Remember me
Abstract
The use of ECPELLA in patients with severe lung disease may result in an unfavorable phenomenon of differential hypoxia. The simultaneous evaluation of three arterial blood samples from different arterial line (right radial artery, left radial artery, ECMO arterial line) in patients at risk of Harlequin syndrome (also called differential hypoxemia (DH)) can localize the “mixing cloud” along the aorta. Focusing the attention on the “mixing cloud” position instead of on isolated flows of Veno-Arterial Extracorporeal Membrane Oxygenation (VA ECMO) and Impella CP makes the decision making easier about how to modify MCSs flows according to the clinical context. Herein, we present two cases in which ECPELLA configuration was used to treat a cardiogenic shock condition and how the ECPELLA-induced hypoxia was managed.
Keywords: Cardiogenic shock, ECMO, ECPELLA, Harlequin syndrome, mechanical cardiovascular support
How to cite this article: Introduction 
Impella CP (Abiomed, Danvers, MA, USA) is a microaxial, non-pulsatile flow pump used for temporary left ventricular assistance during cardiogenic shock.[1]
The simultaneous use of Veno-Arterial Extracorporeal Membrane Oxygenation (VA ECMO) and Impella (called ECPELLA) is becoming increasingly common to support cardiogenic shock condition. The combined action of both devices allows to maintain systemic circulation and to unload the left ventricle (LV) at the same time, providing an alternative mechanism of decompression with less invasiveness compared with surgical venting (pulmonary venting through the right superior pulmonary vein or pulmonary artery trunk, LV apical venting).[2],[3],[4]
However, peripheral VA ECMO may result in anatomic regional differences in oxygen saturation. According to Falk et al.,[5] differential hypoxemia (DH, also called Harlequin syndrome) is a condition where the saturation differs between the upper and the lower parts of the body.[5] In patients with severely impaired lung function, when there is intrinsic cardiac activity, the blood leaving the heart and entering the aorta may be poorly oxygenated due to impaired gas exchange in the lungs and fulminant differential hypoxemia (FDH) may develop: the upper body is perfused with poorly saturated blood, and the lower body by hyper-oxygenated ECMO flow, with subsequent risk of global cerebral and myocardial hypoxia.
In patients with cardiogenic shock and severe lung dysfunction supported by ECPELLA, Impella pulls poorly oxygenated blood from the LV and expels into the ascending aorta, potentially leading to an increased risk of differential hypoxia and hypoxic coronary and cerebral perfusion.
Herein, two cases in which ECPELLA configuration was used to support a cardiogenic shock condition and how the ECPELLA-related hypoxia was managed.
We obtained informed consent to use clinical data from the patients' relatives.
Case report 1
A 44-year-old man (weight 80 kg, height 180 cm) suddenly developed cardiac arrest due to cardiogenic shock from fulminant myocarditis in the cardiology unit. Return to spontaneous circulation (ROSC) was achieved after 95′ of cardiopulmonary resuscitation supported with VA ECMO (with cannulation of the left femoral vein and left femoral artery) and the use of percutaneous intra-aortic balloon (IABP). The low flow time was limited to 30′.
At time of admission to the cardiac intensive care unit (CICU), lactate level was 6.1 mmol/L despite of epinephrine 0.1 mcg/kg/min and norepinephrine 0.12 mcg/kg/min continuous IV infusions. To ensure adequate cerebral perfusion, near infrared spectroscopy (NIRS) monitoring was used through two sensor pads placed bilaterally on the forehead (INVOS™ 5100C, Medtronic). Transesophageal echocardiography (TEE) showed an ejection fraction (EF) of 10%, severe LV distension with concomitant severe interventricular septum hypertrophy. It was then decided to remove IABP on day 1 and to unload the LV with percutaneous Impella CP implantation through the right femoral artery and ECPELLA configuration was started.
During the CICU stay, the patient developed a severe pulmonary edema, secondary to the excessive LV overload, exacerbated by oligo-anuric acute kidney injury (AKI), ultimately leading to the development of DH (right arterial artery PaO2 = 69 mmHg vs left arterial artery PaO2 = 296 mmHg), suggested also by a drop in cerebral right regional oxygen saturation (rS02) >20% form the baseline with a normal value of left rS02.
It was therefore necessary to modify the ECPELLA settings, decreasing Impella flow (anterograde flow) and improving VA ECMO flow (retrograde flow) to move the “mixing cloud” closely to the aortic valve, avoiding perfusion of deoxygenated blood to the cerebral arteries.
