Extracorporeal membrane oxygenation (ECMO) is used in patients with acute respiratory failure (ARF) not responsive to conventional management. The practice of awake ECMO for patients with ARF has become an area of interest but with limited experience. There have been reports of awake ECMO in adults in various clinical scenarios, with the most experience reported in bridging to lung transplantation.1,2 More recently, awake ECMO has been reported in the pediatric population.3
The practice of awake ECMO in neonatal patients has been reported only as an exceptional measure in rare cases3; traditional management of these patients predominates, characterized by variable ventilator rest strategies and need for sedation to provide the required support and care safely. We present a series of eight neonates who were electively extubated while on ECMO for ARF, as well as expanded discussion of two representative cases (Table 1). This case series was approved by the Institutional Review Board at our institution, with waiver of consent.
A male neonate delivered by urgent cesarean section for fetal bradycardia at 39 weeks gestation and birthweight of 2,950 g presented with acute hypoxic respiratory failure secondary to meconium aspiration syndrome and bilateral pneumothoraces. He was placed on venoarterial (VA) ECMO on day of life 2 due to progressive hypoxic respiratory failure not responsive to conventional therapies. The infant was cannulated with a 12 Fr venous cannula in the internal jugular vein and an 8 Fr arterial cannula in the carotid artery. Despite three thoracostomy tubes and protective ventilator settings, significant air leak persisted. He was extubated on ECMO day 2 to eliminate positive intrathoracic pressure with the intent to allow healing of air leak while providing full ECMO support for respiratory failure. During this time, he was maintained on ECMO flow of 100 ml/kg/min, sweep 0.1 L/min, and sweep fraction of inspired oxygen (FiO2) of 0.40–0.50. Within 48 hours, the left pneumothorax had resolved. Within 96 hours, complete resolution of bilateral pneumothoraces was achieved (Figure 1). The infant received only minimal necessary sedation with morphine at 0.020 mg/kg/hr for comfort and to minimize excessive movement. The infant tolerated enteral feeds and participated in low-volume oral feeding trials (Figure 2) of 10 ml per feeding. We experienced no cannula or circuit-related complications as a result of extubation. After 4 days, ECMO day 6, the infant achieved adequate lung recruitment with only spontaneous breathing and had a successful trial off ECMO on nasal cannula 2 L/min, FiO2 0.80. Because cannulation had been performed by open-cut down, the infant was reintubated for surgical decannulation and was again extubated 12 hours later to humidified high-flow nasal cannula (HHFNC) at 4 L/min, FiO2 0.30. Five days postdecannulation, he was weaned off respiratory support entirely.
This was a male infant delivered by cesarean section at 39 weeks and birthweight of 2,910 g with left congenital diaphragmatic hernia, pulmonary hypoplasia, and left lung malposition and malformation. On the day of life 7, he was placed on venovenous (VV) ECMO using a bicaval cannula in the right atrium for worsening hypoxia due to pulmonary hypertension despite aggressive conventional management. He underwent diaphragmatic hernia repair on ECMO day 6 secondary to the development of a thoracic compartment-like syndrome that was complicating ECMO flow. His early ECMO course was complicated by pulmonary hemorrhage, resulting in complete airway and lung opacification. Attempts to remove airway debris and obstruction were unsuccessful. Given the nature and fragility of his pulmonary malformations, as well as anticipated prolonged ECMO course, every effort was made to manage his airway and lungs atraumatically. Continued ventilator support was felt to be nontherapeutic in the presence of ongoing airway obstruction, so he was extubated on ECMO day 8 to nasal cannula. He required ECMO flow of 120–140 ml/kg/min with sweep flow 0.8–1.2 L/min at FiO2 of 1.0. He was reintubated on ECMO day 13 for the administration of perflubron for atraumatic airway clearance and lung recruitment. This resulted in adequate lung recruitment and return of native lung function. The infant was successfully decannulated on ECMO day 16. His post-ECMO course was complicated by progressive pulmonary hypertension despite aggressive treatment. The parents requested redirection of care, and all medical and surgical therapies were discontinued. During his awake, extubated course on ECMO, his parents were able to interact with and establish a positive bond with him, despite his severity of illness and ultimate demise.
