Predicting the viability of liver allografts procured from non–heart-beating donors: The role of histopathology^†

Successful liver transplantation is dependent on many factors, one of which is the quality of the donor organ. Until recently, non–heart-beating donors (NHBD) were not considered suitable for liver transplantation because of the higher risk of graft nonfunction due to severe ischemia/reperfusion injury and late complications related to intrahepatic biliary strictures. However, an increasing shortage of donor livers has led to efforts to extend the donor pool by the use of extended criteria donors, including NHBD livers.1-3

Abbreviation

NHBD, non–heart-beating donor.

Human NHBDs are patients with brain injury incompatible with recovery whose condition does not meet formal criteria for brainstem death and whose cardiopulmonary function ceases before organs are retrieved.4 It includes 2 groups: controlled and uncontrolled donors. Controlled donation takes place when the death occurs within an intensive care unit/hospital setting and there is planned withdrawal of therapy by the patient's medical team. Uncontrolled donation takes place when death occurs outside the hospital or within the emergency room and resuscitation is continuing and unpredictable. Under either scenario, cessation of blood supply before organ retrieval causes liver tissue damage (warm ischemic injury), the severity of which depends primarily on the length of the interval between circulation cession and organ recovery. The warm ischemic injury suffered before and during organ recovery, plus additional injury occurring during cold preservation (cold ischemia), often causes significant tissue damage after reperfusion (ischemia/reperfusion injury), and this, in severe cases, can lead to rapid posttransplant organ failure.

The major risk of NHBD liver grafts is severe ischemia/reperfusion injury.1-4 By definition, ischemia/reperfusion injury includes the culmination of donor organ damage that occurs during warm ischemia, cold ischemia, and reperfusion. Warm ischemia refers to suboptimal perfusion with blood when the organ is at body temperature, which can occur before or during harvesting and during rewarming after implantation (warm ischemia preferentially targets hepatocytes and endothelial cells for injury).5, 6 NHBD livers suffer the most severe form of warm ischemic insult; that is, the unpredictability of cessation of circulation following withdrawal of life support results in various degrees of suboptimal perfusion and poor oxygenation in a normothermic setting, activation of the coagulation cascade, and release of vasoconstrictive mediators in the terminal phases of life. Cold ischemia refers to the tissue injury that occurs when the donor organ is stored in preservation fluid and immersed in an ice bath;7 it preferentially damages sinusoidal endothelial cells (electron microscopy examination is needed to reliably assess the extent of injury). Loss of sinusoidal microvascular integrity and function and suboptimal microvascular blood flow after revascularization are the major determinants of subsequent graft viability and function. Reperfusion injury refers to tissue damage that occurs during reperfusion of the liver with recipient blood.5-8 Vascular congestion in the intestines during the anhepatic phase contributes to endotoxin leakage and tumor necrosis factor-α–induced activation of Kupffer cells.7, 9 Activated Kupffer cells, in turn, release reactive oxygen species and produce cytokines that contribute granulocyte sludging within the sinusoids, which can undergo degranulation and cause further tissue damage. Sequential hypoxia and reoxygenation also activate complement factors, further contributing to tissue damage.

The biliary tree is receiving increasing attention as a target of ischemia/reperfusion injury.10 Damage to the biliary tract occurs primarily because of preservation injury, which leads to ischemic cholangitis or deep wounding of the bile duct wall. Ischemic injury damages the microvasculature of the peribiliary plexus; loss of microvascular integrity predisposes to microvascular thrombosis after reperfusion. This results in deep wounding of the bile wall. Subsequent partial wound healing and wound contraction, but failed restitution of the biliary epithelial cell lining, result in biliary tract strictures that cause progressive biliary fibrosis, increased morbidity, and decreased organ half-life.

Because of the high rate of ischemic/reperfusion injury and biliary complications from NHBD liver grafts, it would be ideal if the viability of such livers and the associated biliary complications could be predicted prior to transplantation. This is attempted by the assessment of the key clinical parameters, including the time of warm ischemia, the liver enzyme function tests, the presence or absence of additional donor risk factors, the time of cold ischemia, and the risk of recipients, among others. However, currently there are no objective criteria to reliably predict NHBD liver graft outcome. Histological examination of biopsied tissues from NHBD livers plays an important role in this process, largely by exclusion of other risk factors or liver diseases, such as macrovesicular steatosis, bridging fibrosis, chronic active hepatitis, or parenchymal necrosis. However, because NHBD livers often show no significant parenchymal necrosis under light microscopy prior to revascularization to the recipient's circulation system, histopathological examination in general cannot reliably predict posttransplant function of NHBD liver grafts.

In this issue of Liver Transplantation, Monbaliu et al.11 show that the extent of vacuolation in porcine NHBD livers prior to cold preservation correlates with the length of warm ischemia and can be used as an independent parameter to predict the posttransplant graft function and hepatocellular damage. They retrospectively reviewed pretransplant biopsies from pig livers exposed to incremental periods of warm ischemia to quantitate the extent of vacuolation by the pathologist's semiquantitative score, stereological point counting, and digital image analysis; the produced data were then correlated with hepatocellular damage and graft survival. The authors observed that hepatocyte vacuolation in pretransplant liver biopsies increased according to the length of warm ischemia (0, 15, 30, 45, and 60 minutes) and that the degree of vacuolation determined by stereological point counting and digital analysis scoring was able to predict pig liver graft dysfunction and hepatocellular injury. Their findings raise the issue of whether hepatocellular vacuolation in pretransplant human NHBD liver biopsies might be quantitated to predict posttransplant graft function in patients. This has very important clinical implications because currently there are no reliable objective criteria for predicting the posttransplant outcome of NHBD livers in patients.

