Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Graft Failure
Graft failure is a serious life-threatening condition for transplant recipients, and when left untreated, it results inevitably in death of the recipient. It is important to differentiate between acute and chronic graft failure. Although acute graft failure occurs early in the posttransplantation period and is mainly related to donor organ quality or less frequently to technical problems, chronic graft failure is caused by recurrent liver disease, biliary complications, or chronic rejection.
The large gap between organ demand and availability has forced the use of marginal donor organs for liver transplantation. Especially, the epidemic of obesity and diabetes in the United States has resulted in a higher number of donors with steatotic donor organs. In addition, the increased medical acuity of transplant recipients is another significant factor that has an important impact on graft function. Even the best organ might fail when it is transplanted in a very sick environment with marginal perfusion. The objective of this chapter is to review all topics of acute graft dysfunction and failure during the early postoperative period. This chapter should guide transplant surgeons and physicians to recognize early this serious condition and to take appropriate measures to avoid futile outcome. Chronic graft failure is not the objective of this chapter and is discussed in Chapter 64, Chapter 79 , and 80 dealing with retransplantation, recurrent disease, and chronic rejection.
Definition
Although primary nonfunction (PNF) is an exclusive diagnosis of early graft failure in the absence of any casual factor, this statement needs to be revised because it is well documented that PNF is associated with several risk factors such as age and cold ischemia time. On the other hand, graft failure due to any identifiable technical problems such as hepatic artery or portal vein thrombosis is not included in the definition of PNF. Graft failure due to cardiac failure in the recipient is also not included in the diagnosis of PNF.
The term primary nonfunction is best defined as graft failure soon after graft reperfusion with no discernible cause that leads to either retransplantation or death of the patient in the early postoperative phase. We at the University of California Los Angeles (UCLA) use the term PNF when graft failure occurs within the first 7 days after transplantation ( Table 75-1 ). The condition of graft failure after 7 days with initially normal or poor function is named delayed PNF . However, the definition of both PNF and delayed PNF require either retransplantation or death.
TABLE 75-1
Proposed Descriptors of Initial Graft Function After Liver Transplantation
| Graft Function | Abbreviation | Definition and Clinical Features |
|---|---|---|
| Normal initial function | NIF | Allograft with normal liver function or complete recovery. |
| Initial poor function or early graft dysfunction | IPF or EGD | Impaired initial allograft function with high peak serum transaminase and persistent high bilirubin levels. Definitions are presented in Table 75-2 . |
| Primary nonfunction | PNF | Death or retransplantation within the first posttransplantation week. |
| Delayed primary nonfunction | dPNF | Allografts with initial normal or poor function that develop nonfunction resulting in death or retransplantation within postoperative days 8 to 30. |
| Small-for-size dysfunction | SFSD | Same definition as IPF or EGD in the presence of a graft weight–to–body weight ratio of <0.8%. |
| Small-for-size nonfunction | SFSNF | Same definition as PNF in the presence of a graft weight–to–body weight ratio of <0.8%. |
All other conditions of early graft dysfunction that do not result in retransplantation or death we then define as initial poor function (IPF) or early graft dysfunction (EGD). There are several proposed definitions of IPF/EGD, which all include elevated levels of aspartate aminotransferase (AST)/alanine aminotransferase (ALT) ( Table 75-2 ). Although risk factors can be identified in the majority of patients experiencing PNF or IPF/EGD, such factors can also be present when the allograft function is normal.
TABLE 75-2
Definitions of Initial Poor Function or Early Graft Dysfunction
| Author, Year | Used Term | Definition Criteria |
|---|---|---|
| Ploeg et al, 1994 | Initial poor graft function |
|
| Ardite et al, 1999 | Severe initial graft dysfunction |
|
| Nanashima et al, 2002 | Initial poor graft function |
|
| Silberhumer et al, 2007 | Initial poor graft function |
|
| Olthoff et al, 2010 | Early allograft dysfunction | One or more of the following criteria
|
ALT , Alanine aminotransferase; AST , aspartate aminotransferase; INR , international normalized ratio; POD , postoperative day.