The practice of awake ECMO in extubated patients has emerged as an option for respiratory management with several advantages. When used as bridge-to-lung-transplantation in patients who experience pretransplant clinical decline, awake ECMO extended transplant candidacy by improving pretransplant clinical status, specifically, neuromuscular conditioning through physiotherapy and improved nutrition.1,2 Furthermore, it yielded greater survival compared with historical controls managed with mechanical ventilation. Awake ECMO has also been reported in adult and pediatric patients with other causes of respiratory failure, as well as cardiogenic shock.3,4 These reports challenge established patient management strategies on ECMO and demonstrate a role and benefit of awake ECMO. Still, common pulmonary management on ECMO continues to rely on mechanical ventilation for native lung support, as well as the use of high doses of sedatives.
Awake ECMO has not been sufficiently explored in the neonatal population. In 2014, we extubated our first neonate on VA ECMO. We have since adopted a strategy of selective extubation and awake management of neonates on ECMO. Our experience demonstrates that, in select patients, extubation and appropriately decreased sedation is safe and does not itself compromise the effective delivery of extracorporeal life support.
The greatest benefit of this strategy has been observed in patients with air leak. Five patients in our series were extubated to expedite the resolution of pneumothorax. Three cases (1, 4, and 8) demonstrated that this strategy is effective in healing air leak and that spontaneous breathing alone can be sufficient to recruit the lungs for successful weaning and decannulation from ECMO.
Pneumothorax complicating ECMO presents a specific and difficult challenge to management. Tension pneumothorax can cause hemodynamic instability secondary to decreased venous return and, subsequently, ECMO flow. Furthermore, pneumothorax may prolong ECMO to allow for resolution and appropriate lung recruitment. In this high risk, anticoagulated population, chest tube placement is associated with a significant risk of complication and mortality.5 For these reasons, chest tube placement is generally reserved for cases in which pneumothorax acutely compromises ECMO flow and patient hemodynamics. We have shown that extubation on ECMO is a safe alternative for neonatal patients with pneumothorax.
Interestingly, cases 5 and 6, who were also extubated due to air leak, were reintubated due to a perceived increase in work of breathing. Allowing patients to remain awake with spontaneous respirations on ECMO may require alternations to ECMO management to relieve work of breathing. Crotti et al.6 demonstrated that it is feasible to allow patients on ECMO to breathe spontaneously and that dyspnea was relieved by CO2 removal through ECMO. In the case of patient 5, arterial blood gases before reintubation demonstrated a respiratory acidosis with a pH 7.26–7.29 and partial pressure of carbon dioxide (pCO2) 52–58 mm Hg. In the case of patient 6, arterial blood gas at the time of reintubation showed a pH of 7.32 and partial pressure of carbon dioxide (pCO2) of 58 mm Hg. It is possible that these cases may also have ultimately achieved adequate lung recruitment with spontaneous breathing over time with a more balanced management of sedation and an altered ECMO respiratory strategy.
Patients 2 and 3 required more invasive intervention to facilitate decannulation from ECMO due to more complicated underlying pulmonary disease. Both had significant airway plugging, necessitating more aggressive pulmonary toilet. Patient 2 required perflubron administration for airway clearance, and patient 3 required therapeutic bronchoscopy. Both patients ultimately died. Case 2 had a left congenital diaphragmatic hernia with recalcitrant pulmonary hypertension and other congenital airway and pulmonary anomalies, which complicated his ECMO course and contributed to his ultimate demise. Case 3, who required ECMO for severe respiratory syncytial virus (RSV) bronchiolitis and pneumonia, was able to be decannulated and showed steady improvement from her respiratory disease; she ultimately died from complications of a procedure unrelated to her ECMO course after returning to the referring hospital. Still, both cases demonstrated benefits to awake ECMO without compromising the provision of extra-corporeal life support. Both cases were able to wean significantly on sedation. This was especially important for case 2, for whom awake ECMO allowed for family interaction and meaningful bonding in a situation where the overall outcome was expected to be poor. Allowing patients to be awake without a noxious endotracheal tube can enable weaning of pharmacologic sedation, facilitating physiotherapy and rehabilitation, and permitting interaction with family and caregivers.
Scant literature exists on the ideal ventilator approach to lung rest during ECMO, and significant variability exists. Limited evidence suggests an advantage of higher positive end expiratory pressure (PEEP) and maintenance of open lung rest.7,8 Traditionally, respiratory management favors open lung rest and maintenance of some PEEP to prevent complete lung collapse. Our experience in neonates suggests that, while spontaneous breathing during ECMO may result in undesired initial alveolar volume loss, respiratory efficiency improves and spontaneous recruitment can occur without any positive pressure as pulmonary disease improves. The patients in cases 1, 4, and 7 demonstrated initial lung collapse, as expected upon extubation. Over time, with pulmonary recovery, diuresis, and improved respiratory efficiency, these patients demonstrated spontaneous alveolar aeration and adequate lung recruitment to allow for ECMO decannulation. This is demonstrated in the progressive chest x-rays in case 1 (Figure 1). We recognize that this observation was made in a small number of select patients with similar pathology and may not be generalized to all patients with respiratory failure receiving ECMO.
