Abstract
Background
The aim of this study was to assess risk factors for postoperative mortality after liver transplantation among patients with Model for End-Stage Liver Disease (MELD) scores ≥35, with special focus on the MELD scores.
Material/Method
Data from 68 primary liver transplantations in patients with MELD scores ≥35 among 1376 liver transplantations performed in the Department of General, Transplant, and Liver Surgery (Medical University of Warsaw) between January 2002 and October 2014 were analyzed retrospectively. Postoperative (90-day) mortality was set as the primary outcome measure.
Results
Postoperative mortality was 29.4% (20 of 68). The overall survival rates after 1, 5, and 10 years were 61.9%, 59.7%, and 59.7%, respectively. According to univariate analyses, MELD (p=0.014), conventional technique of liver transplantation (p=0.049), intraoperative fresh frozen plasma (p=0.040), and red blood cells (p=0.026) transfusions were risk factors for postoperative mortality. MELD score was the only independent risk factor for postoperative mortality (p=0.023) in multivariate analysis. According to receiver operating characteristics analysis, the optimal cut-off for MELD score in prediction of postoperative mortality was ≥43 (Area Under Curve=0.703, 95% Confidence Interval 0.575–0.831). Postoperative mortality was 21.4% and 42.3% among patients with MELD score <43 and ≥43, respectively (p=0.066).
Conclusions
MELD score is an important predictor of early mortality after liver transplantation, even among recipients with high MELD scores. In particular, patients with MELD score ≥43 should be considered as very high-risk candidates for liver transplantation.
MeSH Keywords: Liver Transplantation, Mortality, Treatment Outcome
Background
Liver transplantation (LT) became a widely accepted treatment of patients with chronic severe liver disease and acute liver failure, as well as those with selected liver malignancies [1]. Due to the success of liver transplantation, there is a significant gap between number of patients waiting for liver transplantation and number of donor organs [1–3]. Because of the persistent shortage of donor livers and the high waiting list mortality, numerous centers are searching for the most effective method of allocation of organs, and the Model for End-Stage Liver Disease (MELD) score is frequently used as the basis of liver allocation [1].
MELD was primarily constructed by Malinchoc et al. [4] to predict patient survival after transjugular intrahepatic portosystemic shunts (TIPS). Initially, it involved serum concentrations of bilirubin and creatinine, international normalized ratio for prothrombin time (INR), and the cause of liver dysfunction [4,5]. MELD can be generalized to patients with liver failure due to various etiologies because the exclusion of etiology in the model did not affect its accuracy [6].
MELD was primarily applied for the allocation system in February 2002 by the Organ Procurement and Transplantation network (OPTN)/United Network for Organ Sharing (UNOS) [5], as it was based on 3 widely available objective variables accounting for both liver and renal function and was not associated with the criterion of time waiting on the list [7]. The prospective study of Wiesner et al. demonstrated that MELD score based on bilirubin, creatinine, and INR can be used to estimate 3-month survival in patients with chronic liver dysfunction and may be used in allocation of donor livers to the most ill recipients on the waiting list [7]. MELD continued to be the basis for allocation systems instead of recipient size and time on the waiting list.
The Italian allocation system from 2012 was divided into 3 qualification processes according to urgency principles based on MELD score: regionally for waitlist candidates with MELD <30, by macro-area for waitlist candidates with MELD scores ≥30, and nationally for super-urgent waitlist candidates [8]. In 2013, the ‘Share 35’ policy was introduced in the United States, giving higher priority to patients with MELD scores ≥35 in the region regardless of local recipients list [9,10]. The ‘Share 35’ policy showed a decrease in discard rate, an increase in regional exports, and rate of liver transplantation performed in candidates with high MELD scores on the waitlist, resulting in decreased waitlist mortality among the sickest patients [10,11], and at the same time in worse overall survival after liver transplantation [11].
Although MELD was initially applied for prediction of pre-transplant waiting list mortality, it seems to be a useful tool to estimate the risk of postoperative mortality.
Soon after its introduction as the basis of the allocation system, it was reported that MELD scores are significantly associated with 1-year post-transplant survival [12].
