Meld exceptions and new predictive score of death on long waiting lists for liver transplantation
S. Iacob, L. Gheorghe, R. Iacob, C. Gheorghe, D. Hrehoret, I. PopescuOriginal article, no. 3, 2009
* Center of Gastroenterology and Hepatology
* Center of General Surgery and Liver Transplantation
The continuous growing number of patients included on the waiting lists for liver transplantation (LT) around the world sets pressure upon policy makers and LT professionals to look for better ways to prioritize patients in need for this life-saving treatment. An ideal liver allocation system should ensure the best use of scarce organs in order to reduce waiting list mortality, by including all proved prognostic factors of survival according to etiology, should serve all potential LT candidates equally well and should improve post-LT survival. The current model for end-stage liver disease (MELD) based allocation system, implemented in February 2002, is rapidly available, standardized and objective, accurately predicting short-term mortality for 83-87% of waitlisted candidates. However, there are patients with liver diseases whose survival is dependent upon factors other than liver disease severity (1). There are around 15-20% of patients that are underserved if allocation of deceased-donor livers for transplantation is based only upon the calculated MELD score (1-3). The already recognized MELD exceptions are: hepatocellular carcinoma, hepatopulmonary syndrome, familial amyloid polyneuropathy and primary oxaluria. Patients with these diseases may be assigned additional MELD points above and beyond their calculated MELD scores. But there are many other clinical conditions complicating liver cirrhosis, such as refractory ascites, hepatic encephalopathy, hyponatremia, renal failure, portal hypertensive bleeding, hepatic hydrothorax, malnutrition, recurrent cholangitis, intractable pruritus, which clearly impacts patient survival and are considered “non-recognized exceptional diagnoses”. There are no guidelines for these non-recognized MELD exceptions, but liver transplant centers may give also additional points for these patients with the intention of giving the patient a score that more accurately reflects the perceived degree of medical urgency.
As it was pointed out by Lake J. (4), the transplant rate is substantially different in Romania from that in the United States, the disparity between the number of patients included on the waiting list (WL) and available cadaver organs is even bigger, and it is conceivable that a different system of organ allocation may better serve patients in our country. As we have previously demonstrated (5), MELD score was found to be an excellent predictor of death within 12 months on our WL, characterized by a long waiting time, as well as MELD score variation in the last 3 months proved to be independent predictor of death on large waiting lists like ours (6). However, the mortality rate on the WL in Romania is still very high. In order to add to the continuous endeavors to better prioritize patients to LT, the aim of our study was to identify the predictive value for death on WL for the complications of liver disease which are now considered “non-recognized exceptional conditions”.
Patients and Methods
Waiting list registrants and data collection
A prospective study was conducted gathering data from 372 consecutive adult patients with end-stage liver disease (2B or 3 United Network for Organ Sharing – UNOS – status fullfilling the minimal listing criteria Child-Pugh score of ³7; or one of the following complications no matter what Child Pugh class was: variceal upper digestive bleeding, spontaneous bacterial peritonitis, hepatocellular carcinoma) listed for LT at the referral center between October 2004 and June 2007. All patients had complete data required for MELD calculation in the liver transplantation file. For patients with hepatocellular carcinoma the MELD score was not assigned to be 20 or 24 as proposed by UNOS, but calculated as usual. The allocation of MELD score to patients with HCC was introduced for the first time in the Eurotransplant International Foundation organ procurement system in 2006 (7). Survival analysis has been conducted in December 2007, 6 months after the inclusion of the last patient. The date of the death while on WL, the date of the last contact with our center and the date of the LT have been registered. Patients transplanted during the study interval and patients still alive at the moment of the survival analysis were considered censored cases.
MELD score calculations
The MELD score was calculated using serum creatinine, bilirubin and the International Normalized (Prothrombin) Ratio (INR) (8) according to the following formula currently in use by UNOS (9): MELD score = [9.57 x loge creatinine mg/dl + 3.78 x log bilirubin mg/dl + 11.20 x log INR + 6.43].
