Female sheep (10 months old, 30–35 kg) from the abattoir were included in this study. The protocol is consistent with regulation number 1069/2009 of the European Parliament and Council regarding abattoir animal products for diagnosis and research, which were authorized by the relevant legal animal welfare authorities (Food and Consumer Product Safety Authority). The sheep were terminated employing a captive bolt pistol to the head. Subsequently, the carotid arteries were cut open and the animal was suspended by its hind legs to exsanguinate. During exsanguination, blood was collected and heparinized (5.000 IU/L). A median sternotomy was performed, and the heart and lungs were harvested en bloc.

On the back table, the heart was immediately topically cooled and visually inspected for abnormalities or damage due to the termination process. The heart was dissected from the lungs by transecting the pulmonary artery and veins. Approximately 10 cm of the ascending aorta was retained and canulated by securing the cardioplegia flush line in the aperture of the ascending aorta. Therefore, both coronary sinuses were perfused simultaneously. The cardioplegic solution (Custodiol histidine-tryptophan-ketoglutarate solution, Essential Pharmaceuticals, LLC, Durham, United States) was supplemented with adenosine (200 μmol/L), lidocaine (500 μmol/L) and nitroglycerin (100 mg/L) and administered at a pressure of 80–100 mmHg (1L, 4°C) [14, 15]. Warm ischemic time was defined as the time between the termination of the animal and delivery of cardioplegia. A maximum of 10 min of warm ischemia was accepted.

All hearts were preserved with oxygenated subnormothermic ESHP (15°C) for 2 h. The subnormothermic ESHP system (modified Kidney Assist Transport, XVIVO, Gothenburg, Sweden) consists of a reservoir, an oxygenator (Medos Hilite 1000 XENIOS, Heilbronn, Germany), and a centrifugal pump (Medos DP2 XENIOS, Heilbronn, Germany). Antegrade perfusion of the coronary arteries was delivered in a pressure-controlled and pulsatile fashion by cannulating the aorta root. The aortic valve was tested for insufficiency. Oxygen (100%) was delivered at a flow of 100 mL/min. Hearts were continuously perfused with a hyperkalemic (20 mmol/L), buffered dextran-albumin perfusion solution supplemented with nitroglycerin (25 mg/L). Perfusion temperature was targeted at 15°C. The hearts were prepared for subnormothermic ESHP by cannulating the aorta and by venting the LV to avoid LV distention. Subsequently, the hearts were submerged within the reservoir and the perfusion cannula was attached to the arterial end of the perfusion system. The perfusion pressure was gradually increased to 45 mmHg, aiming at a perfusion flow of >150 mL/min. The right atrium was left open, enabling the free return of the perfusate from the coronary sinus into the reservoir. Blood gases and electrolytes were analyzed to ensure the perfusate composition (pH: 7.1–7.4; HCO₃ ⁻: 20–25 mmol/L; pO₂: 40–60 kPa; pCO₂: 4.7–6 kPa; K⁺: >20 mmol/L; Ca²⁺: 0.18–0.25 mmol/L).

At the end of preservation, the hearts were submerged in cold cardioplegic solution in order to minimize injury during the preparations for normothermic evaluation. First, hearts were flushed with Perfadex (XVIVO, Gothenburg, Sweden) to flush out the hyperkalemic subnormothermic ESHP perfusate. Subsequently, hearts were weighted. For normothermic evaluation, a Langendorff perfusion system comparable to Langendorff perfusion was used. The normothermic ESHP system (modified Kidney Assist, XVIVO, Gothenburg, Sweden) consists of a heater-cooler and a pump housing unit, which must be connected to a disposable set, consisting of a reservoir, an oxygenator (Medos Hilite 2800 XENIOS, Heilbronn, Germany) and a centrifugal pump (Medos DP2 XENIOS, Heilbronn, Germany). The normothermic perfusion is pressure-controlled and delivers a continuous flow to the coronary arteries using the same cannula as for subnormothermic ESHP. Initial reperfusion was initiated in a controlled way, with a gradual increase in both pressure and temperature [16]. Perfusion temperature was initiated at 27°C with 3°C increments every 5 min over 30 min. After 30 min, the pressure was increased from 40 mmHg to 60 mmHg. Rewarming continued for another 30 min. The perfusate was based on albumin solution supplemented with Perfadex (XVIVO, Gothenburg, Sweden), to which washed RBCs (±500 mL) were added. The gas flow of carbogen (O₂ 95%; CO₂ 5%) was adjusted to maintain physiological pH and arterial carbon dioxide pressure (paCO₂). Blood gasses and electrolytes are analyzed to ensure a perfusate composition within the physiological range (pH: 7.3–7.4; HCO₃ ⁻: 20–25 mmol/L; pO₂: 40–60 kPa; pCO₂: 4.7–6 kPa; K⁺: 3.0–5.9 mmol/L; Ca²⁺: 1.0–1.3 mmol/L; Hb: 4.0–4.5 g/dL). Reperfusion was initiated with a low calcium concentration (<0.8 mmol/L). During normothermic ex situ perfusion, hearts were ventricular paced at a rate of 70 beats/min. Normothermic ESHP was performed for a total duration of 2 h.

