Solid-organ transplantation has been successfully established since the first transplantation of a human kidney in 1954 [1]. Since then, heart, lung, liver and kidney transplantation has become the only sustainable therapy for patients with end-stage organ diseases [2]. Further, transplantation of vascularized composite allografts (VCAs), including upper and lower extremities, face, abdominal wall, uterus and penile grafts, has become an available option to recover the quality of life for patients [3, 4]. However, due to increasing demands for allografts, extension of transplant indications, and increasing prevalence of organ diseases in the general population, we face a global shortage of transplantable grafts. At the beginning of 2025, 13,570 patients were awaiting an organ transplantation in the Eurotransplant-area [5], and more than 100,000 in the US [6]. Therefore, strategies are urgently needed to increase utilization of available grafts safely. To address this gap, transplant surgeons have already started extending acceptance criteria by using marginal grafts, including those from aged donors and grafts that were donated after circulatory arrest (DCD). Because these grafts are associated with higher risk of non-anastomotic complications when they are transplanted without being evaluated or resuscitated on a perfusion device, researchers have developed dynamic preservation platforms that allow for graft assessment and reconditioning prior to transplantation [7–9].

Current perfusion concepts all share the core function of supplying oxygen to the donor graft while outside of the body, but differ in terms of temperature, perfusate composition, and time of perfusion [10]. Three core concepts of dynamic preservation have gained increasing acceptance across most allografts: ex situ perfusion in normothermic (35 °C–38 °C) versus hypothermic (4 °C–12 °C) conditions, and in situ normothermic (35 °C–38 °C) regional perfusion (NRP) [11, 12].

Normothermic machine perfusion (NMP) is generally applied either immediately after procurement and during transport, which limits the period allografts are exposed to cold ischemia, or after an initial phase of cold storage, referred to as « back-to-base», at the transplant center (end-ischemic NMP) [10]. NMP typically utilizes an oxygenated blood-based perfusate, either through blood products or donor whole blood, which is pumped through afferent vessels of the allografts. The high demands of metabolically functional allografts necessitate sophisticated perfusion machines that closely mimic the physiological environment in situ [13]. Acellular perfusates with synthetic oxygen carriers have also been used for NMP to avoid demand for blood products and overcome issues related to hemolysis, but these approaches remain the subject of preclinical research [14, 15]. NMP has already been leveraged to perform graft assessment, evaluating parameters, such as bile production in livers [16], weight change in VCAs [17], monitoring oxygenation in lungs [18] or coronary flow and resistance in hearts [19]. While such objective assessment has increased utilization of marginal grafts [20], consensus on biomarkers and their reliability has yet to be established.

In an alternative approach, hypothermic oxygenated perfusion (HOPE) mainly relies on chilled synthetic preservation solutions, typically containing osmotic agents, electrolytes, buffering substances, metabolic substrates, and antioxidants or free radical scavengers [21, 22], that are pumped through afferent vessels at 4 °C–8 °C [23]. In livers, perfusate may further be administered through the portal vein only (HOPE) or both, the portal vein and the hepatic artery (dual or D-HOPE), with similar long-term outcomes [24]. Hypothermic conditions markedly decrease metabolic demands, allowing allografts to rely only on dissolved oxygen in the perfusate without the need for oxygen carriers in the perfusate. Therefore, hypothermic perfusion machines are inherently simpler and less costly than their normothermic counterparts [25]. Core functions include perfusate cooling and oxygenation as well as maintaining a steady perfusate flow. The main advantage of HOPE—typically applied in an end-ischemic setting—lies in restoring aerobic metabolism at lower temperatures, thus reducing toxic metabolite accumulation (i.e., succinate, NADH) and subsequent reactive oxygen species production, with the aim of dampening ischemia–reperfusion injury (IRI) [26–28]. While HOPE, like NMP, allows for the measurement of biomarkers to predict allograft viability [29–31], assessment of allograft function during preservation remains limited in clinical practice.

NRP is initiated immediately after circulatory death with the intention of avoiding prolonged ischemia in the organ procurement process prior to reperfusion—a concept particularly appealing in donation after cardiac death (DCD) and standard practice in several countries, including Italy, Spain, France, and parts of the United Kingdom [32]. Both the abdominal compartment alone and thoraco-abdominal organs combined may be included in NRP circuits [33, 34]. NRP includes clamping either the thoracic aorta or ligating the cerebral vessels to prevent cerebral blood flow [33]. Common extracorporeal membrane oxygenation (ECMO) devices are employed to generate artificial blood flow [35]. This approach offers the earliest opportunity to assess allograft viability, though reflecting a multi-organ environment [32]. While promising from an allograft utilization perspective, this technique still faces ethical and legal barriers related to the dead donor rule in many countries [36–38]. Hence, NRP will not be discussed in detail in this article.

