Current 10- and 20-year survival rates after liver transplantation exceed 65 and 50% respectively1 2 and the number of people who might benefit from transplantation has increased. Over the last decade, the number of liver transplants performed in the UK has fallen whilst the numbers waiting has risen by over 90%.3 The increasing imbalance between the number of patients who might benefit and the number of organs available from deceased donors has been masked by surgical innovations, with use of extended criteria donors, split livers, organs from non-heart beating donors or from living donors adding to the donor pool.3

Because of the shortfall between organ donor supply and recipient need there will have to be some element of rationing. The basis for rationing should be clear, overt and, where possible, evidence based. Criteria for the selection of cases to a waiting list for organ transplantation must be distinguished from criteria for the subsequent allocation of those cases to a specific donor organ. Both should be agreed after public debate, based on sound ethical principles. Whilst transplant clinicians would be responsible for implementation, they must remain publicly and professionally accountable. The principles underlying both selection and allocation of organs from living donors are different from those underlying the use of organs from deceased donors since the former are given to a specific individual whereas the latter are donated to the donor pool.

Balancing the demands of individual autonomy of the recipient and utility of the donor organ on a background of justice and equity are often difficult and in conflict. To provide a rational approach in the UK, a meeting of health care professionals, patient representatives, ethicists and the public met in Edinburgh in 19964 and agreed the following principles:

• Donor organs are a national resource.

• Patients should be selected if the expected survival in the absence of transplantation is 1 year or less, or if the patient had an unacceptable quality of life because of liver disease.

• There is an expectation that those transplanted would have a survival probability of at least 50% at 5 years with the quality of life acceptable to the patient.

Even with those guidelines, patients continue to be denied transplantation not because they would not benefit from that procedure but because they would not benefit enough. With the current system for the delivery of transplant services in the UK (seven designated units to serve a population of 60 million), equity of access both to the transplant list and to grafts must be maintained and open to inspection. With the increasing pressure on the diminishing resources health care professionals in the UK recently met again to develop more precise guidelines to determine the basis for selection to the waiting list and the subsequent allocation of donors to those patients. The group consisted of representatives of all seven designated UK Liver Transplant Units and met under the auspices of the Liver Advisory Group of UK Transplant (the NHS agency responsible to the English Department of Health for the administration of transplantation in the UK). All but one are liver transplant physicians; M.A. is a transplant surgeon. A.G. and J.N. chaired the group. The consensus document was unanimously agreed by all representatives and subsequently endorsed by the Liver Advisory Group and the individual transplant groups. The meetings were financed through the NHS trusts of the representatives. There are no conflicts of interest from any of the authors.

The purpose of this article is to indicate the current new criteria for selection, de-selection and allocation for adult liver transplantation in the United Kingdom and to indicate differences with other health care systems. These criteria apply to adults been considered for liver only grafts from deceased donors.

SELECTION FOR A LIVER TRANSPLANT WAITING LIST

Principles of selection to a waiting list

The number of transplants that can be done using grafts from deceased donors is limited by availability of suitable grafts: until the number of useable grafts increases, any expansion of indications must be balanced by a reduction in other patients who might have received that graft. Although the true need for transplantation is not known, the current system cannot manage the demand since approximately 14% of all patients currently die or become too sick before a graft is available.3 Selection of patients to the waiting list may be done simply on need; that is, would the recipient benefit in terms of either quality or quantity of life? With this approach, the list would increase considerably and many patients would have no hope of receiving a transplant. This approach we believe is ethically unacceptable although we accept that there are other valid views. An alternative approach is to set selection criteria so that those on the list would have a reasonable expectation of being transplanted within a period such that their health would not deteriorate so far that a graft would carry an unacceptably high risk. The criteria could be set at either a fixed number of recipients on the list at any one time or else selecting patients who meet the criteria but the criteria being determined to ensure that the list did not become too large. We have opted for this latter approach.

The principles of selection include:

1. Selection is based primarily on the risk of death without a transplantation

2. Selection is secondarily assessed on the ability of transplant to improve quality of life

3. All cases being regularly reviewed to ensure that they continue to meet criteria and have neither improved nor become too sick to benefit from transplantation

4. Criteria for removal from the list to be agreed and applied to all patients

Selection criteria for super-urgent liver transplantation

The selection criteria, based on published risk factors5 6 for super-urgent transplantation are shown in table 1. The list is stratified by blood group and time on the super-urgent waiting list; these cases can be offered a donor from any region within the United Kingdom.

