Model for end-stage liver disease (MELD) exception guidelines: Results and recommendations from the MELD exception study group and conference (MESSAGE) for the approval of patients who need liver transplantation with diseases not considered by the standard MELD formula

Citations Over TimeTop 10% of 2006 papers

Abstract

Determining need for liver transplantation (LT) can be effectively estimated when the actual liver disease is highly likely to cause death in the near future. However, for many conditions, the liver disease itself does not carry a high risk of short-term mortality, and other factors contribute to defining the need for LT. The existence of these so-called exceptional cases was recognized in the initial development of the Model for End-Stage Liver Disease (MELD)- and Pediatric End-Stage Liver Disease (PELD)-based liver allocation policy,1 most notably in the case of hepatocellular cancer (HCC). In this instance, the driving imperative for LT is not life-threatening liver failure, but the progression of cancer to a point where a high probability of cure is no longer possible. Using the well-established Milan Criteria2 as selection criteria for good outcome, policymakers initially equated the risk of HCC progression beyond Milan Criteria to 15% for stage 1 lesions and 30% for stage 2 lesions within 3 months of listing analogous to MELD-defined mortality risk.3 These initial estimates proved to be too high based on publications citing much lower risks of progression in cohorts of waiting LT candidates with HCC,4, 5 and subsequently, the risk of HCC progression was reestimated to be much lower. The HCC priority policy was revised accordingly. The HCC example illustrates two important principles for allocating livers to patients with low mortality risk who fall into these exceptional diagnosis categories. First, a nonmortality endpoint, namely the risk of progression beyond Milan Criteria, where a patient would be removed from the LT waiting list, was defined. For any such endpoint, patient-specific, objective definitions must be used. In the case of HCC, the risk of tumor progression beyond the Milan Criteria meets such standards because the risk of progression does not depend on extrinsic influences like geography, location of care, or subjective assessment of symptoms, and the Milan criteria are very well described and reasonably measurable. Policymakers may change the HCC endpoint in the future on the basis of recent evidence that there are some cases beyond Milan Criteria that also carry excellent disease-free survival after LT. Doing so, however, would not require a complete revision of the liver allocation system but would only require reestimation of the risk of progression to this new endpoint. Thus, endpoints can be revised over time if they remain appropriate, objective, and patient-specific estimates of need, and continue to define reasonable success rates relative to the majority of waiting candidates. This process is critical for maintenance of equity of access to the limited donor pool. Second, as evidence accumulated, the original, admittedly arbitrary, estimates of risks of achieving the beyond–Milan Criteria endpoint were not as high as originally thought, and again, the assigned priority was changed based on evidence and without the need to completely readjust the entire system. More recently, with additional accumulated experience, investigators have more directly estimated the risk of HCC progression by using waiting list removal data.6 In this effort, factors associated with waiting list removal for candidates with HCC were identified and used to construct a predictive model for waiting-list dropout, which can be used to more accurately estimate risk of waiting list removals. This approach can serve as a framework for all LT candidates where the estimated risk of removal can be calculated once criteria for such removals are established. These removal criteria we term “waiting list endpoints” and can be mortality or some other measure like HCC stage. The risk of meeting the waiting list endpoint then becomes the metric for prioritization. Although the LT community recognized many conditions other than HCC in which mortality risk does not adequately define need for LT,7 few endpoint criteria and even fewer risk models exist to help quantify this need. A regional peer review system was established so that expert opinion could be brought to bear on making these assessments and currently these Regional Review Boards (RRBs) prospectively review every case (other than HCC within Milan Criteria) where the need for LT is thought to be underestimated by the MELD/PELD defined-mortality risk. However, because there are no well-defined criteria and there has been no comprehensive evidence review to establish guidelines to help RRBs make these assessments, these judgments have been largely based on individual RRB members' expert opinions. Several RRBs around the country have adopted their own systems for making these judgements but there is no consistent national protocol for RRB process. In addition, a systematic examination of the data elements required for better risk assessment studies is required to improve the appropriate prioritization of these exceptional cases. As part of ongoing organ allocation policy assessment and development, the United Network for Organ Sharing, the federal contractor for administration of the national Organ Procurement and Transplantation Network, constituted an effort to evaluate the non-HCC exceptional cases under the MELD/PELD liver allocation system. The MELD Exceptional Case Guideline (MESSAGE) Committee and recent MESSAGE conference participants were charged with two main tasks in an effort to improve ranking of candidates with these exceptional diagnoses where the need for LT might be defined by other endpoints. The specific goals in this exercise were: (1) to identify conditions for which a specific, objective, endpoint exists that defines need for LT such that assignment of additional priority can be automatic (without RRB peer review) and recommend the amount of additional priority so assigned, and (2) for those conditions where there is insufficient evidence, to recommend specific, objective data elements to be collected for individual conditions for which there was insufficient evidence for granting increased priority. Since HCC has been well addressed in other forums, the MESSAGE subcommittee was directed to focus on all