Extending Normothermic Ex Situ Heart Preservation and Viability to 48 Hours Via Protein Supplementation and Plasma Exchange

Abstract

Prolonged normothermic ex situ heart perfusion (NEHP) could expand the donor pool by decoupling procurement, transport, and implantation, but clinically used systems rarely exceed 4–6 hours. We evaluated pragmatic protein replacement strategies enabling 48-hour NEHP in adult porcine hearts without biologic cross-circulation. Thirteen juvenile pigs (50.6±5.0 kg) underwent Langendorff NEHP, randomly assigned to SF65 (n=5), continuous infusion of serum fractionated to <65 kDa with hemofiltration; SF120 (n=3), serum fractionated to <120 kDa with hemofiltration; or plasma exchange (PE, n=5), unfractionated plasma with concomitant plasma filtration and intermittent hemofiltration. Hearts were perfused for 48 hours with intermittent working-mode assessments under predefined failure criteria. All hearts completed 48 hours without failure. PE demonstrated the lowest coronary resistance at 37–48 hours (0.13±0.05 vs 0.30±0.13 SF65, 0.69±0.31 SF120 mmHg/mL/min); SF120 exhibited markedly higher aortic regurgitation (38.1±27.7% vs 5.7±4.9% SF65, 5.0±6.7% PE); SF65 showed the greatest edema (IVSd increase 102±23% vs 37±17% SF120, 33±19% PE); echocardiographic indices and composite histologic injury scores were similar among groups. Reproducible 48-hour NEHP of adult porcine hearts is feasible using clinically accessible protein replacement strategies. Plasma exchange approximates the benefits of fractionated serum regimens while avoiding large-protein accumulation and represents a practical path toward translational long-duration heart perfusion.

Graphical Abstract

INTRODUCTION

The first clinically successful application of normothermic ex situ heart perfusion (NEHP) was reported in 2007.¹ Since its introduction, subsequent clinical trials have demonstrated the ability of NEHP systems to prolong ex situ preservation time, assess transplant suitability of marginal allografts, and effectively increase the donor pool size.^1–4 A few cases have been reported where organs were perfused ex situ for longer periods, but primary graft failure (PGF) was observed in all and a period of mechanical support with extracorporeal membrane oxygenation (ECMO) was required for up to 72h.^5,6

Several groups have reported successful NEHP preservation of heart allografts ≥24 hours by implementing whole-blood or plasma cross-circulation with a live paracorporeal animal.^7–10 Through this paracorporeal cross-circulation, the isolated allografts are provided with a continuous source of metabolic substrates, unspecified protein factors, and endocrine signaling. Perfusate cross-circulation additionally supports allograft vitality by filtration of accumulated toxic metabolites through engagement of the paracorporeal animal’s renal and/or hepatic metabolism. The benefit of perfusate filtration has been further demonstrated by successful NEHP prolongation with implementation of perfusate hemofiltration.^11,12

Our group has previously demonstrated that plasma-only cross circulation can extend normothermic heart perfusion to 72 hours, suggesting that the factors vital to NEHP prolongation are humoral components contained within plasma.⁹ From these findings, it was hypothesized that an NEHP-prolonging paracorporeal animal could be substituted by continuous perfusate supplementation with donor plasma and concomitant perfusate filtration. The clinical translation of this approach is difficult due to a multitude of ethical/regulatory issues, etc.; therefore, this experimental question was demonstrated on piglet hearts for 24 hours with continuous plasma exchange of fresh donor plasma without a paracorporeal animal.¹⁰ In the current study, we aim to expand our understanding of the beneficial effects of plasma or serum supplementation with concomitant perfusate filtration to successfully preserve adult-sized porcine hearts by NEHP for 48 hours.

MATERIALS AND METHODS

Animals

Thirteen healthy juvenile Yorkshire pigs (50.6±5.0 kg) were used as heart donors. All animals received humane care in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and protocols were approved by the University of Michigan Institutional Animal Care and Use Committee (PRO00011170).