Starting with June 2005, a computerised registry was implemented in order to prospectively register all patients listed for LT in our Center at each subsequent visit and to store in a database all MELD scores for each recalculation and all MELD score components (INR, serum bilirubin and creatinine) for each recalculation, according to currently recommended recalculation intervals (yearly if MELD score is £10, every 90 days if MELD score is between 11 and 18, every 30 days for MELD score between 19-24, every 7 days for MELD ³25 according to UNOS status). For patients included on the WL before the introduction of the computerised registry, the MELD data were retrospectively recorded in the database, although they were gathered prospectively from inclusion of that patient on the waiting list. The computerised registry has the ability to dynamically sort in real time the WL, showing at any given moment the transplantation priorities according to the current recorded MELD value for each patient.
This study complies with the standards of Declaration of Helsinki and current ethical guidelines. All patient signed inform consent at inclusion on WL for LT.
Definitions
Initial MELD score is the MELD score calculated at baseline (time of listing).
Current MELD score is the most recent MELD score available for each patient.
Complications of liver cirrhosis (refractory ascites, hepatorenal syndrome, hepatic encephalopathy, spontaneous bacterial peritonitis, variceal upper digestive bleeding) as defined by Bass (10), Runyon (11), Fitz (12) and Sanyal (13) were analyzed. Hyponatremia was defined as serum sodium lower than 130 mEq/L at the time of listing, according to the definition of the International Ascites Club (14) and it was also included in the analysis.
Statistical analysis
Continuous data were expressed as mean ± standard deviation (SD). Categorical data were expressed as proportion of the subjects with a specific feature. T Student test was used to compare continuous variables. The survival during follow-up was evaluated using the Kaplan Meier method. To identify potential predictors of patient death at 12 months, univariate and multivariate Cox’s proportional hazards regression models were used. The parameters identified as independent predictors of death were used in a logistic regression equation to form a new model to predict 12 month mortality on our waiting list for LT. The new model was retrospectivelly evaluated and its ability to predict death on the waitlist was compared by c-statistic (the concordance equivalent to the area under the receiver operating characteristic (ROC) curve) to the current MELD score. A c-statistic between 0.8 and 0.9 indicates excellent diagnostic accuracy and a parameter with a c-statistic over 0.7 should be considered clinically useful. All tests were two sided and a p value < 0.05 was considered to indicate statistical significance.
Results
Waiting list characteristics
Three hundred seventy two patients were listed for LT at our Center between October 2004 and June 2007. Demographic characteristics of the patient population are displayed in Table 1. The highest proportion of patients (72.3%) were listed for liver cirrhosis secondary to viral infections (hepatitis C virus, hepatitis B virus (HBV) or HBV- hepatitis delta virus infection), autoimmune liver diseases (autoimmune hepatitis, primary biliary cirrhosis, primary sclerosing cholangitis) (8.3%), alcoholic liver diseases (11.3%) and others (Wilson disease, secondary biliary cirrhosis, cryptogenic, Caroli disease, polycystic liver-kidney disease) (8.1%). Hepatocellular carcinoma (HCC) was present in 40 (10.8%) patients, out of which 26 were within Milan criteria. Adult to adult living related liver transplantation was performed in 8 patients, out of which 2 for HCC. The complications of liver cirrhosis were encountered in the following proportions:
spontaneous bacterial peritonitis (14.5%), refractory ascites (22.8%), hyponatremia (16.1%), hepatic encephalopathy (30.9%), upper variceal bleeding (23.9%) and hepatorenal syndrome (12.6%).
