Update and Application to Liver Biopsy Interpretation in Clinical Practice
Case 6 -
Harvest Injury (Ischemia-Reperfusion/Preservation Injury)
Julia C. Iezzoni
The patient, a 56-year-old Caucasian male, underwent
orthotopic liver transplantation for end-stage liver disease due to alcohol abuse. The liver donor was a
50-year-old Caucasian male who had suffered a sudden, massive cerebral vascular accident. After
declaration of brain death, the donor remained hemodynamically stable, and vasopressor support was not
necessary to maintain organ perfusion. Organ harvest, transport, and transplantation were uneventful,
with a total cold ischemia time of approximately 12 hours. As per protocol, a post-perfusion biopsy of
the liver was performed immediately upon restoration of the allograft's blood supply once in the
recipient (Figure 19 – H&E). For the first day after transplantation, the patient's liver function
tests were characterized by markedly elevated aminotransferases and a high total bilirubin. By
post-transplant day 7, these values had diminished only slightly, and a liver biopsy was performed.
Diagnosis: Harvest Injury (Ischemia-Reperfusion/Preservation Injury)
Pathologic findings: Sections of the post-perfusion biopsy of the
liver allograft show early harvest injury, characterized by scattered clusters of necrotic hepatocytes
with associated neutrophils. The biopsy 7 days after transplantation demonstrates the typical findings
of moderate-to-severe harvest injury, including centrilobular hepatocellular ballooning degeneration,
cholestasis, and bile ductular proliferation, which has periductular acute inflammation ("acute
Harvest injury is the nonimmunologic damage to the liver allograft due to the mechanical
and physiologic consequences inherent in the transplantation procedure. This damage results from two
major insults, ischemia and reperfusion, and these two injuries occur at different steps in the process
of transplantation. First to occur is the ischemic injury. The liver allograft is subjected to ischemia
from the time it is disconnected from its blood supply in the donor to when its blood supply is restored
in the recipient. To reduce this ischemic injury, the organ is kept cold, and this time period is called
the cold ischemia time. Immediately upon restoration of the blood supply, the liver is subjected to
further damage as the result of the initial reperfusion with warm, oxygenated blood. This reperfusion
injury (also known as "warm injury"), both exacerbates the injury already caused by the ischemia and
introduces additional mechanisms of damage. Despite advances in organ preservation and surgical
technique, a certain degree of harvest injury of the allograft remains an inevitable consequence of liver
transplantation and results in variable degrees of allograft damage and clinical dysfunction . As such, harvest injury is a common diagnostic consideration for the
pathologist evaluating liver transplant biopsies.
A basic appreciation of the transplant procedure is useful to understand this type of
liver injury. In brain dead, heart-beating donors, preservation of the allograft is initiated by
infusion of the liver with cold preservation medium prior to its extraction. The preservation medium
used at most institutions is UW solution (so-called because it was developed at the University of
Wisconsin), designed specifically to counter potential mechanisms that contribute to ischemic injury.
The liver is removed from the donor, often flushed again with cold preservation medium, placed in a
sterile plastic bag that is filled with the medium, and then fully immersed and stored in an ice slush
maintained at approximately 0 °C. Prior to revascularization of the liver in the recipient, the cold
preservation fluid in the allograft is flushed out with blood, lactated Ringer's solution, or a medium
formulated specifically to minimize reperfusion injury. After the vascular anastomoses are in place and
the blood supply is restored, a "post-perfusion" biopsy of the liver is obtained. This biopsy provides a
"time zero" data point of the histologic extent of the harvest injury incurred as a result of the
transplantation procedure. Several studies have demonstrated that the extent of damage present in the
post-perfusion biopsy correlates directly with the degree of subsequent harvest injury-induced allograft
dysfunction. As such, the post-perfusion biopsy is a valuable tool for the prediction of ensuing harvest
injury in the early post-transplantation period.