Surgical Procedure and Back Table Preparation

Hearts were recovered under general anesthesia following standard procurement methods as previously described.¹¹ Cold (6°C) del Nido cardioplegia (50 cc/kg; CAPS Inc., Detroit, MI) was infused via the proximal innominate artery following aortic and branch vessel occlusion. Following diastolic arrest, hearts were explanted, weighed, and, on the back table, flushed and submerged in chilled (5–7°C) CoStorSol (Preservation Solutions, Elkhorn, WI) for cannulation and instrumentation. Del Nido provides single-dose diastolic arrest but is not formulated for static cold storage; the back-table CoStorSol flush, administered through aortic root cannula, clears residual cardioplegia and intravascular blood while providing the cold, osmotically balanced milieu required during cannulation. Pressure-transduced cannulas were placed in the left ventricular apex, right ventricle via the right pulmonary artery, left atrium via a left pulmonary vein (monitoring and infusion), and aortic root (Medtronic, Inc., Minneapolis, MN, USA). Cannula positions are depicted in Figure 1. Remaining branches were ligated.

Figure 1. Normothermic Ex Situ Heart Perfusion (NEHP) Circuit Schematics.

(A) NEHP circuit configuration for the SF65 and SF120 experimental groups, consisting of a hard-shell reservoir, oxygenator/heat exchanger, roller pump, and hemoconcentrator with continuous serum fractionate infusion and hemofiltration; (B) Intermittent left atrial (iLA) perfusion configuration used for scheduled hemodynamic and echocardiographic assessments at a fixed perfusion rate of 0.75 cc/g/min. SF65 = 65 kDa serum fractionate; SF120 = 120 kDa serum fractionate.

NEHP System and Perfusate Composition

The NEHP system has been described previously by our group.^10–12 Briefly, it consisted of a hard-shell reservoir (Terumo, Ann Arbor, MI), FX05 Baby Capiox Oxygenator/Heat Exchanger (Terumo, Ann Arbor, MI), M-pump (proprietary roller-style pump, Daris LLC, Ann Arbor, MI), and Prisma flex HF1000 Hemoconcentrator (Baxter, Deerfield, IL) (Figure 1A). In the plasma exchange group, a Plasmaflow OP-05W(A) plasma filter (Asahi Kasei, Tokyo, Japan) was added.

Perfusate

For each study, blood was obtained from a single designated blood-donor pig weighing >90 kg. A dedicated donor was used rather than autologous collection because the volume required for 48-hour perfusion with scheduled exchanges (approximately 5–6 L) exceeded what could be safely obtained from a heart donor. Use of allogeneic blood also models a clinically scalable workflow supported by compatible blood-bank products. Blood was then collected in CPDA-1 and separated via centrifugation at 3,600 RPM at 4°C for 30 minutes. Red blood cells were leukocyte reduced with RS1 filters (Haemonetics, Boston, MA) and washed with Plasmalyte. Perfusate was composed of 90 cc plasma, 110 cc pRBCs (target Hgb 8.0–11.0 g/dL), 250 mg solumedrol, 45 mg magnesium, 50 mg calcium (target 1.15–1.3 mmol/L), 2,500 U heparin, 40 mg gentamicin, and 250 mg nafcillin.

Fractionated Serum Supplement Preparation

Donor plasma was treated with calcium (100 mg/mL), incubated at 4°C for 48 hours, and centrifuged to isolate serum. Serum was sterile-filtered (0.22 μm), then fractionated through 120 kDa or 65 kDa hollowfiber hemofilters (Terumo, Ann Arbor, MI, USA) to create serum fractionate <120 kDa (SF120) and serum fractionate <65 kDa (SF65), respectively. Protein size was confirmed by western blot (data not shown). The 120-kDa hemofilters were obtained from an early-generation Molecular Adsorbent Recirculating System (MARS) platform. Production of hemofilters with this molecular-weight cutoff has since been discontinued, and no equivalent device is currently available.

NEHP Reanimation

The NEHP circuit was primed and de-aired with calcium- and bicarbonate-free perfusate. Adenosine (3 mg) was circulated for 5 minutes prior to heart connection. Following de-airing, Langendorff perfusion was initiated at 0.5 cc/g/min. After tissue reperfusion, progressive rewarming to 37°C and electrolyte correction were performed (Supplemental Figure 1). A perfusate exchange eliminated residual cardioplegia and reperfusion byproducts.

