Selection of suitable recipients for liver transplantation: The evolution of the Model for End-Stage Liver Disease (MELD) score

Review

Hell J Surg. 2024 Oct-Dec;94(4):206–212
doi: 10.59869/24048

Filippos F. Karageorgos, Nikolaos Antoniadis, Georgios Katsanos, Georgios Tsoulfas

Department of Transplantation Surgery, Center for Research and Innovation in Solid Organ Transplantation, Aristotle University School of Medicine, Thessaloniki, Greece

Correspondence:  Georgios Tsoulfas, FACS, FICS, MD, PhD, Chief, Department of Transplantation Surgery, Center for Research and Innovation in Solid Organ Transplantation, Aristotle University School of Medicine, 49 Konstantinoupoleos Street, Thessaloniki 54622, Greece, e-mail: tsoulfasg@gmail.com

ABSTRACT

In 1967, Starlz and his colleagues performed the first successful human liver transplant. The greatest barrier to the widespread use of liver transplantation remains the shortage of cadaveric organs. To bridge the gap between the demand of organs and the supply of them for transplantation, a prioritization system was essential. This led to the development of the Model for End Stage Liver Disease (MELD) score, which is used until today to aid the prioritiation of transplantations. A literature search was conducted using the Scopus database, Pubmed and Google Scholar. Research studies examining the MELD score and its variations are presented. The use of the MELD score and its variations has been useful in transplantation, but their limitations are discussed along with the overall strategy that the transplant centers should follow. Despite the fact that the mathematical evaluation and characterisation of the status of a possible transplant recipient is utilised with the MELD score and its variations, there are more variables to be included in the final decision on who will be the optimum recipient at any given time.

Key Words: MELD score, liver transplantation, organ allocation, Child-Turcotte-Pugh score, MELD-3.0

Submission: 16.11.2024, Acceptance: 29.04.2025

INTRODUCTION

Liver is a crucial organ for the human body, since it is responsible for a variety of functions, including metabolism of nutrients, storage of vitamins and glycogen, synthesis of clotting factors and albumin, production of cholesterol and bile, among other functions. The first successful human liver transplantation (LT) was carried out by Starzl and associates in the 1960s, taking place in the United States of America. Over the years, LT has become the definitive treatment for patients suffering from acute liver failure, end-stage liver disease (ESLD), and, in some cases, hepatocellular carcinoma (HCC). The primary impediment to the increased use of LT is the limited availability of grafts. Until this day, no solution exists for creating a fully biocompatible, fully functional and long-lasting wearable or implantable artificial liver that can be transplanted into a patient and be comparable to a graft used in LT. Nevertheless, efforts have been made [1-3]. In other organs such as kidneys, there are cases of wearable and implantable devices of artificial kidney under development to fully replace renal function in patients with end-stage kidney disease (ESKD) [4]. Some of these devices have been tested in humans and others are close to human clinical trials [4]. Recently, a 2023 publication [1] reported the engineering of a clinical-grade bioartificial liver device. The device was used in pigs with post-hepatectomy liver failure and in humans used after extended liver resection. Furthermore, another recent publication of 2023 [2] presented a study of a wearable artificial liver of approximately 9 kg, marking a significant step toward the development of wearable artificial livers.

This lack of alternative graft options for liver transplants is the reason why mostly cadaveric livers are utilised. Additionally, living donors can provide part of their liver to be transplanted to a patient [5,6]. To balance the difference between the supply of organs and their demand suitable for transplant, a prioritising mechanism was essential. These systems include MELD-related scores for prioritisation. According to the Greek Transplant Organization (site), the criteria for liver transplants include blood group, the somatometric dimensions of the recipient in relation to the donor, waiting time on the transplant list, as well as the severity index of the recipient’s condition (Government Gazette 973/1996) [7]. In addition, Fouzas et al. (2019), mentions that the Child-Turcotte-Pugh (CTP) score, complications of portal hypertension and place on the waiting list are also used for the liver allocation [8].

In this review, MELD-related scores and pre-MELD era scores are presented and discussed. In addition, the limitations of MELD score and necessary additions in the scores are discussed.

