Paediatric extracorporeal membrane oxygenation and extracorporeal cardiopulmonary resuscitation

Learning objectives.

By reading this article you should be able to:

• •Describe the differences between venoarterial and venovenous extracorporeal membrane oxygenation (ECMO) and cardiopulmonary bypass.
• •Identify the basic components of the ECMO circuit.
• •Explain the indications and contraindications for mechanical cardiopulmonary support in infants and children.
• •Recognise the common indications for extracorporeal cardiopulmonary resuscitation.

Key points.

• •Extracorporeal membrane oxygenation (ECMO) is conceptually similar to cardiopulmonary bypass.
• •ECMO is a resource-intensive therapy available in selected centres.
• •ECMO can be used to support a variety of states that cause cardiorespiratory failure.
• •Extracorporeal cardiopulmonary resuscitation is associated with improved survival and neurological outcomes for refractory in-hospital cardiac arrest.

Extracorporeal membrane oxygenation (ECMO) refers to mechanical cardiopulmonary support, a highly specialised service available in a small number of paediatric and adult hospitals.1, 2, 3 Both human and physical resource utilisation are high and a robust support network including intensive care, cardiac surgery, cardiology, perfusion and blood bank services, and rapidly responsive laboratory services are necessary for optimal delivery of care.1, 4 ECMO is conceptually similar to the cardiopulmonary bypass (CPB) performed during cardiac surgery, albeit with some important differences (Table 1).2, 5

Table 1.

Comparison of CPB, VA ECMO, and VV ECMO. SVC, superior vena cava; IVC, inferior vena cava; RA, right atrium; Ao, aorta.

	CPB	VA ECMO	VV ECMO
Purpose	Perioperative	Allow for recovery of diseased myocardium	Allow for recovery of diseased lungs
Cannulation approach	Central by sternotomy	Peripheral by cutdown or Seldinger	Peripheral by cutdown or Seldinger
Cannulation sites	SVC/IVC or RA, Ao	Large systemic vein, large systemic artery	Drainage and infusion cannula in large systemic vein
Haemodynamic support	100%	90–100%	None
Risk of thromboembolism	High	High	Low
Duration of use	Hours	Days	Days–weeks
Recirculation of oxygenated blood	No	No	Yes

Extracorporeal cardiopulmonary resuscitation (eCPR) is the practice of initiating ECMO during active CPR and is an emerging technique in the treatment of patients with refractory cardiac arrest.² However, delivery of this potentially lifesaving treatment remains challenging. To be effective, eCPR must be available within a timely manner, necessitating 24 h access to sterile and primed ECMO circuits, perfusion services, blood products, cardiac or general surgery, and anaesthesia. Recent studies examining the efficacy of eCPR suggest improved survival in patients with refractory cardiac arrest compared with conventional CPR.2, 4, 6

History of ECMO

CPB was first successfully performed by John Gibbon and team in 1953 in the repair of an atrial septal defect.⁵ The original bypass machine used a bubble oxygenator, which was unsuitable for long-term cardiopulmonary support because of the rapid denaturation of red blood cell proteins caused by direct exposure to oxygen.⁷ Following the subsequent development of the membrane oxygenator in 1969, ECMO was adapted from CPB during the 1970s, and is one of the few medical technologies to be pioneered in children and subsequently adapted for adult use.3, 5 The Extracorporeal Life Support Organization (ELSO) is ‘an international non-profit consortium of health care centres and individuals who are dedicated to the development, evaluation and improvement of ECMO … in the neonate, child and adult’. More information can be found at http://www.ELSO.org. Since 1989, ELSO has recorded the use of and outcomes associated with ECMO. Cardiac indications, including eCPR, now account for approximately 50% of cases in the ELSO registry.³ Approximately 60% of children undergoing ECMO survive to hospital discharge, a statistic that has remained fairly stable over time.⁶ Survival is superior in neonates at approximately 70%, and is highly dependent on aetiology of cardiopulmonary failure.⁴ As of 2016, among neonates undergoing ECMO for cardiac indications, overall hospital survival was only 41%, while survival for pulmonary indications was 75%.⁸ Single ventricle physiology, lower weight (<3 kg) and higher Risk Adjustment in Congenital Heart Surgery-1 score are associated with an increased risk of in-hospital mortality.⁸ Amongst children and infants undergoing ECMO for respiratory indications, survival was greatest for children with asthma or bronchiolitis and least for those with pertussis or malignancy.⁹

