The protocol describes a porcine ex vivo heart perfusion system in which direct loading of the left ventricle may serve as an assessment technique for graft health while simultaneously providing a holistic evaluation of graft function. A discussion of the system design and possible assessment metrics is also provided.
Ex vivo machine perfusion or normothermic machine perfusion is a preservation method that has gained great importance in the transplantation field. Despite the immense opportunity for assessment due to the beating state of the heart, current clinical practice depends on limited metabolic trends for graft evaluation. Hemodynamic measurements obtained from left ventricular loading have garnered significant attention within the field due to their potential as objective assessment parameters. In effect, this protocol provides an easy and effective manner of incorporating loading capabilities to established Langendorff perfusion systems through the simple addition of an extra reservoir. Furthermore, it demonstrates the feasibility of employing passive left atrial pressurization for loading, an approach that, to our knowledge, has not been previously demonstrated. This approach is complemented by a passive Windkessel base afterload, which acts as a compliance chamber to maximize myocardial perfusion during diastole. Lastly, it highlights the capability of capturing functional metrics during cardiac loading, including left ventricular pulse pressure, contractility, and relaxation, to uncover deficiencies in cardiac graft function after extended periods of preservation times (˃6 h).
Orthotopic heart transplantation is the current gold standard of care for end-stage heart failure1. Unfortunately, the field is significantly limited by a severe donor shortage crisis, resulting in only 2,000 heart transplants being performed each year when over 20,000 people would benefit from the lifesaving procedure2. This organ shortage is expected to worsen as the prevalence of heart failure in the United States alone is projected to surpass 8 million individuals by 20303. Steady increases in waitlist survival times – as a result of improved medical management, advances in mechanical circulatory support, and amendments to the UNOS allocation policy – have resulted in a further increase in the number of patients in need of transplantation at any given moment4,5.
Ex vivo machine perfusion or normothermic machine perfusion (NMP) is a preservation modality that has facilitated the expansion of the supply pool by allowing the use of organs donated after circulatory death (DCD) while achieving some extension of preservation times5,6,7,8. Unlike static cold storage, the current gold standard for preservation, NMP maintains organs in a metabolically active state, which creates the opportunity for real-time monitoring and graft assessment, becoming the standard preservation method for DCD grafts8,9. However, NMP devices currently used clinically are restricted to the Langendorff perfusion mode, which lacks quantitative metrics to predict transplantation outcomes and is unable to capture functional parameters6. For instance, lactate accumulation during Langendorff perfusion has been denoted as the best metabolic predictor of post-transplantation outcomes and is currently used in the clinical setting as a proxy for cardiac graft health10. However, even as the best assessment biomarker, it fails to reliably anticipate the need for mechanical circulatory support post-transplantation11,12. Similarly, the predictive capabilities of commonly utilized hemodynamic parameters (i.e., aortic pressure and coronary blood flow) are largely limited by the retrograde nature of current clinically used configurations for heart machine perfusion9.
The development of assessment protocols for accurate and precise determination of cardiac graft health during NMP would have an immense impact in the field beyond improving post-transplantation outcomes. Objective predictive tools would enable the reliable evaluation and likely utilization of marginal or extended criteria organs (i.e., prolonged warm (> 30 min) and cold ischemia times (> 6 h), increased donor age (> 55), other comorbidities, etc.) from both DCD and brain death donors (DBD) that are currently rejected for transplantation due to the stringent selection criteria13. By enabling the use of marginal hearts, NMP could facilitate an increase in the organ supply as it is estimated that successful transplantation of half the hearts that currently go unused would be sufficient to eliminate the heart waitlist within 2-3 years14. Hemodynamic measurements obtained from left ventricular loading during NMP have garnered significant attention within the field due to their potential as objective assessment parameters. Previous studies have demonstrated that these parameters, such as left ventricular pulse pressure, contractility, and relaxation, are more indicative of cardiac graft function than metabolic trends15,16,17.
