hr. The infusion is started six hours prior to the anticipated operation and continued up to 24 hours post surgery [54].

The optimal approach to repair of a congenital diaphragmatic hernia (CDH) in a newborn on ECLS has been examined extensively over recent decades. A review from the American Pediatric Surgical Association outcomes task force suggested that early repair, before anasarca develops on ECLS, may have improved survival and shorter ECLS duration (Grade D recommendation) [55]. In contrast, Partridge, et al., reported decreased outcomes in CDH patients undergoing surgical repair while on ECLS compared with repair following decannulation [56].

Postoperative Care

How should patients be transported on extracorporeal life support?

Interhospital transport on extracorporeal life support (ECLS) is defined as either local, regional or long distance. Local transport is performed by ground and is defined as a distance of less than 150 miles. Regional transport encompasses distances of 150 to 1,000 miles and requires either rotary or fixed wing aircraft. Long distance transport over more than 1,000 miles generally requires jet aircraft capability.

The majority of patients referred to an ECLS center can be transported before initiation of ECLS. However, a number of patients miss the window of stability or rapidly deteriorate precluding safe transport. Additionally, some medical centers are located in remote geographic areas without a regional ECLS center making prolonged transport risky even when an early referral is made. As a result, several ELSO centers have developed capabilities for mobile ECLS transport which has become more feasible as simpler, less cumbersome equipment has been developed [57].

Challenges faced with running a mobile ECLS program include expense; delay from time of initial referral until ECLS team arrival; logistical and medicolegal issues with cannulation away from the home institution and substantial resource utilization. Therefore, mobile ECLS is not a substitute for early referral and transport. With regard to some of these issues, our institutional ICU is considered to extend to the referring institution even when in another state, thus avoiding issues with licensing, privileging, and disagreements over management. The patient is, therefore, on our "service" despite being in a referring institution. The survival to discharge with mobile ECLS has been reported to be between 33 and 100%, depending on the series. A recent review reported on 102 critically ill adult patients managed with mobile ECLS prior to transport to the tertiary referral center. Of those, 93% had venovenous ECLS and seven percent had venoarterialECLS. There were no mortalities or major complications associated with the transfer [58].

Complications

What circuit component complications occur?

As the extracorporeal life support (ECLS) circuitry has become more sophisticated, many of the complications associated with ECLS have been reduced. However, equipment malfunctions still occur. Oxygenator failure, tubing rupture, pump and heat exchanger malfunction, and thrombosis and hemolysis are still being reported to the ELSO registry[59]. By early 2015, ELSO noted up to 1600 neonatal and 800 pediatric oxygenator failures. In pediatric cardiac ECLS, nearly 1000 failures were observed [3]. Other complications are less common, but all of them effect patient survival negatively. Using a daily check list may allow identification of equipment issues before circuitry components need to be changed emergently. Predetermined emergency ventilator settings and cardiovascular medications should be immediately available in case of equipment failure.

How is hemolysis assessed?

Hemolysis results from direct erythrocyte trauma in the circuit. High roller pump occlusion pressures and the use of centrifugal pumps contribute to higher rates of hemolysis. Daily assessment of free plasma hemoglobin may allow early identification of hemolysis [60][61]. Development of hyperbilirubinemia or acute renal dysfunction in the setting of rising free serum hemoglobin indicates ongoing hemolysis [62]. A consumptive coagulopathy sometimes occurs due to thrombi within the circuit and may be associated with thrombocytopenia and high free serum hemoglobin levels. Under those circumstances, a circuit change may be necessary.

How does one manage air in the circuit?