To optimize the position of the “mixing cloud,” blood was taken from three arterial sites: right radial artery, left radial artery, and ECMO arterial line, obtaining the results presented in [Table 1].
Table 1: Blood samples from right radial artery, left radial artery, and ECMO arterial line during different phases of ECPELLA support in case report 1ECPELLA was then set to maintain the “mixing cloud” near the aortic valve (Impella CP support level P4 – flow 2.2 L/min), VA ECMO 4.65 L/min, epinephrine 0.05 mcg/kg/min, norepinephrine 0.08 mcg/kg/min) while diuretics, continuous renal replacement therapy (CRRT) and higher levels of PEEP were started to resolve the pulmonary edema.
After pulmonary edema resolution, the anterograde flow (Impella CP and LV ejection) was progressively increased while reducing the VA ECMO flow, to progressively move the “cloud” from the aortic valve to the distal portion of the aortic arch.
On the eighth day from cannulation the VA ECMO was successfully removed and on the ninth day from cannulation Impella support was also weaned.
Unfortunately, the patient later developed a septic shock due to Klebsiella Pneumoniae Carbapenemase-producing (KPC-R) which resulted in his death after 12 days of hospitalization.
Case report 2
In a 64-year-old man, weight 100 kg, height 170 cm, body mass index (BMI) 34.6 kg/m2, long-term smoker (75 packs years); suspected but never treated for chronic obstructive pulmonary disease (COPD), suffering from inferior ST-elevation myocardial infarction (STEMI) complicated by cardiogenic shock an Impella CP was inserted through the left femoral artery before starting high-risk percutaneous coronary intervention (PCI) in another hospital. The procedure resulted in four drug-eluting stents (DESs) on right coronary artery (RCA), left anterior descending (LAD) artery, and circumflex artery (Cx) found chronically occluded.
The day after the procedure, due to the persistence of severe hypoxemia, metabolic acidosis, and oligo-anuria, the patient was transferred to the CICU to evaluate the insertion of VA ECMO (bridge to decision).
At admission, lactate level was 5.3 mmol/L under Impella CP support level P8 (flow 3.5 L/min), epinephrine 0.05 mcg/kg/min and norepinephrine 0.20 mcg/kg/min continuous IV infusions. An EF was 16% associated with hypertrophic and hypokinetic LV and concomitant only mild right ventricle (RV) disfunction was found at TEE. Cerebral perfusion was monitored by NIRS.
It was therefore decided to use ECPELLA with cannulation of the left femoral vein and left femoral artery for VA ECMO and leaving the Impella CP in left femoral artery. VA ECMO flow was initially set at 4.7 L/min and Impella CP was downgraded from P8 (flow 3.5 L/min) to P4 (flow 1.8 L/min) to maintain continual LV unloading.
During the CICU stay, the patient developed a severe lung injury, due to the combination of pulmonary edema and community acquired pneumoniae (CAP) complicated by oligo-anuric AKI and, despite the ECPELLA support, right rS02 decreased below 40 (with left rS02 value around 60) and the PaO2 measured at the right arterial line progressively decreased to 71 mmHg, suggesting the development of FDH.
As in the former case, blood was taken from three arterial sites (right radial artery, left radial artery, ECMO arterial line) to localize the mixing cloud along with the aortic arch and optimize ECPELLA opposite blood flow as shown in [Table 2].
Table 2: Blood samples from right radial artery, left radial artery and ECMO arterial line during different phases of ECPELLA support and after configuration of VA ECMO with LV vent in case report 2Since the first three samples showed a distal localization of the mixing cloud [Video 1; [Figure 1]] to reduce differential hypoxia and maintain adequate cerebral perfusion, Impella CP was switched from P4 (flow 1.8 L/min) to P2 (flow 1.4 L/min) and epinephrine infusion was stopped, reducing left ventricular ejection.
With the new setting, a partial proximal migration of the mixing cloud, still inadequate to maintain brain oxygenation, was observed.
An “Impella CP OFF” test was then performed, downgrading Impella CP to P1 (flow 0.8 L/min) and improving VA ECMO flow to 5 L/min and this new setting made it possible to approach the mixing cloud to the aortic valve.
After further evaluation of the mixing cloud position and considering the severity of lung injury, a switch from ECPELLA to total VA ECMO with trans-apical LV venting through a left mini-thoracotomy to maintain the LV unloading was decided.
