In emergency situations, extracorporeal membrane oxygenation (ECMO) temporary mechanical support adoption for treatment of refractory cardiogenic shock (CS) is currently implemented, because of its simplicity and time-saving.1,2 However, a delicate concern while on peripheral veno-arterial (v-a) ECMO therapy is the retrograde flow in the aorta toward the left ventricle (LV) with increase of LV afterload.1–7 The consequent LV pressure overload may account for LV dilatation, increase in left atrial (LA) pressure, and pulmonary edema. Furthermore, LV overload increases the wall stress thus negatively influencing the process of myocardial function recovery. If the overload is extreme and LV contractile impairment significant, the aortic valve may remain closed permanently, causing blood stasis and predisposing a setting for thrombi formation in the left heart chambers.
A wide variety of LV venting techniques have been reported in the literature1–7 in terms of surgical and percutaneous approaches but all showed high risk of cardiac trauma/lesion, air embolism, bleeding, aortic valve regurgitation, and system displacement while on mechanical support running.
In this article, we describe the LV decompression while on peripheral ECMO running by adoption of a percutaneous insertion of a venous cannula into the pulmonary artery (PA) trunk, in a selected case series of adult patients with profound myocardial depression in which the direct LV apical venting resulted to be not recommended.
Patients Population and Surgical Technique
Between 2017 and 2018, eight patients (three females, mean age: 49.6 years old, and five males, mean age: 58 years old, respectively) suffered refractory CS due to acute myocardial infarction (AMI) (n = 4), acute myocarditis (n = 2), acute decompensation on chronic heart failure (n = 1), and primary graft failure after heart transplantation (Htx) (n = 1), respectively (Table 1). In a single postcardiotomy case, the PA trunk was vented centrally.
In all cases, after a multidisciplinary CS team discussion, it was decided to proceed with femoro-femoral (in six cases) and femoro-axillary (in one post-AMI case since limb ischemia after intra-aortic balloon pump [IABP] support) v-a ECMO placement associated with percutaneous LV venting by usage of a Bio-Medicus NextGen (Medtronic) cannula via right internal jugular vein access to reach and drain the main PA, in the hybrid operating room (OR) (Table 1).
Femoral cannulation was performed, on the right groin after cut-down and vessels exposure, traditionally via Seldinger technique by usage of DLP Bio-Medicus (Medtronic) cannulae (21 Fr, venous drainage and 19 Fr, arterial return) (Figure 1). The axillary artery cannulation was performed, on right side, similarly via Seldinger technique and same Bio-Medicus cannula (19 Fr) (patient #3, Table 1). In a single case, a partial central ECMO cannulation was achieved, as being a postcardiotomy scenario, by inserting a 15 Fr Bio-Medicus NextGen (Medtronic) cannula directly into the main PA trunk (patient #6, Table 1), via Seldinger technique as well. The PLS with Quadrox D oxygenator (Getinge, Maquet) circuit was connected to the cannulae and the pump system (Cardiohelp [Getinge, Maquet] in five and Levitronix CentriMag [Abbott, St. Jude] in three, respectively).
A right IJ venous access was established using direct ultrasound visualization. A Lunderquist guidewire (Cook) was advanced under fluoroscopic guidance into the right atrium. A 15 Fr (50 cm in length) Bio-Medicus NextGen (Medtronic) cannula was then advanced over the Lunderquist wire with its distal tip positioned at the level of tricuspidal valve. The Lunderquist wire was removed and a Swan-Ganz catheter (Edwards Lifesciences, Irvine, CA) was advanced into the Bio-Medicus cannula with its distal tip positioned in the main PA. The cannula was then advanced over the Swan-Ganz catheter to get the PA too. The Swan-Ganz catheter was removed and the cannula connected through a “y” line to the venous drainage of the ECMO circuit (Figure 1).
Extracorporeal Membrane Oxygenation Management
The ECMO blood flow was adequately adjusted during the first 24–48 hours to maintain a cardiac index of 2.6 L/min/m2, by pulse contour cardiac output and noninvasive ClearSight system (Edwards Lifesciences, Irvine, CA) analyses plus Cardiohelp display monitorings (Getinge, Maquet), when adopted, a mixed venous oxygen saturation (SvO2) around 70%, a central venous pressure of 8–10 mm Hg, and a mean arterial pressure of 60–65 mm Hg.
The pulmonary drainage flow was manually adjusted by usage of an external flow-probe device (Spectrum Medical) and a surgical clamp on the transjugular ECMO venous line according to the eventual clinical evidence of significant pulmonary edema, including pinky frothy endotracheal secretions and X-chest ray findings, elevated central venous pressures, transesophageal or transthoracic echocardiographic evidence of LV distension, stasis, contrast “smoke” sign, intracardiac thrombus, reduced LV ejection fraction (EF), and intermittent or absent opening of the aortic valve, refractory ventricular arrhythmias such as ventricular tachycardia.
Before cannulation, all patients received an intravenous heparin bolus (40–80 units/kg). During ECMO support, heparin was administered continuously to achieve an activated clotting time (ACT) of 160–180 seconds and a prothrombin time value of 70–80, with the aim of avoidance of any eventual intracardiac clotting formation around the pulmonary cannula or vascular pulmonary thromboembolic event occurrence.1 Infusion of antithrombin III (AT III) was required if the AT III serum level was below 80%.
