Therapy for Cardiorespiratory Failure in the Neonate

Basic Techniques

The standard ECMO procedure used most frequently in newborns with HRF today is venoarterial (VA) bypass, which provides both pulmonary and cardiac support. The right atrium is cannulated via the right internal jugular vein with a Silastic or polyvinyl chloride catheter (8- to 14-French diameter). Whether a roller head pump or centrifugal pump is utilized, blood is siphoned from the right atrium and circulated through the artificial lung. These pumps are servoregulated to slow or shut down if venous return is not adequate to meet circuit flow demands. As the blood circulates through the artificial lung, gas exchange occurs against a filtered mixture of oxygen, CO ₂ , and nitrogen in ambient air. The artificial lung commonly used in neonates in the United States is a hydrophobic, pediatric-specific, polymethylpentene hollow fiber membrane (Quadrox iD [diffusion membrane oxygenator]) that serves as a blood-gas interface, similar to the alveolar capillary membrane. Gas flows through the hollow fibers and venous blood flows around the outside of these fibers. Oxygen and carbon dioxide flow down their diffusion gradients, and gas exchange occurs across the hollow fiber membrane. Similar to the native lung, blood leaving the membrane lung has a normal to low CO ₂ and a high Pa o ₂ , with oxygen saturation of 100%. Depending on the size of the infant, CO ₂ may need to be added to the gas mixture in order for the blood leaving the membrane lung to have a normal pH and pCO ₂ . Advantages of this newer artificial lung include a low priming volume and reduced blood/foreign surface area contact. This new generation of artificial lung also incorporates a heat exchanger to rewarm the oxygenated blood returning to the patient.

Oxygenation can be regulated by varying blood flow through the ECMO circuit. The higher the volume of cardiac output diverted through the membrane lung, the better the oxygen delivery from the ECMO circuit. Oxygenated blood exiting the artificial lung is returned to the infant via an 8- to 12-French catheter positioned in the ascending aortic arch through a right common carotid artery cannulation.

Blood flow through the ECMO circuit, at a rate of 80 to 100 mL/kg per minute, is usually adequate to provide excellent cardiac and respiratory support with maintenance of adequate blood pressure and oxygenation. Arterial wave dampening with narrowed pulse pressure is noticeable at flow rates approaching the infant's cardiac output, because the ECMO circuit provides relatively nonpulsatile blood flow. Pressor support and vasodilators can usually be stopped while the infant is on venoarterial ECMO support, because perfusion pressure from the pump largely replaces cardiac output. Only mild sedation—allowing the babies to breathe spontaneously, open their eyes, and move their extremities—is recommended on bypass support.

Systemic anticoagulation therapy with unfractionated heparin is administered for the duration of the bypass procedure to prevent clotting in the circuit and possible thromboembolization. Activated clotting times (ACTs) are measured at bedside every 1-2 hours and are maintained within a range of 180-240 seconds. Lower ACT targets are chosen in patients with elevated risk for bleeding (such as preterm infants or postoperative cases), although this increases the risk of thrombus formation in the ECMO circuit. Some centers monitor anticoagulation with antifactor Xa assays, maintaining values between 0.3 and 0.7 unit/mL. Thromboelastography (TEG) can also be used to monitor coagulation status during an ECMO run.

Bivalrudin, a reversible direct thrombin inhibitor, is used instead of heparin for anticoagulation in some centers. Bivalrudin is particularly useful as an anticoagulant in cases where heparin-induced thrombocytopenia (HIT) is suspected. This complication, although not common, can occur in infants as well as adults, particularly in infants who have had cardiac surgery.

