Endothelial and Hemodynamic Function in a Large Animal Model in Relation to Different Extracorporeal Membrane Oxygenation Cannulation Strategies and Intra-Aortic Balloon Pumping

Background: The use of simultaneous veno-arterial extracorporeal membrane oxygenation (ECMO) with or without an Intra-Aortic Balloon Pump (IABP) is a widely used tool for mechanical hemodynamic support. Endothelial function, especially in relation to different cannulation techniques, is rarely investigated in the setting of extracorporeal life support (ECLS). In this study, we analyzed endothelial function in relation to hemodynamic and laboratory parameters for central and peripheral ECMO, with or without concomitant IABP support in a large animal model to gain a better understanding of the underlying basic mechanisms. Methods: In this large animal model, healthy female pigs with preserved ejection fraction were divided into the following groups related to cannulation strategy for ECMO and simultaneous IBAP support: control (no ECMO, no IABP), peripheral ECMO (pECMO), central ECMO (cECMO), pECMO and IABP or cECMO and IABP. During the experimental setting, the blood flow in the ascending aorta, left coronary artery and arteria carotis was measured. Afterwards, endothelial function was investigated after harvesting the right coronary artery, arteria carotis and renal artery. In addition, laboratory markers, such as creatine kinase (CK), creatine kinase muscle–brain (CK-MB), troponin, creatinine and endothelin were analyzed. Results: The blood flow in the ascending aorta and the left coronary artery was significantly lower in all discussed experimental settings compared to the control group. Of note, the cECMO cannulation strategy generated favorable hemodynamic circumstances with higher blood flow in the coronary arteries than pECMO regardless of flow circumstances in the ascending aorta. The concomitant usage of IABP did not result in an improvement of the coronary blood flow, but partially showed a negative impact on the endothelial function of coronary arteries in comparison to the control. These findings correlate to higher CK/CK-MB levels in the setting of cECMO + IABP and pECMO + IABP. Conclusions: The usage of mechanical circulatory support with concomitant ECMO and IABP in a large animal model might have an influence on the endothelial function of coronary arteries while not improving the coronary artery perfusion in healthy hearts with preserved ejection.


Introduction
The implantation of veno-arterial (VA) extracorporeal membrane oxygenation (ECMO) has become a widely used tool for patients with refractory cardiopulmonary or isolated 2 of 13 cardio-circulatory failure and cardiac arrest [1][2][3]. The appropriate cannulation strategy for VA-ECMO application with respect to survival rates and myocardial recovery remains a topic of discussion [4][5][6]. For central access (cECMO) with direct cannulation of the ascending aorta and the right atrium, a sternotomy is mandatory. Peripheral cannulation (pECMO) of the femoral artery and vein is the other predominant implantation technique of VA-ECMO. Further, the simultaneous mechanical support of VA-ECMO with an intra-aortic balloon pump (IABP) leads to significant reduction of intracavitary cardiac pressure, with a better left-ventricular unloading, and generates pulsatile flow conditions [7,8]. Besides the common clinical usage of different VA-ECMO settings, with or without IABP, it is unknown how hemodynamic alterations have an impact on myocardial regeneration and coronary artery endothelial function. Nitric oxide (NO) is a molecule derived from the endothelium of blood vessels orchestrating vascular relaxation in muscular arteries. An endothelial dysfunction, potentially due to mechanical circulatory support via shear stress or circulating inflammatory molecules, can lead to an imbalance and pathologic condition regarding vasodilatation and thrombogenic circumstances. The assessment of NO-mediated vasodilatation provides information about the function and integrity of the endothelium including vascular adaptation [9]. Evaluating the different VA-ECMO therapy settings and related perfusion-mediated endothelial function seems of particular importance to provide a better understanding for cardiac recovery. Consequently, this experimental study was aimed to analyze the effects of different mechanical cardio-circulatory support settings in a healthy porcine model to gain a better understanding for underlying basic mechanisms regarding hemodynamic and endothelial function.

Materials and Methods
The responsible ethical committee for animal care of the University of Cologne (LANUV Northrhine-Westphalia, Recklinghausen, Germany) approved the study protocol, and the animals were treated in compliance with the Directive 2010/63/EU of the European Parliament.