Micro air emboli occur intermittently on ECLS without much consequence. Large air emboli, however, may lead to interruption of flow and potential devastating consequences if they enter the arterial circulation. Air can be introduced when the circuit is accessed for infusions and medications, especially on the venous drainage side where circuit pressures are negative. Another common source of small air bubbles in the ECLS circuit is from air introduced elsewhere into the patient’s venous blood system via intravenous fluids. Infusion sytems with built in filters may reduce the introduction of such air. The majority of air emboli are captured by and dissipate from the membrane oxygenator. As a result, they are prevented from entering the arterial circulation. It is for this reason that infusions into and access to the post-oxygenator, reinfusion segment of the circuit is discouraged, if not against policy, in many institutions. When air is injected into the circuit, more makes it through the oxygenator with a centrifugal pump when compared to a roller pump, perhaps related to the ability for air to collect in the pump head [63].

If air is noted on the arterial side, attempts must be made to aspirate the air before it enters the patient’s circulation. If this cannot be done, emergency ventilator settings should be implemented and the circuit disconnected from the patient by clamping the drainage and reinfusion limbs. The circuit may then be debubbled or replaced. At times, the involved segment of tubing may specifically be clamped and debubbled.

Finally, preoxygenator pressure monitoring prevents high negative pressures on the venous side of the circuit. High negative pressures tend to induce cavitation in the venous drainage portion of the circuit and air bubble formation [6].

What cannula complications are observed and how are they managed?

Migration or kinking of the cannula can result in decreased venous drainage with concomitant loss of support or obstruction of outflow which may result in high circuit pressure and potential circuit rupture. In addition to these complications, during VV ECLS cannula malposition may result in recirculation with an associated decrease in support. Complications of VV ECMO are observed more frequently in patients cannulated with a non-wire wound double lumen cannula. Such cannulae are prone to kinking with positional changes. Currently available double lumen cannulae should be oriented with the arterial side anterior in order to direct the return of oxygenated blood toward the tricuspid valve. Even when a cannula is initially placed in proper position, it can migrate with changes in patient position, volume status, or diaphragm elevation. When cannula malposition is suspected, it can easily be diagnosed with a chest radiograph or an echocardiogram. Echocardiography can also be used to reposition the cannula if needed [64].

Venous or arterial blood vessel or atrial perforation is a devastating cannula complication which most often occurs during cannulation. Failure to obtain flow from a cannula during initiation of bypass should raise suspicion for malposition, including extravascular placement. The presence of a cardiac tamponade with blood in the pericardium should raise the spectre of atrial or ventricular perforation which must be ruled out. In newborns and young children, our experience, and that of others, is that the Avalon cannula has been associated with multiple complications during cannulation, including that of vascular perforation [42]. As a result, we now have a policy that all percutaneous cannulation procedures must be performed with fluoroscopic guidance, preferably in the operating room. In addition, the Avalon cannula, which is inserted via the right internal jugular vein, requires that the tip be placed into the inferior vena cava (IVC). The reinfusion port is only a short distance from the tip; thus, there is only a short segment of the cannula which is in the IVC. There have been a number of cases in children where the tip of the cannula migrated into the hepatic veins, the right atrium, or the right ventricle with potential loss of flow and/or perforation. As a result, we no longer use the Avalon cannula in patients less than two years of age.

How are bleeding and thrombosis managed?

Bleeding, thrombosis and disseminated intravascular coagulation are by far the most common complications seen with ECLS. The interaction of blood with the circuit components is immediate upon initiation of ECLS. This activates thrombogenic pathways leading to platelet activation, aggregation and consumption despite administration of heparin [62]. At the same time, the fibrinolytic pathway is activated leading to clot lysis [62]. As a result of heparin administration, thrombocytopenia, and clot lysis, bleeding complications are frequent in patients on ECLS. Cannula site, gastrointestinal and surgical bleeding occur in up to eight percent of neonates and 18% of children treated for respiratory failure. In cardiac patients, surgical site bleeding is reported in up to 30% of infants and children [3].