Selection criteria for elective adult transplant waiting lists

All cases selected for elective adult liver transplant must have a projected 5-year survival after transplantation of greater than 50%. Those with alcoholic liver disease must comply with previous recommendations concerning abstinence from alcohol, compliance with treatment, subsequent follow-up and ongoing therapy.7 Where appropriate, those cases with prior illegal substance use must also comply with current guidance (available on www.uktransplant.nhs.uk). All cases for selection to the elective adult transplant waiting list must fall into one of three categories.

Category 1. Projected 1-year liver disease mortality without transplantation of >9%

The current 1-year post-liver transplant mortality in the UK is approximately 9% and therefore all cases selected under this category must have an expected mortality without transplantation of greater than that 9%. Several models have been developed to assess prognosis, with and without transplantation. Although these are powerful tools, caution must be used when data obtained from retrospective analyses are applied to other populations; furthermore, models which are valid for populations usually have wide confidence intervals when applied to individuals. For the clinical components (such as encephalopathy), there may be variations in interpretation; for blood analytes, there are potential inherent errors since different results from laboratories may vary although currently all laboratories in the transplant units measure serum bilirubin and creatinine concentrations using the same methodologies.

The UK Liver Transplant Units have collaborated in developing a new scoring system to predict waiting list mortality– UKELD score (United Kingdom Model for end-stage liver disease)– constituted from the patient’s serum bilirubin, INR, creatinine and sodium.8 This score and a calculator is available on the UKT website (www.uktransplant.nhs.uk). The score was developed by analysis of 1103 patients and validated in a separate prospective cohort of 452 patients. A UKELD score greater than 49 predicts a greater than 9% 1-year mortality and is the minimum criteria for entry to the waiting list under this category. It may be that, to ensure that those who are placed on the transplant list do have a realistic chance of receiving a graft, this level will need to be adjusted upwards. As the variables of the score are dynamic and will change with treatment; thus, UKELD score will need to be re-assessed over time.

Cases with alcoholic liver disease, past intravenous drug use or current methadone users must be assessed as recommended in prior Liver Advisory Group documents.7 Professional assessment of the risk of recurrence of that risk behaviour will be undertaken in either a written agreement signed by the patient or a letter sent from the transplanting unit stating the criteria and which transplantation has been offered, the contents of which are agreed with the patient. Cases with porto-pulmonary syndrome must have demonstrated a mean pulmonary artery pressure less than 50 mm Hg, and should have had a clinically significant response to one of Lanoprost, Sildenafil or Bosentan.9

Category 2. Hepatocellular carcinoma

A lesion must be present and in the same location on at least two radiological modalities from ultrasound, triple phase CT scan, MRI scan or hepatic arteriogram. A single lesion must be ⩽5 cm in diameter or if multiple less than or equal to three lesions all less than or equal to 3 cm in diameter. The widest diameter recorded by CT scan is judged to be the maximum tumour diameter. It is recognised that different imaging modalities may identify differences both in number and size of tumour10 but to qualify as an HCC will require a congruent lesion to be seen on a minimum of two different radiological modalities. There must be no evidence of vascular invasion and no distant metastases. It is recognised that some have argued that these criteria could be expanded without affecting long-term tumour-free survival11^–13 and may be modified by the use of some immunosuppressive agents such as sirolimus that have a possible anti-cancer effect14 15 but we feel that given the current organ shortage and the lack of long-term data, no change should be made in the selection criteria. The role of pre-transplant interventions such as embolisation, radio-frequency ablation or ethanol injection remain uncertain16 and should not affect the transplant criteria: we believe that, at present, there is no evidence that down-sizing a tumour should affect the indication. These criteria will remain under review.

Category 3. Variant syndromes

It is recognised that disease severity scoring systems may not adequately describe all patients who might benefit from liver transplantation included in the above criteria. Selection to the adult liver transplant waiting list may also be allowed under one of a number of variant syndromes (table 2).

Whilst the introduction of these variant remains contentious and is expected to cover fewer than 10% of all registrations for an adult elective liver transplant it will be subject to regular review. Diuretic resistant ascites carries a poor prognosis16 and may not be adequately reflected in a scoring systems such as MELD (model for end-stage liver disease).17 18 Chronic hepatic encephalopathy may result in a dramatic deterioration in quality of life independent of any risk of mortality, as does persistent, intractable pruritus. On occasions those may be acceptable indications for liver transplantation. Lethargy is not considered an accepted primary indication for orthotopic liver transplantation since improvement is often small after transplantation.19 Porto-pulmonary hypertension and cystic fibrosis were not considered variant syndromes but could be transplanted under the standard UKELD score criteria. It remains to be established whether indications such as recurrent cholangitis will become indications.

