Objectives: Hypothermic oxygenated perfusion reduces ischemia-reperfusion injury and improves early graft function in extended criteria donor liver transplant. However, use of machine perfusion may be limited because of lack of dedicated perfusion devices. We evaluated feasibility and clinical outcomes of dual hypothermic oxygenated perfusion performed using a standard cardiopulmonary bypass system.
Materials and Methods: We retrospectively analyzed 87 adult recipients who underwent orthotopic liver transplant from extended criteria donors (2018-2024). We compared 34 cases with static cold storage (SCS group; 2018-2022) versus 53 with dual hypothermic oxygenated perfusion using a modified cardiopulmonary bypass system (D-HOPE group; 2023-2024), analyzing early graft function, early allograft dysfunction, postoperative renal function, complications, and short-term survival. Morphological assessment included light and transmission electron microscopy at 3 time points: before preservation, after dual hypothermic oxygenated perfusion, and after static cold storage.
Results: Dual hypothermic oxygenated perfusion was associated with significantly lower aspartate aminotransferase on postoperative day 1 (1450 [1180-2210] vs 2280 [1620-3010] U/L; P = .⁰⁰³) and alanine aminotransferase (1180 [900-2002] vs 1890 [1280-2500] U/L; P = .⁰²⁴). Total bilirubin and international normalized ratio were significantly lower on postoperative days 3 and 7. Early allograft dysfunction occurred in 28.3% of patients in the D-HOPE and 61.8% in the SCS group (P = .⁰⁰⁴). Reduced estimated glomerular filtration rate (<45 mL/min/1.73 m²) was less frequent and total intensive care unit and hospital stay were significantly shorter in the D-HOPE group. Electron microscopy demonstrated better preservation of mitochondrial ultrastructure after dual hypothermic oxygenated perfusion, whereas light histology showed comparable findings across groups.
Conclusions: Dual hypothermic oxygenated perfusion performed with a standard cardiopulmonary bypass system is feasible and safe and associated with improved early graft function in extended criterial donor liver transplant. This conventional platform may facilitate hypothermic oxygenated perfusion in transplant centers lacking dedicated machine perfusion systems. Key words: Early allograft dysfunction, Ischemia-reperfusion injury, Mitochondrial ultrastructure, Organ preservation, Renal dysfunction

Introduction
Liver transplantation remains the only definitive therapy for patients with end-stage liver disease. Despite major advances in surgical techniques and perioperative care, the persistent shortage of donor organs continues to constrain transplant activity worldwide. To meet the growing demand, transplant programs increasingly rely on extended criteria donor (ECD) grafts, including livers from elderly donors, those with substantial macrovesicular steatosis, hypernatremia, prolonged vasopressor exposure, or extended cold ischemia time. However, the use of ECD grafts is associated with a higher risk of ischemia-reperfusion injury (IRI), early allograft dysfunction (EAD), and postoperative complications.¹ Hypothermic oxygenated perfusion (HOPE) has emerged as an effective strategy to mitigate IRI. This technique was first described by Dutkowski and colleagues, who demonstrated that short-term HOPE of donor livers enables controlled reoxygenation prior to implantation, thereby attenuating classical reperfusion injury and facilitating restoration of cellular energy metabolism in experimental liver transplantation.² The protective effect of HOPE is primarily related to oxygen delivery under hypothermic conditions, when metabolic demand is minimal. This environment allows recovery of mitochondrial respiration and ATP resynthesis while limiting the burst of reactive oxygen species typically observed during warm reperfusion. As a result, HOPE provides gradual and controlled graft reoxygenation, reduces oxidative stress, dampens inflammatory activation, and ultimately improves early graft function^3,4 (Figure 1). The clinical benefit of HOPE in liver transplant with ECD grafts has been confirmed in several multicenter randomized studies. In the PERPHO trial, HOPE was associated with lower peak transaminase levels, a reduced incidence of EAD, and shorter hospital stay compared with static cold storage (SCS).