Although this configuration continued to be at potential risk for Harlequin syndrome, due to the maintenance of spontaneous LV systolic ejection, the systematic evaluation of blood samples described above allowed to confirm the approach of the mixing cloud to the aortic valve and, consequently, the differential oxygenation problem resolution.
Unfortunately, despite the high level of cardiovascular support provided, which managed to reach an adequate organ perfusion during the full mechanical cardiovascular support phase, the patient developed a massive acute bowel ischemia, which lead to his death in day 7.
Discussion The combination of VA ECMO and Impella (ECPELLA) has been recently proposed as an effective configuration for supporting patients with severe cardiogenic shock.[3],[4] However, the evidence for the efficacy of this combined technique comes from case reports and small single-center studies.[3],[4],[6],[7] Furthermore, when Impella pumps out blood from the LV in patients with severe lung injury, it may worsen a pre-existent subclinical differential hypoxia and make it clinically relevant, potentially inducing hypoxic coronary and cerebral perfusion.
The development of Harlequin syndrome may be quick or progressive, but when it happens require immediate treatment to prevent cerebral damages. A negative fluid balance, obtained by diuretics administration or CRRT, may help preventing or reducing the worsening of lung function but is usually not determinant, being the pulmonary edema firstly caused by LV overload. Reaching a negative fluid balance is usually part of the treatment for the lung impairment and the resolution of differential hypoxia, but the initial management requires more rapid alternatives.
A possible strategy to solve the problem of possible coronary desaturation under ECPELLA has been recently described in two case reports[8],[9] in which the VA ECMO was converted into veno-arteriovenous ECMO (VAV ECMO) using an additional internal jugular venous (JV) cannula as the second outflow of ECMO. Then a combination of VAV ECMO with femoral artery and jugular vein perfusion and femoral vein drainage and axillary IMPELLA was established. With this configuration, oxygenated blood was sent both to the inferior vena cava and the femoral artery bilaterally to maintain oxygenation in the pulmonary artery. In this way the anterograde flow pumped out by Impella from the LV includes a sufficient amount of oxygenated blood.
We hypothesized a similar configuration for the patient of the second case. However, considering the important BMI (34.6 kg/m2), a body surface area of 2.11 m2 and a predicted flow of 5 L/min, a V-AV ECMO configuration would potentially have resulted in inadequate flow. A second outflow in the SVC creates a certain level of recirculation between the two venous cannulas: the drainage femoral cannula may drain part of the blood returning from the jugular vein cannula, worsening the efficiency of the ECMO system and potentially causing an ineffective drainage of the RV. On the other side, the use of trans-apical LV venting allowed both differential hypoxemia resolution and effective RV and LV unloading (by the femoral and trans-apical cannulas, respectively).
The location of the “mixing cloud,” where fully oxygenated blood from the ECMO circuit will meet blood ejected from the native ventricle, is constantly changing, depending upon the amount of ECMO support provided and the degree of left ventricular ejection and lung function. In this particular condition, a systematic approach based on blood samples taken from the three arterial sites of the circulatory system (right and left radial artery, VA ECMO arterial line) could be useful to determine the localization of the mixing cloud, reducing the risk of perfusing coronary and cerebral arteries with poorly oxygenated blood. In the first case, cloud localization along the aortic arch was useful to ensure an effective cerebral oxygenation during the resolution of the pulmonary edema while guiding weaning from VA ECMO; in the second, closed monitoring of the mixing cloud position along the aortic arch was essential in the management of the Harlequin syndrome. Furthermore, thanks to this approach, in the second case it was possible to confirm the ineffectiveness of the ECPELLA configuration in use and the consequent need to switch to different type of mechanical circulatory support.
Conclusion ECPELLA appears to be an effective approach to support severe cardiogenic shock. However, in patients with severe lung disease, it may result in an unfavorable phenomenon of differential hypoxia, with an increased risk of cerebral hypoxic perfusion.
Future studies are needed to better define the effectiveness of this approach and to investigate more directly how blood interacts in the aortic arch during dual mechanical assistance.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient (s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References 
Correspondence Address:
Matteo Giunta
Department of Anesthesia and Critical Care, Citta della Salute e della Scienza di Torino, University of Turin, Corso Bramante 88-90, 10126, Turin
Italy
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/aca.aca_176_21
Comments (0)