A number of our patients demonstrated additional benefits to extubation on ECMO. Oral feeding was attempted in patients who were receiving a level of respiratory support that permitted safe oral feeding, specifically nasal cannula or high-flow nasal cannula support of 2 L/min or less. This included patients 1, 4, and 7. Those patients in whom feeding was deemed appropriate were bottle feeding while awake on ECMO. Figure 2 illustrates bottle feeding by case 1, and equal success with regular oral feeds was achieved in case 4. Furthermore, all patients extubated on ECMO were receiving regular physical and occupational therapy at the bedside.
The practice of awake neonatal ECMO requires significant team preparation and careful consideration of the balance of potential risks versus benefits on an individualized basis. From our experience, we recommend consideration of the following:
- Duration of anticipated ECMO course
- Stability of the ECMO circuit
- Individualized risks versus benefits of extubation
- Readiness and capacity for emergent reintubation
- Readiness and commitment of the entire team
- Discussion with the family, including informed assent
To our knowledge, there are no published trials comparing awake ECMO to traditional ECMO in neonatal patients. We have shown that awake neonatal ECMO appears safe and may offer significant advantages over traditional management in certain clinical scenarios, particularly in cases of persistent air leak. Advantages include avoidance of ongoing ventilator-induced lung and airway trauma, resolution of air leak while avoiding the hemorrhagic risks associated with invasive drainage, de-escalation of pharmacologic sedation, mitigation of neuromuscular deconditioning, and improving family bonding. As the native lungs heal, particularly with resolution of air leak, spontaneous breathing is often sufficient to recruit the neonatal lungs without aggressive invasive pulmonary maneuvers. Our report is limited by a relatively small sample size and lack of controls for comparison of clinically important outcomes, such as time on ECMO, duration of post-ECMO respiratory support, time to oral feeding, duration of sedation and analgesic support, patient and family experience, and hospital length of stay. More experience is necessary to define clear patient selection criteria to identify patient in whom awake ECMO provides the most benefit. Prospective comparison trials are warranted to further explore the clinical benefits and safety of awake neonatal ECMO.
1. Olsson KM, Simon A, Strueber M, et al. Extracorporeal membrane oxygenation
in nonintubated patients as bridge to lung transplantation. Am J Transplant 2010.10: 2173–2178.
2. Fuehner T, Kuehn C, Hadem J, et al. Extracorporeal membrane oxygenation
in awake patients as bridge to lung transplantation. Am J Respir Crit Care Med 2012.185: 763–768.
3. Schmidt F, Jack T, Sasse M, et al. “Awake veno-arterial extracorporeal membrane oxygenation
” in pediatric cardiogenic shock: A single-center experience. Pediatr Cardiol 2015.36: 1647–1656.
4. Yeo HJ, Cho WH, Kim D. Awake extracorporeal membrane oxygenation
in patients with severe postoperative acute respiratory distress syndrome. J Thorac Dis 2016.8: 37–42.
5. Jackson HT, Longshore S, Feldman J, Zirschky K, Gingalewski CA, Gollin G. Chest tube placement in children during extracorporeal membrane oxygenation
(ECMO). J Pediatr Surg 2014.49: 51–53.
6. Crotti S, Bottino N, Ruggeri GM, et al. Spontaneous breathing during extracorporeal membrane oxygenation
in acute respiratory failure. Anesthesiology 2017.126: 678–687.
7. Schmidt M, Stewart C, Bailey M, et al. Mechanical ventilation management during extracorporeal membrane oxygenation
for acute respiratory distress syndrome: a retrospective international multicenter study. Crit Care Med 2015; 43: 654–664, 39:847–856, 2013.
8. Alapati D, Aghai ZH, Hossain MJ, Dirnberger DR, Ogino MT, Shaffer TH; Extracorporeal Life Support Organization Member Centers: Lung rest during extracorporeal membrane oxygenation
for neonatal respiratory failure-practice variations and outcomes. Pediatr Crit Care Med 2017.18: 667–674.