Once the infant is placed on ECMO, ventilation and airway oxygen support are reduced to avoid further ventilator-induced lung injury (VILI) while maintaining functional residual capacity (FRC). Wide practice variation exists with regard to ventilator settings used for “lung rest” during bypass support. A study was published detailing ventilator practices during ECMO, derived from more than 3000 neonatal cases submitted to the ELSO database between 2008 and 2013. Conventional mechanical ventilation (CMV) was used in 88% of cases while 12% used high-frequency ventilation (HFV) during ECMO. Most practitioners who used CMV for lung rest used a rate of 10 cycles/minute and peak inspiratory pressures (PIP) of 15-20 cm H ₂ O. Fi o ₂ use ranged from 21%-40%. There was far more variation in the use of positive end expiratory pressure (PEEP) than in other parameters: 32% used low PEEP (4-6 cm H ₂ O), 43% used high PEEP (10-12 cm H ₂ O), and 22% used a mid-PEEP range (7-9 cm H ₂ O). The mean time on ECMO was reduced by 22 and 21 hours in the high and median PEEP groups, respectively. This confirms an older report where high PEEP shortened ECMO duration by 32 hours. Lower PEEP may be used if the patient has a pneumothorax or pneumomediastinum. Use of HFV compared to CMV was associated with longer ECMO duration, although these were sicker patients with increased oxygenation indices (OIs), lower pH, and lower blood pressure prior to ECMO initiation. If volume ventilation is preferred, 4-6 mL/kg tidal volume can be used to achieve mild chest rise with each delivered breath. The goal is to maintain FRC and to allow for continued pulmonary toilet. The respiratory status of the infant can be monitored with intermittent arterial blood gases obtained from the umbilical or peripheral arterial line. An oxygen saturation electrode inserted on the venous side of the circuit allows continuous monitoring of the mixed venous saturation at the level of the right atrium. A mixed venous saturation of 70% or greater reflects adequate oxygen delivery. The efficiency of the membrane lungs in use today allows most neonates to have normal mixed venous saturations throughout the ECMO course.

As lung function improves and pulmonary hypertension abates, the mixed venous saturation and partial pressure of arterial oxygen (Pa o ₂ ) rise above the baseline oxygenation provided through the artificial lung. Blood flow through the artificial circuit can then be decreased in small increments (10-20 mL) while mixed venous and arterial oxygenation remain adequate. When ECMO flow has been reduced to 10-20 mL/kg per minute, the infant can usually be weaned from extracorporeal support.

Increased Fi o ₂ and ventilator settings are provided as the patient approaches decannulation, although the support required is usually far less than needed prior to ECMO. The duration of bypass support may be anywhere from 2-60 days with an average need for 9.2 days of support. Two decades ago, the average time on ECMO support was only 5 days. The significant increase in ECMO-run duration and marginal increase in mortality rate among neonatal ECMO patients compared to years past likely reflects a sicker cohort of patients now requiring ECMO consideration. In 1992, 1345 neonates with HRF were placed on ECMO in the United States whereas, in 2017, only 533 infants required ECMO among cases reported to the ELSO Registry. Over the same time period, the number of neonatal cardiac cases increased substantially to 275 in 2017. Continued ventilator support following ECMO can be quite variable (days to weeks) depending on the underlying cause of respiratory/cardiac failure, as well as the degree of barotrauma incurred before initiation of ECMO.

In most ECMO centers, the right carotid artery and internal jugular vein are permanently ligated at decannulation. This approach is technically easier and prevents any concern about acute embolic complications associated with vascular reconstruction. However, some surgeons reanastomose these vessels if the patient has been on bypass less than 7-10 days to prevent future ischemic risks to the developing newborn brain as well as during later adult life. Studies show no disadvantage to this approach. Magnetic resonance angiograms (MRAs) and Doppler flow studies confirm good antegrade flow through the reconstructed carotid artery in the majority of these infants with no embolic sequelae reported. Neurodevelopmental outcome and neuroimaging were equally favorable. However, one study in 2006 reported a 72% occlusion or stenosis of the right common carotid artery (RCCA) at 2 years of age in a select cohort of 18 survivors with CDH. There was no difference in neurologic impairment compared to ligated controls in this study, however. In 2015, a single center retrospective study of 51 ECMO survivors with varying diagnoses who underwent right common carotid reconstruction (RCCR) (at <7 days ECMO support) had Doppler ultrasound or MRA assessment of carotid patency at a median age of 63 days. Thirty-seven of 51 (73%) survivors who underwent carotid repair had follow-up imaging to determine carotid artery patency as well as auditory brainstem evoked response (ABR) testing prior to hospital discharge. Thirty-one of 37 (83%) patients showed good antegrade blood flow through the RCCA. Compared to ligated infants, there was no difference in right-sided brain lesions or failed ABR at discharge among infants with carotid reconstruction.