In the preclinical setting, ex situ perfusion of multiple days was first introduced for livers in 2020 in Zurich [13], demonstrating the feasibility of long-term (i.e., >24 h) preservation. Using a custom-built NMP device, explanted porcine and discarded human livers were preserved for up to 10 days [13]. In contrast to previously known NMP systems, the Wyss–Zurich device allowed for prolonged ex situ perfusion in near physiologic conditions and a functional state [13]. Thanks to automation with feed-back controllers, perfusion parameters could be automatically controlled in a tight physiological range, limiting on-site interaction to a minimum. This approach was subsequently validated in the first-in-human application on compassionate use basis, preserving a liver initially declined for transplantation for 3 days followed by successful implantation [116]. The same platform was later adapted to the needs of resected partial livers, which could be used to preserve partial human livers for a week, showing normal tissue integrity and hepatic function [118]. Besides offering a platform for profound viability assessment and organ function, such advancements have since enabled ex situ treatment including pharmacological defatting of steatotic grafts [119–121], building on previous efforts in short-term preservation [122, 123]. The first RCT is currently ongoing and data is expected to be available soon (ISRCTN14957538).

Other groups focused on adapting currently available NMP devices to meet the requirements of multi-day perfusion [124–127], i.e., nutrition, precise control of acid-base balance and blood gases, glucose control, as well as dialysis [128–131]. Using and modifying commercially available devices comes with the obvious advantage of easy accessibility and reproducibility [132]. The limitation to this approach is the need for constant human intervention. The longest preservation time with this strategy so far was reported by the Italian group, which successfully preserved a declined human liver over 17 days by incorporating an extracorporeal blood purification system into their NMP device [127]. Indeed, hemodialysis was shown to further improve perfusate quality in multiple organs for long-term perfusions [133]. Further, successful prevention of microbial contamination [134, 135] and hemolysis [13] were found to be crucial for long-term perfusions. Besides mere preservation of viability, the Australia group pioneered the ability to perform liver split procedures of 10 whole livers without interrupting perfusion [126, 136]. Their work is based on the seminal work of performing split procedures during short-term perfusion [137–140] and marks a milestone in ex situ liver research. Establishing such long-term perfusion liver models has the immense potential to revolutionize the study of liver injury, repair and regeneration [141–143].

While ex situ heart perfusion has not been performed for longer than 24 h to date, an increasing amount of case reports that document prolonged ex situ perfusions that enabled long distance transport of allografts [144–146]. Notably, a heart was successfully transported across the Atlantic ocean while being perfused ex situ for 16 h, illustrating that NMP can even enable world-wide organ sharing [145]. Collectively, the clinical literature agrees on three consistent conclusions: i. NMP is safe and non-inferior to SCS for standard donors; ii. NMP enables reliable functional assessment that can rescue marginal or DCD hearts; and iii. scaling DCD programs with perfusion platforms can substantially expand transplant activity, especially with prolonged perfusion durations that enable longer transport thus wider organ sharing.

Extending the interval between donor lung procurement and implantation has several important clinical and logistical implications. Prolonged preservation facilitates broader donor-recipient matching across larger geographic regions and allows transplant centers to avoid nighttime surgery, which is associated with increased complication rates and inferior outcomes [147, 148]. More importantly, lengthening EVLP duration transforms the platform from a short-term assessment tool into a therapeutic environment in which injured or initially discarded lungs can be actively rehabilitated [149].

The Toronto lung transplantation program first reported long-term perfusion in porcine lungs and human discarded donor lungs only years after the advent of EVLP [150]. Ever since, the group has extensively investigated prolonged EVLP in porcine models, demonstrating that continuous EVLP for 12–24 h is feasible [151, 152]. Similarly, other institutions report successful extension of porcine EVLP durations for up to 24 h [153–156], the Hannover program transplanting the extended EVLP lungs into healthy porcine recipients with subsequent short periods of graft evaluation [157]. In Minnesota, the 24 h prolonged evaluation was extended to human discarded donor grafts [158]. A maximum preservation time of 3 days has been reported in porcine lungs using a protocol of two short (4 h) normothermic EVLP cycles alternating with cold storage at 10 °C [159].

Despite strong interest in the topic and several pre-clinical reports of long-term EVLP for up to 24 h or more, clinical evidence remains largely limited to case reports with the longest documented clinical normothermic continuous EVLPs being 11.25 h and 15.5 h, respectively [160, 161]. However, a recent report from the Netherlands group describes the first clinical experience with n-EVLP–HOPE, using hypothermic oxygenated machine perfusion (HOPE) after a period of normothermic EVLP (n-EVLP) in a small cohort of human lung transplantation patients [162]. Grafts from the n-EVLP–HOPE group did not differ significantly in early post-transplantation outcomes compared with the control group.

There are ongoing controversial discussions about what parameters are needed for long-term EVLP, such as what temperature to use, perfusate composition, whether the perfusate should be exchanged in the EVLP and in what intervals. However, despite those advances, additional comprehensive research is needed to advance clinical implementation.

