Normothermic Machine Perfusion of Kidney Grafts: Devices, Endpoints, and Clinical Implementation

For the past 50 years, kidney grafts have been flushed with cold preservation solution and stored at 4 °C until transplantation [1‐3]. Referred to as static cold storage (SCS), preservation at 4 °C reduces the impact of oxygen deprivation, as it decreases the metabolic rate in the tissue. Since the late 2000s, a steady shift from SCS as standard preservation mode towards machine perfusion methods has evolved. Hypothermic machine perfusion (HMP) not only cools but also continuously flushes the organ. While not conclusively demonstrated, the continuous flushing of microvessels and removal of toxic metabolites might be beneficial. HMP was first explored in the 1960s, when Folkert O. Belzer and others published the successful storage of canine and later human kidney grafts [4‐7]. The goal was to mitigate the detrimental damage occurring during retrieval, storage, and transplantation. Head-to-head comparisons with SCS at this time, however, indicated similar outcomes between the two storage modalities. Therefore, SCS was adopted in clinical routine since it was the cheaper and easier approach [8]. Hypothermic machine perfusion was revisited after 2000 and numerous studies proved the superiority of HMP with respect to delayed graft function (DGF) and graft survival [9‐11]. Currently available HMP devices differ mostly in terms of their respective operation modes (pressure vs. flow controlled). Arterial perfusion pressures are commonly set to 25–30 mmHg [1, 12, 13] and despite significantly reduced metabolism, grafts still consume oxygen at a rate of 5–10% of their consumption in in situ conditions [14‐16]. To meet this demand, oxygenated HMP (oxHMP) was introduced. Oxygenated HMP may reduce the extent of ischemic injury by restoration of adenosine triphosphate (ATP) and decrease of mitochondrial succinate accumulation during the ischemic phase, as the presence of oxygen allows for aerobic metabolism. While early concern regarding generation of reactive oxygen species (ROS) by addition of oxygen during HMP arose [14], a study underscored the beneficial effects of oxHMP in reduced biopsy-proven acute rejection (BPAR) and severe postoperative complications (Clavien–Dindo ≥ grade III), resulting in a lower overall graft failure rate in a cohort of human donation after cardiocirculatory death (DCD) kidney transplantations [17].

While HMP is a valuable tool for organ preservation, it does not allow for organ assessment. The imbalance between patients on the waitlist and the availability of donor organs is further exacerbated by the high number of organs discarded due to limitations in current technologies for accurately assessing organ quality. Kidney histology, age, gender, ethnicity, body mass index (BMI), diabetes, and hypertension show only a weak correlation with transplant outcomes, yet they remain standard criteria for determining organ suitability [18]. In the United States, the number of kidneys procured but not transplanted rose from 17.9% in 2011 to a notable high of 26.7% in 2022 [19]. Hence, methods to more accurately assess organ quality are needed [18].

Various classifications describing donor risk profiles have been published. Extended-criteria donors (ECD) comprise donors over the age of 65 with elevated terminal creatinine levels, death by cerebrovascular accident or hypertension, and/or DCD organs. Extended criteria donors (ECD) show an elevated risk for graft failure compared to standard-criteria donors (SCD) [20]. Donation after cardiovascular death organs exhibit a higher DGF and graft loss rate [21]. Other donor risk classifications used for clinical decision making include the Kidney Donor Profile Index (KDPI)/Kidney Donor Risk Index (KDRI). Both cover a wide range of donor factors and ultimately summarize the likelihood of graft failure relative to all kidneys recovered in the previous year [22, 23]. The use of marginal donor organs represents an effort to mitigate the organ shortage crisis. Achieving successful outcomes relies heavily on precise donor–recipient matching and the implementation of advanced preservation strategies. Normothermic machine perfusion (NMP) offers significant potential to address these challenges [24]. Recent research on NMP has progressed significantly, providing initial data that demonstrate its safety and efficacy [25]. This review will summarize the basic concepts of kidney NMP devices, describe the different devices currently in use, explore strategies for graft assessment, and provide an update on the clinical use of kidney NMP.

Normothermic machine perfusion devices follow the basic principle of establishing conditions that are similar to those in vivo. The conceptual framework for NMP includes a steady perfusion of the kidney with oxygenated blood or blood-like perfusate, nutrition, and metabolic control. The immediate goals are to avoid any additional injury to the kidney and to assess the function of the organ. Devices for normothermic perfusion commonly feature the following components: an oxygenator, bubble trap, filter, flow and pressure sensor, organ chamber, perfusate reservoir, perfusate pump, and a heating unit. Currently, four kidney perfusion devices are commercially available.