The current organ allocation policy of the United Network for Organ Sharing (UNOS) specifies that the diagnosis of PNF is restricted to 7 days or less from the time of transplantation. According to the Organ Procurement and Transplantation Network (OPTN) policy 3.6, criteria for PNF are clearly defined and include clinically relevant parameters (AST, international normalized ratio [INR], acidosis) of graft function ( Table 75-3 ). When criteria of PNF are present within the first 7 days, patients have the same high priority as patients with acute liver failure. Patients who develop PNF criteria after 7 days cannot be listed as high urgency and go by Model for End-Stage Liver Disease (MELD) score for retransplantation.
TABLE 75-3
UNOS Criteria for Primary Nonfunction According to OPTN Policy 3.6
| Transplanted liver within 7 days as defined by |
|
AST , Aspartate aminotransferase; INR , international normalized ratio; OPTN , Organ Procurement and Transplantation Network; UNOS , United Network for Organ Sharing.
Incidence
In a recently published report of 5347 consecutive liver transplantations performed at UCLA, the incidence of PNF was overall 8.4% with 7.8% in adults and 11.1% in pediatric recipients. In same series the rate of PNF in adult recipients has been significantly reduced after the introduction of the MELD allocation system from 9.1% (pre-MELD era) to 6.0% (MELD era). An analysis of the Scientific Registry of Transplant Recipients (SRTR) database showed that the national rate of PNF in the United States was 5.8% in the MELD era. Another analysis from UCLA demonstrated that PNF in adult recipients was the most common indication for retransplantation at 28%. IPF during the very early postoperative days can progress to PNF or delayed PNF or can completely resolve and return to normal graft function in the majority of cases. The rate of IPF is higher than PNF and depends on its definition (see Table 75-2 ). In two recently and independently published studies, which used a very similar definition of IPF, the observed rate of IPF was 21% in one study and 23% in the other. Most centers, including ours, continue to report PNF rates between 2% and 10% and IPF rates between 16% and 27%. Programs with more aggressive use of donors with extended criteria may experience higher rates, although PNF can occur even in the most optimal circumstances. It is also conceivable that PNF could be significantly underreported. Early postoperative deaths attributed to sepsis, neurological injury, multiorgan failure, or other organic causes could be the indirect effect of a nonfunctioning graft.
Diagnosis
Most clinicians use a composite of clinical, laboratory, and sometimes histological findings to diagnose IPF and PNF in the face of early graft dysfunction. The diagnosis of PNF is generally made early, often within the first 3 postoperative days, when either death or retransplantation intervenes. Often the onset of PNF can be recognized as early as after implantation of the allograft. Signs of PNF during the transplant procedure are (1) hemodynamic instability requiring increasing vasopressor support, (2) worsening negative base excess, (3) hypothermia, and (4) severe coagulopathy. In the most severe circumstances, the compromised allograft appears to perpetuate the acidosis and hemodynamic instability, and the only saving alternative may be to render the recipient anhepatic while listing the patient at the highest priority for retransplantation. However, the scenario of total graft nonfunction during the transplant operation is charitably rare, and a variety of more subtle clinical findings usually reflect the degree of dysfunction. Reversal of acidosis and improving kidney function are signs of good allograft function. The absence of these findings may signify IPF or PNF.
Bile secretion during the transplant procedure itself is an excellent prognostic factor. Bile flow rate has been reported in many studies to be one of the most useful predictors of postoperative allograft function. Because the cellular secretion of bile into the biliary canaliculi represents an active transport process requiring adenosine triphosphate, the production of bile during the early course after implantation mirrors the recovery of adenosine triphosphate synthesis in the allograft. Anecdotally, the color of the bile may be equally important, with a golden brown color held as the ideal. On the other hand, bile that turns into a pale yellowish color often indicates significant disturbances of the biliary secretion process and allograft dysfunction.
