Liver Allocation Patient Q&As

These FAQs should be read in conjunction with the NHSBT Organ and Transplant Directorate (Liver Advisory Group) Liver Allocation Proposals LAG 14(31)a,b,c,d

The questions were proposed by LPTC. Answers from Dr Alex Gimson, 31 December 2014


Model Summary

Model Basis
Model 1 (M1) - need

Greatest risk of dying on a transplant list

4827 adult elective registrations between Jan ’06 and Dec ’12 – then living/domino/ multi-organ transplants were excluded, leaving 4519 registrations.

Model 2 (M2) - utility

Utility: greatest chance of surviving after transplantation

3484 orthotopic only liver transplants between Jan ’06 and Dec ’12 – using both DBD and DCD donors (2925 complete case cohort)

Model 3 (M3) – transplant benefit

Greatest net life years gained

Derived for each potential recipient by subtracting the projected survival without transplantation (M1) from their projected survival after transplantation (M2) for each donor whose liver was donated during 2013

Simulation cohort

Candidates on waiting list from Jan ’13 to Dec 13

(1287 registrations and 629 transplants)



  1. What was the major issue that the simulation tried to address?
  • The primary issue that the simulation tried to answer was what was the best process of offering liver organs for transplantation.
  1. When saying one method of offering is better than another what does better mean?
  • Better was defined as the offering process that resulted in the fewest patients dying before and after a transplant, or put another way the offering process which was predicted to have the most patients surviving overall taken from the point of registration.
  1. What was the main outcome of the simulation?
  • It demonstrated that offering organs based on transplant benefit was predicted to result in the best outcome, with more patients surviving (fewer dying before or after the transplant).  The current unit based allocation was worse than the other three tested models of offering organs.
  1. What is a Group 1 registration?
  • There are two categories of patient: Group 1 and Group 2. Group 1 patients are ordinarily resident in the UK and qualify for NHS treatment or are entitled under reciprocal health agreements to medical treatment in the UK. There are very few Group 2 cases. For more information, see NHSBT website
  1. Why is the cancer cohort from different year groups to the non-cancer cohort?
  • Patients registered between 2006 and 2012 to estimate the ‘non-cancer needs model’, patients registered in 2009-2012 to estimate the ‘cancer needs model’.
  • For the cancer cohort, information regarding the maximum AFP level and number and size of tumours was not collected until September 2007 and data was not fully reported until January 2009. Therefore, the time period for the cancer cohort was moved to include all registrations between 1 January 2009 and 31 December 2012. The time period for the non-cancer cohort remained as registrations between 1 January 2006 and 31 December 2012.
  1. In M1 there were 4827 elective registrations. Why are only 4519 registrations used in the modelling?
  • Living/domino/multi-organ transplants +incomplete info were excluded
  1. In M2, 3484 orthotopic liver transplants took place. Why are only 2925 transplants used?
  • Incomplete case information therefore excluded
  1. Is there an assumption in these models that a national allocation system will be introduced?
  • No. The models can be applied using the existing zonal distribution system. They could also be applied to a national system but if that was the case, then further work would be required to ensure consistency of UKELD component sample analysis between centres.
  1. Will transplant centres have to use the organ they are offered?
  • No, this is just an offering sequence. The models propose to transparently identify the next potential recipient based on the modelling criteria. Each Unit would then make a decision as to whether or not that organ is appropriate. If it isn’t, the organ would be offered to the next proposed patient.
  1. What is UKELD?
  • UKELD is the United Kingdom Model for End Stage Liver Disease. It is a scoring system used to predict prognosis for patients with cirrhosis on a transplant waiting list.