The purpose of the present retrospective study was to assess risk factors for postoperative mortality among liver transplant recipients with high MELD scores (≥35), focusing on the influence of the MELD score.
Material and Methods
A total of 1376 liver transplantations were performed in the Department of General, Transplant, and Liver Surgery at the Medical University of Warsaw in Poland between January 2002 and October 2014. The data were analyzed retrospectively. During that period, there were 76 liver transplantations among patients with pre-transplant biochemical MELD score ≥35. Following exclusion of 8 liver retransplantations, the final study cohort comprised 68 primary liver transplant recipients.
The study was not subject to ethics committee approval due to its retrospective character.
MELD scores considered in the present study were calculated using pre-transplant biochemical parameters: serum bilirubin, creatinine, and INR, with no exception points.
Postoperative (90-day) mortality was set as the primary outcome measure. Continuous and categorical variables are presented as medians with interquartile ranges and numbers with percentages. The chi-square test was used to compare categorical variables between groups.
Logistic regression was used to evaluate risk factors for postoperative mortality. Only statistically significant risk factors in univariate analyses were included in the multivariate analysis. The final model was selected based on the backward elimination method (p>0.05 was used as an exclusion criterion).
Receiver operating characteristics (ROC) curves were constructed to search for the optimal cut-off for MELD scores in the prediction of postoperative mortality. Odds ratios (ORs) and areas under the curve (AUCs) are presented with 95% confidence intervals (95% CIs). The overall survival rate was estimated using the Kaplan-Meier method. The level of statistical significance was set at 0.05. STATISTICA version 12 (StatSoft Inc., Tulsa, OK) was used to conduct statistical analyses.
Results
The baseline characteristics of the study cohort are presented in Table 1. Indications for primary liver transplantations are demonstrated in Table 2. The 3 most common indications for primary liver transplantation involved were: liver failure in the course of Wilson’s disease (16/68, 23.5%), hepatitis B virus (HBV) and/or hepatitis C virus (HCV) infection (14/68, 20.6%), and liver failure of unknown etiology (11/68, 16.2%). A total of 25 out of 68 patients died during the median follow-up period of 64.5 months. Most of the deaths occurred during the 90-day postoperative period (20 of 25, 80.0%) with a postoperative 90-day mortality rate of 29.4% (20 of 68). The overall survival rates after 1, 5, and 10 years were 61.9%, 59.7%, and 59.7%, respectively (Figure 1).
Table 1.
Baseline characteristics of 68 patients with MELD scores ≥35 treated with liver transplantation included in the final study cohort.
| Median or n | Interquartile range or % | |
|---|---|---|
| Age of the recipient (years) | 33.5 | 23.5–49.5 |
| Sex of the recipient | ||
| Male | 34 | 50.0% |
| Female | 34 | 50.0% |
| Acute liver failure | 35 | 51.5% |
| Bilirubin (mg/dL) | 30.1 | 14.1–41.7 |
| Creatinine (mg/dL) | 2.4 | 1.4–3.9 |
| INR | 3.8 | 2.6–5.5 |
| Sodium (mEq/L) | 133.0 | 131.4–136.9 |
| MELD | 41 | 37–44 |
| Red blood cells transfusions (units) | 8 | 6–10 |
| Fresh Frozen Plasma transfusions (units) | 10 | 8–14 |
| Cold Ischemia Time (hours) | 9.0 | 8.0–11.0 |
| Technique of transplantation | ||
| Piggyback | 46 | 67.6% |
| Conventional | 22 | 32.4% |
| Age of the donor (years) | 46.0 | 35.5–53.0 |
| D-MELD | 1854.0 | 1342.5–2265.5 |
MELD – Model for End-Stage Liver Disease; INR – International Normalized Ratio; D-MELD – Donor-MELD.
Table 2.