Hearts were maintained via NEHP for 48 hours at a coronary perfusion rate of 0.60 cc/g/min, temperature of 37°C, and sweep gas (50% O₂, 45% N₂, 5% CO₂) titrated to pCO₂ 40±5 mmHg. Hemoglobin was maintained at 8.5–11 g/dL with perfusate exchanges every 16 hours to offset progressive hemolysis and accumulation of metabolic byproducts. Every 12 hours, the perfusion technique was transitioned to a coronary only LA infusion strategy at a fixed rate of 0.75 cc/g/min (Figure 1B). Hearts were given ~5 minutes to equilibrate to iLA conditions prior to iLA hemodynamic recording (q12 hours) and echocardiography assessment to assess valvular and overall systolic function (q24 hours). The iLA assessment intervals for imaging and hemodynamics were established to limit cumulative time in the non–steady-state iLA configuration.

Perfusion Maintenance and Experimental Groups

A negative control group using unsupplemented, unfiltered NEHP was not repeated in the present study, as prior work from this laboratory^9,11,12 and others², has demonstrated that unsupplemented, unfiltered NEHP does not reliably maintain donor hearts beyond 4–6 hours; in accordance with the 3R principle of Reduction, these data were used rather than duplicating a predictably non-viable arm.

Because this was a feasibility study, no formal sample-size calculation was performed. Based on prior experience in our laboratory, the enrollment target was set at five experiments (n=5) per group. Enrollment in the SF120 group was discontinued after three experiments because additional 120-kDa hemofilters were no longer available, as described above. On each experimental day, before procurement, group allocation was determined by simple random assignment using slips labeled SF65, SF120, or PE drawn from an opaque container (Supplemental Figure 2). Hearts received SF65 (continuous infusion at 30–50 cc/hr, n=5), SF120 (continuous infusion at 30–50 cc/hr, n=3), or PE (continuous unfractionated plasma infusion at 30–50 cc/hr, n=5). SF65 and SF120 groups underwent continuous 22 kDa hemofiltration (1.0 cc/g/hr) with matched fluid replacement. As to limit redundant fluid loss and excessive removal of signaling and metabolic substrates, the PE group underwent continuous plasma filtration matched to infusion rate, with hemofiltration initiated only as needed (K⁺ >5.5, Ca²⁺ >1.40 mmol/L, or glucose >400 mg/dL). Termination criteria included: gross valvular dysfunction, refractory arrhythmia, lactate increase >1.0 mmol/L/hr for two consecutive hours, or lactate >7.0 mmol/L. Vasoactive medications were managed according to a prespecified algorithm: epinephrine and/or norepinephrine were initiated for MAP < 30 mmHg while preserving coronary perfusion. A vasoactive-inotropic score was computed at hourly intervals. The primary endpoint was completion of 48 hours of perfusion without meeting termination criteria. Secondary endpoints comprised the hemodynamic, metabolic, echocardiographic, edema, and histopathologic measures described below.

Data Collection and Histopathology

Data collection was performed as previously described.¹¹ Hemodynamic parameters were recorded continuously; blood gases collected hourly; chemistry panels every 6 hours. Intermittent left atrial (iLA) hemodynamic assessment was performed every 12 hours at 0.75 cc/g/min (Figure 1B).¹² Echocardiography was performed by BH and TN in iLA mode every 24 hours using a M9 (Mindray) platform with a phased-array transducer applied directly to the ex situ heart at standardized short-axis and four-chamber views. Stroke volume (SV), ejection fraction (EF), cardiac output (CO), and interventricular septal diameter (IVSd) were measured at a fixed perfusion rate of 0.75 cc/g/min. In addition to echocardiography, aortic regurgitation (AR) was quantified continuously from circuit flow sensors as the ratio of pulmonary artery (PA) return flow to aortic inflow: AR (%) = 100 × [(PA flow − Aortic flow) / Aortic flow].

Myocardial oxygen consumption (MVO₂) was calculated as: MVO₂ = [(CaO₂ − CvO₂) × Q] / (100 × HW) where CaO₂ and CvO₂ are the arterial and venous oxygen content (mL O₂/dL), respectively, Q is coronary flow rate (mL/min), HW is heart weight (g), and the factor of 100 converts oxygen content from per deciliter to per milliliter.