Back to top

  1. Why do the results tables not show the median UKELD scores?
  • The modelling uses a more sophisticated calculation which analyses the four UKELD components separately as well as considering albumin.
  1. Do the models use UKELD as a predictor of severity of disease?
  • No. They use the separate components of the UKELD score: creatinine, sodium, bilirubin and INR (international normalized ratio). The formula predicts death on the transplant list, uses many factors including those in UKELD and is more accurate than UKELD alone.
  1. Has variability between centres’ sample analysis been considered?
  • Yes, in detail. We’ve concluded that unit-based correction factors for some of the tests will be required for any national liver transplant allocation scheme.
  1. What are LUVS?
  • LUVS are low UKELD variant syndromes. They are rare and account for about 3% of all transplants. In those cases UKELD may not reflect severity of their liver disease or assess the quality of life of some patients who are being transplanted for severe symptoms (poor quality of life) but no or little risk of dying from their disease.  Some adjustment should be made to allow these patients fair access to transplant, because they are disadvantaged by UKELD scoring alone. The conditions we considered are shown in the Table 1.
  1. How will this be corrected?
  • We considered four options for this group of patients:
  1. No correction
  2. Escalating priority UKELD points based on time on waiting list
  3. Adding priority points based on median of all registrants so their predicted time to offer of a donor was the same as the median for all patients
  4. Identifying a proportion of donors equal to the proportion of LUV registrations each year and utilize that donor pool for them.
  • More work needs to be done of this small group of patients to tease out the consequences of different offering sequences for this group, but our initial proposal is considering proportional allocation.


Table 1: Summary of Variant Syndromes and decisions made



Proportional Allocation correction

Additional UKELD points

Option agreed at Nov LAG

(not yet agreed by patients)

Familial Hypercholesterolaemia/ Primary Hyperlipidemia *

Proportional allocation based on % of registrations

Additional Points increasing with time

Proportional allocation based on % of registrations

Intractable Pruritus (disadvantaged due to quality of life, not risk of death)*

Proportional allocation based on % of registrations

Additional Points increasing with time

Proportional allocation based on % of registrations

Cholangiopathy with refractory cholangitis or intrahepatic sepsis


Possible triggers>1 episode of sepsis with MOF; sepsis with multiple antibiotic resistance; dependence on external biliary drainage.

Proportional allocation based on % of registrations

Hepatopulmonary Syndrome (HPS)*



Proportional allocation based on % of registrations

Polycystic Liver Disease (PCLD)*


Agreed malnutrition indicator

Proportional allocation based on % of registrations

Chronic Hepatic Encephalopathy (disadvantaged due to quality of life, not risk of death)

Proportional allocation based on % of registrations

Points increasing after set time


No correction needed

Diuretic Resistant Ascites (DRA)*


Agreed malnutrition indicator

No correction needed

Familial amyloidosis/ Familial Amyloid Polyneuropathy


Transplanted according to UKELD of domino recipient



No correction needed

Sickle Cell Hepatopathy*


No correction needed

Gylycogen Storage Disease

(Other liver disease)

Proportional allocation based on % of registrations

Additional Points increasing with time

Proportional allocation based on % of registrations?

Portopulmonary Hypertension (PPH)

(Other liver disease)


MPAP>25mmHg + reversibility

Proportional allocation based on % of registrations?