The 372 study subjects generated 1094 sets of concurrent laboratory data for recalculation of MELD score, with a mean of 2.9 ±1.4 MELD score determinations per patient. The majority of the patients on the WL are status UNOS 2B and 3 with a mean Child-Pugh score of 7.6 ± 2.1 at listing and a mean MELD score at first evaluation of 13.3 ± 4.9. The median MELD scores at listing and at the time of liver transplantation were 12.8 and 16.9, respectively. Ninety seven (26.1%) patients died while on the waiting list for LT and 35 patients (9.4%) were transplanted during this time interval. There was a significant difference between current MELD scores of the patients still waiting on the WL and patients transplanted (12.4 ± 4.5 vs 14.9 ± 5.6, p=0.01) as well as between patients still waiting on the WL and patients that died on the WL (12.4 ± 4.5 vs 21.7 ± 8.8, p<0.0001). There was also a significant difference between initial MELD scores of the patients still waiting on the WL and patients that died on the WL (12.1 ± 4.3 vs 15.8 ± 5.3, p<0.0001).
Patients with an initial MELD score between 11-18 remained in the same category also at the last evaluation in proportion of 76.1%, while 7.2 % of patients had an improvement of the MELD scores. Patients included in the severity category of 19-24 had a favorable evolution in 16.3% of cases, while 30.6% had a progression of the severity of liver disease. Patients in the most severe disease category, with a MELD score >25 at time of listing, regressed by adequate therapy in a lower severity category (MELD score of 10 to 24) in 27.3% of cases. Four patients were excluded from the list due to tumor progression. For too sick conditions were excluded 5 patients and only 2 patients were excluded due to improved condition (one patient with Wilson disease and one patient with alcoholic liver disease).
Survival analysis
The actuarial survival rate at 12 and 24 months was 76.7% and 63%, respectively (Fig. 1). Median survival was 65.5 months. The mean follow-up interval was 14 months.
The initial and current MELD score, initial and current Child-Pugh score, as well as the complications of liver cirrhosis (HCC, spontaneous bacterial peritonitis, refractory ascites, hyponatremia, hepatic encephalopathy, upper variceal bleeding, hepatorenal syndrome) were evaluated as potential predictors of death while on the WL using univariate Cox regression analysis. The results of the analysis are shown in Table 2. All analysed variables including the MELD score and Child-Pugh score at inclusion and at the last calculation and complications of liver cirrhosis except upper variceal bleeding and HCC, are highly predictive for death on the WL. However, presence of HCC reached marginal significance. In the multivariate survival analysis performed by Cox Proportional Hazards Model, the following factors were found to be independent predictors of death on our WL: refractory ascites (p = 0.002, HR = 2.1) and hepato-renal syndrome (p = 0.002, HR = 2.1). Current MELD score (p = 0.07, HR = 1.04), Child-Pugh score at inclusion (p = 0.07, HR = 1.27) and hepatic encephalopathy (p = 0.07, HR = 1.53) reached marginal statistical significance in the multivariate Cox analysis. Hyponatremia (p=0.25), spontaneous bacterial peritonitis (p=0.66), initial MELD score (p=0.23) and current Child-Pugh score (p=0.6) had no statistical significance. There was a significant correlation between the presence of hyponatremia and the presence of refractory ascites (37.6% compared to 9.8% of patients without refractory ascites, p < 0.0001), as well as between the presence of hyponatremia and hepatic encephalopathy (31.3% vs 9.3%, p < 0.0001). 34% of patients with hyponatremia associated also hepatorenal syndrome type 2 compared to 13.5% that had hyponatremia but without hepatorenal syndrome (p = 0.0003).
The independent predictors of death identified by the survival analysis have been included in a logistic regression equation in order to develop a new prognostic score. The results of the multivariate analysis are depicted in Table 3. The following new risk score for death while on the WL for LT was calculated:
New Score = 1/(1+ exp(-(-4.38 + 1.34 x Refractory ascites + 0.9 x Hepatorenal syndrome + 0.15 x Current MELD), where refractory ascites and hepatorenal syndrome are registered as 1 if the event occurs and current MELD is introduced as numeric value.