METHODOLOGY

Here, for this review a literature search was conducted on three databases (i.e., Scopus, Pubmed and Google Scholar). Articles related to the MELD score and similar variations were selected and information was extracted on the MELD score and the similar scores. In addition, the snowballing technique in selected articles was used to find more MELD score variations.

PRE-MELD ERA

Child and Turcotte in 1964 reported the earliest classification scheme for predicting survival in patients receiving portosystemic shunt surgery who have liver cirrhosis complicated by variceal bleeding [9,10]. It included two laboratory values and three clinical variables. The five variables are:

  • serum bilirubin and albumin.
  • hepatic encephalopathy, ascites, and nutritional status

According to their mortality risk for major surgery, individuals with cirrhosis were categorised into three groups.

Pugh and associates revised the original Child-Turcotte classification in 1973 by changing nutritional status by prothrombin time (PT) and giving each of the five variables a score between 1 and 3 [9,10]. This revised score was named the Child-Turcotte-Pugh score or in brief CTP score.

The CTP score was calculated by United Network of Organ Sharing (UNOS) based on:

  • hepatic encephalopathy, serum bilirubin, albumin, the severity of ascites, and PT/international normalised ratio (INR).

The Child-Turcotte-Pugh score was widely used, but had limitations, (i.e., subjectivity in scoring, the lack of precision in predicting survival). The MELD score was introduced as a more objective and accurate scoring system for assessing liver disease severity.

MODEL FOR END-STAGE LIVER DISEASE (MELD) SCORE

Published in 2001 and used in the United States liver allocation system in 2002, the MELD score can be characterised as the first step towards selecting LT recipients for the patients with the highest mortality risk [10,11]. The Mayo End-Stage Liver Disease model was first developed as a tool for predicting short-term survival (three months) following transjugular intrahepatic portosystemic shunt (TIPS) in a cohort of cirrhosis patients. It was required to incorporate the logarithmic expression of three laboratory values to the calculating formula:

  • serum creatinine,
  • bilirubin, and
  • INR.

The latter (i.e., INR) reflects the extent of liver dysfunction and the underlying aetiology of chronic liver disease, providing an additional adjustment in clinical assessment.

Subsequently, the aetiology of liver disease was removed from the calculation and the scoring was renamed the “Model for End-Stage Liver Disease” (MELD) by the same team of researchers. MELD scores range from 6 to 40. When an organ becomes available, the greater the MELD score, the greater the likelihood of receiving a liver from a deceased donor. The MELD score is determined using the equation below:

MELD score = 3.78 · In(serum bilirubin [mg/dL]) + 11.2 – In(INR) + 9.57 · In(serum creatinine [mg/dL]) + 6.43

In the final score, higher values indicate a greater liver disease severity and an increased risk of mortality. One of the best indicators of how urgently a person needs a transplant is their MELD score. It is not the only factor, though. Other considerations include:

  • Recipient-to-donor body size compatibility
  • blood type matching
  • Availability and demand for deceased donor livers
  • Geographical considerations (i.e. proximity to the donor liver)

According to literature data, the mortality (rounded to the nearest percentage) for a MELD score in the range 0-7 is 30 and 90 days, the perioperative mortality is 6% and 10 %, respectively [12]. In addition, for a MELD score >25, the perioperative mortality remains at 90% for both for 30 and 90 days [12].

LIMITATIONS OF MELD SCORE

The incapacity of the MELD score system to appropriately prioritise patients undergoing LT, who also have a concurrent diagnosis of HCC, has been its most noteworthy drawback [9]. In addition, the MELD score is losing its predictive ability as the epidemiology of liver disease shifts [11]. When HCC patients are first diagnosed, their MELD scores are frequently low, which may lead them to underestimate the urgency for transplantation before the tumour progresses beyond the stage amenable to LT. The MELD score primarily focuses on three laboratory values: serum bilirubin, serum creatinine, and INR. However, it does not include other important factors, such as the aetiology of liver disease, the presence of complications, such as ascites or hepatic encephalopathy, or the patient’s overall clinical condition. The MELD score relies on serum creatinine levels to evaluate renal function. However, creatinine levels can be influenced by factors other than kidney function, such as hydration status and muscle mass. In patients with fluctuating fluid status or those with renal dysfunction not solely attributed to liver disease, the MELD score may not accurately reflect the true severity of liver disease. MELD score is a static measure calculated at a specific point in time. It does not reflect changes in liver disease severity or patient condition over time.