⁵ Similar findings were reported in the French HOPExt study, where HOPE improved early graft function and decreased postoperative complications.^6,7 Dual hypothermic oxygenated perfusion (D-HOPE) is a clinical refinement of this approach and involves simultaneous ex situ perfusion of the liver via both the portal vein and the hepatic artery using an oxygenated perfusate at 4 °C to 10 °C. With the ability to provide oxygen delivery to both vascular inflow pathways, D-HOPE ensures more homogeneous graft oxygenation and has gained increasing acceptance in clinical practice as part of hypothermic machine perfusion strategies. Despite strong clinical evidence, the broader adoption of HOPE remains limited in some transplant centers as a result of restricted access to dedicated machine perfusion systems. In this context, adapting a standard cardiopulmonary bypass (CPB) system for HOPE represents a practical and readily implementable alternative. This approach enables controlled dual inflow perfusion via the portal vein and hepatic artery, with continuous monitoring of temperature, perfusion pressure, and perfusate gas parameters and with use of equipment routinely available in liver transplant centers. In this study, we aimed to assess the clinical outcomes, safety, and reproducibility of ex situ D-HOPE performed with a standard CPB system in ECD liver transplant.

Materials and Methods
This study was conducted at the Department of Organ and Tissue Transplantation, Moscow Multidisciplinary Clinical Center named after S.P. Botkin, Moscow Healthcare Department (Botkin Hospital), between January 2018 and December 2024. Adult patients who underwent orthotopic liver transplant with ECD grafts were included. Donor livers were classified as ECD if at least one of the following risk factors was present: donor age >65 years, intensive care unit stay with mechanical ventilation >7 days, body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) >30, macrovesicular steatosis >40%, hypernatremia (sodium >165 mmol/L), elevated alanine aminotransferase (ALT; >105 U/L) and/or elevated aspartate aminotransferase (AST; >90 U/L), total bilirubin >3 mg/dL (51 µmol/L), vasopressor support with norepinephrine >0.2 µg/kg/min (or an equivalent catecholamine dose) for at least 2 hours before procurement, or cold ischemia time >8 hours.⁸ We conducted a retrospective comparative analysis of 87 recipients who underwent liver transplant with ECD grafts. Patients were categorized into 2 groups according to the preservation method applied. The SCS group (n = 34) included transplants performed after conventional SCS between 2018 and 2022 and was the control group. The D-HOPE group (n = 53) included transplants in which graft preservation was performed using ex situ D-HOPE with a standard CPB system between 2023 and 2024. The D-HOPE method was performed using a standard CPB system modified according to a back-to-base configuration to enable extracorporeal hypothermic oxygenated liver perfusion. Perfusion was conducted through 2 separate circuits (portal and arterial) with roller pumps and a membrane oxygenator (Figure 2 and Figure 3). Custodiol (HTK) solution was used as the perfusate and was modified by adding 20% human albumin to create a colloid component (100 mL of 20% albumin per 1 L of Custodiol). The perfusate was oxygenated with a gas mixture of 95% O2 and 5% CO2. Perfusion parameters were as follows: temperature of 4 to 10 °C, portal pressure of 3 to 5 mm Hg, arterial pressure of 25 mm Hg, and partial oxygen pressure of 400 to 600 mm Hg. Upon completion of perfusion, corresponding to the start of graft implantation, the liver was flushed with cold (4 °C) 5% albumin solution to remove residual perfusate. The same 5% albumin