Experimental Setting
Thirty-three healthy female pigs (Deutsche Landrasse Pietrain) with a body weight of 60 kg were divided into four groups: control, peripheral ECMO (pECMO), central ECMO (cECMO), pECMO and IABP or cECMO and IABP. The control group did not receive any cardiac assistance as displayed in Table 1. Animals were premedicated with combinatory intramuscular application of xylazine (2 mg/kg) and zoletil (10 mg/kg). Additionally, pigs received atropin (0.02 mg/kg i.m.). After initiation of sedation with propofol (2 mg/kg i.v.), animals were intubated endotracheal. Analgo-sedation was maintained with fentanyl (0.012-0.025 mg/kg), midazolam (0.96-1.2 mg/kg) and propofol (4-6 mg/kg) as continuous infusion. Ventilation of pigs was performed using a volume-controlled ventilator (Fabius GS, Dräger, Lübeck, Germany) with the aim to keep arterial blood gas parameters in a physiologic range (pO2 > 100 mmHg). Achievement of full anesthesia was continuously examined using the interdigital claw reflex. After sterile washing and covering, a urinary catheter, a central venous catheter (8.5 Fr, ZVK Arrow International, Reading, PA, USA) through the right jugular vein and an arterial catheter (20 G Leadercath, Vygon, Aachen, Germany) through the right carotid artery were placed surgically. Central venous pressure was kept constant via continuous saline infusion (5-10 mL/kg/h). Mean arterial pressure was adjusted to achieve physiological organ perfusion (>60 mmHg). Moreover, the contralateral carotid artery was used for measurement of cerebral blood flow with transit time flow probes (Transonic Systems Inc., New York, NY, USA). Afterwards, a sternotomy was then performed and the heart was exposed. Further, a pulmonary catheter (20 G Leadercath, Vygon, Aachen, Deutschland) and 5 Fr pressure transducer-tip catheters (Model SPC-350S, Millar Instruments Inc., Houston, TX, USA) were placed into the pulmonary artery and the left and right ventricles. Transit time flow probes (Transonic Systems Inc., New York, NY, USA) recorded cardiac output and coronary blood. Further, the blood flow was recorded by transit time flow probes (Transonic Systems Inc., New York, NY, USA), which were placed at the ascending aorta and the left anterior descending artery, respectively. The right femoral artery and vein were exposed after systemic anticoagulation (heparin 300 IU/kg), and in the case of peripheral ECMO (pECMO) treatment were cannulated in a wire-guided manner. For central cannulation (cECMO), the arterial cannula was directly placed into the ascending aorta. A venous cannula was placed in the right atrium in both scenarios. ECMO circulation was performed for 10 h with a flow index of 50 mL/kg/min/m 2 . ECMO equipment consisted of a console (Bio-Medicus 540, Medtronic, Dublin, Ireland), a centrifugal pump (BPX-80 Bio-Pump, Medtronic, Dublin, Ireland), a membrane oxygenator (CAPIOX ® FX, Terumo, Tokyo, Japan), a heat exchanger (Bio-Cal 370, Medtronic, Dublin, Ireland), an extracorporeal circuit (Tubing set, Terumo, Shibuya, Japan) and arterial/venous cannulas (Biomedicus 17 Fr art./21 Fr ven., Medtronic, Dublin, Ireland). IABP (97e, Datascope Corp., Fairfield, NJ, USA) were implanted in the aortic arch and placed distally. After the initiation of extracorporeal circulation, animals were examined for 10 h. Euthanasia was commenced with the application of pentobarbital (80 mg/kg). Measurements were performed and recorded using a 16-channel hemodynamic set-up, all data were digitized at a rate of 500 Hz and were subsequently analyzed (Hugo Sachs Elektronik-Harvard Apparatus GmbH, March-Hugstetten, Germany). Quantification of Arterial blood samples were withdrawn 10 h after ECMO initiation for measurement of parameters of systemic inflammation and endothelial injury (endothelin-1, ET-1). Plasma was separated after centrifugation for 10 min at 2000× g rpm and samples were stored at -80 until assay preparation. Enzyme-linked immunosorbent assays were performed in accordance with manufacturers' recommendations (ET-1, R&D Systems Inc., Minneapolis, MN, USA; MPO, USCN Life Science Inc., Wuhan, China). Further analysis of blood samples, including creatine kinase (CK), creatine kinase muscle-brain (CK-MB), troponin and creatinine, was conducted in the central laboratory of the university hospital (University Hospital Cologne, Cologne, Germany) as previously described by our research group in "Concomitant Intra-Aortic Balloon Pumping Significantly Reduces Left Ventricular Pressure during Central Veno-Arterial Extracorporeal Membrane Oxygenation-Results from a Large Animal Model" [10].

