Viability assessment of the liver during ex-situ machine perfusion prior to transplantation

Purpose of review In an attempt to reduce waiting list mortality in liver transplantation, less-than-ideal quality donor livers from extended criteria donors are increasingly accepted. Predicting the outcome of these organs remains a challenge. Machine perfusion provides the unique possibility to assess donor liver viability pretransplantation and predict postreperfusion organ function. Recent findings Assessing liver viability during hypothermic machine perfusion remains challenging, as the liver is not metabolically active. Nevertheless, the levels of flavin mononucleotide, transaminases, lactate dehydrogenase, glucose and pH in the perfusate have proven to be predictors of liver viability. During normothermic machine perfusion, the liver is metabolically active and in addition to the perfusate levels of pH, transaminases, glucose and lactate, the production of bile is a crucial criterion for hepatocyte viability. Cholangiocyte viability can be determined by analyzing bile composition. The differences between perfusate and bile levels of pH, bicarbonate and glucose are good predictors of freedom from ischemic cholangiopathy. Summary Although consensus is lacking regarding precise cut-off values during machine perfusion, there is general consensus on the importance of evaluating both hepatocyte and cholangiocyte compartments. The challenge is to reach consensus for increased organ utilization, while at the same time pushing the boundaries by expanding the possibilities for viability testing.


INTRODUCTION
Liver transplantation is a lifesaving treatment for patients with end-stage liver disease, acute liver failure and primary liver cancer.Unfortunately, with increasing rates of liver disease, the number of transplantable organs available is unable to meet demand, and mortality on the waiting list is high [1].The shortage of donor livers, in combination with altering organ donor demographics, such as ageing population and increased obesity, has led to the increased utilization of livers from extended criteria donors (ECD) of suboptimal quality [2,3].ECD livers provide an increased risk of developing early allograft dysfunction (EAD), primary nonfunction (PNF), or posttransplant ischemic cholangiopathy [4][5][6][7][8][9][10][11].Predicting the postreperfusion viability of ECD livers prior to transplantation is one of the most challenging calls to make in organ transplantation.For recipient safety reasons, high-risk donor livers are, therefore, often declined for transplantation [12].
Ex-situ machine perfusion now provides a unique opportunity to assess ECD liver function before transplantation.Viability assessment has mainly been described during normothermic machine perfusion (NMP).However, there are new approaches for assessment during hypothermic machine perfusion (HMP).To add value to known donor and recipient risk factors for worse posttransplant outcome, reliable criteria to determine donor liver viability and predict save use need to be established.This article reviews currently used and promising new viability criteria during ex-situ liver machine perfusion.

VIABILITY ASSESSMENT
For long-term graft survival, both liver synthetic and detoxification functions and biliary integrity have to be preserved.Therefore, viability criteria have been chosen to test the hepatocyte and cholangiocyte compartment.Most viability criteria can only be assessed during NMP, but some tests are also possible during HMP (Fig. 1).Table 1 provides an overview of the viability criteria currently used in clinical practice, which are further discussed below.

VIABILITY ASSESSMENT DURING HYPOTHERMIC MACHINE PERFUSION
At 4-10 8C, metabolic activity is significantly reduced to 6-15% of normal, posing challenges for real-time assessment of metabolism and function [13].Nevertheless, several perfusate analyses can be performed at 7-10 8C, at which HMP is typically performed.

Flavin mononucleotide
Flavin mononucleotide (FMN) can be assessed during both NMP and HMP.FMN, typically bound to mitochondrial complex I under physiological