In effect, efforts have been dedicated to the development and identification of optimal loading methods to maximize assessment accuracy. Through these efforts, other groups have identified the most relevant mode of aortic perfusion during loading, whereby a stronger correlation between hemodynamic parameters and post-transplantation function was seen when implementing a passive afterload (i.e., no retrograde perfusion to the aorta during loading) when compared with pump-supported afterload (i.e., retrograde perfusion to the aorta during loading)18. This indicates that assisted coronary perfusion likely masks functional deficiencies. Previous studies have successfully incorporated passive afterloads into perfusion setups by implementing systems that mimic the Windkessel effect18,19,20. The Windkessel effect aids in dampening the fluctuation in blood pressure, maintaining continuous blood flow to the tissues and improving coronary perfusion. This protocol achieves the Windkessel-based passive afterload using a modified intravenous (IV) bag enclosed in two spring-loaded plates, where coronary perfusion is exclusively dependent on heart ejection (patent pending).
The use of passive left atrium (LA) pressurization (i.e., gravity-dependent pressurization) during loading, although common practice in small animal heart perfusions, is rarely utilized in the loading of large hearts21,22,23. Instead, the large majority of methods reported in the literature rely on secondary pumps for LA pressurization18,24,25,26,27,28. The pressurization of the LA through a gravity-dependent reservoir, rather than by pump, significantly simplifies the implementation of loading protocols. The use of gravity provides a fixed and constant pressure source, which greatly decreases the need for complicated control systems to achieve and maintain adequate LA pressurization. Moreover, through this pressurization approach, the requirement for a secondary pump is eliminated, facilitating the incorporation of loading capabilities into current Langerdoff setups, as only an extra reservoir is needed. The integration of loading capabilities into clinically utilized machine perfusion systems would amplify the application of cardiac NMP devices by facilitating detailed assessment of cardiac grafts during the preservation period. In effect, maximizing the utility of a system that poses significant financial commitment for patient care due to transportation and device utilization29.
This protocol demonstrates the feasibility of employing both passive afterload and passive LA pressurization during left ventricular loading. Through the validation of passive afterload/LA pressurization as a loading method, this protocol also provides an easy and effective manner of incorporating loading capabilities into established Langendorff perfusion systems. Importantly, it highlights the capability of functional assessment to uncover differences in viable versus failing hearts after extended periods of preservation (˃6 h).
This study was conducted in accordance with the Institutional Animal Care and Use Committee (IACUC), Massachusetts General Hospital, and Jove's animal guidelines. Hearts (170 – 250 g) were harvested from Yorkshire pigs (30 – 35 kg, age 3-4 months, mixed sex) using a model of donation after brain death and perfused retrogradely (Langendorff) for 6 h prior to loading. All grafts were exposed to a cold ischemia time of approximately 1h during instrumentation.
1. System design
2. Perfusate system preparation
3. Cardiac graft procurement
4. Graft preparation
5. Cardiac graft revival
6. Cardiac graft loading
7. End of perfusion
Hearts from 4 Yorkshire pigs (30 – 35 kg) were harvested and preserved via Langendorff NMP for 6 h prior to 4 h of continuous loading. This experimental condition was chosen since 6 h is the average clinical preservation duration (5.1 ± 0.7 h)34. Through the addition of 4 extra hours of continuous loading (total of 10 h ex vivo time), some degree of heart failure was expected as a clear correlation between perfusion time and myocardial function decline has been previously reported35. As such, this provided the opportunity to feature the importance of functional metrics for graft assessment and their ability to discern viable versus failing hearts. In this capacity, most Langendorff-based metrics indicated no discernable difference in viability between cardiac grafts, with vascular resistance (CVR, Figure 2C), oxygen uptake rate (OUR, Figure 2D), lactate accumulation (Figure 2E), glucose consumption (Figure 2F), and potassium accumulation (Figure 2G) being non-statistically different during the entire perfusion time. Similarly, no difference was seen in the weight gained by the grafts (Figure 2H). However, a steady decline in heart rate was seen for failing grafts after 2 h of loaded time (Figure 2A), with the area under the curve being non-statistically different until 3 h of perfusion (Figure 2B). Additionally, loading-dependent metrics featured in Figure 3 suggested a loss of cardiac function as early as 30 min into loading time, with the measure for contractility (dP/dtmax) being the first metric to indicate loss of function (Figure 3C). Similar to heart rate, a steady decline in left ventricular pulse pressure (Figure 3A) was seen after 2 h of Loaded perfusion, with statistically significant differences between grafts' pulse pressure (Figure 3B) and relaxation (dP/dtmin, Figure 3D) seen after 1.5 h of perfusion.