MELD VARIATIONS AND EVOLUTION

MELD-Na

A study from the United States of America (USA), published in 2006 [13], used data from an end-stage liver disease database of patients listed for liver transplantation at six USA transplant facilities. This led to the development of a MELD-Na score that incorporated serum Na into the existing MELD formula [13]. This model for end-stage liver disease-sodium (MELD-Na) score modified the existing model for end-stage liver disease score or MELD score equation only in the case of patients with serum Na between 120-135 mEq/L. The following equation defines the MELD-Na score [13]:

MELD – Na score = MELD + 1.59 · (135 – Na)

Another study from the USA, published in 2008, gives a different MELD-Na equation [14]. In this study, the authors used data gotten from all adult candidates for primary liver transplantation that were registered with the Organ Procurement and Transplantation Network in 2005 and 2006 [14]. The equation is for cases where the serum sodium concentration is bound in the range 125 and 140 mmol/L, and is given by:

MELD – Na score = MELD – (Sodium mmol/L) – [0.025 · MELD · (140 – Sodium mmol/L)] + 140

Then in 2015 found that the survival benefit of LT increased significantly with decreasing serum Na values when candidates’ MELD scores were >11. Thus, the new agreed-upon equation for MELD-Na calculation [15] is given by:

MELD – Na score = MELD + 1.32 · (137 – Na) – [0.033 · MELD · (137 – Na)]

where the serum Na concentration is bound between the range 125-137 mmol/L, [16]

MESO

Researchers from Taiwan developed the MELD to serum sodium ratio (MESO), in a retrospective analysis of 213 cirrhotic patients [17]. The MESO index was developed so as to enhance the opposing effect of serum sodium and MELD on outcome prediction [17]. It is uded to estimate the three-month mortality ratio and has limited clinical application [17]. The MESO index was introduced in 2007 and is calculated using the following equation:

MESO = (MELD Score) / (Serum Na in mEq/L) · 10

iMELD

The integrated MELD score (iMELD), which integrates serum (Na) and the age of the recipient is an alternative to the MELD formula. It was developed through a retrospective analysis of 310 patients who underwent elective transjugular intrahepatic portosystemic shunt (TIPS) [18]. The iMELD model was better than MELD in the prediction of the 12-month mortality [18]. The iMELD score was introduced in 2007 as the MESO score. The MELD score, age (years), and Na (mmol/L) served as the foundation for the iMELD equation [18]. It is calculated using the following equation:

iMELD = MELD + (0.3 · age[years]) – (0.7 + Na[mmol/L]) + 100

UKELD

In the United Kingdom, the UK Liver Transplant Units have collaborated to develop of a new scoring system aimed at predicting waiting list mortality [19]. This was called the UKELD score or the UK Model for end-stage liver disease. This score is calculated from patient’s serum bilirubin, INR, sodium and creatinine. A separate prospective cohort of 452 patients served as validation for the UKELD score, which was created through examination of 1,103 patients [19]. Based on Barber et al. (2011), the score is calculated from the patient’s INR, serum creatinine, serum bilirubin and serum sodium using the following equation [20]:

UKELD = (5.395 · In(INR)) + (1.485 · In(creatinine)) + (3.13 · In(bilirubin)) – (81.565 · In(Na)) + 435

MELD-XI

Also known as the “MELD excluding INR,” this variation of the MELD score excludes the INR variable [21,22]. This modification was introduced since the MELD score is a good predictor of short-term mortality in cirrhosis, but it can overestimate risk when INR is artificially elevated by anticoagulation [22]. The equation for this adjusted MELD score is as follows [22]:

MELD – XI = 5.11 · In(Total bilirubin) + 11.76 · In(Creatinine) + 9.44

MELD-Albumin

MELD-albumin score is another variation of MELD score, but it includes albumin rather than INR [21]. It is given by the following equations depending on the level of albumin [21]:

If albumin ≥4.1 g/dL

MELD – Albumin = 11.2 · In(1) + 3.78 · In(Total bilirubin) + 9.57 · In(Creatinine) + 6.43