KEY POINTS
Ex-situ normothermic machine perfusion is increasingly used for viability assessment prior to transplantation.
During hypothermic machine perfusion the mitochondrial injury marker FMN has been proposed a general viability marker.
During normothermic machine perfusion both the hepatocyte and cholangiocyte compartment can be assessed.
If both cellular compartments are successfully tested as viable, a low incidence of primary non-function and ischemic cholangiopathy can be expected, increasing extended criteria donor utilization.conditions, is released from damaged complex I during ischemia and subsequent reoxygenation.The Zurich group analyzed FMN in 54 livers during oxygenated-HMP and demonstrated a strong correlation between FMN release, coagulation factors levels and peak transaminases post-transplantation, and showed predictive value for EAD [14].In 2020, the group form Zurich measured FMN in the perfusate of 50 livers, with elevated concentrations of FMN in the perfusate correlating with a high rate of graft loss [15].The authors advise against transplantation if FMN concentration exceeds 8000AU after 30 minutes of oxygenated-HMP.In their most recent work, which includes all perfusions from previous work, they suggest a FMN threshold of 6000AU and a NADH threshold of 8000AU after 30 minutes of perfusion [16].Confirmation of these single center observations in other centers is eagerly awaited.

Perfusate analysis
Patrono et al. [17] have analyzed perfusate in 50 DBD livers during oxygenated-HMP and assessed levels of aspartate aminotransferase (AST), alanine transaminase (ALT), lactate dehydrogenase (LDH), glucose, lactate and pH at 90 min of perfusion in relation to postoperative outcomes [17].All parameters, except lactate, correlated with EAD.However, macrovesicular steatosis was the only independent predictor of EAD, as all perfusate parameters were closely correlated to the severity of steatosis.

VIABILITY ASSESSMENT DURING NORMOTHERMIC MACHINE PERFUSION: HEPATOCYTE VIABILITY
NMP is typically performed at 35-37 8C, enabling assessment of a metabolically active liver.The criteria for hepatocyte viability are based on perfusate analysis, perfusion parameters and bile production.

Lactate levels
During NMP, three phases in lactate dynamics can be typically identified; initially increasing, peaking at 1 h, followed by a rapid clearance within 2 h, ending in a stable state characterized by consistently low lactate levels during the remainder of the perfusion [18,19].Nondecreasing lactate levels in the perfusate are generally considered a sign of poor graft function.
As shown in ].

Glucose levels
The liver is a key metabolic organ with a major role in glucose metabolism, including glycogenolysis, gluconeogenesis and glycolysis [38].Glycogenolysis, an ATP-independent process, continues during static cold storage (SCS) driven by the lack of ATP, which is demonstrated by increasing glucose levels at the beginning of NMP.In viable livers, high levels of glucose should prevent glycogenolysis and trigger glycogenesis, and as a result, glucose levels will decrease.Normal levels of glucose during NMP can reflect minimal ischemia, but it can also point to glycogen exhaustion or extensive liver injury [35].A glucose challenge test can rule out liver injury: a viable liver metabolizes glucose and levels in the perfusate will drop [25,39,40].Only Watson et al. [35] included a glucose challenge in their viability criteria.Both Reiling et al. [30] and Mergental et al. [25] included a decreasing glucose trend.

Acid-base homeostasis
The liver is an important regulator of acid-base homeostasis.

Perfusion parameters
In liver perfusion, flow in the hepatic artery and portal vein is determined by perfusion pressure and vascular resistance.Prolonged ischemia can damage the microcirculation of the liver, leading to increased vascular resistance and reduced perfusion at fixed pressure, resulting in impaired function [44,45].Steatotic livers, with narrower sinusoids, exhibit lower flow, leading to secondary hypoxia and reperfusion injury [46].Elevated vascular resistance during machine perfusion indicates liver injury and later liver dysfunction [45].Whereas many groups suggest necessity for stable hepatic artery and portal vein flows [20,28,30,31], some added specific target flows [25][26][27]29], all livers reached the target flows.Quintini et al. [29] further specified target flow per unit liver weight, but even here, all livers met the target flows.This indicates that currently flow measurements alone cannot distinguish between viable and nonviable livers.