Indications for liver transplantations among 68 patients with MELD scores ≥35 included in the final study cohort.
| Indication | n | % |
|---|---|---|
| Wilson’s disease1 | 16 | 23.5% |
| HBV and HCV infection2 | 14 | 20.6% |
| Liver failure of unknown etiology | 11 | 16.2% |
| Alcoholic liver disease3 | 5 | 7.4% |
| Amanita phalloides poisoning | 4 | 5.9% |
| Paracetamol poisoning | 4 | 5.9% |
| AIH | 3 | 4.4% |
| HCC4 | 2 | 2.9% |
| Budd-Chiari Syndrome | 2 | 2.9% |
| PBC | 2 | 2.9% |
| PSC | 1 | 1.5% |
| Liver Cystic disease | 1 | 1.5% |
| Dubin-Johnson Syndrome | 1 | 1.5% |
| Volatile compounds poisoning | 1 | 1.5% |
| Haemophagocytic syndrome | 1 | 1.5% |
One patient with HBV infection;
one patient with HIV infection;
one patient with HBV infection;
one patient with HBV infection and ALD, and one patient with HBV and HCV infection.
HBV – hepatitis B virus; HCV – hepatitis C virus; AIH – autoimmune hepatitis; HCC –hepatocellular cancer; PBC – primary biliary cirrhosis; PSC – primary sclerosing cholangitis, HIV – human immunodeficiency virus.
Figure 1.
Overall survival of 68 patients with Model for End-Stage Liver Disease (MELD) scores ≥35 treated with liver transplantation included in the final study cohort.
Univariate analyses revealed that higher MELD scores (p=0.014, OR 1.17 per 1 point increase, 95% CI 1.03–1.32), conventional technique of liver transplantation (p=0.049, OR 3.00, 95% CI 1.01–8.95), intraoperative fresh frozen plasma (p=0.040, OR 1.08 per 1 unit increase, 95% CI 1.00–1.16), and red blood cells (p=0.026, OR 1.09 per 1 unit increase, 95% CI 1.01–1.18) transfusions were significant risk factors for postoperative mortality (Table 3). According to multivariate analysis, higher MELD score was the only independent risk factor for postoperative mortality (p=0.023, OR 1.16 per 1 point increase, 95% CI 1.02–1.31). According to ROC analysis, the optimal cut-off for MELD scores in prediction of postoperative mortality was ≥43 (AUC=0.703, 95% CI 0.575–0.831) (Figure 2). The established cut-off was associated with sensitivity of 55.0%, specificity of 68.8%, accuracy of 64.7%, positive predictive value of 42.3%, and negative predictive value of 78.6%.
Table 3.
Risk factors for postoperative 90-day mortality after liver transplantation among patients with MELD scores ≥35.
| Factors | Univariable analysis | ||
|---|---|---|---|
| p | Odds ratio | 95% Confidence Interval | |
| Age of the recipient | 0.457 | 1.01 | 0.98–1.05 |
| Sex of the recipient | |||
| Male | 0.595 | 0.75 | 0.26–2.15 |
| Acute liver failure | 0.876 | 0.92 | 0.32–2.61 |
| Bilirubin | 0.920 | 1.00 | 0.97–1.03 |
| Creatinine | 0.359 | 1.11 | 0.89–1.37 |
| INR | 0.703 | 0.96 | 0.75–1.21 |
| Sodium | 0.481 | 0.97 | 0.88–1.06 |
| MELD | 0.014 | 1.17 | 1.03–1.32 |
| Intraoperative Red blood cells transfusions | 0.026 | 1.09 | 1.01–1.18 |
| Intraoperative Fresh Frozen Plasma transfusions | 0.040 | 1.08 | 1.00–1.16 |
| Cold Ischemia Time | 0.370 | 1.13 | 0.86–1.48 |
| Technique of transplantation conventional | 0.049 | 3.00 | 1.01–8.95 |
| Age of the donor | 0.399 | 1.02 | 0.98–1.06 |
| D-MELD | 0.119 | 1.08 | 0.98–1.18 |
MELD – Model for End-Stage Liver Disease; INR – International Normalized Ratio; D-MELD – Donor-MELD. Odds ratios are given per: 1 year increase for age of the recipient; 1 mg/dL increase for bilirubin; 1 mg/dL increase for creatinine; 1 for INR; 1 mEq/L increase for sodium; 1 point increase for MELD; 1 unit increase for blood and plasma transfusions; 1 hour increase for duration of cold ischemia; 1 year increase for age of the donor; 1 point increase for D- MELD.
Figure 2.