Following 48 hours, hearts were weighed and full-thickness sections from five regions were fixed in 10% neutral buffered formalin. Sections were processed and stained with hematoxylin and eosin (H&E) by the ULAM Pathology Core at the University of Michigan (RRID:SCR_018823). H&E slides were scored by a board-certified veterinary pathologist (ILB) blinded to experimental groups, using ordinal severity scoring (0–3) for myofiber degeneration, hemorrhage, interstitial edema, and endothelial changes (Supplemental Table 1). Scores from the five sampled regions were averaged within each injury category, and the average composite injury score was calculated as the mean of the four category scores.

Statistical Analysis

Continuous variables are presented as means ± standard deviations. Comparisons were conducted using two-way ANOVA with Tukey’s multiple comparisons. Histopathologic data were compared using the Kruskal–Wallis test with Dunn’s post-hoc pairwise comparisons. A p < 0.05 value was considered significant.

RESULTS

Heart Characteristics and Viability

There were no significant differences in initial heart weights or cold ischemic times across groups (Supplemental Table 2). Among the 13 hearts, 7 (53.8%) were female and 6 (46.2%) were male with the following breakdown amongst groups: PE (3F/2M), SF65 (2F/3M), SF120 (2F/1M). All 13 hearts completed 48 hours of NEHP without meeting predetermined termination criteria. Coronary perfusion flow was maintained at approximately 0.60 cc/g/min (±0.05 cc/g/min, SD) in all groups, though the SF120 group required lower aortic flow rates due to elevated coronary resistance and aortic regurgitation (Supplemental Table 3).

Hemodynamic Characteristics

At initiation of NEHP (h 2–6), there was no difference in coronary resistance (CR), mean arterial pressure (MAP), or aortic regurgitation (AR) across groups (Figure 2A–C). Over time, the SF120 group displayed progressive hemodynamic deterioration requiring escalating vasoactive support (Figure 2D), with a significant increase in CR (h 2–12: 0.31±0.32 vs h 37–48: 0.69±0.31 mmHg/mL/min, p < 0.001) while SF65 and PE remained stable. At h 37–48, CR was lowest in the PE group (0.13±0.05) and highest in SF120 (0.69±0.31 mmHg/mL/min; all pairwise p < 0.001). MAP followed a similar pattern, with the lowest values in PE and highest in SF120 at late time points (Supplemental Table 3). AR increased across time only in the SF120 group (38.1±27.7% vs SF65 5.7±4.9% and PE 5.0±6.7%; SF120 vs both p < 0.001) (Figure 2C; Supplemental Table 3).

Figure 2. Hemodynamic Characteristics During NEHP.

(A) Coronary resistance; (B) Mean arterial pressure; and (C) Aortic regurgitation depicted as continuous variables segmented into 12-hour intervals for statistical comparison; (D) Vasoactive-inotropic score depicted as continuous mean. Statistically significant differences between groups (p < 0.05) are depicted by * (SF65 vs SF120), dagger (SF65 vs PE), and double dagger (SF120 vs PE). Statistically significant change over time is depicted by * on the Y axis. SF65 = 65 kDa serum fractionate; SF120 = 120 kDa serum fractionate; PE = plasma exchange.

Metabolic Characteristics

During initial perfusion (h 2–12), oxygen consumption was comparable across groups (Figure 3A). All groups showed a significant decrease in oxygen consumption between h 2–12 and h 13–24, corresponding to resolution of initial hypercontractility. After the first phase, the SF120 and PE groups maintained this reduced oxygen consumption through h 37–48, whereas the SF65 group demonstrated a subsequent increase (h 37–48: 0.018±0.006 mL/g/min, SF65 vs SF120 p < 0.001, SF65 vs PE p < 0.001) (Supplemental Table 4). These dynamics occurred despite higher vasoactive-inotropic scores in the SF120 group (Figure 2D).

Figure 3. Metabolic Characteristics During NEHP.

(A) Oxygen consumption; (B) Carbon dioxide production; and (C) Perfusate lactate concentration depicted as continuous variables segmented into 12-hour intervals for statistical comparison; (D) Net lactate consumption depicted as mean ± standard deviation. Statistical annotations as in Figure 2.

CO₂ production was initially highest in SF65 (h 2–12, p < 0.001 vs both groups) (Figure 3B). Over time, CO₂ production increased in SF120 and decreased in PE, such that at h 37–48 both SF65 and SF120 exceeded PE (Supplemental Table 4). The PE group displayed a more linear CO₂ production trend.

