Extracorporeal membrane oxygenation (ECMO) is a potentially lifesaving therapy for patients suffering from cardiopulmonary failure. Extracorporeal membrane oxygenation for primary cardiac failure has increased in the past 2 decades, now exceeding the proportion of children supported for respiratory failure.1 Extracorporeal membrane oxygenation cannulation in children with complex congenital heart disease (CHD) may be complicated by limited access because of vessel injury and thrombosis. Transhepatic access for cardiac catheterization and central venous line (CVL) placement is well described, though its use for ECMO cannulation has not been reported.2–4
We report successful use of transhepatic cannulation for venovenous (VV) ECMO, providing extended support to 15 month-old female with bidirectional Glenn (BDG) anatomy and intractable hypoxia.
A 15-month-old female with hypoplastic left heart syndrome status-post Norwood procedure and BDG was admitted to cardiovascular intensive care unit with increased work of breathing and significant hypoxia. Admission echocardiogram revealed normal systolic function and severe tricuspid valve (TV) regurgitation. Cardiac catheterization demonstrated multiple VV collaterals (subsequently coil occluded) and severely diminished flow into the right pulmonary artery (PA) from atrial compression. Despite aggressive diuresis and afterload reduction, hypoxemia worsened, and she developed progressive lactic acidosis, and intermittent myocardial ischemia. Full mechanical ventilation with inhaled nitric oxide, deep sedation, and paralysis were required to maintain arterial oxygen saturation (SaO2) 50–60%. Venovenous (VV) ECMO was proposed to provide stability by improving systemic oxygen delivery in addition to reversing myocardial ischemia and pulmonary hypertension. Veno-arterial (VA) ECMO using the carotid artery was planned if reversal of hypoxemia with VV ECMO did not improve hemodynamics enough to sustain adequate oxygen delivery.
Central venous access was limited because of vessel occlusion from thrombosis and existing vascular access, including peripherally inserted central catheter in left subclavian vein (SCV). Both the femoral veins, right SCV, and left internal jugular vein (IJ) were occluded. In effort to prevent superior vena cava (SVC) syndrome and preserve right IJ for future catheterizations, VV ECMO was established through a hepatic vein. Extracorporeal membrane oxygenation cannulation was performed at bedside because of patient’s instability. To identify and measure the straightest pathway through liver, ultrasound (US) imaging was performed using 7 cm, 13 MHz linear probe (GE Healthcare, Wauwatosa, WI); targeted hepatic vein was 1.1 cm in diameter. Under US guidance, right hepatic vein was entered percutaneously through long-axis approach with 21 G, 7 cm echogenic needle (Angiodynamics, Queensbury, NY). Four Fr micropuncture catheter (Angiodynamics, Queensbury, NY) was used to upsize guidewire and dilate tract and vessel for placement of 15 Fr, 10 cm long dual lumen ECMO cannula (Origen, Austin, TX). Cannula connected to ECMO circuit and flow titrated to 600 ml/min (cardiac index 1.18 L/min/m2). Chest radiograph (Figure 1) and echocardiogram confirmed tip of ECMO cannula in right atrium.
With ECMO initiation, SaO2 increased from 40%’s to upper 70%’s and SVC saturation improved from teens to upper 40%’s. Over next 5 days, systemic oxygen delivery stabilized and she extubated to noninvasive nasal positive pressure. Patient movement while extubated did not compromise ECMO flow. Extracorporeal membrane oxygenation day 10, in attempt to improve systemic cardiac output, decrease end diastolic pressure, and improve PA flow, TV replacement (25 mm, Epic, St. Jude Medical, St. Paul, MN), and right atrial reduction were performed. Because of hemodynamic instability, she returned from operating room on VA ECMO with 10 Fr aortic cannula (Medtronic, Minneapolis, MN) through open sternum and same transhepatic venous cannula used preoperatively. Over next 5 days, hemodynamics stabilized, aortic cannula removed, and sternum closed, effectively transitioning patient to transhepatic VV ECMO. The following week, she developed sepsis and progressive multiorgan failure. Support was withdrawn on ECMO day 22 because of her irreversible and terminal condition.
We present the first described use of transhepatic VV ECMO in a child with complex CHD and intractable hypoxemia. She presented distinct challenges for ECMO management: limited vascular access, instability requiring bedside procedure, and BDG physiology.
