Anesthetic Management of a Patient on Venoarterial Extracorporeal Membrane Oxygenation Undergoing Transcatheter Aortic Valve Implantation

A 41-YEAR-OLD male with a complex history including a bicuspid aortic valve with aortic stenosis, a background of inherited cardiomyopathy, and an implantable cardiac defibrillator for prevention of malignant arrhythmias presented to the cardiology department of our hospital with signs and symptoms of congestive heart failure. He previously had undergone a full workup for transcatheter aortic valve implantation (TAVI), including multidetector computed tomography for annular sizing. A transthoracic echocardiogram performed on this patient showed a left ventricular ejection fraction of 15% to 20% with low-flow/low-gradient severe aortic stenosis (aortic valve area: 0.9 cm², peak gradient: 27 mmHg). The multidisciplinary team believed that an urgent aortic valve intervention was warranted and decided to expedite the TAVI.

The cardiac decompensation resulted in cardiogenic shock complicated by acute kidney and liver injury. Due to lack of response to inotropic therapy, peripheral venoarterial extracorporeal membrane oxygenation (VA-ECMO) was instituted via a 17F left femoral arterial cannula, an 8F ipsilateral backflow cannula for distal leg perfusion, and a 25F right femoral venous cannula.¹ Hemodynamic stability on VA-ECMO was established for 48hours to allow time for improvement in organ function before the procedure.

Before induction of anesthesia, the heparin infusion for ECMO anticoagulation was discontinued and the baseline activated clotting time (ACT) was 185 seconds. The implantable cardiac defibrillator was deactivated and external defibrillation pads were applied. General anesthesia was induced with targeted controlled infusions of propofol and remifentanil targeting a bispectral index value between 40 and 60, and rocuronium was given for muscle relaxation. After tracheal intubation, ventilation was managed with lung protective strategies, and the VA-ECMO sweep flow was adjusted to ensure normocapnia (arterial partial pressure of carbon dioxide 35-45 mmHg). Oxygen saturations were measured with a pulse oximeter placed on the right index finger to enable detection of upper body hypoxaemia² and remained >99% throughout the case. Periods of hypotension were treated with fluid boluses and by titrating a norepinephrineinfusion, targeting a mean arterial pressure of 60 mmHg. Immediately after TAVI, cannulation with a 14F sheath was performed and a heparin bolus of 5,000U was administered, yielding an ACT >400 seconds. VA-ECMO was managed by the perfusionist, and flows were kept constant until the time of TAVI deployment, when the VA-ECMO flow was temporarily reduced to 25% of the baseline. A 26mm Sapien 3 valve (Edwards Lifesciences, Irvine, CA) was deployed, followed by a brief run of pulseless ventricular tachycardia, managed with external defibrillation. In line with our routine practice, fluoroscopy alone was used to guide valve deployment, obviating the need for transesophageal echocardiography. At the end of the case, an intra-aortic balloon pump was inserted through the TAVI puncture site in the right femoral artery to prophylactically decompress the left ventricle. On the same evening the patient required femoral artery repair due to bleeding from the insertion site of the intra-aortic balloon pump, which was removed. The patient was extubated on the third postoperative day and was weaned off VA-ECMO 4days after the TAVI. He was referred to and accepted by the regional cardiac transplant center on the basis of improving multiorgan function.

The use of mechanical circulatory support in the form of VA-ECMO has the potential to reduce complications associated with high-risk TAVI procedures^3, 4, 5 but also presents unique challenges to the anesthesia team.

It is feasible to perform TAVI through the transfemoral route with the patient under sedation combined with local infiltration of the peripheral cannulation sites, with evidence from large registries and meta-analyses demonstrating substantial benefits with sedation only.^6, 7, 8 We considered this option for the specific patient, but we elected to administer a general anesthetic because the patient’s dyspnea and mild confusion precluded easy adoption of the supine position and his immobility for the duration of the case.

There is limited evidence as to the pharmacokinetic changes expected in adults on ECMO.^9, 10, 11 Issues include the increased volume of the circuit leading to hemodilution; the sequestration of lipophilic drugs within the circuit tubing; and the absorption of proteins, especially albumin, onto the circuit, which can result in increased free drug. Organ dysfunction in critical illness can compound these changes. Therefore, we aimed to deliver a cardiostable induction with short-acting drugs and believe that remifentanil and propofol target controlled infusions best allowed for this close titratability. We started at deliberately low doses and further reduced maintenance anesthetic doses in order to achieve the bispectral index targets mentioned previously.

With regard to periprocedural hemodynamics, VA-ECMO provides a safety net because it can minimize hemodynamic perturbations during certain phases of the TAVI procedure. In that context, the significance of the underlying disease is greatly reduced compared with situations without extracorporeal circulation. Nevertheless, the collaboration between the anesthesia and perfusion teams remains important because any reduction in the ECMO flows can indicate hypovolemia, tamponade, or high systemic vascular resistance, and the manipulation of the flows can restore an adequate mean arterial pressure when the administration of fluids or vasopressors fails to improve hemodynamics.²

Anticoagulation management is of crucial importance¹² because exposure of blood to the artificial components of the ECMO circuit triggers the coagulation pathway, and TAVI carries a significant thromboembolic risk. Unfractionated heparin is the most widely used systemic anticoagulant in both settings, with ACT being the most common monitor of anticoagulation. In the absence of specific guidance on the perioperative management of anticoagulation during ECMO, our objective was to reconcile the need to minimize bleeding risk at arterial puncture sites with the ongoing requirements for anticoagulation. Discontinuing the intravenous heparin infusion before induction of anesthesia allowed for safe femoral vessel cannulation, and a subsequent heparin bolus provided an adequate intraprocedural level of anticoagulation. Protamine was not administered at the end of the procedure because the circuit and membrane gas exchanger are heparin-bonded and therefore very likely to thrombose acutely if used. It is possible that alternative methods of anticoagulation monitoring, such as thromboelastography, could have been used to provide a more accurate perioperative anticoagulation management,¹³ although lack of immediate availability precluded its use in our case (Table1).

In summary we present a report on the anesthesia management of a VA-ECMO patient undergoing TAVI. Given the uncommon nature of these procedures thus far and the relative paucity of similar reports in the literature, we believe that our approach may prove helpful in teams managing similar patients in the future.