Perfusate lactate levels remained lower in PE at h 2–24 but subsequently increased, ultimately exceeding SF65 and SF120 at h 37–48 (Figure 3C). However, SF65 and SF120 groups underwent continuous hemofiltration with active lactate removal. Net lactate consumption (venous minus arterial lactate) demonstrated consumption in the PE group (−0.017 ± 0.07 mmol/dL) versus production in SF65 and SF120, though differences were not statistically significant (Figure 3D; Supplemental Table 5).

Biochemical Analysis

Perfusate albumin and total protein concentrations were comparable across groups at hour 2 (Figures 4A–B). Over time, the SF120 group displayed progressive accumulation of albumin, total protein, and other large serum proteins (>50 kDa) including alanine transaminase, amylase, and total bilirubin (Figures 4C–E), consistent with large-protein supplementation without concomitant large-protein clearance. Conversely, soluble phosphorous (<22 kDa) accumulated preferentially in the PE group (Figure 4F), reflecting the absence of continuous hemofiltration. Exploratory ELISA analysis of myocardial strain and mitochondrial integrity markers was limited by sample availability and is presented in Supplemental Figure 3.

Figure 4. Perfusate Biochemical Concentration Trends Over 48 Hours of NEHP.

(A) Albumin; (B) total protein; (C) alanine transaminase; (D) amylase; (E) total bilirubin; and (F) phosphorous. Values reported as mean ± standard deviation. SF65 (blue circles), SF120 (red squares), PE (black triangles). SF65 = 65 kDa serum fractionate; SF120 = 120 kDa serum fractionate; PE = plasma exchange.

Interstitial Edema

Interstitial edema, measured by IVSd percent increase from baseline, was comparable across groups at hour 1 (Figure 5A). The SF65 group displayed progressive IVSd increase, reaching 102±23% at hour 48 versus 37±17% (SF120) and 33±19% (PE) (SF65 vs SF120 p = 0.009, SF65 vs PE p < 0.001). A majority of IVSd increase in PE occurred between hours 1 and 24. Gross weight increase followed a similar pattern (SF65 64.4±8.3% vs SF120 39.2±8.7%, and PE 41.1±7.1%; Figure 5B; Supplemental Table 6).

Figure 5. Interstitial Edema Assessment.

(A) Echocardiographic measurement of percent increase in interventricular septal wall diameter (IVSd) from in vivo baseline at hours 1, 24, and 48. IVSd was measured at the mid-septum at end-diastole. (B) Gross heart weight percent increase at hour 48. All echocardiographic measurements were acquired during intermittent left atrial perfusion at a fixed rate of 0.75 cc/g/min. Values reported as mean ± standard deviation. *p < 0.05, **p < 0.01, ***p < 0.001.

Echocardiography and Histopathology

Echocardiographic assessment during iLA perfusion demonstrated no significant difference in SV, EF, or CO across groups at hour 48; however, there is a notable downward trend (Figures 6A–C). Tricuspid regurgitation was absent in all hearts; aortic and mitral regurgitation remained trivial-to-mild in SF65 and PE, whereas SF120 hearts overall developed mild/moderate aortic and mitral regurgitation (Supplemental Table 7). Histopathology showed no difference in myofiber degeneration, hemorrhage, or endothelial damage. Interstitial edema was significantly greater in PE (1.9±0.17) by histological assessment than SF65 (0.80±0.38) or SF120 (0.33±0.33; both p < 0.001). However, average composite injury scores were not significantly different across groups (Figure 7).

Figure 6. Echocardiographic Measurements During Intermittent Left Atrial Perfusion at Hours 0, 24, and 48.

(A) Stroke volume (mL/beat); (B) ejection fraction (%); and (C) cardiac output (LPM). All measurements were acquired in intermittent left atrial (iLA) mode at a fixed perfusion rate of 0.75 cc/g/min. Values reported as mean ± standard deviation.

Figure 7. Histopathologic Injury Scoring Following 48 Hours of NEHP.