The Kidney Assist device (Organ Assist, Groningen, the Netherlands) is a device with a Conformité Européenne (CE) mark and has been widely used [26‐33]. Kidney Assist is pressure-steered and offers temperature ranges between 12 and 37 °C, thereby enabling hypo-, subnormo-, and normothermic perfusion (Supplementary Table 1). Hypothermic preservation is supported for 24 h, normothermic perfusion for up to 6 h. Pressure, flow, and temperature are monitored in real time and direct sampling is possible. A rotary pump applies pulsatile perfusion with 60 bpm. A hollow fiber oxygenator allows for efficient oxygen transport. The kidney is protected and humidified in a chamber, where it is easily accessible throughout. It is not suitable for transport but does offer a battery life of 20 min [34].

OrganOx Ltd. (Oxford, United Kingdom) has developed the OrganOx Metra K, which is a portable kidney NMP device designed for prolonged periods of perfusion. This machine offers automated gas inflow and during the first clinical cases it was set to maintain an arterial pO₂ of 250 mmHg for their first 24 cases and reduced to 190 mmHg in the last 12 cases [36‐38]. While an arterial pressure of 90 mmHg was initially aimed for, this was subsequently lowered to 75 mmHg after experiencing excessive renal blood flow of > 750 mL/min in the phase I study [38]. A centrifugal blood pump, continuous inline blood gas sensor, and a hollow-fiber oxygenator are part of the pressure-controlled device. Continuous recording of arterial and urinary flow, pressure, temperature, and arterial pH are carried out. The device features arterial, venous, and ureter cannulation and integrated infusion pump drivers are available [36]. Safety and feasibility have been shown in a phase I, non-randomized, single-center study as well as a correlation between ex situ biomarkers (GST-Pi delta) and post-transplant outcomes (12-month estimated glomerular filtration rate (eGFR)) [38].

EBERS Medical Technology SL (Zaragoza, Spain) developed the ARK Kidney® device and is currently performing a multicenter, prospective, and open-label clinical investigation that explores the viability, performance, and safety of kidney grafts after being subjected to their device [39]. The portable, pressure-controlled ARK Kidney® applies continuous perfusion in a closed perfusion circuit. Perfusion parameters like hemoglobin concentration, oxygen saturation, temperature, flow, pressure, renal resistance index, and urine production are tracked throughout perfusion. Using a historical control group of SCS kidneys, their chosen primary endpoint is defined as adverse events, while DGF, PNF, graft renal function, patient and graft survival, and others are considered secondary endpoints [39, 40]. Using an uncontrolled DCD graft, this device was able to preserve its viability for 73 h [40].

PerKidney®, by the Italian company Aferetica s.r.l. (San Giovanni in Persiceto, Italy), is a CE-certificated multifunctional customizable device that envisions serving as a complete treatment mode for kidney ex vivo conditioning. Their incorporated PerSorb® (CytoSorbents Inc., New Jersey, United States) filters the perfusate of inflammatory mediators, as it adsorbs hydrophobic molecules up to 55 kDa during ex situ perfusion [41]. The device has been used for HMP of porcine kidneys with subsequent NMP to mimic transplantation [42]. Progress on clinical trials and preclinical data on NMP applying PerKidney® have not yet been released.

In addition to these commercially available devices, custom-made devices are used for experimental human graft perfusion by research groups in Toronto [43], the UK [24, 44‐52], the USA [53], and Australia [54, 55]. In contrast to commercial devices, the latter mainly support normothermic application. More information on custom-made perfusion devices can be found in Supplementary Tables 2 and 3.