Bile production
Bile production is considered one of the higher levels of liver functions, as it requires considerable ATP content.Bile is produced by hepatocytes, and cholangiocytes lining the bile duct alter the composition of bile.Therefore, bile production is associated with hepatocyte viability, whereas bile composition is associated with cholangiocyte viability.All studies in Table 1 have listed production of bile as a viability marker [21,22,25,[29][30][31]33,35,36], demonstrating consensus on the negative implications of absence of bile production.Only two groups have stated a specific cut-off value for bile production.Quintini et al. [29] defined a production of more than 2 ml/h as a minimum, Van Leeuwen et al.
[33] defined a production of more than 10 ml after 2.5 h of NMP, of which at least 4 ml in the last hour.However, graft loss has been described despite proper bile production, and successful transplantations have been reported without proper bile production [35,36,39,47].As an example, Zhang et al. [36] transplanted four livers that did not meet the Groningen group criteria, yet all showed immediate function.It should be noted that malposition of the biliary drain may lead to false-negative absence of bile production.

VIABILITY ASSESSMENT DURING NORMOTHERMIC MACHINE PERFUSION: CHOLANGIOCYTE VIABILITY
As cholangiocytes play a crucial role in modifying bile composition by re-absorbing solutes and secreting water and bicarbonate, the viability of the cholangiocytes can be assessed by analyzing the composition of bile.

Bile composition
Low levels of biliary glucose, high levels of bicarbonate and an alkalotic pH are indications of viable cholangiocytes [24,48].A biliary pH of more than 7.45 was defined by Cardini et al. [22] and Van Leeuwen et al. [33].In the cohort of Watson et al.
[35], 3 out of 16 transplanted livers were unable to achieve a biliary pH greater than 7.40, and developed ischemic cholangiopathy.In addition to low pH, these livers also had similar glucose levels in bile and perfusate, and a low biliary bicarbonate.Therefore, Watson et al. [35] stated that the difference between bile and perfusate glucose should at least be 10 mmol/l, and Van Leeuwen et al. [33] suggested a delta of at least 5 mmol/l.In addition to the biliary pH and the delta glucose, Van Leeuwen et al. [33] also included a delta pH and a delta bicarbonate.Importantly, the VITTAL trial showed that meeting hepatocellular viability criteria is not sufficient to predict ischemic biliary damage [25].All transplanted livers met the hepatocyte criteria, but nevertheless, 45% of the recipients developed irregularities in the bile ducts, with 18% requiring retransplantation for ischemic cholangiopathy, suggesting nonviable cholangiocytes and bile ducts, despite viable liver parenchyma.

PROMISING VIABILITY CRITERIA
In addition, clinically implemented viability criteria, new criteria, such as coagulation factors and imaging techniques have been studied in preclinical setting.

Coagulation factors
The liver is crucial for regulating coagulation and fibrinolysis [49].NMP circuits are heparinized, yet production of coagulation factors can serve as a viability marker.The long-term NMP study of Eshmuminov et al. [40] revealed significantly higher coagulation factor-V levels in the perfusate of functioning versus nonfunctioning livers at 48 h of perfusion, although this difference disappeared thereafter.Weissenbacher et al. [50] showed significantly higher levels of von Willebrand factor antigen in the perfusate of EAD livers.Van

Imaging techniques
Orita et al. performed a 18F-fluorodeoxyglucose (FDG)-PET/CT to assess viability in two discarded human livers after 1 week of NMP [52].With FDG-PET/CT, the glucose metabolism can be followed: glucose is taken up by the GLUT-2-transporter, metabolized into FDG-6P and dephosphorylated back to FDG, which then leaves the cells and returns to the circulation [53].Homogenous FDG uptake was observed in all livers and therefore indicated intact transport, metabolism, and excretion of glucose.Fodor et al. [54] utilized hyperspectral imaging (HIS) to analyze the liver parenchyma, monitoring oxygen saturation levels (StO 2 ), tissue hemoglobin index (THI), near-infrared perfusion (NIR), and tissue water indices (TWI).HIS was performed during SCS and at 1, 6, 12 h, and end NMP.During NMP, the StO 2 , THI and NIR perfusion indices significantly increased, whereas the TWI drastically decreased.A significantly higher THI was seen in discarded livers compared with transplanted livers end NMP.Kneifel  ] analyzed the indocyanine green (ICG) plasma-disappearance-rate (PDR) and the ICG-perfusion with a near-infrared camera during 7 days of NMP.The ICG-PDR was significantly higher on day 0 in grafts that survived at least 7 days of NMP.ICG perfusion, the distribution of ICG in the liver during NMP, was significantly different at day 0 between long-surviving and short-surviving grafts.