Laboratory values can also contribute to the diagnosis of PNF. Serum transaminase levels in the tens of thousands or levels that are steadily increasing imply severe organ injury and unlikely recovery. Peak AST serum levels greater than 5000 units/L are reported to result in a PNF rate of 41%, as opposed to a rate of 10% in those with peak AST serum levels of 2000 to 5000 units/L. Others have identified initial AST levels above 2000 units/L or levels that are slow to resolve as being predictive of PNF. The significance of high AST levels for PNF is reflected in the UNOS definition in which an AST level of 3000 units/L or higher is a mandatory criterion defining PNF within the first 7 posttransplant days (see Table 75-3 ). Elevated ALT levels and prothrombin times may have a similar predictive value. Persistent lactic acidosis, hypoglycemia, hyperkalemia, increasing hyperbilirubinemia, and severe hypoprothrombinemia are all obvious signs of poor function. Rather than an absolute cut-off value of any test, it is always the trends in values that are of greatest importance.
Other clinical indicators of initial allograft dysfunction include the patient’s mental status, urinary output, and pulmonary status. Multiorgan failure is the inevitable result of a nonworking allograft. The best correlation with poor outcome does not appear to be the failure of any particular individual organ system, but rather the absolute number of organ systems involved.
Five specific criteria proposed by John Hopkins have been advocated in the evaluation of IPF: (1) rising serum transaminase levels, (2) poor synthetic function with elevated INR despite continuous administration of fresh-frozen plasma, (3) “minimal” bile production, (4) impaired metabolic clearance with hyperammonemia, and (5) patent hepatic vessels by Doppler ultrasonography. Another criterion of impaired allograft clearance is the development of toxic tacrolimus levels during the first posttransplant week. Some authors recommend the use of dynamic metabolic tests for the assessment of initial allograft function. The plasma disappearance rate of indocyanine measured during the first 5 posttransplant days was predictive for early postoperative complications, including PNF, in a French study, whereas others advocated for measurement of the maximal enzymatic liver function capacity by the LiMAx in the early posttransplant period. Most patients with poor, but reversible, function begin to improve by the third posttransplant day, whereas those with PNF will continue to worsen.
It is essential that with any evidence of graft dysfunction, vascular or other technical abnormities must be considered and excluded. It is important to emphasize that IPF and PNF exclude hepatic artery, portal vein, or hepatic venous outflow abnormalities, as well as other systematic processes such as abdominal compartment syndrome and right heart failure. Although vascular abnormalities can usually be excluded by noninvasive studies (computed tomographic angiography, Doppler ultrasonography), surgical exploration is often the most expeditious manner in which to exclude a wide variety of vascular or mechanical factors and allow “hands-on” assessment and safe biopsy of the graft.
Etiology
A number of studies over the years have attempted to analyze a multitude of donor and recipient variables to determine the cause of graft failure. A comprehensive list of factors implicated in PNF and/or IPF is provided in Table 75-4 . Knowledge as well as avoidance of these risk factors is probably the best strategy to prevent IPF and PNF. The underlying mechanisms leading to graft dysfunction in each of these risk categories are also likely to have considerable overlap.
TABLE 75-4
Risk Factors for Primary Nonfunction and Initial Poor Function
| Donor Factors |
|
| Procurement Factors |
|
| Recipient Factors |
|
For purposes of discussion, factors influencing graft function can be sequentially categorized within the transplant process as being related to (1) the donor, (2) procurement, and (3) the recipient. The most important issues associated with graft failure include donor age, donor steatosis, small-for-size syndrome (SFSS) in the living donor liver transplant (LDLT) setting, and cellular and molecular events of ischemia and reperfusion. The last is of such significance that it warrants its own chapter ( Chapter 105 ) and will therefore not be considered in detail here.
Donor-Related Factors
Various studies have identified important factors to be independently associated with allograft dysfunction, including PNF and IPF. The most important variables are donor age, steatosis, and duration of cold ischemia time. In LDLT the functional liver volume of the partial graft is a strong predictor for SFSS. Although also other factors have been identified, the discussion of donor-related factors is focused on the main variables mentioned earlier.