Figure 1: Perfusion System Setup. (A) Perfusion setup. Black lines indicate the flow of perfusate during both perfusion modalities. The blue dashed line indicates the flow of perfusate during Langendorff perfusion only, and the orange lines indicate the flow of perfusate during load perfusion. (B) Location of the aortic (solid line) and atrial (dashed line) pressure sensors. Since the atrial pressure sensor is located spatially below the atrium, the height difference was taken into consideration for proper LA pressurization. Please click here to view a larger version of this figure.
Figure 2: Conventional assessment metrics. Metrics are denoted as conventional if their acquisition is possible during Langendorff perfusion. (A) Heart rate was calculated from the pressure waves obtained using the left intraventricular pressure sensor. However, as this metric can easily be calculated from EKG traces, also available during Langendorff, it is considered a conventional metric. The solid line is the median, and the shaded region is the interquartile range (IQR). (B) Area under the curve (AUC) of heart rate data for every 30 min of perfusion time. (C) Vascular resistance (CVR) of the grafts. (D) Oxygen uptake rate (OUR). (E) Lactate accumulation. (F) Glucose consumption. (G) Potassium accumulation. (H) Percent weight gain as a proxy for edema. All data is expressed as median ± interquartile range (IQR, n = 4). Repeated measures two-way ANOVA and Tukey-Kramer HSD post-hoc analysis on standard least squares means. * = p <0.01. Please click here to view a larger version of this figure.
Figure 3: Non-conventional assessment metrics. Metrics are denoted as non-conventional when their acquisition is only possible during loaded perfusion. (A) Left Ventricular (LV) pulse pressure plotted over time. The solid line is the median, and the shaded region is the IQR. (B) The area under the curve (AUC) of the LV pulse pressure for every 30 min of perfusion time. (C) AUC of cardiac contractility was obtained from the maximum derivative of the LV pulse pressure for every 30 min of perfusion. (D) AUC of cardiac relaxation was obtained from the minimum derivative of the LV pulse pressure for 30 min of perfusion time. All data is expressed as median ± interquartile range (IQR, n = 4). Repeated measures two-way ANOVA and Tukey-Kramer HSD post-hoc analysis on standard least squares means. * = p <0.01, ** = p <0.05, *** = p < 0.001, **** = p <0.0001. Please click here to view a larger version of this figure.
Ion | Concentration (mmol/L) |
Na+ | 135-145 |
K + | 3.5-6.0 |
Ca +2 | 1.0-1.3 |
Cl – | 96-106 |
Table 1: Acceptable range of ion concentrations in the perfusate.
Supplementary Figure 1: Perfusion system setup. Please click here to download this File.
Normothermic machine perfusion is a powerful modality for organ preservation and assessment that has greatly impacted the field of cardiac transplantation by expanding the donor pool of adult hearts36. This expansion is the result of the ability to currently utilize a small pool of hearts previously considered unsuitable for transplantation. Normothermic machine perfusion preserves cardiac grafts in a beating state, offering the opportunity for both functional and metabolic assessment. However, despite its potential, the current application of NMP remains constrained to a limited subset of marginal organs (i.e., organs donated after circulatory death with < 30 min warm ischemia time from donors < 55 years of age and no comorbidities). This limitation is caused by the lack of accurate and precise assessment techniques, which hinder the reliable evaluation of a broader range of marginal hearts.
Left ventricular loading during NMP has emerged as a promising assessment technique, as hemodynamic parameters obtained during loading can be highly correlated to post-transplantation outcomes15,16,18. In effect, the incorporation of cardiac loading during NMP as an assessment technique could facilitate the evaluation and likely utilization of extended criteria organs, further expanding the donor pool. Despite the immense potential impact, loading capabilities are not currently available in clinical NMP systems37. To simplify and standardize the incorporation of loading capabilities into established NMP systems, this protocol demonstrates the feasibility of passive left atrium (LA) pressurization (i.e., gravity-dependent pressurization), an approach that is rarely utilized in large heart perfusion. This pressurization method eliminates the need for secondary pumps, significantly simplifying the implementation of cardiac loading. However, as gravity is the major determinant of LA pressure, cognizance of the spatial placement of the loading reservoir and the accompanying pressure sensor (atrial pressure sensor) is critical for successful cardiac loading. Extreme care must be taken to account for height-induced pressure differences, as both over and under-pressurizing of the LA results in suboptimal loading conditions and can lead to graft failure through different mechanisms.