MicroRNA
Matton et al. [57] investigated microRNAs as sensitive, specific, and stable markers for cell function, stress, and injury during NMP.Elevated miRNA-122 was significantly associated with liver injury, although there was no threshold value identified for irreparable damage [58][59][60].

Liver maximum capacity test
Our group employed the liver maximum capacity (LiMAx) test to assess liver function during NMP [61].This is a clinically validated substrate-based cytochromal breath test, which measures 13 CO 2 production.The advantage is that all livers can be subjected to an equivalent dose, making direct comparison between results possible.Schurink et al. demonstrated a significant correlation between LiMAx values and lactate clearance, and significant inverse correlation between LiMAx values and histological injury.However, no correlation was found between LiMAx values and microRNA-122 or FMN values in the perfusate.

DISCUSSION AND FUTURE PERSPECTIVES
In this review, we have summarized the emerging consensus on parameters that indicate liver viability during NMP.Viability assessment during HMP could prove to be an important contribution to the field.During NMP lactate clearance, pH stabilization and bile production are uniformly considered as biomarkers of adequate liver function prior to acceptance.There is also growing agreement that besides hepatocellular criteria, cholangiocyte criteria should be considered to prevent biliary ischemic complications post-transplant.In both domains, no consensus was found on specific cut-off values, which might be hampered by considerable differences in perfusion protocols, acceptance criteria and recipient populations.In most viability assessment studies, livers were transplanted into recipients with relatively low Model for End-Stage Liver Disease (MELD) scores.This raises the question if these tested ECD livers are equally suitable for recipients that require a retransplantation and recipients with high MELD scores, as these patients are more at risk for complications and graft loss.So far, recipient factors have not been combined with viability assessment, opposed to the risk scores that were developed recently to balance donor and recipient risks (e.g.UK-DCD score, Balance of Risk score, ET-DRI score).
Considering the differences in length of the viability assessment protocol, the 2.5 h decision moment described by Van Leeuwen et al. [33] is swift but proved to be a safe decision moment, without any PNF in their cohort.In their series, about two of three of all tested livers were ultimately transplanted.The short assessment period might, however, be too strict for slowly recovering livers, discarding potentially viable organs meeting the criteria shortly after the decision moment.Other protocols take up to 6 h to reach a decision moment, which might yield a higher ECD liver utilization rate.This indicates that there is potential to further explore the lower limits of organ viability and even push the boundaries more.In addition, organ perfusion and viability assessment can lead to a better understanding of essential hepatobiliary injury and repair processes, opening doors for regenerative medicine applications and potentially repair currently declined organs.

CONCLUSION
In conclusion, current viability assessment protocols during ex-situ liver perfusion enable well tolerated use of ECD livers.Efforts should be taken to reach consensus on requirements for acceptance to increase wider application of these organs and ultimately reduce waitlist mortality.

Table 1 .
Viability criteria used in the clinical practiceLiver viability assessment during machine perfusion Groen et al.
Table 1, all clinical studies incorporated perfusate lactate clearance [20-36].It is noteworthy that cut-off values range between 1.7 and 4.5 mmol/l, and lactate is measured at different time points during NMP.Watson et al. [35] integrated liver weight in the equation for lactate clearance, leading to a more accurate estimation of clearance per gram liver tissue.Seidita et al. [31], Cardini et al. [22] and Bral et al. [21] state that lactate should decrease over time, without further specification.In addition to the varying cutoff values, controversy exists on whether lactate clearance is an accurate marker of graft viability.The Zurich group showed lactate clearance up to 24 h of NMP in four explant cirrhotic livers, questioning the predictive value of lactate clearance in early assessment [37 et al. also used HIS during NMP at 1, 2 and 4 h of perfusion and showed that StO 2 and THI predicted lactate values at 1 and 2 h of NMP, THI also predicted lactate values at 4 h [55