Donor Age
The factor “donor age” has been demonstrated in many studies to be associated with outcome. Although several studies have shown that the use of livers from older donors is safe, there is established evidence that advanced donor age is a risk factor for early graft dysfunction, including PNF. This is especially important to keep in mind because there is a changing trend to use more organs from older donors to expand the donor pool. A recent large SRTR database analysis revealed that increasing donor age translates in an incremental risk for developing PNF. In this study, allografts from donors older than 60 years had a 1.57 higher odds ratio of developing PNF compared to organs from donors younger than 40 years. This observation is supported by a finding of another study in which older donor age (>50 years) was a risk factor for early retransplantation. Another recently published study that validated early allograft dysfunction found that donor age above 45 years was an independent risk factor for allografts to develop early dysfunction ( Fig. 75-1 ). Donor age had the highest odds ratio among all risk factors in this study. This observation is consistent with findings from previous studies.
Donor age is also a mandatory element of established risk scores such as Donor Risk Index (DRI), survival outcomes following liver transplantation (SOFT), and balance of risk (BAR) score. Although these risk scores focus on survival and not on the incidence of IPF or PNF, both types of graft dysfunction are associated with an increased risk for early posttransplant death.
Steatosis
Fatty liver or steatosis can be classified into microvesicular (MiS) or macrovesicular steatosis (MaS). A more recent definition of MaS further divides this entity into small-droplet (sd-MaS) and large-droplet MaS (ld-MaS). The term ld-MaS is used when a single fat vacuole is larger than the half of the cell displacing the nucleolus to the cell periphery, whereas sd-MaS describes fat vacuoles that are smaller than the half of the cell and do not displace the nucleolus. MaS is associated with alcohol abuse, diabetes mellitus, hyperlipidemia, obesity, metabolic syndrome, and the use of certain drugs. The epidemic of diabetes and obesity in the United States during the past 2 decades has led to an increased number of donors with hepatic MaS. MiS is a condition characterized by innumerable tiny lipid vesicles. This rare type of steatosis is found in pathological conditions often associated with mitochondrial injury, such as Reye’s syndrome, acute viral or drug injury, long-term total parenteral nutrition, sepsis, and some genetic disorders. Of note, sd-MaS is often incorrectly labeled MiS in the surgical literature.
The incidence of hepatic steatosis in the general population was variably reported between 6% and 24% in the past. In a more recently published prevalence study, almost one third of an urban population in the United States had hepatic steatosis. There was a significant ethnic difference in the prevalence of hepatic steatosis with 45% in Hispanics, 33% in whites, and 24% in blacks. The higher prevalence in Hispanics was related to the higher prevalence of obesity and diabetes in this population. The prevalence of steatosis in deceased donors ranged by report from 13% to 26% in the past and is most likely higher now.
An association between fatty infiltration and PNF has long been appreciated in clinical liver transplantation. A study from the 1990s looked at the predictive value of donor liver biopsy results and found that livers with MaS had a higher rate of PNF. A large multivariate analysis of risk factors found that hepatic steatosis was independently associated with both IPF and PNF. Another study reported that organs with 30% or greater steatosis had a PNF rate of 13%, as opposed to only 2.5% in those that were nonfatty. However, higher rates of PNF have been reported with increasing degrees of steatosis.
Although various studies have reported that select steatotic grafts can be used in appropriate recipients with acceptable outcome, and that steatosis is reversible after orthotopic liver transplantation (OLT), MaS greater than 30% is an established risk factor for graft failure in a large UNOS study. Although this and other studies do not discriminate between sd-MaS and ld-MaS, the large-droplet type appears to be the most significant condition that has to be considered for risk assessment. To use organs with MaS safely, an appropriate donor-recipient match is of paramount importance for these high-risk organs. The policy of many transplant centers is to allocate steatotic organs to low-MELD recipients or patients with hepatocellular carcinoma. The concept behind this policy is that high-acuity recipients with high MELD scores may not tolerate IPF or are even at significantly higher risk for developing PNF. A combined large liver transplant database study from the United States and Europe showed that the use of grafts with MaS of greater than 30% was safe for recipients with a low risk score, whereas allocation of those organs to recipients with higher risk scores resulted in inferior outcome. As with many other donor factors, graft steatosis alone is unlikely to determine the outcome of liver transplantation. Except in the case of severe ld-MaS, many fatty grafts can be used successfully if careful attention is paid to contributory donor variables, recipient selection, and limitation of ischemia time.