Exposing the LA to abnormally high pressures results in elevated cardiac filling pressures (i.e., mitral valve opening before complete relaxation of the left ventricle), a phenomenon known to cause cardiac graft damage and poor prognosis38. Additionally, and less evident, high pressures expose the LA to non-physiological stretching, leading to abnormal mechano-modulation and a potential increase in the secretion of atrial natriuretic peptides (ANPs)39. Both mechano-modulation and ANPs are known causatives of atrial fibrillation40,41. On the other hand, non-physiologically low left atrial pressures may result in inadequate filling of the left ventricle during diastole, leading to decreased cardiac output. As this protocol utilizes passive afterload (i.e., no retrograde perfusion to the aorta during loading with complete dependence of coronary blood flow on antegrade perfusion from the left ventricle), decreased cardiac output can result in reduced myocardial perfusion and relative ischemia. Over/under pressurization of the LA can also occur during the loading process, making it critical to pre-fill the loading reservoir and carefully verify its pressure before closing the atrium around the loading cannula. Incorrect atrial pressurization is the most probable source of technique-dependent graft injury. It was also the most challenging section to execute correctly, with numerous iterations being conducted to determine the best-performing sequence and the ideal size of the loading cannula.
Another particularly important element for protocol success and relevance is the use of a passive afterload that mimics the Windkessel effect seen in in vivo circulation. Under physiological conditions, the Windkessel effect refers to the ability of large central arteries, particularly the aorta, to act as a reservoir during systole and a conduit during diastole. During systole, the elastic walls of the aorta expand to accommodate a surge in blood volume42. This stored blood is then gradually released during diastole, aided by the elastic recoil of the arterial walls. This mechanism reduces the temporal variability of blood flow, ensuring somewhat continuous organ perfusion43. This is especially important for the heart as43 unlike other organs, myocardium perfusion occurs almost exclusively during diastole44. As such, incorporating the Windkessel effect into ex vivo perfusion systems provide a more physiological source of coronary perfusion, beneficial during both retrograde and loaded perfusion modes. Moreover, the implementation of this system into loaded perfusion for assessment purposes seems to augment predictive capabilities with the cardiac performance of grafts perfused with Windkessel-based afterload more strongly correlating to post-transplantation outcomes, as shown by others18. The Windkessel effect in this setup is achieved using a modified IV bag enclosed between two acrylic plates connected via a screw-in spring (Supplementary Figure 1), a relatively simpler system than other reported Windkessel apparatus18,19,20. Furthermore, the ability to tighten or untighten the loading spring provided immense control over the recoil pressure, facilitating the fine-tuning of experimental settings.
Using our system, the perfusion method can be easily transitioned back and forth between Langendorff and loaded mode. Loaded mode enabled the acquisition of functional data from the left ventricle of the grafts (i.e., left ventricular pulse pressure, contractility, relaxation, Figure 3). This functional data revealed a distinct difference in graft viability that was not apparent in biochemical trends (Figure 2), including lactate (Figure 2B), the clinical standard for cardiac graft assessment in ex vivo perfusion systems10. Furthermore, perfusion in Loaded mode, as described in this protocol, enables the possibility of measuring cardiac output (CO = heart rate * stroke volume), providing an additional and more conventional metric of cardiac function45,46. Stroke volume, the more challenging variable to acquire from the current setup, could be obtained by quantifying coronary flow, which, along with atrial and aortic flow, can be utilized to calculate end-diastolic and end-systolic volumes. Similarly, stroke volume can be obtained via echocardiogram measurements or pressure-volume loops. It is important to highlight that this setup is only capable of loading the left side of the heart, resulting in functional data of the left ventricle. This method provides no information regarding the right ventricular function. This is likely the largest limitation of the current protocol, particularly as right ventricular dysfunction is a more common occurrence in transplanted than left ventricular dysfunction47. However, right ventricular dysfunction is reported to resolve after a couple of weeks post-transplantation, perhaps de-prioritizing its importance47.