The immediate goal of kidney NMP is to improve transplant outcomes and increase organ utilization. This can be achieved through improved kidney preservation, kidney assessment, and/or kidney treatment (Fig. 1). Kidney NMP exhibits great potential to serve these goals, but currently, no conclusive evidence exists that this can be achieved. In view of the clinical benefits of subnormothermic and normothermic machine perfusion in the setting of liver, lung, and heart transplantation, however, it seems only a matter of time until refined protocols and technology enable broader clinical application for the kidney as well. This gaining momentum is demonstrated by numerous kidney NMP clinical trials that are currently registered (Table 1). In view of these ongoing trials, it is important to note that the establishment of suitable endpoints is paramount to identify the benefits of this novel preservation strategy. Established primary composite endpoints for clinical trials in organ transplantation include recipient death, graft failure, BPAR, and graft dysfunction. As outcomes after kidney transplantation are already excellent with conventional preservation strategies, a high number of patients need to be enrolled into clinical studies to provide the power to accurately determine a difference or the superiority of these primary endpoints [56]. In consequence, many investigators use surrogate endpoints including (but not limited to) DGF, primary non-function (PNF), eGFR, and BPAR as well as postoperative complications and duration of hospital stay, as they can be detected within a much shorter timespan than traditional hard endpoints. Not all of these, however, have meaningful translation into clinical outcomes after transplantation and a cautious approach should thus be maintained in their interpretation. Table 2 contains a list of primary and secondary endpoints chosen in past and present clinical kidney NMP trials.

With increasing interest in enhancing organ preservation through technological innovation, several research groups have focused on advancing the clinical application of kidney NMP. These efforts started in the early 2000s and progressed from large animal experiments to case studies and eventually to randomized controlled trials (RCT). In summary, the safety and logistic feasibility of kidney NMP were demonstrated; however, the superiority of this technology has, except for secondary endpoints and in anecdotal reports, not yet been proven.

A recent large RCT including 277 kidneys found comparable DGF rates between kidneys subjected to SCS plus 1‑hour NMP compared to SCS alone. NMP was found to be safe and suitable for clinical use, as no differences in DGF, occurrence of infectious complications, or any other adverse events were seen, but also no clear benefit could be detected [48]. In prior non-randomized studies, the same group was able to show a benefit of kidney NMP with regards to the incidence of PNF, patient and graft survival, eGFR, serum creatinine, graft function, and the incidence of acute rejection [24, 44, 46, 48, 49, 51]. In 2015, 50% of a UK cohort that were deemed unusable were transplanted successfully without complication during a clinical trial, indicating superiority of kidney graft utilization for this preservation and assessment method [24].

The OrganOx Ltd. group shared their preliminary results of a phase I trial already in 2023 where they compared 36 kidney grafts that were subjected to NMP for up to 24 h with a matched historical cohort [37]. Similar to the experience of Hosgood et al., they could demonstrate safety and feasibility without significant changes in graft function and outcome. However, with significantly prolonged preservation times in their study translating into a marked reduction of nighttime procedures, these results are encouraging.

The Toronto group reported on their clinical experience of 13 kidney grafts that were normothermically perfused for 1–3 h after HMP [43]. Similarly to the results in the UK, they were able to demonstrate safety and feasibility, yet no differences in postoperative graft function nor patient or graft survival between the NMP and control group were found [43]. The Rotterdam group also concluded safety and feasibility of kidney NMP, reporting comparable rates of PNF, early rejection, graft infection, and thrombosis [26]. In a more recent trial, the same group presented preliminary data comparing HMP with subsequent NMP to HMP alone. Of the 15 kidneys subjected to HMP and subsequent NMP before transplantation, six showed immediate graft function, while seven exhibited DGF and two PNF [30]. Especially in DCDs, NMP translated into beneficial outcomes compared to SCS alone [48, 49]. Most recently, a brief report stated that kidney NMP was feasible for 3 days after uncontrolled DCD [40]. The latest clinical experience with DCD kidney NMP was reported by Hameed et al. They perfused a total of 16 kidneys for at least 1 h normothermically before transplantation. No case of PNF or graft loss within 12 months occurred. Compared to the contralateral conventionally stored partner kidneys, grafts undergoing NMP displayed a lower incidence of DGF (23.5% versus 64.7%) while exhibiting similar eGFR slopes over time [57]. A first clinical series of short pre-implantation controlled oxygenated rewarming (COR) of six ECD kidney grafts that were compared to matched cold-stored controls demonstrated higher creatinine clearance (CrCl) and normalized fractional excretion of sodium (FENa) by postoperative day 7 [27].