Frequent laboratory monitoring and repletion of platelets and clotting factors minimizes the risk of bleeding. The ideal monitoring of anticoagulation is still debatable, but many institutions follow activated clotting times (ACTs), thromboelastography or both. Minor bleeding may be addressed with topical agents and the repletion of platelets and coagulation factors. Excessive bleeding may necessitate adjustment in ACTs or even cessation of heparin; administration of aminocaproic acid and possible surgical control. Reduction in ACTs to the 170-190 second range or cessation of heparin is frequently associated with thrombus formation in the circuit. As a result, it is recommended that a primed backup circuit be available when ACTs are lowered beyond the 170-190 second range. The development of less thrombogenic circuits which do not require anticoagulation are ongoing, but human data have not been produced to date [65].

What neurologicsequelae are observed in patients treated with extracorporeal life support?

Patients treated with ECLS remain at risk for neurologic complications. Often, these precede the initiation of ECLS due to the severity of hypoxia and cerebral hypoperfusion associated with the critical illness that necessitated the escalation of care. Neonatal patients experience the highest rate of complications which range from minor developmental delay to seizures, hypoxic-ischemic encephalopathy, stroke and intracranial hemorrhage (ICH). The rates of each individual complication are approximately four percent with 11% of patients on ECLS experiencing at least one of these complications [66]. While ELSO Registry studies have suggested that carotid artery cannulation is associated with increased rates of neurologic injury in patients < 18 years of age, the stroke rate in those on ECLS is only approximately 1.4% higher in the group with carotid artery ligation when compared to those with non-carotid artery cannulation [7][67][68]. It has been suggested, therefore, that carotid artery cannulation should be the primary access in young children and in those with marked physiologic instability when VA ECLS is required. It should be noted that the stroke rate with carotid artery cannulation is not altered with age over the first 18 years of life.[7]

A recent study of neonates on ECLS demonstrated an increased risk for ICH in infants with persistent pulmonary hypertension who also had thrombocytopenia and low fibrinogen levels. In addition, neonatal hypertension was associated with an increase in the rate of ICH [69]. As a result, the authors suggest that maintaining fibrinogen and platelet counts within normal ranges is crucial to prevent ICH in newborn patients on ECMO.

neurologic outcome after ECMO

Visual abstract courtesy of Francois Luks

What renal sequelae are observed in patients treated with extracorporeal life support?

Acute kidney injury(AKI) and renal failure complicate up to 15% of ECLS runs and serve as an independent predictor of mortality with a rate of 27% [70]. AKI results from hypoperfusion associated with the patient’s primary disease and complications such as hemolysis while on ECLS. Younger patients and those who experience cardiac arrest prior to cannulation have a higher rate of AKI[71].

Diaphragmatic hernia(CDH) infants on ECLS experience AKI and renal failure more frequently than other neonates on ECLS. A 2011 review by Gadepalli, et al., demonstrated a 71% rate of AKI in CDH infants requiring ECLS over a ten year period with 22% meeting criteria for injury and 49% falling into a failure category based on RIFLE criteria. Survival was decreased to 27.3% in CDH patients with renal failure on ECLS which compares to a survival of 80% in such newborns without renal failure[72].

Recent trends have advocated earlier initiation of continuous renal replacement therapy (CRRT) for pediatric patients on ECLS, suggesting better survival if therapy is initiated prior to greater than 10% fluid overload [73]. While the feasibility of CRRT and ECLS has been documented, the data on the effect of CRRT on mortality in this group of patients is currently lacking [74].

Outcomes

How have mortality rates for patients on extracorporeal life support changed?

In the newborn population, the criteria for extracorporeal life support (ECLS) has expanded to include premature neonates, congenital diaphragmatic hernia (CDH) patients with severe pulmonary hypoplasia, those with IVH, those with septic shock, patients with severe forms of congenital heart disease, and, in general, more complicated cases. Simultaneously, as the ability to support neonates with respiratory failure has advanced, "straightforward" newborn cases that required ECLS in the past, e.g., those with meconium aspiration syndrome and PPHN, are now reasonably supported by conventional measures, including iNO. As such, more complex patients are now being managed with ECLS[3][7] which adversely affects outcomes, including survival. In fact, the most recent statistics show a downward trend from 80% survival to discharge in the early 1990s to 66% over the last few years.