If albumin <4.1 g/dL

MELD – Albumin = 11.2 · In(1 + (4.1 – Albumin)) + 3.78 · In(Total bilirubin) + 9.57 · In(Creatinine) + 6.43

uMELD

The updated MELD (uMELD) score was introduced in 2008 [23]. It MELD assigns a higher weight to bilirubin, while reducing the weight of creatinine and the international normalised ratio [23]. The updated MELD predicts the waiting list mortality more accurately. Candidate ranking on the waiting list would change if liver allocation was done using the updated MELD. The uMELD score is calculated as [23,24]:

uMELD = 1.266 · In(1 + creatinine) + 0.94 · In(1 + bilirubin) + 1.658 · In(1 + INR)

MELD-Plus

Introduced in 2017 [25], the MELD-Plus is a novel risk score that accurately stratifies, after hospital admission, the short-term mortality rate of patients with cirrhosis [25]. MELD-Plus is able to estimate the 90-day mortality rate after a cirrhosis-related admission, as it is considered as an unbiased and well-validated score [25]. The MELD-Plus includes a nine-variable risk score that calculates the L factor, where these variables include total bilirubin, creatinine, INR, sodium, total cholesterol, albumin, white blood cells, length of stay and age [25,26]. The L factor is introduced in the following equation to find the 90-day mortality [25].

           MELD – Plus = p(90 day mortality) = exp (L)/1+ exp(L)

MELD-3.0

It was initially published in 2021 and includes two new variables (female sex and serum albumin), a reduction in the maximal serum creatinine of 4.0 mg/dL to 3.0 mg/dL, and two new interaction terms (i.e., between bilirubin and sodium and between albumin and creatinine) [16]. It is cacuated using the following equation [16]:

MELD 3.0 = 1.33 (if female) + [4.56 · In (bilirubin)] + [0.82 · (137 – Na)] – [0.24 · (137 – Na) · In(bilirubin)] + [9.09 · In(INR)] + [11.14 · In(creatinine)] + [1.85 · (3.5 – albumin)] – [1.83 · (3.5 – albunin) · In(creatinine)] + 6

THE MELD SCORE IN NON-TRANSPLANT SURGERIES

The MELD was initially created by Mayo Clinic researchers to predict patient survival after elective TIPS placement. Further studies demonstrated that the MELD can be used to predict mortality risk in non-transplant surgeries [27,28]. These cases may include hepatectomy for HCC, or surgeries other than LT in patients with liver cirrhosis due to alcoholic liver disease, viral aetiology or non-alcoholic steatohepatitis, among others [27,28]. Moreover, studies [29,30] have reported that the MELD score could be applied before considering patients for cardiac surgery (e.g., isolated tricuspid valve surgery or cardiac surgery with cardiopulmonary bypass).

DISCUSSION

This study focuses on the variations of the MELD score from its creation to this day (see Figure 1).

Figure 1. The MELD score and its variations found in the current paper.

Even though all the variants of the MELD score improve in a way the mortality prediction, it is true that more predictors (i.e., patient’s genetics, race, living conditions, climate conditions living in and even predictors not accurately measurable from person to person) should be taken into account to redefine the mortality and finally get a good, personalised predictor for all patients.

In addition, there is no consensus on the upper limit of the MELD score in which an LT would not benefit the patient. The MELD score has a top value of 40 and patient with this value could also be transplanted. LTs taking place in high values of MELD (i.e., 35-40) exist, but the outcome of the LT can be complex due to the patient’s condition. Overall, there is no MELD cut-off above which the LT is considered unsafe. However, the decision to proceed with an LT is more multidimensional, taking into account the transplant centre’s policy, the overall clinical condition of the patient, and the availability of donor grafts (whether limited or not). This decision should be made on a case-by-case basis.

The limitations of this study include its narrative review format, which implies that these reviews are non-systematic, meaning that there is a hierarchy of evidence identifying narrative reviews below other review forms [31]. In addition, more databases could be utilised (e.g., Web of Science) to include a broader range of relevant papers.

CONCLUSIONS

The MELD score and its variants have remained reliable predictors of mortality on the short-term waiting list. The MELD score continues to be improved, for as with the MELD 3.0, which actively addresses limitations such sex-based disparities in waitlist mortality. In general, more predictors should be utilised, such as a patient’s genetics, race and living conditions to create a better representative score to achieve a more personalised medicine approach.