Physiology of ECMO

ECMO can be used to support a variety of states resulting in cardiopulmonary failure, including cardiogenic shock, severe acute lung injury, intoxication, sepsis, failure to wean from CPB during cardiac surgery, profound hypothermia, and others.1, 2, 3 The particular mode of ECMO chosen is dependent on the physiological state of the patient requiring support.1, 2 Cardiac failure always requires venoarterial (VA) ECMO support (drainage cannula in a large systemic vein or right atrium, infusion cannula in a large systemic artery).1, 2, 3 In this case, the patient receives near total mechanical haemodynamic support, and the patient's lungs participate minimally in gas exchange. VA ECMO poses considerable risk of systemic embolic complications, both from thrombus or air.1, 3 The main differences between CPB and VA ECMO are as follows: (i) cannulation to CPB is nearly always central via sternotomy, whereas a peripheral cannulation site is preferred for ECMO except in the postoperative setting; (ii) CPB is typically used for short term cardiac procedures, while ECMO can be used for several days or weeks; (iii) ECMO is utilised to allow recovery of damaged/non-functional respiratory or cardiac system; (iv) the CPB circuit allows for scavenging of surgical blood, while ECMO does not; (v) higher levels of anticoagulation are required for CPB; (vi) CPB results in greater cellular destruction than ECMO; and (vii) CPB circuits use a prepump blood reservoir to protect against sudden interruptions of venous inflow, while most modern ECMO circuits do not use such a reservoir.5, 7

Pure respiratory failure can be supported using venovenous (VV) ECMO, in which both the drainage and infusion cannulae are placed in large systemic veins.1, 2 Drainage and infusion cannulae can either be separate, or a single cannula with both drainage and infusion lumens (Avalon® cannulae, Gettinge AB, Gothenburg Sweden) can be placed in the same large systemic vein. Systemic embolic phenomena are considerably less common with this mode of ECMO. Disadvantages of VV ECMO include lack of haemodynamic support, and mixing of oxygenated and deoxygenated blood (re-circulation) at the site of cannulation, which reduces the overall efficiency of extracorporeal oxygenation.1, 2, 3 This method of support is not appropriate if significant pulmonary hypertension is present, as left ventricular filling is dependent on right ventricular output.¹⁰ In the absence of a patent ductus arteriosus systemic perfusion will be impaired, as the right ventricle is unable to overcome the load imposed by the hypertensive pulmonary vasculature.¹⁰

For VA ECMO, vascular access can be obtained via the neck (typical in small children and infants) in the carotid artery and internal jugular vein or via the groin (more common in larger children and adults) in the common femoral artery and vein.1, 2 This can be achieved either by cut-down and direct visualisation or percutaneously by Seldinger technique.1, 2 If the patient is status postcardiac surgery and unable to be weaned from CPB, ECMO can be initiated via the arterial and venous cannulae already in situ via the ascending aorta and right atrium. Additional indications for direct central cannulation are recent sternotomy, or current open chest.1, 3, 4

The failing ventricle may be unable to eject against the increased afterload imposed by ECMO flows.² Thus, persistent elevation of end-diastolic pressure (EDP) may lead to increased myocardial oxygen demand, elevated left atrial pressure (LAP), pulmonary hypertension, and oedema, all of which may lead to delayed myocardial recovery.2, 11 To unload the failing ventricle atrial septostomy may be performed during central cannulation, or a left atrial vent may be placed in order to drain the left atrium via thoracotomy.2, 11