*existing variant syndromes

Back to top

  1. What are the advantages/disadvantages of proportional allocation for LUVS?
  • It is fairer because if the proportion of LUVs as a % of all patients waiting for a transplant goes up or down, so will the proportion of organs offered to them.
  1. What are the advantages/disadvantages of additional UKELD points for LUVS?
  •  Would allow LUVs to have a better chance of receiving an organ
  • The disadvantage is knowing what number of extra points to give them; 10, 15, 20. How would you decide how many points to give them?
  1. Why are you recommending proportional allocation for LUVS?
  1. The data in Table 5 of the Results section (LAG 14(31)d) show several groups who get less organs compared to the current scheme, eg: patients under 50, cancer, PSC, AID and a cohort of ‘other liver disease’. How do you explain this?
  • The Tables all show the number of organs that would be offered in the first 12 months only. This does not mean that they will not get a transplant simply that they don’t get it in the first year. The reason they don’t get it in the first year is because there were other cases with a much higher risk of death or transplant benefit – so those cases were offered the organ first.  The models have been created to reduce death on the waiting list and to increase the number of life years gained. The patients not selected in the first 12 months were not as likely to die/would gain more life years than the ones not selected.
  • Some groups would wait a shorter period for their transplant (so reducing their risk of death) and some as above may wait longer. Importantly their risk of dying on the transplant list for all patients is lower than in the current system. 
  • The Liver Allocation Review was instigated because the current system was not considered to be transparent and that alternative models could reduce deaths on the waiting list and increase the overall life years gained.
  1. Because these patients are competing against newly registered patients all the time, what will happen to these patients after 12 months?
  • New patients onto the waiting list have variable severity of liver disease. Some will come on the list with a lower risk of death compared to those already on it, and some a higher risk.
  • We agreed at the outset that the time spent waiting for a transplant was less important than preventing people dying on the waiting list. Reducing the latter was the main goal not reducing the time spent waiting for a transplant.
  • If you prioritise offers on the basis of who has waited longest then some very sick people newly on a transplant list will die because they have not waited as long as some much less sick candidates.
  • Importantly in the transplant benefit or needs based system each patient’s risk of dying without a transplant is the same. The models aim to equalize that risk across all patients and causes of liver disease.

Back to top

  1. How have the prognoses and number of life years gained been calculated?
  • All the prognostic models take aetiology, the cause of the liver disease, in to account so for instance PSC is specifically identified in the model. In fact PSC cases have a low risk of dying on the transplant list as was shown in the tables.
  • Life years are calculated by estimating survival time with a transplant (M2) and then subtracting the survival time without a transplant (M1) to give the net life years gained.
  1. Is there flexibility in the proposed model to adapt when say, certain diseases become stratified or when curative treatment becomes available for specific diseases?
  • If new curative treatments become available that will be taken in to account because we intend to build into the process three reviews over the next 5 years when such alteration might be introduced.
  1. Are both DCDs and DBDs used in the simulation?
  • Yes. It is important to include all options for patients waiting on the list.
  1. What if an organ was offered to a centre, and the centre didn’t feel the organ was appropriate for the candidate identified?
  • The organ would then be offered to the next candidate identified by the model, not by the centre.
  • In a zonal allocation system, the organ would be offered to the next identified patient on the zonal waiting list, irrespective of the patient’s unit. In a national allocation system, the organ would be offered to the next identified patient on the national waiting list, irrespective of the patient’s unit.
  1. How can we be sure the statistical methodology is correct and any assumptions made in the modelling are reasonable?
  • The modelling will be peer reviewed by in medical journals.
  • Professor David Collett who leads the Statistics and Audit team at NHSBT is a recognized international expert in this area and has written a number of books on this topic.
  1. What effect will the use of the new, effective antiviral therapies for Hepatitis C have on the modelling? Is there a mechanism in the proposed model to allow for changes in cohorts and for reviews to take place?
  1. Is the current IT system able to adequately handle the proposed model?
  • An IT system will be developed to handle the offering sequence.  This has been successfully done with the kidney allocation scheme which handles bigger numbers than the liver allocation scheme will.
  1. Will patients be able access information on how long they are likely to wait for their transplant, based on their current condition?
  • It will be possible to tell patients an average waiting time depending on the state of their liver blood tests and other parameters. That might change with time as their blood tests change.

Back to top

Additional Glossary Terms




Cold Ischemic Time


Donation after brain death


Donation after circulatory death


Fixed Term Working Unit


Hepatocellular Carcinoma


Liver Advisory Group


Low UKELD Variant Syndromes


Multidisciplinary team


United Kingdom Model for End Stage Liver Disease


Weblinks in this document

LAG (Liver Advisory Group)

Liver Allocation Proposal and Background LAG 14(31)a,b,c,d

Liver Patients’ Transplant Consortium

NHSBT Group 1


Back to top