At the conclusion of the 48-hour perfusion protocol, hearts were explanted and full-thickness tissue samples were collected from five regions: the right atrium (RA), right ventricle (RV), left atrium (LA), left ventricle (LV), and interventricular septum (IVS). Each full-thickness sample was assessed across the entire wall thickness. Sections were fixed in 10% neutral buffered formalin, stained with hematoxylin and eosin, and scored by a board-certified veterinary pathologist blinded to experimental group. Injury categories include myofiber degeneration, myocardial hemorrhage, interstitial edema, and endothelial damage, each graded on an ordinal scale from 0 (absent) to 3 (most extensive). The composite injury score represents the mean of these category scores across all sampled regions. Results are reported as mean injury score ± standard deviation. ***p < 0.001, ****p < 0.0001.

DISCUSSION

Heart failure is a rising global epidemic with heart transplantation considered the gold standard for treatment of end-stage disease.^13–17 Current constraints of ex situ allograft preservation time to 6 hours remain a limiting factor for continued donor pool expansion.^18,19 The only clinically available normothermic ex situ heart perfusion device, the OCS Heart (Transmedics, Andover, MA, USA), has demonstrated non-inferior or improved post-implantation outcomes compared to standard static-cold storage.^2,20–22 Prolonged ex situ allograft preservation could mitigate temporal and geographic limitations of organ donation, improve organ matching, and allow for further assessment or treatment of marginal hearts. Our study demonstrates that 48-hour NEHP of adult-sized porcine hearts is achievable without biologic cross-circulation by combining protein supplementation with targeted perfusate filtration.

Protein Supplementation

In this series of 13 hearts, three distinct strategies were evaluated: small-protein serum supplementation (SF65), larger-protein serum supplementation (SF120), and continuous plasma exchange (PE), all of which supported 48 hours of NEHP without triggering predefined failure criteria. Across groups, echocardiographic indices at 48 hours were preserved and histopathology demonstrated no observed endothelial damage or coronary thrombosis in any group.

Despite this shared feasibility, the strategies diverged in their hemodynamic and metabolic profiles. Large-protein supplementation without concomitant large-protein clearance (SF120) led to progressive accumulation of albumin and other high-molecular-weight proteins, higher bilirubin levels, and increasing coronary resistance with hemodynamic lability that required escalating vasoactive support. In contrast, PE balanced ongoing protein supplementation with large-protein clearance and maintained the lowest coronary resistance and a more stable perfusion profile. Current results suggest that ideal perfusate supplementation is multifactorial: while the addition of small protein supplement (<65 kDa) is sufficient to prolong NEHP to 48 hours, large protein supplement (molecules <65 kDa and molecules 65–120 kDa) provides more optimal preservation characteristics (Supplemental Table 8).

Protein Size, Endothelial Integrity, and Metabolic Efficiency

Our comparison of SF65 and SF120 suggests that proteins in the 65–120 kDa range contribute meaningfully to edema control and metabolic behavior during prolonged perfusion. Hearts receiving large-protein supplementation exhibited less gross weight gain and smaller increases in interventricular septal thickness compared with SF65, effects that may reflect improved endothelial barrier function mediated by albumin and other mid-sized plasma proteins.

Although histologic edema was highest in PE, SF65 had the highest IVSd and weight gain, reflecting a distinct fluid-compartment phenomenon. IVSd and gross heart weight capture bulk wall expansion dominated by intramyocyte and perivascular fluid, whereas the histologic interstitial edema score reflects focal microscopic separation between myofiber bundles. Continuous unfractionated plasma exchange maintains colloid oncotic pressure and limits bulk swelling, but the ongoing transcapillary flux inherent to the exchange strategy may generate the focal interstitial separation visible on H&E. For transplant relevance, bulk swelling is the parameter that would impair graft fit and function and is best reflected by IVSd and gross weight. Importantly, composite histologic injury scores did not differ across groups (Figure 7).

Metabolically, the SF65 group demonstrated higher oxygen consumption at late time, consistent with less efficient oxidative metabolism. Both SF120 and PE maintained lower oxygen consumption and CO₂ production, compatible with more efficient mitochondrial function. Although our ELISA data were underpowered, trends toward lower circulating flavin mononucleotide in the large-protein groups support a hypothesis of more preserved mitochondrial integrity with broader protein supplementation (Supplemental Figure 3).