other areas of exceptional diagnoses and RRB requests for priority upgrades and state where there was no data and identify data elements to be collected. The MESSAGE subcommittee presented the results of their literature review at an international meeting of experts in the field held March 1 and 2, 2006, in Chicago, IL, at which final recommendations were formulated. The preceding articles in this supplement summarize these deliberations. In this article, we summarize the conclusions reached for individual exception conditions regarding LT waiting-list endpoints, the priority magnitude recommended, and the data required to develop endpoints in cases where no clear endpoint has been defined. We will offer some conclusions for future development of policy to address exceptional cases. Current evidence and expert opinion regarding exceptional conditions where MELD/PELD alone may be inadequate for prioritization for LT are listed in Table 1. Waiting list endpoints are indicated for all exceptional case diagnoses. For most diagnoses, a risk of mortality is deemed most appropriate, but since the evidence suggests that the MELD/PELD score does not adequately measure wait-list removals as too ill, mortality risk or risk of no liver disease end organ damage for these cases, additional data are needed. Automatic assignment of priority is indicated based on sufficient evidence, expert opinion, or established practice. Biggins et al.8 cite studies indicating that mortality risk is the appropriate endpoint when considering ascites as an additional criterion for assigning priority for LT. These studies indicate that ascites is associated with increased mortality risks but subjective measurements, even those defined by the International Ascites Club, are not dependent on intrinsic patient characteristics and are more a reflection of physician practice patterns/preferences and reporting biases. Serum sodium in association with MELD score has shown improved accuracy for estimating mortality risk, especially for patients with low MELD scores, although the absolute increase in predictive value is small.9 The United Network for Organ Sharing/Organ Procurement Transplantation Network waiting list data collection process has accumulated a large amount of serum sodium values and these data should be analyzed for prospective validation in association with the MELD score. At present, in the absence of a clear contribution of objectively measured ascites to the mortality risk defined by MELD, no automatic assignment of additional priority points can be made. There is little additional objective evidence that ascites changes the MELD-defined mortality risk, especially at the higher range of MELD scores where liver allocation is most likely to occur, and most measures of severity of ascites remain physician-specific. Therefore, there is little evidence to support additional priority for patients with severe ascites and those referred for RRB approval should be extremely rare and unusual. Ham et al.10 rightfully point out that there are no patient-specific, well-documented objective measures of hepatic encephalopathy (HE), although there are data suggesting that HE (if consistently measured) can be associated with increased mortality, independent of the MELD score. Potential objective variables to be assessed for future refinement of the contribution of HE to mortality risk are: endotracheal intubation for airway protection in severe HE, and/or increased intracranial pressure. The West Haven Criteria are subject to observer bias and are based on subjective assessments of mental status. Endotracheal intubation is also dependent on physician practice patterns. Thus, although there is evidence for increased mortality risk for patients with chronic liver disease and HE, the lack of objective methods for quantification of its severity makes it extremely difficult to accept increased priority for this condition in a consistent and equitable manner. Thus, the available evidence does not support automatic increases in priority and case-by-case approvals by the RRBs also should be extremely rare unless supported by intracranial pressure measurements or endotracheal intubation justifications. Polycystic liver disease rarely carries an increased mortality risk, although these patients can have severe deterioration in their quality of life.11 There are many nontransplant options that may temporarily or rarely, permanently, relieve symptoms. In extreme cases, malnutrition resulting from inanition and early satiety can develop and profoundly impair affected patients' immune responses to infection and their ability to survive surgery. There is insufficient evidence to warrant automatic priority increases for patients with polycystic liver disease. Individual prospective requests to RRBs for increased priority should include on and of based on be supported because there is no with quality of have assigned a increase in priority to cases where severe can be by liver and A initial score can then be increased by 2 or 3 points every 3 months the patient This approach these cases some increase in but does not for transplantation in such cases without waiting making it for all patients from increased pressure is a of chronic liver disease and has been associated with increased mortality in these Although there are highly and for there are also to of these there are no prospectively measures of the severity of or objective measures for for although serum may be a for success of the majority of cases of can be with nontransplant and there are no objective, patient-specific measures of no automatic increase in priority can be at this Individual prospective to RRBs for increased priority based on should include of to and of over a of of and for endotracheal is in patients with MELD/PELD liver allocation policy increased in because the of this disease a risk of out that was beyond that calculated by the MELD score et indicate that there are data suggesting that in patients with and is associated with increased mortality, but these results are from to be an appropriate waiting list in that it is patient specific, and objectively and has been associated with The to which this is independent from MELD only defined risk to be established and with MELD might be on these data and the with the policy for increased priority for patients with and with no other cause for the disease expert opinion an automatic increase in MELD priority to the severity