flush was performed in the SCS group immediately before implant. All transplants were performed using a standard orthotopic liver transplant technique with vascular anastomoses and duct-to-duct biliary reconstruction. Postoperative management and immunosuppressive protocols were standardized across both groups. A conventional immunosuppressive regimen was applied in all cases. Induction therapy consisted of 20 mg basiliximab administered intraoperatively and repeated on postoperative day 4. Glucocorticoid therapy included intravenous methylprednisolone (10 mg/kg) administered immediately before graft reperfusion, followed by tapered dosing on postoperative days 1 to 3, with reduction to 30 mg/day and complete discontinuation on day 4. Maintenance immunosuppression in most recipients consisted of tacrolimus monotherapy, initiated on postoperative day 2 at 0.05 mg/kg/day, targeting trough levels of 7 to 10 ng/mL. In cases of tacrolimus-related nephrotoxicity or other adverse effects, mycophenolic acid (1000 mg twice daily) was introduced, and the target calcineurin inhibitor trough level was reduced to 4 to 7 ng/mL. Patients with autoimmune liver diseases received oral methylprednisolone when clinically indicated. Graft function was assessed by serial biochemical measurements during the first postoperative week, including ALT, AST, total bilirubin, and international normalized ratio (INR). Early allograft dysfunction was defined according to the criteria of Olthoff and colleagues⁹: peak AST >2000 U/L, bilirubin ≥10 mg/dL, and/or INR ≥1.6 on postoperative day 7. The severity of postoperative complications was evaluated with the Comprehensive Complication Index. Morphological assessment included transmission electron microscopy and light histology. Immediately after arrival of the donor liver in the operating room and before initiation of perfusion, a baseline tissue sample was obtained from the graft for morphological analysis. In addition, a 2 × 2-cm biopsy fragment was excised from the liver edge. This fragment was not connected to the perfusion system and was immediately placed in cold (4 °C) Custodiol solution and maintained under SCS conditions. After completion of D-HOPE, an additional tissue sample was obtained from the perfused graft. A corresponding sample was also taken from the previously excised biopsy fragment that had been stored in Custodiol under SCS conditions. All specimens were subsequently processed for electron microscopic and histological analysis according to the study design shown in Figure 4. We used IBM SPSS version 26.0 for statistical analyses. We presented continuous variables as median and interquartile range (IQR) and presented perfusion parameters as mean ± SD due to their stable distribution and protocol-driven nature. We used the Mann-Whitney U test for group comparisons and analyzed categorical variables with the χ2 test or the Fisher exact test, as appropriate. We assessed correlations with the Spearman rank correlation coefficient. P < .⁰⁵ was considered statistically significant. The study protocol was approved by the institutional ethics committee of Botkin Hospital and conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all recipients before transplant.

Results
The study groups were comparable with respect to baseline donor and recipient characteristics (Table 1). Median donor age was similar between groups, 56 years (IQR, 28-64 y) in the SCS group and 55 years (IQR, 28–67 y) in the D-HOPE group (P = .⁵⁴). Median recipient age was 49 years (IQR, 38–57 y) in the SCS group and 50 years (37-59 y) in the D-HOPE group (P = .⁶⁷). Sex distribution, BMI, Model for End-Stage Liver Disease score, and cold ischemia time did not differ significantly between groups (P > .⁰⁵). More than half of donors had BMI >25, and approximately 20% had BMI >30. Hypernatremia (sodium >155 mmol/L) was observed in 38% of donors. Macrovesicular steatosis of >40% was identified in 70.6% of grafts in the SCS group and 79.3% in the D-HOPE group. Two or more ECD risk factors were present in 44% and 47% of cases, respectively. Median cold ischemia time was 5.6 hours (IQR, 3.8-7.4 h) in the SCS group and 5.4 hours (IQR, 3.6-7.2 h) in the D-HOPE group (P = .