The mechanisms underlying the increased sensitivity of steatotic grafts to transplantation remain incompletely understood. The importance of this area of research is obvious in the context of an expanding donor organ shortage and increasing incidence of obesity and diabetes in donors. However, our understanding of the mechanisms associated with IPF/PNF in steatotic grafts has progressed from an association through observation of histological changes to molecular mechanisms in the study of ischemia-reperfusion injury. There is growing experimental and clinical evidence that steatosis exacerbates the mechanisms of injury related to ischemia-reperfusion.
Small-for-Size Syndrome
Even before the era of LDLT, it was known that the use of an allograft of inadequate size for a given recipient could result in graft dysfunction or failure. With the increasing importance of adult-to-adult living donor transplantation, the importance of relative graft-to-recipient body weight has become more evident. When Emond et al explored the feasibility of living donor adult-to-adult transplantation; they found that graft size strongly correlated with function in the recipient. This group was perhaps the first to use the term small-for-size syndrome (SFSS) when describing the pattern of dysfunction that occurred when allograft volume was less than 50% of the recipient’s expected liver volume.
Allografts with a size of less than 40% of a normal liver or a graft weight–to–body weight ratio (GWBWR) of less than 0.8% meet the criteria defining a small-for-size graft. A graft below this size threshold that develops signs of liver failure with jaundice, coagulopathy, encephalopathy, and ascites within the first week after transplantation is termed SFSS . Like IPF and PNF, SFSS can be divided into small-for-size dysfunction and small-for-size nonfunction in the presence of a GWBWR of less than 0.8% (see Table 75-1 ). Small-for-size dysfunction is defined as dysfunction during the first postoperative week requiring the presence of two of the following on 3 consecutive days: total bilirubin greater than 100 μmol/L (>5.8 mg/dL), INR greater than 2, and encephalopathy grade 3 or 4. Small-for-size nonfunction is defined according to PNF as graft failure resulting in retransplantation or death of the recipient within the first week after transplantation.
The incidence of SFSS reported in the living donor literature ranges from 2.9% to 12.5%. Although SFSS can also occur in whole-organ grafts, this syndrome is a feared problem in partial grafts from living donors or from splits in deceased donors. As with IPF and PNF, the occurrence of SFSS is far from predictable, and many small-for-size grafts have been used successfully. SFSS can demonstrate the same spectrum for clinical features and severity as IPF and PNF. It is likely that a number of both graft- and recipient-related factors are important in determining when SFSS will occur. It also now seems likely that not only anatomical graft size but also functional size can be of relevance. The latter may be influenced by some of the same parameters associated with IPF and PNF, such as age and steatosis.
Early on, Emond et al speculated on a mechanism of SFSS that included portal hyperperfusion and an overwhelmed metabolic capacity of the small graft. Many others now agree, and the prevailing theory for a mechanism of SFSS is that of portal hyperperfusion with sinusoidal injury resulting from portal pressure exceeding sinusoidal compliance. Histological studies of partial grafts with SFSS revealed portal vein and periportal sinusoidal endothelial denudation and focal hemorrhage into the portal tract connective tissue, as well as signs of poor hepatic arterial flow and vasospasm. Experimental studies have suggested that these events result in failure of regeneration, which is the most important contributing factor in SFSS. There is clinical evidence in LDLT that portal flow rates above 260 mL/min/100 g graft weight are associated with pronounced hyperbilirubinemia and poor outcome. A smaller graft represents a higher vascular resistance for the portal draining splanchnic blood. According to Ohm’s law, the pressure gradient is directly proportionate to the vascular resistance at a given flow rate:
where R is the vascular resistance, Δ P is the pressure gradient, and Q is the flow rate.