Notably, the four grafts presented within the results of this manuscript were procured and perfused using Langendorff for 6 h in the exact same manner. This experimental condition was selected as 6 h is the average preservation duration of normothermic machine-perfused grafts, and the addition of an extra 4 h of continuous loading (total of 10 h ex vivo time) was expected to result in some degree of graft failure. This expected failure provided the opportunity to highlight the necessity for functional metrics and demonstrated their assessment capability when compared to Langendorff-based metrics. Indeed, the results in this manuscript demonstrate the capability of some functional parameters acquired via passive LA pressurization and passive afterload loading to unmask viability differences of DBD cardiac grafts with extended preservation times. Although not demonstrated here, it is reasonable to assume this assessment effectiveness may be replicated to determine the viability of grafts with other known injuries and/or functional deficiencies (i.e., DCD, older donors, comorbidities), but further research is required to determine the functional characteristics and differences within these grafts. Similarly, further work is also required to determine how functional metrics correlate with transplantable versus non-transplantable cardiac grafts, as well as other possible sources of viability indices. For instance, the difference between aortic and atrial loading pressures can be a reliable viability index, as the difference between these pressure readings can indicate a lack or presence of cardiogenic shock48.
The authors have nothing to disclose.
We gratefully acknowledge funding to SNT from the US National Institute of Health (K99/R00 HL1431149; R01HL157803; R01DK134590; R24OD034189), the National Science Foundation under Grant No. EEC 1941543, the Claflin Distinguished Scholar Award on behalf of the MGH Executive Committee on Research, and the Polsky Family Award for Leaders in Surgery. We acknowledge research funding to AAO from the Hassenfeld Family Foundation, the MGH Executive Committee on Research, and the MGH Center for Diversity and Inclusion. We acknowledge research funding to GO from the Sarnoff Cardiovascular Research Foundation.
4- way Stopcock | Smiths Medical | MX9341L | |
4-0 Prolene sutures | Ethicon | 8711 | |
5-0 Suture | Fine Scientific Tools | 18020-50 | |
Aortic Connector | VentriFLO Inc | Custom Made | |
Aortic root cannula | Medtronic Inc | 10012 | |
Bovine Serum Albumin | Sigma | A7906 | |
Calcium Chloride | Sigma | C7902 | |
Cell Saver | Medtronic Inc | ATLG | |
Cell Saver cartridges | Medtronic Inc | ATLS00 | |
Dextran | Sigma | 31389 | |
EKG epicardial leads | VentriFLO Inc | Custom Made | |
Equipment stand and brackets | VentriFLO Inc | Custom Made | |
External Pace maker | Medtronic Inc | 5392 | |
Falcon High Clarity 50mL conical tubes | Fisher Scientific | 14-432-22 | |
Flow Probes | TranSonic Sytems inc | 1828 | |
Heparin sodium Injection | Medplus | G-0409-2720-0409-2721 | |
Hollow fiber oxygenator and Venous Resevior | Medtronic Inc | BBP241 | Affinity Pixie, 1L |
HTP 1500 Heat Therapy Pump | HTP | 6826619 | |
Insulin | Humulin R | MGH Pharmacy | |
Iworx Data Acquisition System | Iworx | IX-RA-834 | |
Krebs-Henseleit Buffer | Sigma | K3753 | |
Leukocyte Filter | Haemonetics | SB1E | |
Organ Chamber | VentriFLO Inc | Custom Made | |
Pacing Wires Biopolar | Medtronic Inc | 6495 | |
Penicillin-Streptomycin | ThermoFisher Scientific | 15140122 | |
Pressure Trasnducers | Iworx | BP100 | |
Pulsatile Pump | VentriFLO Inc | 2100-0270 | |
PVC Tubing | Medtronic Inc | HY10Z49R9 | |
Right Angle Metal Tip Cannula 20F | Medtronic Inc | 67318 | |
Sodium Bicarobonate | Sigma | 5761 | |
Standard PHD ULTRA CP Syringe Pump | Harvard Aparatus | 88-3015 | |
Tourniquet kit 7in | Medtronic Inc | 79006 | |
Transonic Flow box | TranSonic Sytems Inc | T402 | |
Venous Resevior | Medtronic Inc | CB841 | Affinity Fusion, 4L |
WIndKessel Bag | VentriFLO Inc | Custom Made | |
Y adapter | Medtronic Inc | 10005 |