Clinically, harvest injury presents 1 to 3 days after transplantation. It is
characterized by a sharp elevation in transaminases, a sustained high total bilirubin level, and in cases
with more severe injury, an elevated prothrombin time. Diminished bile flow also may be noted. The time
frame for functional and histologic resolution of harvest injury depends on the severity of the damage.
Minimal-to-mild harvest injury usually is associated with good organ function, with functional and
histologic reversion to normal as soon as one week after transplantation. In cases of modest harvest
injury, complete functional and histologic recovery occurs in most cases, but it may require up to 6
weeks to normalize. In severe cases, abnormalities may persist for up to 6 months post-transplantation;
sometimes, severe harvest injury may result in permanent damage to the allograft. Overall, most cases of
harvest injury resolve within 3 weeks after transplantation.
The most severe clinicopathologic manifestation of harvest injury is primary nonfunction
(PNF). Primary nonfunction is characterized by the immediate and irreversible metabolic failure of the
newly implanted allograft. Clinically, PNF is characterized by rapidly rising serum transaminases, lack
of bile formation, and severe coagulopathy. The clinical course progresses quickly to hypoglycemia,
hepatic encephalopathy, acute renal failure, disseminated intravascular coagulation, and death unless
urgent re-transplantation is performed. Despite advances in preservation technique, PNF occurs in 5% to
15% of liver transplant patients.
Several factors are associated with an increased risk of the development of harvest
injury in the liver allograft. These include cold ischemia time exceeding 10 hours, donor episodes of
hypotension (especially if treated with vasopressors), increased donor age (especially when greater than
50-years-old), and severe macrovesicular steatosis of the liver allograft.
Of particular relevance to the pathologist is the risk factor of macrovesicular
steatosis, because the pathologist is called upon routinely to assess the extent of macrovesicular fatty
change in a liver being considered for transplantation. Typically, this evaluation is performed on a
frozen section of a biopsy of the potential allograft. Several studies have documented that donor livers
with severe macrovesicular steatosis (>60%) are more susceptible to harvest injury, including PNF,
than those devoid of or with mild (<30%) macrovesicular steatosis. The outcome of livers with
moderate macrovesicular steatosis (30-60%) is variable and may depend on the presence of additional risk
factors. As a general rule, donor livers with greater than 50% macrovesicular steatosis are considered
unsuitable for transplantation. When assessing a potential donor liver for fatty change, it is essential
to evaluate only the macrovesicular, not microvesicular, steatosis. As will be discussed below,
microvesicular steatosis is a common finding in ischemic allografts, and while studies are pending,
currently microvesicular steatosis has not been shown to be irreversibly deleterious to the allograft.
Since available histochemical stains for fat (such as Oil red O) are frought with technical difficulties
and inter-observer variability, we and others recommend that the extent of macrovesicular steatosis be
assessed on routine H&E stained tissue sections. This approach helps to avoid the misintrepretation
of microvesicular steatosis as macrovesicular steatosis, which potentially could result in discarding a
Mechanistically, harvest injury is a complex, multi-factorial cascade of pathophysiologic
processes that involves various cell types (e.g. hepatocytes, sinusoidal endothelial cells, activated
Kupffer cells, neutrophils), platelets, and a diverse variety of bioactive molecules, some of which are
cytotoxic (e.g. reactive oxygen intermediates, acute reactant cytokines, proteases, prostaglandins).
While a detailed description of these mechanisms is beyond the scope of this discussion, it is notable
that the ischemia and reperfusion induced injury of the sinusoidal endothelial cells is a central step in
the chain of events that results ultimately in the clinical and histological manifestations of harvest
injury. In addition to neutrophil margination and platelet activation, this sinusoidal endothelial cell
injury causes microcirculatory blood flow disturbances, which leads to severe alterations and impairments
in local tissue oxygenation. In conjunction with the reperfusion-induced activation of Kupffer cells,
this impaired oxygenation results in hepatocellular damage. Depending on the severity of the damage,
this injury causes variable degrees of allograft dysfunction or even nonfunction.