The main absolute contraindications to ECLS are technical, including small size (<1.5 kg or <32 weeks gestational age), or lethal underlying disease, such as chromosomal abnormalities (trisomy 13 or 18), uncontrolled haemorrhage, or irreversible brain damage.1, 2, 3 Additional relative contraindications are centre-specific, such as less significant extremes of weight or prematurity (<36 weeks or 2.5 kg in some centres), non-fatal intracranial haemorrhage, irreversible organ failure in a patient ineligible for transplant, or prolonged mechanical ventilation (>2 weeks) before ECMO initiation.1, 2, 3

Technology of ECMO

The ECMO support apparatus consists of drainage and infusion cannulae, a mechanical blood pump, a membrane oxygenator, monitoring lines and electronic monitoring devices, heat/moisture exchanger, pressurised gas source, and tubing (Fig 1).1, 2, 3, 4 There are two types of mechanical blood pumps in use today, roller pumps and centrifugal pumps. In the current era, centrifugal pumps are preferred because of the lower risk of haemolysis and thromboembolic complications.1, 3, 4 Circuit pressures (prepump, pre- and postmembrane) and pre- and postmembrane saturations are typically measured continuously. Alterations in circuit pressures or saturations serve to alert clinicians to the need for preventative measures, such as volume administration, membrane oxygenator replacement or afterload reduction.1, 3

Fig 1.

Typical representation of VA ECMO circuit. This circuit consist of arterial and venous cannulae, infusion lines, monitors, blood pump, membrane oxygenator and medical gas supply, heat/moisture exchanger and bridge between drainage and infusion lines. (Reproduced from Cooper and colleagues, 2007)⁴.

An additional feature of most modern ECMO circuits is an ultrasonic flow meter. This measures the flow in the postmembrane infusion limb of the circuit. This is an accurate reflection of the actual blood flow to the patient (usually between 70 and 100 ml kg⁻¹ min⁻¹).¹ Actual patient blood flow can be affected by alterations in pump speed, membrane oxygenator resistance, total patient/circuit volume (preload), and patient afterload.1, 3, 4

An important feature of the ECMO circuit is the so-called ‘bridge’. This is a route by which the patient can be removed from the ECMO circuit in an emergency or during weaning, while still maintaining continuous flow through the ECMO circuit across the bridge. Emergent indications to remove a patient from ECMO may include: contamination of ECMO circuit by air embolus or thrombus; uncontrolled life-threatening haemorrhage requiring reversal of anticoagulation; or in order to change any of the ECMO circuit components.² In an emergency or when weaning ECMO, the bridge is opened and the patient is disconnected from ECMO by clamping the tubing on the patient side of the bridge.²

Anaesthesia considerations

The role of the cardiac anaesthetist, paediatric intensivist, or both, in initiating ECMO cannot be overstated, and a thorough understanding of the effects of commonly used anaesthetic drugs on haemodynamics, and the effects of the addition of the ECMO circuit on pharmacokinetics and pharmacokinetics is critical.

The cardiac anaesthetist, anaesthesia trainee or intensivist may be called upon to manage a patient's airway before initiation of ECMO or eCPR and manage the patient's haemodynamics during cannulation to, or decannulation from ECMO. Typically patients cannulated to ECMO are receiving maximal cardiac and respiratory support, generally including vasoactive and inotropic medication and maximal safe ventilation settings.1, 2 Patient ECMO flow is generally started gradually and increased to the maximal flow that the membrane oxygenator will allow. Subsequently, the flow is decreased to maintain adequate output to maintain superior vena cava (SVC) saturations of 70% when infusion side saturations are 95% or higher.1, 2 ECMO does not provide pulsatile blood flow, and therefore upon initiating ECMO the arterial pressure wave form will flatten, and both systolic and diastolic blood pressure will approach the mean.¹ Inotropic support should be weaned, and vasodilators such as milrinone or nitroprusside may be necessary to optimise ECMO flow.1, 3