Supplementation with Hemofiltration

The SF120 group highlights the risks of unbalanced large-protein supplementation. Continuous infusion of higher-molecular-weight proteins without a dedicated large-protein filter led to progressive accumulation of albumin, liver enzymes, and bilirubin, alongside rising coronary resistance and more frequent need for vasoactive adjustments. These mechanisms may also have contributed to the aortic regurgitation observed in the SF120 group. Because the present feasibility study was not designed to determine the etiology of this finding, the following possibilities are proposed as testable hypotheses rather than definitive mechanisms:: (1) progressive coronary edema and rising coronary resistance in SF120 elevated diastolic root pressure and created a hemodynamic substrate for diastolic backflow; (2) accumulation of high-molecular-weight proteins in the absence of a matched large-protein clearance step likely altered perfusate oncotic and viscosity properties; and (3) progressive interstitial expansion within the aortic root and valvular apparatus may have contributed to functional leaflet malcoaptation. Importantly, histopathology did not show frank endothelial destruction or widespread thrombosis, implying that the observed hemodynamic changes were driven by vascular tone rather than irreversible structural damage. These findings support that small-solute hemofiltration alone is unlikely to be sufficient once large-protein supplementation is introduced; pairing protein delivery with selective clearance (e.g., plasma filtration) appears necessary to prevent accumulation of vasoactive or pro-inflammatory factors.

Clinical Translation

From a translational perspective, the most immediately actionable finding is that continuous plasma exchange provides a practical route to 48-hour NEHP. Plasma exchange uses components already embedded in cardiac surgery and transplant workflows and does not require custom fractionation or specialized serum processing. By contrast, fractionated serum supplementation (SF65 and SF120) is mechanistically informative but currently less scalable. PE appears to offer a more favorable overall balance between edema control, hemodynamic stability, and metabolic efficiency. In the context of recent preservation strategies, this platform offers a substantively longer preservation window, continuous metabolic waste clearance, and functional assessment of a loaded heart, positioning NEHP as a reconditioning and organ-directed-therapy platform rather than a transport modality (Supplemental Table 9).

Study Limitations

This study has several limitations. First, sample sizes were modest and unequal (n = 5, n = 3, and n = 5 for SF65, SF120, and PE, respectively), limiting the strength of between-group inferences. Accordingly, findings specific to SF120 should be considered preliminary. Second, supplementation and filtration rates were chosen pragmatically rather than formally optimized. Third, the progressive 48-hour decline in echocardiographic systolic indices across all groups, although not statistically significant, needs to be noted and without a transplantation endpoint, our findings should be interpreted as evidence of the feasibility of 48-hour perfusion rather than evidence of preserved transplantability, which remains untested. No single marker of transplantation suitability currently exists.^23,24 Additional investigations to validate current findings with a transplantation model are ongoing. Lastly, perfusate lactate interpretation is confounded in this model. Recent clinical data indicate that perfusate lactate profiles do not reliably predict post-transplant primary graft dysfunction.^25,26 Additionally, continuous hemofiltration in SF65 and SF120 actively cleared lactate while PE received only intermittent hemofiltration; PE nonetheless demonstrated net lactate consumption on arteriovenous sampling.

Conclusions

In an adult porcine model, we demonstrate that 48-hour normothermic ex situ heart perfusion is feasible without biologic cross-circulation when protein supplementation is paired with appropriate perfusate filtration. These data support plasma-based, filtration-coupled NEHP as a practical foundation for future studies aimed at extending preservation, rehabilitating marginal grafts, and enabling organ-directed therapies.

Abbreviations

Ao

aortic root

AR

aortic regurgitation

CBC

complete blood count

CO

cardiac output

CR

coronary resistance

EF

ejection fraction

H&E

hematoxylin and eosin

HR

heart rate

iLA

intermittent left atrial perfusion

IVSd

interventricular septal wall diameter

LA

left atrium

LV

left ventricle

LVEDV

left ventricular end diastolic volume

LVESV

left ventricular end systolic volume

LVOT-d

diameter of the left ventricular outflow tract

LVOT-VTI

velocity time integral of the left ventricular outflow tract

MAP

mean arterial pressure

NEHP

normothermic ex situ heart perfusion

PA

pulmonary artery

PE

plasma exchange

RA

right atrium

RV

right ventricle

SF120

120 kDa serum fractionate

SF65

65 kDa serum fractionate

SV

stroke volume