Longer term follow-up of this cohort is of interest to pediatric surgeons and neonatologists alike. Unfortunately, carotid reconstruction is time sensitive and the longer ECMO runs reported over the past 5 years may preclude arterial reanastomosis in most. Beyond 7-10 days, scar tissue encasing the carotid artery at the cannulation site, a part of the natural healing process, often makes reanastomosis impossible.

Venovenous Extracorporeal Membrane Oxygenation

A growing experience supports the ability of venovenous ECMO to provide adequate oxygen delivery in patients with serious respiratory failure but adequate heart function. In adults, this technique usually involves draining desaturated blood from the right atrium and returning oxygenated blood through the femoral vein. Similarly, the majority of pediatric ECMO patients with HRF are supported with venovenous (VV) ECMO, utilizing a double lumen catheter in the right atrium or a two-vein cannulation.

VV bypass in newborn infants has been greatly improved by the development of double-lumen catheters of varying sizes (13, 16, 19 French) that allow bypass support with cannulation of the right atrium alone. The 13 French catheter is appropriate for term newborns >2.5 kg birth weight. Desaturated blood is withdrawn from the right atrium through the outer fenestrated venous catheter wall. Oxygenated blood from the ECMO circuit is returned to the inner arterial, single lumen catheter, which is positioned to direct blood across the tricuspid valve. Higher flow rates are required to maintain adequate oxygen delivery with VV bypass, owing to the obligatory mixing of fully saturated blood returning from the ECMO circuit with desaturated systemic venous return within the right atrium.

Advantages of VV bypass include avoidance of carotid artery cannulation, which is highly desirable, and maintenance of normal pulmonary blood flow. The major disadvantage is that, unlike VA bypass, VV ECMO does not provide cardiac support. Oxygen delivery in VV bypass remains entirely dependent on adequate native cardiac output. The severity of cardiac dysfunction in ECMO-sick patients is highly variable, and often the heart function will improve once adequate oxygen content is restored on VV ECMO. Out of 3052 neonatal ECMO cases reported to the U.S. ELSO Registry between January 2013 and January 2018, 1010/3052 (33%) were supported on VV bypass with 75% survival; survival among infants supported on VA bypass, in comparison, was 61%. VV ECMO complication rate and length of bypass support also compare favorably with VA bypass. Only 10%-12% of VV cases require conversion from VV to VA support should the patient's cardiac output prove to be subpar. Therefore, VV bypass should always be the first consideration in infants whose primary diagnosis is pulmonary disease.

Personnel Needs

ECMO is the most labor-intensive procedure in the neonatal intensive care unit (NICU). Specialists trained in managing patients on ECMO must remain at the patient's bedside for the duration of bypass. In addition to monitoring the infant's respiratory status, specialists must regulate the rates of blood and gas flow through the extracorporeal circuit to meet the infant's metabolic and respiratory demands. They must adjust anticoagulation by frequently measuring the activated clotting time or antifactor Xa assays in the blood and titrate the heparin infusion accordingly. Additionally, they must evaluate the patient for bleeding and replace losses appropriately. The specialists can be physicians, nurses, perfusionists, or respiratory therapists who have completed extensive training in ECMO support.

An ECMO physician trained in the clinical management of bypass patients must always be readily available for consultation, especially in case of mechanical failure or acute clinical decompensation. Additional support services required include 24-hour availability of personnel trained in radiology and ultrasonography, pediatric surgery, neurology, genetics, cardiology, and cardiothoracic surgery. The expertise and personnel needed to support these patients are extensive and costly.

Criteria for Patient Selection

The cumulative experience of many ECMO centers has resulted in the establishment of guidelines that are currently used to decide whether ECMO support is appropriate. These are described in Box 70.1 .