No clear rules for evaluating the viability and functionality of a kidney graft during NMP have been established. The quality assessment score (QAS) published in 2015 [24] suggests evaluating macroscopic appearance, renal blood flow, and total urine output for kidney quality assessment. When applied in a cohort of primarily from transplantation excluded kidneys undergoing NMP, 81% of the kidneys were deemed suitable for transplantation [24]. However, passing the QAS does not guarantee graft function, as highlighted by PNF of two kidney transplants from a young uncontrolled DCD donor [50]. Further to the criteria set forth in the QAS, perfusion parameters including renal blood flow (RBF), intrarenal resistance (IRR), urine output, FENa, acid–base homeostasis, oxygen consumption, injury markers, histology, and imaging have been considered as assessment tools for kidneys during NMP [35, 58].

In a preclinical study, IRR, acid–base balance, and lactate clearance were found to be predictive for renal function after transplantation [59]. Another group reported a correlation of oxygen consumption, renal blood flow, and lactate with the severity of kidney injury during porcine NMP [60]. In addition, the dynamics of RBF, oxygen consumption, perfusate pH, perfusate oxygen and carbon dioxide partial pressures, bicarbonate, lactate, and potassium may yield valuable information [46, 48]. The protocol of an RCT in the Netherlands enforced perfusion dynamics, microcirculation, oxygenation, damage markers, perfusate sodium/urine sodium ratio, oxygen consumption, urine production, amount and characteristics of extracellular vesicles, and the QAS for assessment of kidney quality [32]. Additionally, most protocols include histopathological assessments to describe the renal constitution before, during, and after NMP. Hematoxylin and eosin (H&E) and periodic acid–Schiff staining (PAS) have most frequently been described in the literature [35, 62]. Others reportedly include endothelial nuclei and fibrin staining for more specified information [63].

Of note, perfusate composition does, irrespective of organ function, influence perfusion dynamics. It was shown, for instance, that increasing blood hemoglobin levels of up to 5 mmol/L were correlated with better CrCl and FENa and levels of < 4.5 mmol/L were associated with worse overall kidney function [61].

Feasibility without apparent deleterious effects on kidney or recipient after NMP was already established in 2011 [44]. The crux of the matter lies within the definition—a benefit can be gained, if an unmet need and defined endpoint criteria are met. Conceptually, patient and graft survival rates in kidney transplantation are high. Three prominent limitations in the field are DGF, late graft loss, and high organ discard rates; hence, endpoints need to relate to any of these and target outcomes respected by the Food and Drug Administration (FDA) and European Medicines Agency (EMA). It is expected that viability and functional assessment during NMP yield information with a stronger predictive value compared to conventional scores like KDRI or cold ischemia time [27]. The current reality and existing evidence do not convincingly demonstrate the benefit of kidney NMP. The feasibility and safety have been attested to in various RCTs and case reports applying NMP before transplantation. NMP durations of up to 24 h have been reported for human kidney grafts prior to transplantation [26‐28, 30‐32, 37, 44‐49, 51]. Even though comparable rates of DGF, PNF, and eGRF to control cohorts have been seen in phase I/II clinical trials, the safety and feasibility of this technology has been demonstrated. In addition, NMP has been shown to increase organ utilization by allowing for pre-transplant assessment, although evidence for this is limited to case series and non-randomized trials. And finally, in a recent trial by OrganOx Ltd., the first evidence of prolongation of total preservation times using NMP technology exists, potentially impacting clinical practice as priorly seen in liver transplantation [37]. Preclinical data suggest that prolonged kidney preservation through NMP might be feasible within the next years: 73 h of NMP have been reported—the longest experimentally perfused kidney grafts to date [98]. Multiple benefits of extended kidney NMP exist: organ storage under preserved metabolism, improved reconditioning, use as a platform for novel therapeutics, and potentially organ regeneration can be facilitated with the application of this technology [99, 100].

The most optimal perfusion strategy and perfusate composition for kidney NMP have not been explicitly defined to date. Linear or pulsatile perfusion, gradual or immediate rewarming, gradual pressure increase or fixed pressure profiles and different preservation solutions have been explored. While NMP is established in other organs, kidney NMP has not yet been implemented in clinical routine. In view of the potential benefits of prolonged perfusion, a device with automated flow, gas supply, and nutrition might relieve logistic and operational hurdles and enable prolonged kidney perfusion. Adding to the known existing prototypes by Organ Assist, OrganOx Ltd., EBERS Medical Technology S.L. and Aferetica s.r.l., a modular and configurable automated NMP system by a group in Aachen, Germany was described [101]. Further preclinical evidence to fully understand all aspects of kidney NMP and carefully designed clinical trials are needed to establish NMP as a meaningful tool in kidney transplantation.