The area under the receiver operating characteristic curve for this model was 0.85, compared to 0.8 for current MELD score, indicating a very good prediction of death while on the WL for LT (Fig. 2). A cut-off value of 0.45 for the new score had a sensitivity of 56.8%, a specificity of 92.7%, a positive predictive value (PPV) of 72.9% and a negative predictive value (NPV) of 95% for predicting death while on the WL for LT. A cut-off value of 19.23 for current MELD score had a sensitivity of 55.7%, a specificity of 91.7%, a PPV of 69.7% and a NPV of 94.9% for predicting death while on the WL for LT.
|
|
Discussion
The MELD score is an objective and robust predictor of short-term mortality in patients with decompensated liver cirrhosis (15) and had a positive effect on organ allocation, but it is not the perfect tool, the holy grail of organ allocation (16). In a recent systematic review (17) that included 11 studies, only four studies (4512 patients) demonstrated a
statistical superiority of the MELD in comparison with the Child-Turcotte-Pugh (CTP) system, whereas seven studies (8020 patients) showed no statistical difference. However, both scores have several drawbacks. The major criticisms of CTP score consisted in the limited number of medical urgency cathegories (classes A, B, C), the subjectivity and variation upon certain medical intervations of some
parameters such as hepatic encephalopathy and ascites, as well as the lack of an assessment of renal function, which is a well established prognostic marker in cirrhosis (18, 19). Within the CTP cathegories for disease severity, waiting patients were ordered according to time on the waiting list. MELD score was successful in de-emphasizing waiting time as a major factor in prioritizing patients (20), as well as in eliminating subjective assessments. One limitation of MELD score is related to variables used in the equation that may be subject to some variability, depending on the
laboratory methods used. Another limitation is related to some complications of liver cirrhosis that are not well served by MELD score, like HCC, hepatic encephalopathy, esophageal varices bleeding, spontaneous bacterial peritonitis.
Taking into account the above mentioned facts, the high mortality due to liver cirrhosis in Romania (second country in Europe after Hungary – Ascione A, personal communication 2006), the large number of patients included on our WL, the long waiting time to LT, the huge disparity between the need for LT in Romania and the available cadaver organs, we tried to find a better allocation system of organs between patients with same MELD score and a new score that would better predict mortality on our WL.
As we have previously demonstrated (5, 6), MELD score was once again confirmed as an independent and a good
predictor of death (c-statistic = 0.80) within 12 months on our WL. Because MELD is a continuous scale and it lacks the
ceiling or floor effect, it should be maintained for sure in the allocation system. But, the current MELD may need fine-
tuning to accommodate patients with the serious complications that are well known predictors of poor outcome in patients with cirrhosis. In this way, patients with hepatic encephalopathy, severe ascites, spontaneous bacterial peritonitis, hyponatremia might not receive LT in a timely manner if the MELD score alone were to be used for organ allocation. Patients with these complications may not have a higher
baseline MELD and they could be down-staged in the MELD era, as it was recently proved (21). The authors found that the predictive ability of the complications was fairly comparable with that of the MELD system. In addition, the occurrence of complications (porto-systemic encephalopathy, esophageal varices bleeding, spontaneous bacterial peritonitis) was shown to be an independent predictor of poor prognosis associated with a decreased survival. Spontaneous bacterial peritonitis, hepatic encephalopathy and MELD score proved to be predictors of poor survival also in our analysis. Variceal bleeding did not negatively influenced survival of our patients because in the majority of cases it is well managed endoscopically (either emergency band ligation as well as primary or secondary
prophylactic band ligation) (22).