The novel XVIVO ex situ PV loop system consists of a cylinder-piston combination that is placed on the annulus of the mitral valve, a balloon, a motor, and a control unit (Figure 1). The balloon is placed in the LV through the mitral valve, secured with a transapical line, and filled with saline. As stated before, the LV is vented. This assures that there will be little to no fluid surrounding the intra-ventricular balloon. To make sure that the balloon does not protrude from the LV to the left atrium, the mitral valve annulus is closed by placing two polyamide nylon tapes (Ethicon 3 mm × 70) behind the chordae tendineae of both the posterior and anterior leaflet of the mitral valve. Subsequently, the nylon bands are secured using a tourniquet kit, which ensures that the mitral valve annulus completely encloses the mitral valve cannula. A WheatStone bridge pressure catheter is placed in the balloon for real-time pressure measurements. The balloon is connected to the mitral valve cannula, which forms the connection to the piston-cylinder combination (Figures 1A, B). As a result, the volume within the balloon is in continuity with the volume within the cylinder-piston combination, providing a fluid-tight system with a constant and preset fluid volume. As the LV contracts, the volume is pushed out of the balloon into the cylinder, moving the piston (Figure 1A). This movement of the piston is transmitted through a Bowden cable (Figure 1C) which is connected to the motor pulley (Figure 1D), registering the volume changes in the LV during the cardiac cycle. Both the motor and pressure sensor are connected to the control unit which includes a computer with XVIVO software (Figure 1E).

During the cardiac cycle the software switches between the four phases of the cardiac cycle (Figure 2). The software detects the changes in the cardiac cycle by an algorithm dedicated to every phase of the cardiac cycle. Change in the cardiac phase is detected by the software as follows. The software switches from the diastolic filling phase to the isovolumetric contraction phase when a small outward volume displacement from the ventricular cavity is detected. It switches from the isovolumetric contraction phase to the systolic ejection phase when a certain predetermined threshold pressure is reached. The systolic ejection phase changes to the isovolumetric relaxation phase when a small inward volume displacement to the ventricular cavity is detected. And lastly, the software switches from the isovolumetric relaxation phase to the diastolic filling phase when the ventricular pressure is lower than the predetermined preload pressure.

Furthermore, the volume within the cylinder-piston combination is pressurized by force generated by the motor. Therefore, the internal volume can be pressurized by changing the motor power. This enables the simulation of a virtual physiologic load on the ventricle (pre- and afterload). Both the power and speed delivered by the motor can be controlled with the XVIVO software. The software reads the pressure measured in the LV, compares this with the desired preload pressure, and subsequently increases or decreases the motor power to adjust the LV pressure. The desired preload pressure can be adjusted in the XVIVO software after every cardiac cycle. As a result, the LV pressure will increase to the desired preload pressure within one cardiac cycle, enabling the ability to effortlessly execute a preload challenge to assess LV function.

In the systolic ejection phase, the software controls the afterload. The afterload is calculated from the measured ventricular outward flow and a virtual impedance based on the analogue described by Kung and colleagues [17]. The simplified virtual impedance is constructed out of four components (Figure 3; Table 1). The inertia component (L) counteracts a flow change. Compliance (C), primarily the aortic compliance, is the stretching capacity of the arteries. R_p is the virtual flow resistance of the aortic valve, and R_d is the virtual flow resistance of the distal track/the systemic vascular resistance (SVR).