There are several histopathological changes associated with increasing lengths of warm ischemia in porcine12 and rat13 NHBD livers, including hepatocellular vacuolation, sinusoidal congestion, and focal hepatocyte cell dropout. Monbaliu and colleagues11 showed that, among these features, the extent of vacuolation after warm ischemia highly correlated with primary nonfunction and hepatocellular damage. It is of note that although stereological point counting and digital analysis scoring were able to predict pig liver graft dysfunction and hepatocellular injury, the pathologist's semiquantitative score was less effective. In the authors' experience, semiquantitative scoring (cutoff > 50%) could predict graft viability and recipient survival, but this score could not discriminate the extent of vacuolation between the different experimental groups with more than 15 minutes of warm ischemia. The latter findings are not entirely unexpected because such a histological quantitation system is known to be associated with interobserver and intraobserver variations, although it allows clinical application during the reasonable time frame of the transplantation procedure. The drawback of stereological point counting and digital analysis scoring is that they are unlikely to be available in a logistically acceptable time frame under clinical circumstances.

The study by Monbaliu et al.11 showed that vacuolation could be used as a meaningful and quantifiable parameter only when the biopsies were obtained from livers exposed to warm ischemia alone. After an additional 4 hours of cold ischemia, the extent of hepatocellular damage increased in the groups exposed to warm ischemia when the stereological point counting method was used, and the degree of vacuolation no longer contained the same predictive value as immediately after warm ischemia. Because vacuolation was not observed after exposure to cold ischemia alone, warm ischemia appears to be a prerequisite before cold ischemia can have an additional effect on hepatocyte vacuolation (the extent of vacuolation might reflect an accumulation of both warm and cold ischemia effects). Therefore, the authors' observations suggest that a liver biopsy be obtained immediately after organ recovery to assess vacuolation (prior to cold preservation).

With electron microscopy and Oil Red O staining, Monbaliu and colleagues11 have shown that the vacuoles in hepatocytes do not represent steatosis. However, the possible contribution of microvesicular steatosis cannot be entirely excluded, given that electron microscopy and Oil Red O staining have their own pitfalls (very small sized sample for electron microscopy; requirement of alcohol fixation for Oil Red O staining). The exact mechanism for hepatocyte vacuolation during warm ischemia has not been clearly defined. Several possibilities have been postulated, including lysosomal degradation (also called autophagic vacuolation), internalization of apical cell membranes due to adenosine triphosphate depletion, and anoxia in general with an influx of plasma into the cells. Another potential mechanism is the high intrahepatic venous blood pressure that develops after cessation of circulation, which can manifest as plasma entry into hepatocytes and/or sinusoidal congestion/dilatation. Because cytosol or cellular organelles were not observed in the vacuoles under electron microscopy, the autophagic process appears unlikely, at least in the pig model.

In animal models of NHBD liver transplantation, the length of warm ischemia could be well controlled. As the duration of warm ischemia is the best predictor of posttransplant graft function, it is not surprising that an incremental increase in the length of in situ warm ischemia was found to be associated with increased risk of developing hepatocellular damage and primary graft nonfunction in animal models.11-13 However, in the true clinical setting for both controlled and uncontrolled NHBDs, the exact duration of warm ischemia is frequently unknown. Therefore, there is a practical need to determine whether hepatocyte vacuolation in human NHBD livers could serve as a useful parameter for predicting liver graft outcome in patients.

Several important issues should be considered for the histological evaluation of human NHBD liver biopsies. It may be difficult to attribute the observed histopathological changes (vacuolation, sinusoidal dilation/congestion, or hepatocyte dropout) to warm ischemia versus other preexisting conditions in human livers, such as microvesicular steatosis, macrovesicular steatosis, or frozen section artifact. For example, donor-derived steatosis (microvesicular steatosis in particular) may mimic hepatocyte vacuolation, and it may be difficult to make a clear distinction during frozen section evaluation. Although steatosis can be confirmed by positive Oil Red O staining, it is not practical to obtain good-quality Oil Red O staining during the short time period of frozen section assessment. Even if an Oil Red O stain becomes available, the nature of patch staining may still prevent confident assignment of fat versus vacuolation. In addition, the scoring for vacuolation and steatosis by pathologists is associated with both interobserver and intraobserver variability, which may prevent reproducible clinical application. Furthermore, the architectural and cytological distortion associated with the frozen section process often prevents detailed histological assessment. It is of note that vacuolation is a frequent freezing artifact, and it may be very difficult or impossible to detect areas of vacuolation versus steatosis or necrosis on frozen sections. All these could potentially limit the use of scoring for vacuolation to predict the outcome of NHBD liver grafts in real clinical settings. It is conceivable that some of these obstacles may be partially overcome by stereological point counting or digital image analysis, but these methods need to be further validated in NHBD livers biopsies, and it remains a challenge to use these techniques in the limited time frame of clinical transplantation.