No inotropes were given while on ECMO running. All ECMO support was conducted under normothermia. Closed heart examinations by transesophageal echocardiography (TEE) were performed to assess left and right ventricle (RV) motion and unloading, daily.
Vascular ultrasonography assessments of the tibial artery flow and pulsatility were performed every 2 days to evaluate and provide an adequate distal leg perfusion. In two patients (Table 2), due to the absence of both anterior and posterior tibial artery flow and a mean pressure of the superficial femoral artery < 50 mm Hg, a distal perfusion by adoption of a percutaneous 8 Fr. Straight catheter (Teleflex, Arrow), through a T line coming off the main arterial femoral ECMO cannula, was placed to avoid the deleterious consequences of limb malperfusion. In all remaining cases, the arterial cannula (17 Fr) was judged to be not fully occlusive and a distal perfusion catheter insertion was considered to be not necessary.
All patients needed blood transfusions to achieve a hematocrit of 28–30%, and platelet infusions were given when the platelet count was <50,000–60,000.
Mechanical ventilation was continued throughout ECMO support in all patients The ventilator setting was commonly set at a tidal volume of 8 ml/kg, 8 breaths/min, positive end expiratory pressure < 10 cm H2O, and an FiO2 of 0.40–0.60. An IABP was employed in all patients to reduce the afterload and to improve coronary perfusion and maintain a pulsatile flow (Table 1).1
No need of oxygenator or circuit change occurred during the ECMO support period in all patients’ population. No significant hemolysis occurred while on ECMO running.
At our institution, no attempts to wean off ECMO were usually considered during the first 72 hours. Criteria for weaning include an SvO2 ≥ 70%, a hematocrit of 28–30%, the absence of bleeding or tamponade, a left ventricular EF ≥ 40% with an aortic time–velocity integral >10 cm on echocardiography and absence of LV distension after externally clamping the ECMO PA drainage line, acceptable RV contraction with the absence of moderate to severe tricuspid regurgitation, normal blood lactate levels (<1.5 mmol/L), and a normal urine output (>80 ml/h).
A gradual weaning by reducing the ECMO flow by 10% every ~12 hours was our main strategy, under strict TEE examinations and full systemic intravenous heparization. Once an ECMO flow of 1.5 L/min/m2 was reached, in the presence of two or more consultant surgeons, the pump flow was radically reduced at 0.5 L/min/m2 for ~20 minutes with the PA drainage line clamped and using an IABP support set at 1:1. If the hemodynamics in terms of systemic arterial pressure (mean pressure > 60 mm Hg), LV contractility (EF ≥ 40%), central venous pressure (10–15 mm Hg), and SvO2 (>70%) showed no significant changes without the addition of new inotropes, the ECMO support was removed surgically within the next 1 hour in the OR, in most of the cases for safety, or at bedside, in a couple of cases. In all patients, the IABP support was maintained for at least 5 days after ECMO removal.
Both procedure and study have been approved by our institutional review board.
Mean ECMO support time was 8.5 days (Table 2). Seven (87.5%) patients were successfully weaned from ECMO since full recovery of myocardial function and the ECMO explanted in the ICU in the majority of cases. One patient (patient #4) was successfully bridged to heart transplantation (Htx). ECMO-related adverse events are listed in Table 2. All patients were successfully discharged home or joined a rehabilitation center after ECMO treatment except for a single case (patient #7) who died due to severe sepsis (Table 2).
The most important concern in the current widespread use of peripheral v-a ECMO therapy is the high risk to develop severe LV distension which compromises myocardial function recovery and leads to pulmonary edema.1–7
Prevention and treatment of LV overload in v-a ECMO may, therefore, prove useful in a number of situations, but, at the same time, the optimal technique and the target patient population who will actually benefit from venting procedures remain unclear. In addition, all well-known procedures reported are invasive, thus increasing bleeding risk and all eventual ECMO-related complications rate, and therefore the risk/benefit ratio should be carefully assessed preoperatively.2
According to the classification of unloading techniques, the most common locations of unloading result to be the left atrium (31%), followed by the Aorta/IABP (27%) and the transaortic approach by usage of microaxial-flow pumps (27%). The LV itself (11%) and the PA (4%) are also adopted for unloading.2
The PA drainage is currently adopted during less invasive aortic valve surgery (AVR) procedures while on cardio-pulmonary bypass (CPB) running in case of not technically feasible LV venting traditionally through the right pulmonary vein or in the case of surgeon preference in terms of technical minimally-invasive setting.8 This provides a bloodless field during AVR and an adequate LV unloading while on CPB. Such kind of a strategy mirrors the setting of any eventual peripheral (femoro-femoral or femoro-axillary) v-a ECMO support associated with a pulmonary venting for indirect LV unloading.2,7,8
Sternal sparing and central ECMO support avoidance may be useful to reduce the risk of bleeding and infectious complications. LV apex sparing while on ECMO may be useful in terms of avoidance of surgery on a frail LV tissue or in case of significant negative clinical influence on the potential ventricular recovery which preferably imposes a “no-LV apical touch” strategy2,7 (Figure 1).