Based on this rule, a smaller graft is exposed to a higher pressure gradient. The negative effect of elevated portal venous pressure on outcome was reported in an LDLT study. In this study an elevated portal venous pressure of 20 mm Hg or higher within the first 3 postoperative days was associated with SFSS and inferior survival.
The inflow resistance of a small graft can be further elevated in the setting of outflow obstruction. It is evident that portal hyperperfusion in the setting of outflow obstruction results in a higher sinusoidal pressure and a higher risk for developing SFSS. Experimental studies using a combined model of portal hyperperfusion and outflow obstruction have demonstrated that additional outflow obstruction resulted in congestion, confluent centrilobular necrosis, and reduced proliferation.
Although graft size is perhaps the most important factor to be considered in transplantation of partial grafts, other donor and recipient factors have been associated with SFSS, including recipient’s illness, donor age, steatosis, and duration of ischemia. A study on adult-to-adult LDLT showed that small grafts in sicker recipients have an additive negative impact on early graft survival. In this study the overall incidence of SFSS was 12.5%, but it was observed exclusively in recipients with Child-Turcotte-Pugh B and C status. In that group the incidence of SFSS was 80% (four of five).
Other Donor-Related Factors
Hypotension is a well-described cause of shock liver. As such, it seems intuitive that prolonged hypotension or cardiac arrest in the deceased donor may add to the risk for graft dysfunction. The use of multiple or high-dose vasopressors has in fact been identified as a risk factor for early graft dysfunction. Vasopressors are well known to cause splanchnic vasoconstriction, which may add to the “harvest injury” frequently invoked when an organ fails to function. However, hypoxia, hypotension, and cardiac arrest are often surprisingly well tolerated by the liver, and grafts from such donors may still have excellent function. There is evidence from various studies that allografts from donation after brain death donors with cardiac arrest perform as well as those from donors without cardiac arrest. In these studies the rate of PNF, as well as the initial allograft function reflected by peak ALT levels, serum total bilirubin level, and prothrombin time, was not significantly different in transplant recipients with organs from donors with and without cardiac arrest. Previous cardiac arrest with successful resuscitation has been hypothesized to induce a form of protective ischemic preconditioning that may partly protect the graft during the transplant process. Although the use of high-dose and multiple vasopressors is clearly a risk factor for IPF and PNF, a large SRTR database analysis demonstrated that prerecovery use of dobutamine or inotropes was not a predictor for PNF in the multivariate analysis. Therefore cardiac arrest or high vasopressor use should not preclude transplantation of livers from these donors.
A number of studies suggest that donor hypernatremia can affect graft function. Diabetes insipidus and poor fluid management are the most common causes of this condition. Although the exact mechanism is not known with certainty, it is postulated that hypernatremia leads to increased intracellular osmolality, with the subsequent cellular edema incurred at reperfusion leading to graft dysfunction. Totsuka et al looked at peak donor serum sodium level as well as corrected sodium level at the time of procurement. They found a significant correlation between uncorrected hypernatremia (>155 mEq/L) and both IPF and PNF. The incidence of PNF was 18.5% in the uncorrected hypernatremia group versus 3.4% in those with normal sodium levels. When serum sodium level is corrected before procurement, the increase in PNF is no longer found. We at UCLA address donor hypernatremia (>160 mEq/L) at the time of procurement by infusing a 5% dextrose-free water solution through the cannulated inferior mesenteric vein during the donor preparation (see Chapter 43 ). However, a recently published study investigating severe hypernatremia in deceased liver donors found no association between hypernatremia and early transplant outcome. In this study the rate of graft failure within 7 days or within 30 days after liver transplantation was not significantly different between the groups with donor serum sodium levels of less than 160, 160 to 169, and 170 or higher mEq/L. Despite the conflicting reports, there is not enough evidence that livers from donors with hypernatremia should be precluded from liver transplantation.
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here