Receiver operating characteristics curves for MELD scores in prediction of postoperative mortality. AUC – area under curve; 95% CI – 95% confidence interval; MELD – Model for End-Stage Liver Disease.
Postoperative mortality was 21.4% (9/42) in patients with MELD scores <43 as compared to 42.3% (11/26) among patients with MELD scores ≥43 (p=0.066).
Discussion
The best method of liver allocation and liver recipient prioritization is still under debate due to the persistent shortage of donor organs, extension of donor criteria, and substantial waiting list mortality with the risk of worse postoperative outcomes.
This study corroborates that pre-transplant MELD score is an important predictor of postoperative mortality, even among liver transplant recipients with high MELD scores.
The MELD cut-offs for defining high MELD score liver recipients vary between numerous studies, ranging from 21 to 36 [13–17]. The differences may be the result of various population characteristics [16]. Risk factors including MELD score among liver transplant recipients with high MELD scores need to be estimated. In the study cohort, transplantations performed among liver transplant recipients with MELD scores ≥35 accounted for 5.5% (76 of 1376) of all liver transplantations conducted in the Department of General, Transplant, and Liver Surgery during the study period. Therefore, the applied cut-off for identifying liver transplant recipients with high MELD scores for the purposes of the present study appear to be justified.
Kaltenborn et al. [13] identified MELD scores ≥30 as a significant independent risk factor for 90-day post-transplant mortality, as well as for long-term mortality, but they did not further analyze whether pre-transplant MELD score is a risk factor for postoperative death among this high-risk sub-population. Moreover, liver recipients with high MELD scores face the risk of significantly longer stay, both overall and at the Intensive Care Unit, along with a higher number of surgical revisions [13].
Similarly, Jacob et al. [15] found that the waiting time was associated with the length of decompensation time among liver transplant recipients with MELD scores ≥36, but they did not search for MELD scores cut-off for postoperative mortality among these high-risk patients. The analysis of post-transplant survival in the United Kingdom and Ireland showed that liver transplant recipients with MELD scores ≥36 are characterized by significantly lower survival outcomes [15]. According to Nekrasov et al., who described the influence of implementation of the Share 35 allocation policy, the MELD score at the time of transplantation was not a significant risk factor in postoperative mortality among patients with MELD scores ≥40 [11].
The results of the present retrospective study may contribute to identification of patients with high MELD scores who are most likely to benefit from liver transplantation despite their severe status. Even among liver recipients with high MELD scores, MELD was a significant risk factor for postoperative mortality, and its value should be considered in qualifying patients for liver transplantation and stratifying postoperative risk. Although the difference between mortality among patients with MELD scores <43 and ≥43 was on the verge of statistical significance, it is most probably due to patient numbers. Most importantly, pre-transplant MELD score was the only independent risk factor for postoperative mortality in the analyzed cohort.
Due to the shortage of liver grafts and increasing waiting times, higher-risk grafts are used to expand the donor pool [18]. Utilization of grafts procured from older donors seems to be one of the most important strategies. Although grafts from older donors can be safely used for older recipients, their allocation to recipients with high MELD scores is considered to particularly compromise post-transplant outcomes [19]. D-MELD, which is the product of 2 variables (recipient MELD and donor age in years), was suggested as a useful tool for allocation of grafts procured from extended-criteria donors to avoid the donor/recipient matches of the highest risk [20]. However, in the present study, D-MELD score failed to predict early postoperative mortality among patients with high MELD scores, similar to the results of a study by Schrem et al. [21]. Therefore, the role of D-MELD score in the allocation process appears to be limited in patients with high MELD scores.
Interestingly, as far as postoperative mortality is concerned, the vast majority of patients died in the early postoperative period, perhaps because liver transplantation and the subsequent postoperative period are crucial for patients with severe liver dysfunction. The overall 2-year survival rate remains constant, presumably due to the younger age of patients (a median of 35.5 years) and associated lower risk of long-term cardiovascular complications and post-transplant malignancies. According to Fussner et al., age is an independent risk factor for development of cardiovascular disease in liver transplant recipients [22]. Haagsma et al. reported that age >40 years was significant risk factor for de novo malignancies after liver transplantation [23]. According to the review by Fung et al., all liver transplant recipients over 45 years old should be screened for malignancies regularly during follow-up period [24]. In addition, 12 out of 16 patients who underwent liver transplantation for Wilson’s disease, which was the most common indication in the studied group, survived for over 2 years. Wilson’s disease is associated with high overall long-term survival rates [25–27], which is the most probable reason for lack of late mortality.