Neonates with meconium aspiration syndrome (MAS) have the best survival by far (nearing 95%), followed by respiratory distress syndrome and PPHN (84 and 77%, respectively). CDH babies requiring ECLS have the worst survival among neonates which has remained at approximately 51% for the past five years.Neonatal Respiratory Runs by Diagnosis However, successful outcomes should focus on not just survival, but also quality of life. While neonates have the highest rate of neurologic complications, their long term outcome is quite good and the rate of seizures and developmental delay is comparable to children with respiratory failure treated with conventional measures. As earlier disabilities resolve, they are found to have more subtle developmental issues such as learning disabilities. In one series, 50% of newborns managed with ECLS had an entirely normal neurologic outcome at school entry. The remaining patients had either epilepsy, developmental delay, or cerebral palsy (12.5%). Lower gestational age and birth weight are associated with an abnormal outcome as is the diagnosis of septic shock [75]. The development of seizures appears to carry a worse prognosis with lower IQ at school age and a trend toward higher rates of cerebral palsy [76][77]. Early referral to therapy programs, including screening for sensorineural hearing loss, may improve overall development and ultimate outcomes.

The outcomes of pediatric ECLS vary by age and primary disease.Pediatric Respiratory Runs by DiagnosisELSO Data on Cardiac ECLS By Diagnoses in Patients 0-30 Days of AgeELSO Data on Cardiac ECLS By Diagnoses in Patients 30 Days to < 16 Years of Age Survival in pediatric ECLS has been stable over the last 5 years for both respiratory and cardiac diagnoses, in the range of 62 and 57%, respectively [3].

What is the outcome for extracorporeal cardiopulmonary resuscitation?

Extracorporeal cardiopulmonary resuscitation (ECPR) is increasingly applied to both adult and pediatric patients during cardiac arrest. To date, more than 1100 neonates and 2300 pediatric patients have received ECPR. Of those, 64% of neonates and 55% of children survived ECLS and 40% of both groups survived to discharge [3]. In-hospital cardiac arrest (IHCA) has the best neurologic outomes and survival to discharge for all patients who receive ECPR [15]. The return of spontaneous circulation (ROSC) rate was 2.24 times better in CPR of IHCA patients than in CPR of out-of-hospital cardiac arrest (OHCA) patients (p=0.0012) [78]. A recent retrospective single institution study of pediatric ECPR recipients reported favorable short term neurologic outcomes in 80-90% of patients [79]. However, robust long term data on neurologic outcomes are lacking.

Multiple centers in the United States and other countries have developed programs for ECPR in patients with OHCA.[80] Favorable neurological outcomes were observed in 11.2% of the ECPR group and 2.6% of the conventional resuscitation group at 6 months (P=0.001). Another study suggested that survival with OHCA was inversely proportional to age and time of CPR.[78]

Research and Future Directions

While extracorporeal life support (ECLS) has provided life saving therapy to tens of thousands of newborns, premature infants are often not ECLS candidates due to their size and risk of bleeding. The artificial placenta [81] concept allows provision of gas exchange via an artificial lung while maintaining a physiologic fetal environment for the extremely low gestation premature newborn. The umbilical vessels or internal jugular vein may be used with or without a pump. PaO2 is maintained at in utero levels (~30 mmHg) as fetal circulation is maintained. While the need for anticoagulation remains a concern, other research efforts are exploring the use of surfaces coated with nitric oxide donors to prevent thrombus formation in the circuit for the artificial placenta, as well as for ECLS in general [82].

For those patients that remain on ECLS for prolonged periods, work is being performed to develop wearable or implantable artificial lungs.[83] The hope is that these oxygenators will be reliable for long periods and will serve as a bridge to transplantation or even as a long term lung replacement "destination" device.