AUTHOR CONTRIBUTIONS

Conceptualisation, G.T.; methodology, F.F.K., A.N., G.K and G.T.; formal analysis, F.F.K.; investigation, F.F.K., A.N., G.K. and G.T.; data curation, F.F.K.; writing—original draft preparation, F.F.K.; writing—review and editing, F.F.K., A.N., G.K. and G.T.; supervision, G.T.; project administration, F.F.K.. All authors have read and agreed to the published version of the manuscript.

ACKNOWLEDGMENTS

None.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

INSTITUTIONAL REVIEW BOARD STATEMENT

Not applicable.

INFORMED CONSENT STATEMENT

Not applicable.

FUNDING

The current research received no funding.

DATA AVAILABILITY STATEMENT

Data sharing is not applicable since no new data is generated.

REFERENCES
  1. Wang Y, Zheng Q, Sun Z, Wang C, Cen J, Zhang X, et al. Reversal of liver failure using a bioartificial liver device implanted with clinical-grade human-induced hepatocytes. Cell Stem Cell. 2023 May;30(5):617-31.e8.
  2. Wang Y, Wang S, He X, Li Y, Xu T, Xu L, et al. A breakthrough trial of an artificial liver without systemic heparinization in hyperbilirubinemia beagle models. Bioact Mater. 2023 Feb;20:651–62.
  3. Karageorgos F F, Neiros S, Karakasi KE, Vasileiadou S, Katsanos G, Antoniadis N, et al. Artificial kidney: Challenges and opportunities. World J Transplant [Internet]. 2024 Mar;14(1):89025 Available from: https://doi.org/10.5500/wjt.v14.i1.89025
  4. Karageorgos FF, Alexiou M, Tsoulfas G, Alexopoulos AH. Hydrogel-Based vascularized organ tissue engineering: A systematized review on abdominal organs. Gels [Internet]. 2024 Oct;10(10):653. Available from: https://doi.org/10.3390/gels10100653
  5. Krendl FJ, Bellotti R, Sapisochin G, Schaefer B, Tilg H, Scheidl S, et al. Transplant oncology–Current indications and strategies to advance the field. JHEP Reports [Internet]. 2024 Feb;6(2):100965. Available from: https://doi.org/10.1016/j.jhepr.2023.100965
  6. Karageorgos FF, Karakasi KE, Kofinas A, Antoniadis N, Katsanos G, Tsoulfas G. Evolving transplant oncology: evolving criteria for better decision-making. Diagnostics. 2025 Mar;15(7):820.
  7. Liver transplantation – Greek Transplantation Organization [in Greek] [Internet] [Accessed: 2024 Sep 8]. Available from: https://www.eom.gr/metamoscheysi-ipatos/
  8. Fouzas I. Introduction to organ transplantation [in Greek]. Greece: Parisianou Press; 2019.
  9. Machicao VI. Model for End-Stage Liver Disease–Sodium Score. Clin Liver Dis. 2017 May;21(2):275–87.
  10. Ruf A, Dirchwolf M, Freeman RB. From Child-Pugh to MELD score and beyond: Taking a walk down memory lane. Ann Hepatol [Internet]. 2022 Jan;27(1):100535. Available from: https://doi.org/10.1016/j.aohep.2021.100535
  11. Trivedi HD. The Evolution of the MELD Score and its implications in liver transplant allocation: A beginner’s guide for trainees. ACG Case Rep J [Internet]. 2022 May;9(5):e00763. Available from: https://doi.org/10.14309/crj.0000000000000763
  12. Jaffer AK, Smetana GW, Cohn S, Slawski B. Perioperative Medicine Update. J Gen Intern Med. 2009 Jul;24(7):863-71.
  13. Biggins SW, Kim WR, Terrault NA, Saab S, Balan V, Schiano T, et al. Evidence-Based incorporation of serum sodium concentration into MELD. Gastroenterology. 2006 May;130(6):1652–60.
  14. Kim WR, Biggins SW, Kremers WK, Wiesner RH, Kamath PS, Benson JT, et al. Hyponatremia and mortality among patients on the liver-transplant waiting list. N Engl J Med. 2008 Sep;359(10):1018–26.
  15. Kalra A, Wedd JP, Biggins SW. Changing prioritization for transplantation. Curr Opin Organ Transplant. 2016 Apr;21(2):120–6.