General anaesthesia for patients anticipated to be cannulated to ECMO should be induced using medications with as little haemodynamic effect as possible. Fentanyl and ketamine have such characteristics.¹² During cannulation, the patient should be completely anaesthetised and neuromuscular block complete to eliminate negative pressure ventilation, which can result in air embolism.¹

In general, the addition of the ECMO circuit increases volume of distribution and clearance of drugs commonly used in critical illness.¹³ In particular, commonly used lipophilic sedatives and analgesics are absorbed by the polymer-based components of the ECMO circuit, dramatically increasing their apparent volume of distribution.¹³ Organ dysfunction further contributes to altered pharmacokinetics, with renal failure decreasing the elimination of drugs and hepatic failure decreasing their deactivation and conjugation to water-soluble metabolites.¹³ In addition to the usual phenomenon of tachyphylaxis, patients on ECMO appear to have increasing analgesic and sedative needs over time. Daily review of sedation and analgesia scores and titration of sedative medications are strongly recommended.¹³

Following establishment of complete haemodynamic support, ventilator support should be reduced to minimise ventilator-associated lung injury.1, 2, 3 A high PEEP, low tidal volume strategy is preferred.³ It is important to appreciate that in VA ECMO the coronary arteries may be still supplied by native cardiac output and therefore it is beneficial to maintain normal pulmonary venous oxyhaemoglobin saturation if possible.³

Direct patient monitoring and infusion lines are extremely useful and should ideally include measurement of right atrial (RA) pressures by catheterisation of the neck veins. In young infants, cannulation to ECMO is typically obtained via the right side of the neck, as it provides the most direct route to the SVC/RA. In this setting, central venous access via the left neck or groin is preferred. It is ideal to establish an arterial catheter before the initiation of ECMO, preferably in the right radial artery.² Additional lines that may be placed during cannulation to ECMO, either via sternotomy at the time of operation or by thoracotomy in non-operative circumstances, include direct LAP measurement.1, 3, 4 The benefit of direct LAP measurement is that it allows real-time assessment of EDP in the left ventricle (LV), thereby alerting the clinician to improving or worsening LV function.¹¹

Immediately after cannulation to ECMO, systemic anticoagulation with heparin is initiated. A bolus dose of 50 U kg⁻¹ of heparin is usually given, followed by continuous infusion.¹ There is no standard infusion rate; however, a starting dose of 10 U kg⁻¹ h⁻¹ is typical. Anticoagulation is monitored at the bedside using activated clotting time, to a goal of 1.5× normal (typically 180–200 s).¹ Standard heparin assay (Anti-Xa concentration) is also commonly used and measures direct heparin effect.³ The therapeutic goal is 0.4–1 μmol ml⁻¹. Haemorrhage is the most common serious complication on ECMO and occurs in up to 30% of patients.1, 3

Routine antibiotics are not recommended by ELSO guidelines; however, many centres use prophylactic antibiotics when the chest is open.1, 3

eCPR

An estimated 16 000 children in the USA experience out-of-hospital cardiac arrest (OHCA) with an approximately 10% survival rate.⁶ Survival after in-hospital cardiac arrest (IHCA) in children is much better, at approximately 40%.⁶

Initial reports of the use of eCPR emerged in the 1990s and have been increasing steadily in incidence ever since. The use of eCPR now accounts for approximately 9% of recorded cases of ECMO use in the ELSO registry.⁶ Of 2226 infants and children having undergone eCPR in the ELSO registry, survival rate was 39.3%.3, 6

No randomised trial comparing conventional CPR with eCPR exists. Nevertheless, the American Heart Association now recommends that eCPR be considered in children with IHCA and an underlying cardiac diagnosis.¹⁴ There are now a number of large meta-analyses in adults and at least one large observational study in children demonstrating the superiority of eCPR to conventional CPR for refractory IHCA.6, 15, 16, 17, 18, 19 Three recent meta-analyses in adult patients with a mixture of IHCA and OHCA patients found an approximately two-to three-fold improvement in both survival to hospital discharge and good neurological outcome with eCPR compared with prolonged conventional CPR.15, 16, 18 Two of these studies found an increase in the effect size for OHCA when compared with IHCA; however, the third found no beneficial effect of eCPR on OHCA.