In Vitro Endothelial Function
For measurements of endothelial function, we followed an established protocol [11]. Hearts were excised postmortem and the right coronary artery was isolated. Simultaneously, the arteria carotis and arteria renalis were harvested. The arteries were carefully isolated from surrounding tissue and placed in cold (4 • C) modified Krebs-Henseleit solution (143 mmol/L Na + , 5.9 mmol/L K + , 1.6 mmol/L Ca 2+ , 1.2 mmol/L Mg 2+ , 126 mmol/L Cl − , 25 mmol/L HCO 3 − , 1.2 mmol/L H 2 PO 4 − , 1.2 mmol/L SO 4 2− , 5.1 mmol/L glucose). For measurements, vessels were cut transversely into rings (4 mm width) using a no-touch technique. For each run, 4 rings of the specific artery were mounted on steel hooks with a constant load of 1 g and incubated in isolated organ baths (Hugo Sachs Elektronik) with 10 mL preoxygenated Krebs-Henseleit buffer at 37 • C and 95% O 2 /5% CO 2 . After equilibration for 30 min, isotonic alterations of vascular ring diameter were constantly recorded (Multipen Recorder; Rikadenki Kogyo Co., Tokyo, Japan). Potassium chloride (60 mmol/L) was added to trigger a maximal contraction and rinsed three times. Indomethacin (10 µmol/L) was used to block the cyclooxygenase pathway (prostaglandin synthesis). Either L-ARGININ or N-nitro-L-arginine (LNNA, 300 µmol/L, Sigma-Aldrich, Burlington, MA, USA) was added to induce or block NO production, respectively. Stable precontraction was then induced by prostaglandin F2 a (10 µmol/L) and endotheliumdependent relaxation to bradykinin was recorded by adding accumulating concentrations (0.0001-10 µmol/L). The mean vessels endothelial function was determined by calculating the average percentage change of vasorelaxation following bradykinin administration [10].

Statistical Analysis
Statistical analyses were performed with GraphPad-Prism9 software (GraphPad Software, La Jolla, CA, USA). The data are presented as (1) boxes with violins (min to max) for flow measurements; (2) points and connection line with error bars (mean and error ± standard error of the mean) for measurement of the endothelial function as loss of contraction; and (3) scatter plots with bars (mean ± standard error of the mean) for laboratory analysis. Data were analyzed using one-way ANOVA or two-way ANOVA with Tukey's multiple comparisons test for comparison of more than two groups. Intergroup differences between two groups were analyzed using the unpaired Student's t-test. A p-value of less than 0.05 was considered statistically significant.

Discussion
Our study significantly contributes to the understanding and delineation of the physiological and pathophysiological processes taking place under mechanical circulatory support.
The main findings of this study are: (I) An improved NO-dependent EDR of the coronary artery through a better perfusion under cECMO therapy. (II) Simultaneous IABP support in addition to ECMO leads to a comparable coronary artery perfusion without showing any benefit regarding the NO-dependent EDR. (III) Perfusion of the arteria carotis is significantly improved in cECMO compared to pECMO therapy, independent of simultaneous IABP support. (IV) Different mechanical circulatory support settings showed no impact on renal artery EDR in our pig model.

Endothelial Function
Endothelial injury and dysfunction following cardio-pulmonary bypass can cause vascular cell impairment and end-organ damage triggered by acute inflammatory response [12]. Exposure of blood to the extracorporeal circuit leads to contact system activation and promotes humoral and cellular inflammation, mimicking systemic inflammatory response syndrome (SIRS) [13]. Endothelial dysfunction, mediated by SIRS, plays a major role in critical illness and is associated with higher mortality [14,15]. Most investigations regarding endothelial function have been assessed in cardio-pulmonary bypass of patients undergoing cardiac surgery, and systematic investigations on this topic for mechanical circulatory support with ECMO are lacking. Therefore, we are the first to systematically investigate the different perfusion profiles of pECMO ± IABP and cECMO ± IABP settings with regard to NO-related endothelial function in the relevant end organs. In this study, we show a positive impact of central versus peripheral cannulation for ECMO implantation on coronary NO-dependent endothelial-derived relaxation. This finding is in line with an improved coronary perfusion in the cECMO setting. Interestingly, clinical experiences with a centrally implanted mechanical life-support system did not outperform peripheral implanted ECMO in patients with or without post-cardiogenic shock (PCS) when assessed for weaning from ECMO or all-cause mortality [1,16]. The addition of an IABP to either cECMO or pECMO circulatory support reduced the coronary artery perfusion and the positive impact of central versus peripheral ECMO cannulation on EDR declined in the discussed experiments. Further, higher serum levels of cardiac markers were detected in the ECMO + IABP groups, representing a potential myocardial damage. Nevertheless, these findings might be due to the effect of an unphysiological blood flow mediated by the ECLS in combination with the experimental setting of healthy hearts and unaffected left-ventricular ejection fraction. In contrast, some clinical studies showed a higher weaning rate from ECMO for cardiogenic shock patients with concomitant IABP circulatory support but contrary outcomes regarding short-term mortality [7,8,17,18]. Differentiation of the patient collective unveils that patients with a recent myocardial infarction under ECMO support might experience higher benefit from additional IABP implementation than those without [19].