The morphologic features of harvest injury differ relative to the time interval
post-transplantation - i.e. the amount of time since the insult. The post-perfusion (or "time zero")
biopsy shows the immediate damage; biopsies thereafter demonstrate the residual sequelae of and recovery
from this injury. As such, liver allograft biopsy interpretation should be performed in conjunction with
knowledge of the post-transplantation interval. In addition, the greater the degree of harvest injury,
the longer time is necessary post-transplantation for histologic (and functional) recovery.
At time zero (i.e. the post-perfusion biopsy), harvest
injury is characterized by several morphologic features, which vary in extent relative to the degree of
severity of the injury. The characteristic morphologic changes are: 1) intrasinusoidal neutrophils; 2)
microvesicular steatosis; and 3) hepatocellular necrosis (varies from necrosis of individual or small
clusters of hepatoctyes to zonal necrosis, depending on the severity of the harvest injury).
Virtually all allografts, even those with minimal harvest injury, demonstrate at least
mild intrasinusoidal neutrophils and focal microvesicular steatosis at time zero. These features are
more extensive in cases with greater degrees of harvest injury. Intrasinusoidal neutrophils are rarely,
if ever, present in biopsies taken before reperfusion, and they are probably of recipient origin.
Microvesicular steatosis likely represents a hepatocellular mitochondrial response to ischemia, but it
may also be related to donor management. As described above, while primary allograft dysfunction has
been associated with severe macrovesicular steatosis, microvesicular steatosis currently has not been
shown to be irreversibly deleterious to the allograft.
Hepatocellular necrosis is present at time zero in allografts with moderate-to-severe
harvest injury. With moderate injury, this manifests as necrosis of individual or small clusters of
hepatocytes; these punctate necrotic foci often are distended with neutrophils. With severe injury,
zonal hepatocellular necrosis, usually involving the centrilobular zone, is seen. This zonal injury has
the appearance of "infarct-type" necrosis, and it may be due to end-arteriolar vasospasm, or
alternatively due to microcirculatory blood flow disturbances caused by sinusoidal endothelial cell
injury. Zonal necrosis is seen particularly in allografts from donors maintained on vasopressors for
Post-transplantation Days 1-21
Biopsies from days 1-21 show a
somewhat different constellation of findings from those at time zero. As before, however, these features
vary with the severity of the injury.
During this period, allografts with minimal-to-mild degrees of harvest injury routinely
show: 1) hepatocellular ballooning degeneration (centrilobular); 2) cholestasis (centrilobular); 3)
individual hepatocyte necrosis; and 4) persistent microvesicular steatosis.
Hepatocellular ballooning degeneration in the transplant setting is believed to result
either from dilatation of the rough endoplasmic reticulum and/or impairment of the secretory capacity of
the injured hepatocytes, which in turn causes intracellular fluid accumulation. As a result, the
ballooned hepatoctyes are swollen and enlarged (two times normal size), with "diluted" cytoplasm, which
appears finely granular or reticulated. The ballooning change usually affects hepatocytes in the
centrilobular region. Cholestasis, either intracellular or intracanalicular, often occurs, and it also
has a centrilobular predominance. In addition, individual hepatocyte necrosis ("acidophilic bodies") may
be seen, either scattered throughout the lobule or mainly centrilobular in distribution. Persistent
microvesicular steatosis usually is present.
Liver allografts with moderate-to-severe degrees of harvest injury are characterized by:
1) extensive hepatocellular ballooning degeneration; 2) extensive cholestasis; 3) lipopeliosis; 4) bile
ductule proliferation with periductular acute inflammation ("acute cholangitis pattern"); and 5) focal
and/or zonal hepatocytic necrosis.
With increasing severity of harvest injury, the hepatocellular ballooning degeneration
becomes more extensive, and in severe cases, it may be panacinar in extent. Intracellular and
intracanalicular cholestasis may be severe, and there may be associated bile duct plugging.