of data on will be in better defining the of waiting list and LT success for these has been associated with increased rates of mortality and the survival is The to which the severity of liver disease and the severity of the contribute to the and after liver LT but MELD scores to well with the of recommend that for with should be required as many cases of are only at the time of for LT. In addition, is required to accurately the so that measurements of and can be pressure is as a waiting list since it is objective and patient-specific and may be with mortality, but more data to be accumulated the of in the and waiting list opinion that increased priority is for patients with pressure to with an estimated MELD points However, since there is clear evidence that patients with have extremely every effort should be to this of to LT. where the should priority only in extremely since the results with LT remain very for these for increased priority for cases should include of the and and the of any to the to LT. there is little data on the magnitude of increased risk to quantification of increased need on the basis of at this for LT for patients with fall into two categories. patients develop and in of hepatic In rare cases, can with of hepatic where is The available evidence suggests that patients with chronic are well by the MELD score as the associated with mortality in studies those in the MELD with should be by using liver such as the Thus, exception points should not be for patients with There have been to RRBs increased priority on the waiting list to this can be an there are no data with waiting list mortality or from the may be a patient-specific there are no measures for objectively this For all of these we that there is no evidence to support automatic or prospective RRB approval for increased priority based on symptoms. of in affected that results in In this of the in the liver and to the results in that can in and other in to the and mortality risk is not defined by their liver disease and is not by their MELD/PELD score. need for LT is defined by The development of in patients with has been associated with increased mortality as has extremely but there are no studies risk factors for out from the waiting Since the of the MELD/PELD or patients with have liver opinion has that of liver alone patient-specific the is sufficient to automatic of increased priority the calculated MELD/PELD score without prospective RRB The amount of increased priority for patients with however, is not based on a assessment of increased risk. In addition, because patients with established as defined by the need for have increased mortality risk, expert opinion automatic of additional increased without evidence for the magnitude of increased priority. than of with this condition have the mortality risk waiting and should be an automatic of The very few cases and rare of with will make validation of risks is that results in of into In the most there is a point in a hepatic that results in into and that is Liver transplantation and can the These patients' mortality risk is not dependent on their liver disease and the MELD score is not of their need for LT. There are no good of factors associated with mortality or waiting-list in these although and and the score have the to define a risk of removal analogous to MELD/PELD for other candidates because they are patient-specific and reasonably objective These diagnoses are patient-specific, reasonably objective, and these variables would be good candidates for future liver allocation policy The liver allocation policy recognized the lack of data for patients with and the of risk of for these patients with the MELD risk of a MELD priority to 15% risk of mortality was The expert opinion that this be automatic if the evidence of and of the most but that this amount of increased risk is completely There are no for other data such as variables or or scores that are associated with deterioration beyond a and no endpoints defining stage have been results in and liver disease. The liver disease chronic patients with liver disease develop all of the of chronic liver disease and their need for LT is reasonably estimated by their mortality risk defined by However, the evidence suggests that patients with liver disease who also have disease increased risks of death LT and some will require an objective, patient-specific has been patients with disease and can be used to waiting LT candidates. collection of the data has not been by the United Network for Organ Sharing/Organ Procurement Transplantation Network data collection system but the evidence is to automatic of increased priority for patients waiting for LT alone with disease to of the amount of increased priority is at this time and patients in need of transplantation should be accurately will require many more cases to be accumulated can be The of new also to be to if those the need for any additional automatic MELD points to be assigned on the basis of liver disease and its associated that include for to the hepatic have improved the success rates for LT in patients with this However, these results have only been of Current evidence suggests for patients with in is not LT a reasonable at cure and that have little In addition, tumor rates are high for patients protocol criteria who to LT longer than after not of cancer other than waiting The results from these evidence for of increased priority for patients in these especially in of the results when patients waiting the of increased priority has not been to the risk of as this should as more data is a in patients with disease and can cause severe However, there is no evidence for associated with mortality who have and or its may be to LT and may be at increased risk of dropout, although this has not been For this expert opinion that prospective RRB be for such cases and that an increased priority can be if specific criteria are the amount of increased priority is not based on endpoint since an endpoint the risk factors have been established. There are of non-HCC or for which LT has been However, these are very few in and there is no data on risk for these cases. For this automatic of increased priority be where the disease has not been have very and should not increased priority prospective cases should individual the RRB process. opinion that hepatic to the liver and hepatic in disease are cases for prospective RRB and The extremely of these cases makes it that data