In children, eCPR for IHCA>10 min was associated with improved survival to hospital discharge and good neurological condition when compared with conventional CPR alone (40% and 27% vs 27% and 18% respectively), and this relationship persisted after adjusting for covariates and in propensity matched cohorts.¹⁹ It is worth noting that in the above study there was a much higher proportion of patients with surgical heart disease in the eCPR group. However, even after eliminating these patients from the eCPR cohort in a post hoc sensitivity analysis, eCPR was still associated with an improved probability of survival and good neurological outcome (Odds ratio 3.1 and 2.8, respectively).¹⁹

Several prognostic factors have been identified.⁶ Shorter duration of CPR (<30 min), primarily cardiac diagnosis, hypothermic cardiac arrest, and intoxication confer improved survival and neurological outcome.⁶ Nevertheless, survival with intact neurological condition has been reported even after prolonged CPR (up to 50 min).⁶

Implementation of eCPR

The structure of the eCPR programme at a given institution must be individualised given the location's unique requirements. Nevertheless, there are some basic features of a successful eCPR programme that should be given consideration. Institutions offering eCPR should designate an eCPR code team leader with the authorisation to request eCPR.6, 11 Indications and contraindications to eCPR should be discussed ahead of time and based on institutional preferences. A generally available protocol for the initiation of eCPR should be agreed upon before the implementation of the institutions eCPR program.¹¹ Generally accepted indications for eCPR include:11, 14, 15

• (i)aetiology of cardiac arrest deemed ‘easily reversible’;
• (ii)witnessed in hospital cardiac arrest without delay of CPR;
• (iii)no return of spontaneous circulation after ≥10 min of high quality CPR;
• (iv)no co-existent contraindication to ECMO.

Necessary equipment and personal should be readily available, indeed survival after ECPR is most successful after cardiac arrest in the catheter laboratory, suggesting that timely access to skilled personnel and proper equipment enhances survival.⁶ The protocol from the University of Arkansas, in the paper by Fiser and Morris¹¹ is simple and easy to follow. It begins with recognised IHCA. The code team leader then requests that the ECMO team be batch paged for an ECMO code (all members receive notification at the same time). The team consists of the code team leader, cardiac surgeon, perfusionist and cardiac anaesthesia, and on-call cardiac operating room staff. Once the ECMO team agrees to proceed to eCPR, rapid deployment of the primed, sterile ECMO circuit is initiated at the bedside.¹¹

Conclusion

ECMO is a highly specialised and resource intensive therapy that can support a variety of physiological states resulting in cardiopulmonary failure. It has a high morbidity and mortality, and requires specialised skills and knowledge for effective delivery. eCPR has been shown to improve mortality and neurological outcomes in refractory cardiac arrest. The effective delivery of eCPR depends on the rapid availability of trained personnel and equipment.

Declaration of interest

None declared.

MCQs

The associated MCQs (to support CME/CPD activity) will be accessible at www.bjaed.org/cme/home by subscribers to BJA Education.

Biographies

David Rosen MD FRCPC is an attending paediatric anaesthesiologist and chief of the anaesthesia department at the Children's Hospital of Eastern Ontario (CHEO) at the University of Ottawa. He is the head of the cardiac anaesthesia division and practices both general and cardiac intensive care at CHEO and the Hospital for Sick Children at the University of Toronto. His clinical and research interests include cardiac anaesthesia, acute pain management and emergence agitation, and professionalism in anaesthesia.

Colin Meyer-Macaulay BSc MD FRCPC is a general paediatrician and paediatric critical care medicine fellow at the Children's Hospital of Eastern Ontario (CHEO) at the University of Ottawa. His major clinical and research interests include the care of patients with congenital heart disease and the management of prolonged mechanical ventilation in paediatric patients.

Matrix codes: 1A01, 2C03, 3G00