The afterload is calculated using the following equation: P = L d Q d t + Q   ⋅   R p + 1 C ∫ Q − Q R d d t + P C 0

In which P is the target pressure in mmHg. The software will control the motor force in such a way that the measured pressure in the ventricle will equal this targeted pressure or afterload. Q is the flow (mL/min) out of the ventricle and Q_Rd is the flow through the virtual flow resistance of the SVR. The inertia component (L) is dependent on the flow change (dQ) over time (dt, in seconds) out of the ventricle. R_p is the flow resistance of the aortic valve. Pressure generated by the compliance component (C) is determent by the volume in the aorta. Volume in the aorta can be calculated by summation of the flow from the ventricle (Q), minus the flow to the body (Qrd). P_c0 is the pressure (in mmHg) in the aorta at the start of the systolic ejection phase. The targeted pressure of the ventricular cavity is recalculated at intervals of dt, and dt is defined as to be smaller than 4 milliseconds.

The equation above is valid in the systolic ejection phase, as in the other phases the aortic valve would be closed. If the aortic valve is closed, the aorta is still loaded and stretched as a result of the systolic ejection phase. During unloading of the aorta, the aortic pressure falls. The formula for decreasing the pressure of the aorta, is based on a resistance-capacitor circuit, the electronic equivalent of resistance and compliance: P c = P c e s   ⋅   e − t R d   ×   C  

In which P_c is the virtual pressure in the aorta, P_ces is the end systolic pressure compliance in the aorta and t is the progression in time since closure of the aortic valve in seconds, C is the compliance and R_d is the SVR. When a heartbeat is in the isovolumetric systolic phase, P_c is the threshold pressure for switching to the ejection phase.

Data is presented as mean ± standard deviation (SD). A repeated measures ANOVA was performed to analyze differences over time. p-value <0.05 was considered statistically significant. Data analysis was performed using GraphPad Prism 9.4 (GraphPad Software, CA, United States).

A total of 5 female sheep of 10–12 months old (25–35 kg) were included in the study. The functional warm ischemic time was 8.4 ± 0.9 min. The cold ischemic time during preparations for subnormothermic ESHP was 5 ± 0.5 min. Hearts were perfused with subnormothermic ESHP for 135 ± 13.3 min and evaluated during normothermic ESHP for 122 ± 1.6 min. The cold ischemic time during preparations for normothermic ESHP was 9 ± 3.4 min. The baseline weight of the heart was 328.4 ± 3 g. Weight gain during preservation was 71 ± 12.7 g (p = 0.0005) and during normothermic ESHP 41 ± 16 g (p = 0.010).

All hearts were perfused with a pressure of 45 mmHg. Coronary flow increased non-significantly over time until 60 min of preservation, and stabilized hereafter (Figure 4A). paO₂ and paCO₂ concentration during subnormothermic ESHP are shown in Figures 4B, C, respectively. Glucose concentration decreased over the course of preservation (Figure 4D), baseline: 11.8 ± 0.3, end of subnormothermic ESHP: 10.4 ± 0.3 (p = 0.005). A significant increase in lactate was seen in the first hour of preservation (Figure 4E). However, after 60 min of preservation, lactate remained rather stable. Lactate dehydrogenase after 5 min of preservation was 426 ± 35 U/L (Figure 4F). At the end of preservation, lactate dehydrogenase decreased to 400 ± 43 U/L (p = 0.3).

Mean pressure during normothermic ESHP was 62 ± 6 mmHg. Changes in flow during normothermic ESHP, are shown in Figure 5A. Changes in both arterial and venous pO₂ and pCO₂ are shown in Figures 5B, C. In the first 30 min of normothermic ESHP, lactate increased significantly (baseline: 0.28 mmol/L, 5 min: 0.92 mmol/L, 30 min: 1.66 mmol/L, p = 0.02). Hereafter, a non-significant increase in lactate was seen between each timepoint (Figure 5D). At the end of normothermic ESHP, lactate dehydrogenase level was 333 ± 37 U/L (p = 0.005, Figure 4D).

PV loop analyses were performed at 60 and 120 min of normothermic ESHP. A simulated output from one of the PV loop assessments with increasing pre-load is shown in Figure 6. Dp/dt-min and -max and maximally generated pressure values are shown in Figure 7. End-systolic elastance (Ees) at 60 min and 120 min of normothermic ESHP was 2.8 ± 1.8 mmHg/mL and 2.7 ± 0.7 mmHg/mL (p = 0.9), respectively (Figure 7C). Tau, the exponential decay of the ventricular pressure during isovolumic relaxation, is shown in Figure 7E.