Those cases that did not satisfy any of the above criteria where the clinician feels that transplantation is appropriate, may be referred to a national appeals panel.

De-selection criteria

As liver disease progresses, the chances of a patient benefiting from a transplant to comply with the 50% 5-year survival criterion will fall. Data from the US and UK suggests that this point is reached when the MELD score approaches 4020 21 but this is based on past experience and careful donor/recipient matching so extrapolation to the non-transplanted situation must be done with caution. Until evidence-based guidelines can be developed, it is proposed not to propose de-selection criteria except for those with liver cell cancer where the tumour exceeds agreed criteria.

ALLOCATION OF DONOR ORGANS TO WAITING LIST RECIPIENTS

Principles of allocation

The rationale of this allocation system is both to maximise the utility of the organ and to reduce the mortality of those on the waiting list. These two aims may be in contradiction as those who are very sick and thus at greatest risk of death may have a worse outcome and will use more resources. The outcome of a transplant depends on a number of factors. Whilst splitting livers may benefit two recipients it has the unwanted consequence that an optimal liver will become for the adult recipient a liver with a potentially worse prognosis and with the consequent risk of increased morbidity, graft loss and mortality.22 Despite that, the smaller, usually paediatric recipient, would benefit and be at a lower risk of dying without a graft.23 There is also the need to balance competing demands between the requirements of a person dying from liver disease against another whose liver function is relatively well preserved but whose life is intolerable because of intractable pruritus.

Matching donor and recipient is important to the success of the procedure.24 Not all organs function well immediately after grafting and those extended donor criteria grafts fair even worse when the cold ischaemia time is prolonged. A number of models predicting the outcome based on both donor and recipient factors have been developed25 but confidence intervals are wide and the clinical judgment of the transplanting team has been considered to remain key. Thus any system for allocation will need to include a mechanism for ensuring that the responsible clinician has the final decision in determining whether to use the given graft.

Allocation of donor to recipient

To maximise the use of donated livers, the first priority will be to decide whether to split the liver. A donor liver will be considered for splitting if the donor is aged less than 40 years, weighs more than 40 kg and has been in intensive care for fewer than 5 days. It is, however, the responsibility of the retrieval surgeon to decide whether livers from donors who meet these criteria will be split as this will depend on other factors, such as the degree of steatosis. The surgeon will also decide whether the recipients will be an adult and a child or two adults.

A donor liver will be offered first to the case with the greatest predicted short-term need. Where there are two cases of identical transplant need the liver will be allocated to the candidate who entered the waiting list first. The chance of receiving a donor organ should be the same for all cases of similar disease severity irrespective of which centre has registered the patient. If a unit does not wish to use a liver offered for an individual case then it will pass to the case next on the ranked priority list. The decision to split a liver for the benefit of a paediatric and an adult recipient should be taken prior to the offering of one of the segments to each recipient. A top band of elective recipients will be identified using their UKELD score such that those with the very highest risk of short term mortality can utilise a larger donor pool than that available to those on each unit’s separate waiting list.

If there is no suitable super-urgent recipient, then the organ is offered to the unit in the zone served by that unit; if there is no suitable recipient, then the organ is offered to other units. The responsible surgeon (usually after discussion with the physician) will decide on the optimal recipient(s), after assessing all the donor issues, such as ischaemic times, donor age, donor risk and recipient issues. The current system works well; modelling clearly demonstrates that those with malignancy have no greater chance of dying awaiting a graft (or being removed from the list because of tumour growth) than those with chronic liver disease (Gimson, Dawwas and van der Meulen, unpublished data). The system also results in those who are listed because of intolerable quality of life have a reasonable expectation of receiving a graft within 12 months.

Future system of allocation

In order for all cases on a transplant waiting list to have an equal chance of receiving a graft independent of the severity of their liver disease there will need to be a similar fixed relationship between the number of registrants on a waiting list each year and the number of donors available to that recipient pool.