will be for regarding priority ranking for these patients on the basis of their of these conditions are which makes development of on these patients such as and disease in liver and and in and In these cases, the MELD/PELD score is an reflection of the need for LT. For other conditions where there is no liver mortality endpoints are not appropriate and other such as risk of risk of failure, and HCC risk should be into There is insufficient evidence to automatic by RRBs for these conditions and they should be on a case-by-case Several of these conditions for to priority. after transplantation of a liver that is of insufficient liver for the or has a of to In cases, severe which has been associated with increased risk of There are no data available indicating these patients have risks of relative to other cases listed for or the MELD/PELD score these patients' mortality risk relative to other waiting LT candidates. is however, that mortality risk is the waiting list endpoint for these cases. In the absence of these expert opinion has that cases where a has been in which and/or ascites is in the early should automatic increase in priority to mortality risk. There are no data indicating that this estimate accurately the of increased risk these and such data to be prospectively collected. Although also as is a the liver is the of These cases can develop high and from the in the There are a few of LT for this condition with success rates to other and of disease. studies also indicate that although some cases to other cases with or without or Thus, mortality risk progression may be the waiting list endpoint for but the lack of risk factors makes of this at this For this there is insufficient evidence to automatic increases in MELD/PELD score priority and these cases should be by RRBs on an individual Since these patients not develop liver failure, it is that their MELD/PELD scores will adequately estimate their mortality risk. However, since these are rare cases, there is no available evidence on which to a quantification of mortality risk at this liver MELD, Model for End-Stage Liver Pediatric End-Stage Liver HCC, hepatocellular Regional Review MELD Exceptional Case HE, hepatic especially of the magnitude of liver would be with the quality evidence from and would be consistent However, this of evidence is not in the field of liver The MELD/PELD allocation system for cases is based on 2 evidence, from and prospective and More as as policy making is are these so-called of these conditions, as so that studies or results that are will be and prioritization policy will have to on expert opinion However, as indicated for some conditions, there are endpoints that might be addressed and data that might be accumulated to a more objective assessment of these patients' need for For many of these cases, there may be where mortality risk may serve as a reasonable prioritization endpoint and could be equated to the MELD/PELD score mortality risk estimates for the cases. there is a near complete lack of data from which risk factors that would accurately define the mortality risk, or list removals for for these conditions might be and data collection by with in specific should be as in the case of the in hepatic disease effort by the of in the majority of these cases, factors other than those in the MELD/PELD score are likely to a and a effort to these data will be data elements to be addressed are in Table 1. For most cases a risk of mortality is deemed to be most appropriate, but since the evidence suggests that the MELD/PELD score does not adequately measure mortality risk for these cases, additional data are needed. Automatic assignment of priority is indicated based on sufficient evidence, expert opinion, or established practice as For a few of the conditions MELD/PELD might be a reasonable estimate of mortality risk and these should be analyzed and For other conditions, mortality risk is not an appropriate of LT need and risk for waiting list removals and progression of must be We have endpoints indicated in the literature that might serve to these For using the which is analogous to the MELD/PELD score in patients with might be a more and for assigning priority of the methods we currently some measure of with the appropriate into a MELD score for patients with might also be appropriate and more of need than the arbitrary, we currently is clear is that endpoints alone are not sufficient for prioritization because they are subjective and be equated to mortality In all of the cases where automatic priority increases are recommended, a amount of MELD/PELD points to be has been These and are not of the of disease and rates of progression and not adequately the need for of the endpoint used. For these arbitrary, more of LT need must be to more the actual of these however, require a much more data collection system. a national review process that data collection methods and and that priority on a consistent basis to be to these with this approach can consistent priority be assigned, by RRBs national standards to the or using a national system that for regional The MESSAGE has an excellent of the available data for these but as it has the evidence that these may not be for some of the conditions, but patients a much more approach to most of these cases. This can only be with a approach on a national that the more exceptional for data to the HCC example Waiting list endpoints need to be and risk factors for progression to these endpoints must be The MESSAGE and the of the Liver and Committee who were in this would like to that the Regional Review Boards the country have held on which conditions should additional MELD points if and on much priority should be for exceptional case The of the MESSAGE was to a approach to listing of patients with additional MELD points the United Network for Organ

Related Papers

• → Model for end-stage liver disease (MELD) and allocation of donor livers(2003)2,508 cited
• → Systematic review: the model for end‐stage liver disease – should it replace Child‐Pugh's classification for assessing prognosis in cirrhosis?(2005)419 cited
• → Is severe progressive liver disease caused by alpha‐1‐antitrypsin deficiency more common in children or adults?(2016)45 cited
• → Liver disease and anesthesia(1993)1 cited
• → A Comparison of Child-Turcotte-Pugh Classification with Four Models of End-Stage Liver Disease-Based Prognostic System for Spontaneous Bacterial Peritionitis in Cirrhosis(2011)