  16. Kim WR, Mannalithara A, Heimbach JK, Kamath PS, Asrani SK, Biggins SW, et al. MELD 3.0: The Model for End-Stage Liver Disease updated for the modern era. Gastroenterology. 2021 Dec;161(6):1887-95.e4.
  17. Huo T, Wang Y, Yang Y, Lin H, Lee P, Hou M, et al. Model for end‐stage liver disease score to serum sodium ratio index as a prognostic predictor and its correlation with portal pressure in patients with liver cirrhosis. Liver International. 2007 May;27(4):498–506.
  18. Luca A, Angermayr B, Bertolini G, Koenig F, Vizzini G, Ploner M, et al. An integrated MELD model including serum sodium and age improves the prediction of early mortality in patients with cirrhosis. Liver Transplantation. 2007 Aug;13(8):1174–80.
  19. Neuberger J, Gimson A, Davies M, Akyol M, O’Grady J, Burroughs A, et al. Selection of patients for liver transplantation and allocation of donated livers in the UK. Gut. 2008 Feb;57(2):252–7.
  20. Barber K, Madden S, Allen J, Collett D, Neuberger J, Gimson A. Elective liver transplant list mortality: Development of a United Kingdom End-Stage Liver Disease Score. Transplantation. 2011 Aug;92(4):469–76.
  21. Liao S, Lu X, Cheang I, Zhu X, Yin T, Yao W, et al. Prognostic value of the modified model for end-stage liver disease (MELD) score including albumin in acute heart failure. BMC Cardiovasc Disord [Internet]. 2021 Dec;21(1):128. Available from: https://doi.org/10.1186/s12872-021-01941-7
  22. Heuman DM, Mihas AA, Habib A, Gilles HS, Stravitz RT, Sanyal AJ, et al. MELD-XI: A rational approach to “sickest first” liver transplantation in cirrhotic patients requiring anticoagulant therapy. Liver Transplantation. 2007 Jan;13(1):30–7.
  23. Sharma P, Schaubel DE, Sima CS, Merion RM, Lok ASF. Re-weighting the model for End-Stage Liver Disease Score Components. Gastroenterology. 2008 Nov;135(5):1575–81.
  24. Bernardi M, Gitto S, Biselli M. The MELD score in patients awaiting liver transplant: Strengths and weaknesses. J Hepatol. 2011 Jun;54(6):1297–306.
  25. Kartoun U, Corey KE, Simon TG, Zheng H, Aggarwal R, Ng K, et al. The MELD-Plus: A generalizable prediction risk score in cirrhosis. PLoS One [Internet]. 2017 Oct;12(10):e0186301. Available from: https://doi.org/10.1371/journal.pone.0186301
  26. Ge J, Kim WR, Lai JC, Kwong AJ. “Beyond MELD” – Emerging strategies and technologies for improving mortality prediction, organ allocation and outcomes in liver transplantation. J Hepatol. 2022 Jun;76(6):1318–29.
  27. Subramanian KKK, Tandon M, Pandey CK, Jain P. Patients with cirrhosis of liver operated for non-transplant surgery: A Retrospective Analysis. J Clin Transl Hepatol. 2019 Mar 28;7(X):1–6.
  28. Delis SG, Bakoyiannis A, Biliatis I, Athanassiou K, Tassopoulos N, Dervenis C. Model for end-stage liver disease (MELD) score, as a prognostic factor for post-operative morbidity and mortality in cirrhotic patients, undergoing hepatectomy for hepatocellular carcinoma. HPB. 2009 Jun;11(4):351–7.
  29. Pathare P, Elbayomi M, Weyand M, Griesbach C, Harig F. MELD-score for risk stratification in cardiac surgery. Heart Vessels. 2023 Sep;38(9):1156–63.
  30. Färber G, Marx J, Scherag A, Saqer I, Diab M, Sponholz C, et al. Risk stratification for isolated tricuspid valve surgery assisted using the Model for End-Stage Liver Disease score. J Thorac Cardiovasc Surg. 2023 Nov;166(5):1433-41.e1.
  31. Sukhera J. Narrative reviews: flexible, rigorous, and practical. Journal of graduate medical education. 2022 Aug;14(4):414-7.