She had prolonged convalescence following a Norwood operation as well as multiple procedures and catheterizations resulting in occlusion of all vessels except right IJ and left SCV. Cannulation of right IJ was avoided, as it was only remaining great vessel draining cerebral circulation. Given our interventional cardiologists’ positive experience with transhepatic access for cardiac catheterization and CVL placement, this route for ECMO cannulation was considered feasible option.
Transhepatic access for cardiac catheterization and CVL placement in children has been described for over 20 years. Largest pediatric series was described by Qureshi et al. (124 transhepatic procedures); 8% of patients experienced major complications including bleeding and heart block.2 Because traditional landmark location does not always identify largest vein with most direct pathway to heart, and in this case bedside fluoroscopy was not available, US guidance was essential for transhepatic cannulation.4 It is critical to avoid the portal circulation and access the hepatic vein with minimal liver trauma, decreasing potential bleeding complications once child is anticoagulated for ECMO. We and others have previously shown challenging access is more readily obtained using real-time US guidance with long-axis approach via linear transducer.5,6
We were able to achieve ECMO flows of 600 ml/min (60 ml/kg/min), improving oxygenation and lactic acidosis. Attempts at higher flows were associated with increased negative return line pressures, likely related to kinking “memory” in cannula caused when cannula was rotated 180° immediately after insertion process (cannula inside body did not rotate when external ports were rotated). Therefore, close attention to orientation of cannula at time of insertion is necessary.
Extracorporeal membrane oxygenation support of patients with BDG physiology is particularly challenging because of separation of systemic venous return and has historically been associated with poor outcomes. Review of six patients with BDG circulation supported with ECMO by Booth et al. had one survivor and noted vessel occlusion impacted site of ECMO cannulation in 50% of patients.7 Recently published query of 103 BDG patients using Extracorporeal Life Support Organization registry found more encouraging results, with 41% of patients surviving to hospital discharge.8
Our patient’s ECMO course was uncomplicated. Improved systemic oxygenation enabled stabilization and opportunity for surgical intervention. Before surgery, she was extubated and eating by mouth. Unfortunately, the multifactorial causes of her hypoxemia were not reversible, including persistent severe diastolic dysfunction compromising PA flow. Despite supporting oxygenation with ECMO for 22 days, she ultimately died as result of multiorgan failure exacerbated by sepsis. Although she was never decannulated, plans to do so included placement of vascular plug device after exchanging ECMO cannula for 10 F introducer sheath.
Transhepatic access for ECMO cannulation is feasible and may be safely used for prolonged ECMO support. Utilization of this cannulation strategy in complex patients with limited central venous access as an alternative to central cannulation warrants further investigation.
1. Brown KL, Ichord R, Marino BS, Thiagarajan RR. Outcomes following extracorporeal membrane oxygenation in children with cardiac disease. Pediatr Crit Care Med. 2013;14(5 Suppl 1):S73–S83
2. Qureshi AM, Prieto LR, Bradley-Skelton S, Latson LA. Complications related to transhepatic venous access in the catheterization laboratory–a single center 12-year experience of 124 procedures. Catheter Cardiovasc Interv. 2014;84:94–100
3. Qureshi AM, Rhodes JF, Appachi E, et al. Transhepatic Broviac catheter placement for long-term central venous access in critically ill children with complex congenital heart disease. Pediatr Crit Care Med. 2007;8:248–253
4. Shim D, Lloyd TR, Cho KI, Moorehead CP, Beekham RH. Transhepatic cardiac catheterization in children. Circulation. 1995;92:1526–1530
5. Alten JA, Borasino S, Gurley WQ, Law MA, Toms R, Dabal RJ. Ultrasound-guided femoral vein catheterization in neonates with cardiac disease. Pediatr Crit Care Med. 2012;13:654–659
6. Law MA, Borasino S, McMahon WS, Alten JA. Ultrasound- versus landmark-guided femoral catheterization in the pediatric catheterization laboratory: a randomized-controlled trial. Pediatr Cardiol. 2014;35:1246–1252
7. Booth KL, Roth SJ, Thiagarajan RR, Almodovar MC, del Nido PJ, Laussen PC. Extracorporeal membrane oxygenation support of the Fontan and bidirectional Glenn circulations. Ann Thorac Surg. 2004;77:1341–1348
8. Jolley M, Thiagarajan RR, Barrett CS, et al. Extracorporeal membrane oxygenation in patients undergoing superior cavopulmonary anastomosis. J Thorac Cardiovasc Surg. 2014;148:1512–1518