An interesting area of work is the application of ECLS to augment successful organ harvest in patients providing donation by cardiac death (DCD) as opposed to brain death.[84] In this scenario, support is withdrawn in patients with a poor prognosis who do not meet brain death criteria and, after electrical cardiac function has ceased for a defined period, the organs are harvested. Initation of ECLS at that time point allows for augmented recovery of donor organs.[85] In addition, other areas of research have centered on warm extracorporeal perfusion of lungs, hearts, and limbs for transplantation with the intent of preventing the ischemic injury and resulting non-function which results from cold storage.

One of the substantial advances that have been made in ECLS over the past decade has been the development of devices which have simplified the circuit and procedure. For example, older centrifugal pumps were associated with hemolysis and hyperbilirubinemia after a few days of ECLS because of the friction and heat associated with the pump heads spinning on ball bearings. The newer centrifugal pump heads are levitated on magnets with minimal friction and heat generation and, as a result, can be used safely for prolonged ECLS. Thus, the safer and simpler centrifugal pumps, which will not as easily allow circuit overpressurization or cavitatation, are now used for ECLS. Likewise, earlier hollow fiber lungs leaked plasma after a few days of ECLS. The current polymethylpentane coated fibers have a low tendency to do so. Thus, low resistance hollow fiber lungs may now be used for ECLS. Finally, double lumen cannulae allow easy venous access for support. All told, the circuit is much safer and simpler than in the past and requires less oversight: some have even added pump flow servoregulation based on PaO₂ and PaCO₂ so that the ECLS circuit, in essence, runs itself.

Other therapies are focusing on use of stem cells for organ development that could ultimately be used in autologous lung transplant or tissue regeneration for infants already affected by bronchopulmonary dysplasia (BPD). A chronic effect of prematurity and respiratory distress syndrome, BPD affects a remarkable number of premature infants [86]. There are no effective therapies for BPD. Ex vivo lung bioengineering with decellularized whole lung scaffolds offers an innovative solution [86]. The process of decellularizing an organ removes the cellular debris and theoretically renders the scaffold immunologically inert. Reseeding a decellularized scaffold with autologous stem cells may enable regeneration of lung tissue that can eventually be utilized for transplantation. An optimally bioengineered lung tissue would mimic the patient’s own cell type (i.e be autologous), easily expand in culture and have the ability to differentiate into the different phenotypes of distal airway cells. Investigators have differentiated pluripotent stem cells [87][88][89][90] and amniotic fluid stem cells [91] into lung precursors that can recellularize a 3D lung tissue scaffold.

differentiation of lung stem cells

Differentiation schema to develop distal airway cells from pluripotent stem cells.

In addition, the ability to create three dimensional biomimetic matrices by decellularizing a lung can enable models for drug testing or possibly enable recellularization with eventual autologous lung transplant.

creation of three dimensional lung

Decellularized lung reseeded with stem cells to create a three dimensional lung.

Perspectives and Commentary

To submit comments about this topic please contact the editors at think@apsapedsurg.org.

Additional Resources

Extracorporeal Life Support Organization

High frequency oscillator ventilation (HFOV)

The oscillator is an effective method of ventilation when it comes to oxygenating a patient. This modality works by delivering very low tidal volumes (1 to 2 mL/kg) at a very high rate (3 to 15 breaths per second). This maintains airway pressures within a constant range allowing more efficient oxygenation. The parameters on the oscillator that can be adjusted are the amplitude, hertz and MAP. Oscillator tidal volumes are less than that of the physiological dead space, thereby minimizing ventilator induced volutrauma. The decreased, but sustained, MAP of oscillatory ventilation reduces barotrauma by avoiding high peak inspiratory pressures [92].

The OSCILLATE trial showed that the use of oscillators early on in the clinical course of ARDS did not reduce, and may increase, in-hospital mortality in the adult population [93]. However, the oscillator has been used successfully in managing patients with ARDS and other pulmonary conditions both in the neonatal and pediatric age groups with favorable outcomes. Many studies have shown decreased surfactant use, days on the ventilator, need for supplemental oxygen and incidence of death, bronchopulmonary dysplasia and neurological disabilities with use of the oscillator when compared to the conventional ventilator [92][93][94].