Use of machine perfusion has been reported in vivo to salvage free flaps after thrombosis of vascular anastomosis. Wolff et al. first reported manual rhythmic perfusion of 3 fibula flaps with heparinized red blood cells at 38 °C for 10–12 days, resulting in flap survival and neovascularization [163]. The same group reported using a closed-loop, low-flow circuit (Novalung MiniLung) at 37 °C, to perfuse thin anterolateral thigh flaps and radial forearm flaps for 4–6 days in vivo in five high-risk patients. Stable coverage was achieved in four of five cases. One subtotal flap was lost due to infection, two cases due to complete epithelial loss, and one case featured a venous congestion [164]. Since then, there has been a decade of incremental experimental progress, but NMP of VCAs beyond 24 h remains uncommon [114, 165–167], with most experiments failing by 30 h of perfusion [15, 166, 168]. The longest reported perfusion durations were 72 h for human upper extremities [169] and 144 h for human fasciocutaneous flaps [170].

Similar to liver transplantation, prolonged NMP promises safe inclusion of additional kidney grafts for transplantation. Therefore, protocols and perfusion devices have been refined to extend preservation times beyond 1 day. Notably, the first clinical report was recently published by Dumbill et al. who perfused kidneys for up to 24 h before transplantation and showed correlation between NMP biomarkers and 12-month graft function [171]. Discarded human kidneys could be further perfused for 48 h, while maintaining urine excretion [172]. The longest ex situ kidney preservation was reported by de Haan et al. who perfused human kidneys for 4 days at sub-normothermic conditions, maintaining a metabolically active state [173].

Despite recent advances, substantial differences persist in achievable preservation periods across graft types. While organ-specific factors contribute to these variations, maintaining perfusate quality represents a fundamental limitation that constrains perfusion duration across all graft types. The longest perfusion durations were currently achieved for liver grafts, ranging up to 2 weeks [119, 127, 132]. This achievement reflects the liver’s unique capacity to actively maintain perfusate quality through its inherent metabolic and detoxification functions. While soluble waste products and toxins can be easily removed with hemodialysis [133], many metabolic byproducts rely on hepatic clearance. Thus, only perfusate exchange can effectively mitigate waste and toxin accumulation for non-liver grafts.

Maintenance of adequate oxygen delivery presents another critical challenge related to perfusate quality. Erythrocyte supplementation has proven essential for NMP of kidneys, VCAs, livers, and hearts [39, 49, 66, 114]. Even in lung perfusion, where acellular Steen solution is routinely utilized in some EVLP protocols, erythrocyte supplementation has demonstrated improved tissue preservation through enhanced oxygen transport and reduced reactive oxygen species generation [86, 157, 174]. However, hemolysis resulting from shear forces in tubing and pumps, combined with suboptimal environmental parameters [175], necessitates ongoing erythrocyte supplementation to maintain sufficient hematocrit levels.

Beyond waste products and oxygen carriers, perfusate further carries a variety of signaling molecules. Suboptimal management of IRI, inflammatory activation triggered by unphysiological environmental parameters, and endothelial injury induced by oncotic pressure fluctuations, supraphysiological shear forces, and IRI itself collectively limit perfusion duration and initiate release of pro-inflammatory molecules [176, 177]. Although cytokine filters have been employed clinically in kidney, liver, and lung, as well as experimental heart perfusion [178–181], extended perfusion periods require comprehensive pharmacological interventions to minimize IRI and attenuate cellular damage responses.

While donor characteristics, such as age or cause of death, as well as procurement parameters, including warm ischemia time, are routinely collected it is challenging to base transplant decisions on these parameters alone [182]. Therefore, machine perfusion has been increasingly used as a platform to perform rigorous graft assessment. As there is no consensus on assessment parameters, we summarized a collection of assessment parameters across different graft types, categorized based on invasiveness and evaluation time [113] (Table 2). Importantly, metabolic function, and associated biomarkers are temperature dependent, leading to higher values for NMP compared to HOPE. For example, the prominent biomarker for mitochondrial injury, flavin mononucleotide (FMN), was established for HOPE in liver transplantation [29–31, 218], but requires different thresholds during NMP [30, 219]. Importantly, many assessment parameters may not be diagnostically conclusive during the first hours of reperfusion as the graft is exposed to IRI and undergoes a temporary reperfusion phase [117]. Hence, all assessment parameters must be evaluated in the context of time, which is currently not standard practice [113]. Another major limitation of current assessment strategies is the lack of normalization for perfusate-borne biomarkers, which should be normalized by perfusate volume and graft size. Given this, transplant communities ideally transition to systematic reporting that allows for comprehensive data collection and identification of indicative assessment parameters with corresponding thresholds, which can be later implemented in national guidelines. Given these, NMP is a platform that substantially improves objective graft assessment and has the potential to safely allow for transplantation of additional grafts.