⁴²) (Table 1). During D-HOPE (Figure 5), perfusion parameters remained stable throughout the procedure. Mean portal pressure was 4.1 ± 0.6 mm Hg, and mean arterial pressure was 25.3 ± 1.4 mm Hg. Perfusate temperature was maintained at 6.0 ± 1.1 °C, with a mean partial oxygen pressure of 480 ± 55 mm Hg. The median duration of perfusion was 105 minutes (IQR, 75-120 min). In the early postoperative period, recipients undergoing liver transplant with D-HOPE demonstrated improved biochemical graft function compared with the SCS group. Peak transaminase levels recorded on postoperative day 1 were significantly lower after D-HOPE compared with the SCS group, with AST of 1450 U/L (IQR, 1180-2210 U/L) versus 2280 U/L (IQR, 1620-3010 U/L) (P = .⁰⁰³) and ALT of 1180 U/L (IQR, 900-2002 U/L) versus 1890 U/L (IQR, 1280-2500 U/L), respectively (P = .⁰²⁴). On postoperative day 3, total bilirubin levels were lower in the D-HOPE group compared with the SCS group, at 102 µmol/L (IQR, 78-134 µmol/L) compared with 156 µmol/L (IQR, 112-198 µmol/L) (P = .⁰¹²). The INR was also reduced following D-HOPE versus SCS, showing 1.46 (IQR, 1.28-1.63) versus 1.82 (IQR, 1.48-2.15) (P = .⁰¹⁵). By postoperative day 7, total bilirubin remained significantly lower in the D-HOPE group compared with the SCS group, showing 88 µmol/L (IQR, 64-112 µmol/L) vs 132 µmol/L (IQR, 94-178 µmol/L) (P = .⁰⁰⁹), and INR continued to be reduced (1.34 [IQR, 1.22-1.53] vs 1.74 [IQR, 1.48-2.02]; P = .⁰¹¹). Clinical outcomes were likewise more favorable in the D-HOPE group (Table 2). Median intensive care unit stay was shorter after D-HOPE versus after SCS, at 3 days (IQR, 2-4 days) versus 5 days (3-9 days) (P = .⁰³⁶). Total hospital stay was also reduced (16 days [IQR, 14-26] vs 21 days [IQR, 17-35]; P = .⁰¹²). Early allograft dysfunction occurred in 15 patients (28.3%) in the D-HOPE group versus 21 patients (61.8%) in the SCS group (P = .⁰⁰⁴). Estimated glomerular filtration rate (eGFR) was <45 mL/min/1.73 m² in 7 patients (13.2%) following D-HOPE and in 13 patients (38.2%) after SCS (P = .⁰⁰⁹). The need for renal replacement therapy, predominantly among patients with marked eGFR reduction, was lower in the D-HOPE group (4 [7.5%]) than in the SCS group (6 [17.6%]), although this difference did not reach statistical significance (P = .¹⁸). A trend toward a lower incidence of acute kidney injury and reduced renal replacement therapy requirement was observed with D-HOPE. Nonspecific surgical complications occurred in 5 patients (9.4%) in the D-HOPE group compared with 11 patients (32.3%) in the SCS group (P = .⁰¹⁰). The Comprehensive Complication Index was significantly lower after D-HOPE (0 [IQR, 0-22.6]) versus after SCS (27.6 [IQR, 0-100]; P < .⁰⁰¹). Light histological examination of graft specimens did not reveal marked differences between the D-HOPE and SCS groups (Figure 6). In both groups, samples after preservation demonstrated hepatocellular dystrophy, pericellular edema, and small foci of coagulative necrosis. The type and extent of these changes were comparable between groups on standard histological assessment. In contrast, transmission electron microscopy demonstrated differences in the degree of hepatocellular ultrastructural injury. In the SCS group, mitochondrial membrane disruption, cristae fragmentation, and matrix swelling were observed. In the D-HOPE group, mitochondrial ultrastructure was better preserved: both outer and inner membranes remained intact, cristae maintained their lamellar organization, and elements of the endoplasmic reticulum were preserved (Figure 7). The incidence of late biliary complications was 11.4% in the D-HOPE group and 26.5% in the SCS group (P = .⁰⁶). In-hospital mortality was 5.9% (2 of 34) in the SCS group and 1.9% (1 of 53) in the D-HOPE group (P = .⁴²). Three-month graft survival was 94.3% in the D-HOPE group and 85.3% in the SCS group (P = .⁰⁷).