The long-term aim is to develop a model of allocation based on the greatest transplant benefit which would take into account both the likelihood of dying without a transplant as well as the likelihood of dying following a transplant. Such a transplant benefit model may take some time to develop and the current proposals represent an interim accommodation. Further consideration will need to be given as how to balance the needs of those being grafted because of poor survival with those being grafted for an intolerable quality of life. In those systems where allocation of deceased donor livers is based solely on the risk of dying, those being grafted for quality of life issues will not be given a graft unless there is the possibility of receiving additional “UKELD points”. Similar consideration will need to be given to those with liver cell cancer where the risk of dying or being withdrawn from the list because of tumour growth is not dependent on the liver function. It is possible that we will develop a system similar to that in the US where additional points are given to give an equivalent non-transplant survival.

PARAMETERS FOR MAINTENANCE AND MONITORING OF THE POLICY

In order to monitor this policy and ensure that it meets the principles of both selection and allocation, there will be close monitoring of waiting list mortality between all units, crude and risk adjusted post-transplant mortality at 90 days, 1, 3 and 5 years and disease specific patient and graft survival at those times.

THE NEW SYSTEM IN CONTEXT

Comparison with other national liver selection and allocation systems

The optimal approach to selection and allocation will depend on many different factors: these will include the burden of liver disease, the requirements for and provision and funding of liver transplantation, the availability of different types of donor livers and national geography, as well the philosophy underling the aims of the selection/allocation process.

In the United States minimal criteria for listing were agreed in 199726 but despite these, the list grew so, in 2002 and after full public discussion and debate, the United Network for Organ Sharing (UNOS) introduced a system that allocates organs to identified adult patients based on a disease severity score, MELD score, which is a validated marker of risk of early death without transplantation. The aim was to reduce the mortality of patients on the waiting list and has been effective.27 28 The system is similar to that described here although a different scoring system, will be utilised. UNOS has recommended certain disease categories as having mortalities not adequately reflected by MELD and so being able to receive additional scoring points, similar to the proposal above relating to variant syndromes. Repeated evaluation of the UNOS system has suggested that it has reduced pre-transplant mortality and there has been little increase in post transplant mortality.2 Some concern remains as to whether the shorter waiting time for those with liver cancer may result in more recurrence.29 Under the UNOS system the proportion of grafts offered to recipients at the top of a MELD stratified list, which are turned down is also very considerable which implies that some transplant teams may have concerns as to whether a system based upon MELD alone is the best way of allocating donors to recipients. Nevertheless the current system is significantly more proscriptive as to which patients may enter the waiting list in the first place.

The practice of selection and allocation varies greatly between different systems outside North America. In some systems, such as Australia, Greece and the Scandinavian countries there are, as far as we can ascertain after discussion with colleagues in these countries, no formal agreements about listing criteria and, with the exception of super-urgent patients, allocation is left to the discretion of the transplant clinicians in the relevant centre. Germany has a mixed system. In Switzerland and Spain,30 there are nationally agreed and published criteria for selection but allocation is left to the discretion of the local centre. Some have adopted the MELD system in some regions. Several other countries, such as France, Israel, the Eurotransplant countries and Brazil31 32 systems have followed the example of the United States with nationally agreed selection criteria and central allocation with the graft being offered to identified patients, usually selected according to blood group, size and MELD score.

Potential concerns with the proposed system

There are several concerns with the new system about potential problems. The introduction of nationally agreed and clear criteria for both selection and listing may lead to an increased bureaucracy without improving equity, utility or benefit. Implementation of selection criteria may not allow for an individualised approach to the patient and some people who would benefit may not be accepted for listing: for this reason, we have introduced an appeals system. Allocation of a liver to an individual required careful matching of the donor and liver and there are many factors that might affect the decision to select one individual over another. Experience in the US has shown that the liver is accepted for the first-offered patient in only 45% of patients.2 Prognostic models which are derived from retrospective studies of patients who have been transplanted may be valid when applied to populations become less accurate when applied to potential transplant candidates who have not been transplanted and may receive a pattern of care different to those from which the test population were drawn. Other concerns are that a formal, protocol-based system will inhibit the development of new approaches and innovation. We believe that this problem can be circumvented by allowing, after full discussion and agreement by relevant interested parties, transplantation in different indications according to a defined and audited protocol.

Current, limited evidence suggests that transplant surgeons are best able to determine the optimal matching of donors and recipients.

These concerns are valid and need recognition. Introduction of this system will require frequent review, modification and monitoring to show not only that transplant clinicians are adhering to the guidelines, but there is improved equity and utility. Finally, it must be stressed that the need for selection criteria for liver transplantation is due primarily to the lack of donor organs so every effort must be made to make full use of all potential organs for transplantation.