The strategy herein described was already reported and confirmed by Avalli et al.4 who reported the case of a successful ventricular recovery in a woman, 43 years old, supported by a peripheral v-a ECMO associated with an endovascular PA venting through a 15 Fr venous cannula inserted percutaneously since the presence of a huge LV apical clot. Similarly, Fouilloux et al.5 documented the successful case of a girl, who was 2 years old, on ECMO in whom the LV unloading was obtained with a 10 Fr cannula inserted into the pulmonary trunk through the inferior vena cava with a femoral approach. It was von Segesser et al.6 who first reported the potential advantages of pulmonary cannulation in animal models. The placement of a pulmonary arterial cannula allows decompression of the pulmonary circulation without abolishing the right ventricular ejection. The risk of complications is small and the decompression of the pulmonary circulation represents an active process. Finally, the procedure is performed percutaneously, without the need for surgical manipulation, with the associated risk for bleeding, heart manipulation, and heart injury, and it is less expensive than the application of axial pumps, easier, and faster than all well-known methods.2
Such a venting method might be criticized to be not effective as the routine direct LV unloading, even if flow-probe drainage monitorings were satisfactory in our studied cohort (Table 2), but it clearly provides sufficient adjunctive venting to right atrial drainage thus avoiding the need for left cardiac chamber-related access or procedures (right superior pulmonary vein cannulation, septostomy, or cardiac apex cannulation) which may result in fatal complications.2
Although not observed in our study cohort, PA cannulation might induce ventricular arrhythmias or significant pulmonary valve insufficiency. Additionally, by increasing the number of cannulas settings and connections, the risk of thrombi formation might be high thus forcing the need of higher ACT and PT times than routinely while on ECMO running, in terms of anticoagulation management.1
However, in summary, as confirmed by our series, in case of not recommended usage of direct LV apical venting (e.g., LV apical thrombi, recent antero-apical AMI, acute myocarditis, and graft failure after Htx), the adoption of v-a peripheral ECMO support associated with percutaneous PA drainage enables the rapid onset of extracorporeal life support with an effective biventricular unloading. The fluoroscopy is needed for pulmonary cannula insertion and the hybrid OR may be the correct location for appropriate cannulae placement as we did for all our patients. The usage of a Swan-Ganz catheter as a guide for pulmonary cannula advancement may be less traumatic and more confident for the operator who is usually represented by an anesthesiologist or an interventional cardiologist. Thereafter, in the setting of an eventual postcardiotomy case, the central direct cannulation of the PA results to provide a satisfactory LV unloading similarly to the routine minimally-invasive AVR surgery CPB installation procedures.8
This is a retrospective observational study. Although our cohort consists of eight patients, it is the largest case series to report on such an ECMO configuration, so far. Further investigations are still required to conclusively confirm the benefits of adjunctive PA cannulation in v-a ECMO patients. However, safety and efficacy of PA cannulation was demonstrated in a very high-risk patient, affected by severe biventricular dysfunction (Table 1).
In the presence of preoperative or intraoperative biventricular dysfunction, adjunctive PA cannulation may allow a better patient management, allowing additional RV drainage and satisfactory LV unloading.
1. Loforte A, Marinelli G, Musumeci F, et al. Extracorporeal membrane oxygenation
support in refractory cardiogenic shock: treatment strategies and analysis of risk factors. Artif Organs 2014.38: E129–E141.
2. Meani P, Gelsomino S, Natour E, et al. Modalities and effects of left ventricle unloading on extracorporeal life support: a review of the current literature. Eur J Heart Fail 2017.19(suppl 2): 84–91.
3. Rupprecht L, Flörchinger B, Schopka S, et al. Cardiac decompression on extracorporeal life support: a review and discussion of the literature. ASAIO J 2013.59: 547–553.
4. Avalli L, Maggioni E, Sangalli F, Favini G, Formica F, Fumagalli R. Percutaneous left-heart decompression during extracorporeal membrane oxygenation
: an alternative to surgical and transeptal venting in adult patients. ASAIO J 2011.57: 38–40.
5. Fouilloux V, Lebrun L, Macé L, Kreitmann B. Extracorporeal membranous oxygenation and left atrial decompression: a fast and minimally invasive approach. Ann Thorac Surg 2011.91: 1996–1997.
6. von Segesser LK, Kwang K, Tozzi P, Horisberger J, Dembitsky W. A simple way to decompress the left ventricle during venoarterial bypass. Thorac Cardiovasc Surg 2008.56: 337–341.
7. Lorusso R. Are two crutches better than one? The ongoing dilemma on the effects and need for left ventricular unloading during veno-arterial extracorporeal membrane oxygenation
. Eur J Heart Fail 2017.19: 413–415.
8. Lamelas J, Mawad M, Williams R, Weiss UK, Zhang Q, LaPietra A. Isolated and concomitant minimally invasive minithoracotomy aortic valve surgery. J Thorac Cardiovasc Surg 2018.155: 926–936.e2.