Lipopeliosis, a condition in which the sinusoids become engorged by rounded fat globules, which follow
the release of fat from necrotic hepatocytes, occurs in approximately 5% of liver allografts and is a
manifestation of harvest injury. Bile ductule proliferation with acute inflammation ("acute cholangitis"
pattern), morphologically may mimic extrahepatic biliary tract obstruction or ascending cholangitis.
This bile ductular proliferation likely is a response to ischemic injury of the bile duct epithelium.
Focal or zonal hepatocellular necrosis often is present and is generally centrilobular in location.
The entities in the differential diagnosis are included not only for their histologic
similarities to harvest injury, but also because they typically occur in the early post-transplantation
period, as does harvest injury.
1) Pathologically documented cases of hyperacute
(antibody-mediated, humoral) rejection are vanishingly rare. On a statistical basis alone,
therefore, this diagnosis is highly unlikely. Hyperacute rejection occurs due to the presence within the
allograft recipient of preformed antibodies directed against donor antigens. Clinically, patients with
hyperacute rejection experience a relentless rise in serum aminotransferases, coagulopathy, and fulminant
hepatic failure within 1-2 weeks after transplant. Biopsies from post-operative day 1 show focal
hepatocellular necrosis and intravascular thrombi. Immunohistochemical or immunofluorescence studies may
demonstrate sinusoidal, venous, or arterial deposition of immunoglobulins and complement. Biopsies taken
2-3 days after transplantation demonstrate extensive coagulative and hemorrhagic necrosis and numerous
sinusoidal neutrophils. Thereafter, the organ undergoes progressive hemorrhagic infarction, and the
entire scenario occurs over a period of 7-10 days. Severe harvest injury initially may resemble
hyperacute rejection due to the hepatocellular necrosis and the sinusoidal neutrophils. However, harvest
injury will not show intravascular thrombi or progressive infarction.
2) Hepatic arterial thrombosis, alone or in combination with portal vein
thrombosis, occurs in a minority (<5%) of patients after liver transplantation. Arterial thrombosis
often occurs within several days of transplantation. Depending on the severity and extent of the
occlusion, hepatic artery thrombosis may cause ischemia or infarction of part or all of the allograft.
As such, the biopsy findings are highly subject to "sampling effect" and may show normal tissue,
centrilobular ballooning with cholestasis, irregular zonal necrosis or complete infarction. These
histologic features are very similar to those of harvest injury. Because of the highly variable extent
and distribution of the histologic findings in hepatic artery thrombosis, this diagnosis is one of
exclusion and requires radiologic confirmation of the vascular occlusion.
- Angelescu M, Hofmann W, Zapletal C, Bredt M, Kraus T, Herfarth C, Klar E. Histomorphological analysis of preservation injury as determinant of graft quality in clinical liver transplantation. Transplant Proc 1999;31:1074-1076.
- Canelo R, Braun F, Dopkens K, Ramadori G, Ringe B. Characterization and influence of risk factors on initial liver function after transplantation. Transplant Proc 2000;32:60-61.
- Cha IC, Bass N, Ferrell LD. Lipopeliosis: An immunohistochemical and clinicopathologic study of five cases. Am J Surg Path 1994;18:789-795.
- Gaffey MJ, Boyd JC, Traweek ST, Ali MS, Rezeig M, Caldwell SH, Iezzoni JC, McCullogh C, Stevenson WC, Khuroo S, Nezamuddin N, Ishitani MB, Pruett TL. Predictive value of intraoperative biopsies and liver function tests for preservation injury in orthotopic liver transplantation. Hepatology 1997;25:184-189.
- Markin RS, Wisecarver JL, Radio SJ, Stratta RJ, Langas AN, Hirst K, Shaw Jr BW. Frozen section evaluation of donor liver before transplantation. Transplantation 1993;56:1403-1409.
- Selzner M, Clavien P-A. Fatty liver in liver transplantation and surgery. Sem Liver Disease 2001;21:105-113.
- Serracino-Inglott F, Habib NA, Mathie RT. Hepatic ischemia-reperfusion injury. Am J Surgery 2001;181:160-166.