Several limitations of the present study should be acknowledged. Firstly, this is a retrospective analysis of single-center experience. However, over 50% of liver transplantations performed among adult recipients every year in Poland are conducted at our center. Secondly, the lack of a significant difference in postoperative mortality between liver transplant recipients with MELD <43 and ≥43 is most probably due to patient numbers.
Conclusions
In conclusion, pre-transplant MELD score is a significant risk factor for postoperative mortality, even among liver recipients with high MELD scores. In particular, MELD scores ≥43 should be considered as very high-risk liver transplantations.
Footnotes
Source of support: Departmental sources
References
- 1.Dutkowski P, Linecker M, DeOliveira ML, et al. Challenges to liver transplantation and strategies to improve outcomes. Gastroenterology. 2015;148(2):307–23. doi: 10.1053/j.gastro.2014.08.045. [DOI] [PubMed] [Google Scholar]
- 2.Johnson RJ, Bradbury LL, Martin K, Neuberger J UK Transplant Registry. Organ donation and transplantation in the UK-the last decade: A report from the UK national transplant registry. Transplantation. 2014;97(Suppl 1):S1–27. doi: 10.1097/01.TP.0000438215.16737.68. [DOI] [PubMed] [Google Scholar]
- 3.Klein AS, Messersmith EE, Ratner LE, et al. Organ donation and utilization in the United States, 1999–2008. Am J Transplant. 2010;10(4 Pt 2):973–86. doi: 10.1111/j.1600-6143.2009.03008.x. [DOI] [PubMed] [Google Scholar]
- 4.Malinchoc M, Kamath PS, Gordon FD, et al. A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts. Hepatology. 2000;31(4):864–71. doi: 10.1053/he.2000.5852. [DOI] [PubMed] [Google Scholar]
- 5.Kamath PS, Kim WR Advanced Liver Disease Study Group. The model for end-stage liver disease (MELD) Hepatology. 2007;45(3):797–805. doi: 10.1002/hep.21563. [DOI] [PubMed] [Google Scholar]
- 6.Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients with end-stage liver disease. Hepatology. 2001;33(2):464–70. doi: 10.1053/jhep.2001.22172. [DOI] [PubMed] [Google Scholar]
- 7.Wiesner R, Edwards E, Freeman R, et al. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology. 2003;124:91–96. doi: 10.1053/gast.2003.50016. [DOI] [PubMed] [Google Scholar]
- 8.Cillo U, Burra P, Mazzaferro V, et al. A multistep, consensus-based approach to organ allocation in liver transplantation: Toward a “Blended Principle Model”. Am J Transplant. 2015;15(10):2552–61. doi: 10.1111/ajt.13408. [DOI] [PubMed] [Google Scholar]
- 9.Feng S, O’Grady J. Share 35: A liver in time saves lives? Am J Transplant. 2015;15(3):581–82. doi: 10.1111/ajt.13102. [DOI] [PubMed] [Google Scholar]
- 10.Massie AB, Chow EK, Wickliffe CE, et al. Early changes in liver distribution following implementation of Share 35. Am J Transplant. 2015;15(3):659–67. doi: 10.1111/ajt.13099. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Nekrasov V, Matsuoka L, Rauf M, et al. National outcomes of liver transplantation for MELD ≥40: The impact of share 35. Am J Transplant. 2016 doi: 10.1111/ajt.13823. [Epub ahead of print] [DOI] [PubMed] [Google Scholar]
- 12.Saab S, Wang V, Ibrahim AB, et al. MELD score predicts 1-year patient survival post-orthotopic liver transplantation. Liver Transpl. 2003;9:473–76. doi: 10.1053/jlts.2003.50090. [DOI] [PubMed] [Google Scholar]