Hemodynamic Assessment
Besides the endothelial function, our study group assessed the different canulation strategies for ECMO therapy and the concomitant mechanical support via IABP with respect to hemodynamic outcomes. Nevertheless, there is an ongoing debate as to whether ECMO therapy should be supported by an IABP or not, as the concept of ventricular unloading with simultaneous IABP has already been described and is frequently used, although IABP support is an indirect method used to unload ventricular cavities. Our study group also demonstrated an analysis of n = 172 ECMO patients with post-cardiotomy cardiogenic shock showing that, independent of ECMO type or canulation strategy, additional IABP support might increase weaning rates off ECLS support with ECMO without affecting in-hospital mortality rates [8]. Furthermore, Madershahian et al. showed that pulsatility induced by IABP improved diastolic filling index and mean coronary bypass graft flows by reducing coronary vascular resistance during peripheral ECMO in post-cardiotomy cardiogenic shock patients [20]. In contrast, our presented data did not show an increase in coronary or carotid blood flow with the additional usage of an IABP in healthy hearts of pigs. Further, the reduced blood flow in the coronary arteries compared to control (but not for the cECMO group) may be a result of the repetitive aortic occlusion induced by the balloon and the unphysiological retrograde flow in peripheral canulated ECMO settings, diminishing the blood flow available for the aortic root. However, the setting with healthy animals and no reduction in cardiac function may have led to a bias that perfusion of the carotidal arteries was better with cECMO, regardless of simultaneous IABP support, when compared to peripheral cannulation. These findings are in line with the prior literature confirming better hemodynamic and O2 supply to the brain [21]. Nevertheless, there exists no general statement when interpreting the flow modularity of ECLS with ECMO ± IABP as the cardiac output and the lung function plays a crucial interfering role. Hence, only an idealized interpretation with regard to flow patterns and flow distribution can be assessed [22].

Limitations
The artificial scenario implanting ECMO and IABP in healthy pigs does not represent a real-life scenario. Moreover, our study protocol in healthy pigs did not aim to produce a pathologic state. Therefore, the investigated data are difficult to transfer onto clinical situations. However, our focus was the hemodynamic and endothelial changes caused by different strategies of extracorporeal life support so as to gain a better understanding of the basic mechanisms in these settings of ECLS with ECMO and/or IABP. Further experiments with ischemic hearts or in a setting of ischemia and reperfusion are needed to systematically asses the influence of the discussed ECLS strategies after a cardiac event. In this regard, all results should be viewed with caution, and transfer of these results into a human clinical scenario might be difficult. The examined animal model, with only a small number of pigs authorized for investigation, limits this study. It remains unclear whether findings can be transferred in a 1:1 manner to human patients. Further, when interpreting these data, the distinct experimental setting has to be considered in terms of an unaffected and preserved ejection fraction. Therefore, the comparison of the used animal model may only partially reflect the hemodynamic settings of humans in need of extracorporeal mechanical support with PCS or acute heart failure. In addition, the impaired blood flow in the ascending aorta can only partially explained by an unphysiological flow produced by the ECLS, but was validated by re-positioning of the flow measurement and does not reflect the overall perfusion generated by ECMO as this was computed and determined beforehand. The study was not designed to analyze outcome measures, and the design of the study does not allow a reasonable interpretation of the reported results. In order to analyze the background and underlying mechanisms of our findings, further experiments are needed, including studies with ischemic hearts and/or reduced ejection fraction. As for now, this study can only be considered an observational study.

Conclusions
Mechanical circulatory support with ECMO and plus/minus IABP might have an impact on the endothelial function, as investigated in this large animal model. Further, hemodynamic measurements showed conflicting results while using mechanical circulatory support. In order to properly assess the impact of ECMO and concomitant IABP support, long-term experimental settings are needed.