High frequency jet ventilation (HFJV)

HFJV is commonly used in patients when there is difficulty with gas exchange or in the setting of a bronchopulmonary fistula or tracheoesophageal fistula. This mode of ventilation works on the principle of very small tidal volumes delivered with high frequency. HFJV does require the conventional ventilator to provide adequate PEEP. In ARDS, as the essential problem lies with difficult oxygenation, there has been a mixed response to the use of HFJV [23][95]. Smith, et al., showed 69% survival in patients placed on HFJV after being on the conventional ventilator for an average of four to ten days. They found a benefit in utilizing HFJV in patients with evidence of barotrauma from the conventional ventilator (e.g., pulmonary interstitial emphysema, pneumothorax, pneumomediastinum, pneumopericardium, pneumoperitoneum) which supports the use of HFJV in the presence of airway trauma secondary to high peak inspiratory pressures [23].

Inhaled pulmonary vasodilators

see Pulmonary Hypertension Medical Treatment

Surfactant

Surfactant is most commonly used in the premature neonate. Studies have also suggested a decrease in need for ECLS in full term neonates in whom surfactant was administered.[96] Definitive data does not exist to support the use of intratracheally delivered surfactant in both the pediatric and adult population. Commonly seen side effects are hypotension, bradycardia and hypoxemia/desaturation. In a retrospective study, Shein, et al., demonstrated improved lung compliance in the first twelve hours following surfactant adminstration in pediatric patients on ECMO for non-cardiac disease, although there was no statistically significant difference in days on the ventilator or overall outcomes. The study did show that there was an increase in the number of days on the ventilator after surfactant administration [97][98]. Further studies and data supporting use of surfactant in the older population have yet to be conducted and published.

see also Respiratory Care and Respiratory Care in the Neonate

Discussion Questions and Cases

To submit interesting cases which display thoughtful patient management please contact the editors at think@apsapedsurg.org.

What are common diseases that cause respiratory failure in a neonate and child?

What physiologic parameters support the need for ECLS?

What are contraindications for providing ECLS support?

In a neonate, what preoperative work up is needed prior to institution of ECLS?

What physiologic parameters support Venoarterial vs. venovenous cannulation?

What effects does extracorporeal flow have upon a neonate? a child?

What is a typical ECLS circuit composed of?

In prenatal consultation of a fetus with congenital diaphragmatic hernia, what information would you discuss with the parents?

If a patient is on ECLS support and the perfusionist tells you the pump is "cutting out"- what steps can you take?

What if the patient becomes hypoxic?

What if there is air in the circuit?

Additonal questions are in SCORE Extracorporeal Life Support conference prepCongential Diaphragmatic Hernia conference prep

Pulmonary Hypoplasia/Hypertension conference prep

References

Media

Origen venovenous cannulas

guidelines to venoarterial cannula

These are suggestions for venoarterial cannula size as a function of patient weight. (from Michael McMullen, M.D. Seattle Children’s Hospital)

guidelines to venovenous cannula

These are suggestions for venoveous cannula size as a function of patient weight. (from Michael McMullen, M.D. Seattle Children’s Hospital)

cardiac ECLS By diagnoses 0-30 days

These data demonstrate the number of ECLS cases by diagnosis among patients age 0 to 30 days of age with cardiac failure. (from the ELSO International Registy report, January, 2016)

cardiac ECLS by diagnosis 30 Days to < 16 Years

These data demonstrate the number of ECLS cases by diagnosis among patients age 30 days to less than 16 years with cardiac failure. (from the ELSO International Registy report, January, 2016)

neonatal respiratory runs by diagnosis

These ELSO data demonstrate the number of ECLS cases in neonatal respiratory failure patients as a function of diagnosis. (from the ELSO Registry Report, January, 2016)

pediatric respiratory runs by diagnosis

These ELSO data demonstrate the number of ECLS cases in pediatric respiratory failure patients as a function of diagnosis. (from the ELSO Registry Report, January, 2016)