- 13.Kaltenborn A, Hartmann C, Salinas R, et al. Risk factors for short- and long-term mortality in liver transplant recipients with MELD Score ≥30. Ann Transplant. 2015;20:59–69. doi: 10.12659/AOT.892322. [DOI] [PubMed] [Google Scholar]
- 14.Rauchfuss F, Zidan A, Scheuerlein H, et al. Waiting time, not donor-risk-index, is a major determinant for beneficial outcome after liver transplantation in high – MELD patients. Ann Transplant. 2013;18:243–47. doi: 10.12659/AOT.883924. [DOI] [PubMed] [Google Scholar]
- 15.Jacob M, Copley LP, Lewsey JD, et al. Pretransplant MELD score and post liver transplantation survival in the UK and Ireland. Liver Transpl. 2004;10(7):903–7. doi: 10.1002/lt.20169. [DOI] [PubMed] [Google Scholar]
- 16.Brandão A, Fuchs SC, Gleisner AL, et al. MELD and other predictors of survival after liver transplantation. Clin Transplant. 2009;23(2):220–27. doi: 10.1111/j.1399-0012.2008.00943.x. [DOI] [PubMed] [Google Scholar]
- 17.Habib S, Berk B, Chang CC, et al. MELD and prediction of post-liver transplantation survival. Liver Transpl. 2006;12(3):440–47. doi: 10.1002/lt.20721. [DOI] [PubMed] [Google Scholar]
- 18.Busuttil RW, Tanaka K. The utility of marginal donors in liver transplantation. Liver Transpl. 2003;9:651–63. doi: 10.1053/jlts.2003.50105. [DOI] [PubMed] [Google Scholar]
- 19.Dudek K, Kornasiewicz O, Remiszewski P, et al. Results of liver transplantation from old donors. Transplant Proc. 2014;46(8):2762–65. doi: 10.1016/j.transproceed.2014.09.022. [DOI] [PubMed] [Google Scholar]
- 20.Halldorson JB, Bakthavatsalam R, Fix O, et al. D-MELD, a simple predictor of post liver transplant mortality for optimization of donor/recipient matching. Am J Transplant. 2009;9(2):318–26. doi: 10.1111/j.1600-6143.2008.02491.x. [DOI] [PubMed] [Google Scholar]
- 21.Schrem H, Reichert B, Frühauf N, et al. The Donor-Risk-Index, ECD-Score and D-MELD-Score all fail to predict short-term outcome after liver transplantation with acceptable sensitivity and specificity. Ann Transplant. 2012;17(3):5–13. doi: 10.12659/aot.883452. [DOI] [PubMed] [Google Scholar]
- 22.Fussner LA, Heimbach JK, Fan C, et al. Cardiovascular disease after liver transplantation: When, what, and who is at risk. Liver Transpl. 2015;21(7):889–96. doi: 10.1002/lt.24137. [DOI] [PubMed] [Google Scholar]
- 23.Haagsma EB, Hagens VE, Schaapveld M, et al. Increased cancer risk after liver transplantation: A population-based study. J Hepatol. 2001;34(1):84–91. doi: 10.1016/s0168-8278(00)00077-5. [DOI] [PubMed] [Google Scholar]
- 24.Fung JJ, Jain A, Kwak EJ, et al. De novo malignancies after liver transplantation: A major cause of late death. Liver Transpl. 2001;7(11 Suppl 1):S109–18. doi: 10.1053/jlts.2001.28645. [DOI] [PubMed] [Google Scholar]
- 25.Catana AM, Medici V. Liver transplantation for Wilson disease. World J Hepatol. 2012;4(1):5–10. doi: 10.4254/wjh.v4.i1.5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Peedikayil MC, Al Ashgar HI, Al Mousa A, et al. Liver transplantation in Wilson’s disease: Single center experience from Saudi Arabia. World J Hepatol. 2013;5(3):127–32. doi: 10.4254/wjh.v5.i3.127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Medici V, Mirante VG, Fassati LR, et al. Liver transplantation for Wilson’s disease: The burden of neurological and psychiatric disorders. Liver Transpl. 2005;11(9):1056–63. doi: 10.1002/lt.20486. [DOI] [PubMed] [Google Scholar]