Discussion
The present study demonstrated that D-HOPE performed using a standard CPB system significantly improved early graft function and reduced the severity of IRI in recipients of ECD livers. These findings are consistent with previously published clinical studies reporting the protective effects of D-HOPE in ECD liver transplant.^1,4 In our cohort, D-HOPE was associated with a lower incidence of EAD (28.3% vs 61.8%; P = .⁰⁰⁴) and a more rapid normalization of biochemical markers of liver function, including AST, ALT, bilirubin, and INR. The protective mechanisms of D-HOPE are thought to involve restoration of mitochondrial function and attenuation of oxidative stress. Hypothermic oxygenated perfusion promotes the clearance of accumulated succinate, which is a key substrate in reactive oxygen species generation, and facilitates recovery of aerobic mitochondrial respiration with subsequent normalization of ATP synthesis.^2,10 In our study, the reduction in peak AST and ALT levels in the D-HOPE group versus SCS group (1450 U/L [IQR, 1180-2210 U/L] vs 2280 U/L [IQR, 1620-3010 U/L] and 1180 U/L [IQR, 900-2002 U/L] vs 1890 U/L [IQR, 1280-2500 U/L], respectively) reflected a less pronounced reperfusion injury and better preservation of graft metabolic integrity. Morphological findings further supported the biochemical results. Light microscopy did not show substantial differences between groups: both baseline samples and preserved samples demonstrated comparable degrees of hepatocellular dystrophy, pericellular edema, and small foci of coagulative necrosis. However, transmission electron microscopy revealed clear differences in ultrastructural preservation. After SCS, mitochondrial membrane disruption, cristae fragmentation, and matrix swelling were observed. In contrast, after D-HOPE, mitochondrial ultrastructure was largely preserved, with intact membranes, maintained lamellar cristae organization, and preserved endoplasmic reticulum structures. These observations are consistent with experimental data demonstrating improved mitochondrial integrity and reduced IRI following HOPE.^2,10 Recipients in the D-HOPE group also showed improved renal function in the early postoperative period. A decline in eGFR of <45 mL/min/1.73 m² was less frequent in the D-HOPE group compared with the SCS group (13.2% vs 38.2%; P = .⁰⁰⁹), and the need for renal replacement therapy was 7.5% and 17.6%, respectively (P = .¹⁸). Similar findings in previous studies have been interpreted as secondary effects associated with reduced hepatic IRI and a lower incidence of early graft dysfunction, potentially resulting in attenuation of the systemic inflammatory response in the early postoperative period.^10,11 Emerging evidence has also suggested that HOPE may influence the risk of hepatocellular carcinoma recurrence after liver transplant. Both clinical and experimental data have indicated that HOPE may reduce recurrence rates in high-risk patients, possibly through modulation of inflammatory signaling, stabilization of microcirculation, and suppression of hypoxia-related pathways such as HIF-1α and vascular endothelial growth factor in the early posttransplant period.^12-14 A key aspect of the present study is the practical implementation of D-HOPE using a standard CPB platform. Most previously published clinical series have relied on dedicated commercial perfusion systems (eg, Liver Assist, LifePort Liver Transporter), whereas our protocol uses equipment routinely available in transplant centers. The adapted CPB system provides perfusion control comparable to specialized devices, including independent regulation of portal and arterial circuits as well as continuous monitoring of temperature and perfusate gas composition, allowing reproducible application of D-HOPE in routine clinical practice. The findings of our study suggest that D-HOPE performed with a standard CPB system is an effective, safe, and reproducible method of dynamic preservation for ECD grafts. The technique was associated with improved early graft function, attenuation of IRI, better postoperative renal outcomes, and a potential reduction in the risk of hepatocellular carcinoma recurrence after transplant.