Another study (23) showed a poor correlation between the MELD score and the presence of hepatic encephalopathy or ascites. This study also suggests that many patients with
moderately severe hepatic encephalopathy or severe ascites might not receive LT in an expedited manner, unless they develop renal insufficiency, if the MELD scores alone were to be used to prioritize organ allocation. There is a good correlation between hepatic venous pressure gradient (HVPG), a measure of portal hypertension, and large varices, bleeder
status and ascites. Although higher HVPG reflects more severe liver disease (24), MELD score did not show any correlation with the severity of ascites (23). HVPG proved to have an independent effect on survival in addition to the MELD score. Although inclusion of HVPG and age in a survival predicting model improved the calibrative ability of MELD, its discriminative ability was not significantly increased (c-statistic = 0.71 for MELD score compared to 0.76 for MELD score variables, age, and HVPG) (25). Refractory ascites was an independent predictor of death while on WL in our study population and one of the components of the new predictive score. It is known that 50% of those with refractory ascites die within 6 months (13). There were developed other prognostic models (includes Child-Pugh score, ascitic fluid protein concentration, previous spontaneous bacterial peritonitis episode, and history of heavy alcohol consumption) for patients with refractory ascites (26) that were also validated and was suggested to be used to
distinguish between patients with refractory ascites with a high risk of short-term mortality, and who should, therefore, be given priority in the allocation of a liver or who might be offered a liver donor graft, from those with a less urgent
indication. This can be help to differentiate between patients within the same MELD category and can be used in association with our prognostic model.
Hepatorenal syndrome type 2 and hyponatremia often superimposes refractory ascites. Hepatorenal syndrome may reverse by treatment with albumin and terlipressin (27), thus prolonging their survival until LT and having a lower MELD score at LT. Anyway, adding refractory ascites and hepatorenal syndrome in antecedents of the patient to the current MELD score improved the accuracy of MELD score in predicting survival at 12 months on our WL. Both current MELD and new score proved to be excellent diagnostic tests with a high specificity and predict survival on the WL with high accuracy due to the high negative predictive values.
Hyponatremia was not an independent predictor of death in our series in contrast to other findings (28-30). However, in our study, hyponatremia was significantly correlated with refractory ascites, hepatorenal syndrome and hepatic encephalopathy, thus confirming that dilutional hypo-natremia should be regarded as an early prognostic marker of poor outcome, announcing the development of more serious complications. Hyponatremia was shown to independently predict the survival in cirrhotic patients, even in the
individual category of the ascites, hepatic encephalopathy, variceal bleeding, spontaneous bacterial peritonitis or renal failure, with a hazard ratio of 5.9–16.5 (30).
Despite the high efficacy of MELD to predict waitlist mortality in our study, incorporation of refractory ascites and hepatorenal syndrome into the formula significantly increased the c-statistic from 0.80 to 0.85. This was due to the ability of these two complications, but not MELD score alone, to identify a subgroup of high-risk patients who died during the study period. The new score can be used to
prioritize patients within the same MELD category who would have a worse prognosis in case they are not transplanted in the context of severe shortage of organs in Romania. These new findings need to be validated and assessed in further prospective studies, especially in liver transplant centers with a higher dynamic of LT.
There are no conflicts of interest and no source of
funding for this study.
References
1. Biggins, S.W., Bambha, K. - MELD-based liver allocation. Who is underserved? Seminars in Liver Disease 2006, 26:211.
2. Llado L, Figueras J, Memba R, Xiol, X., Baliellas, C., Vázquez S, Ramos E, Torras J, Rafecas A, Fabregat J, Lama C, Jaurrieta E. - Is MELD Really the Definitive Score for Liver Allocation? Liver Transpl., 2002, 8:795.
3. Freeman, R.B. - Mortality Risk Versus Other Endpoints: Who Should Come First on the Liver Transplant Waiting List? Liver Transpl., 2004, 10:675.
4. Lake, J. - MELD - an Imperfect, but Thus Far the Best, Solution to the Problem of Organ Allocation. J Gastrointestin Liver Dis., 2008, 17:5.
5. Gheorghe, L., Popescu, I., Iacob, R., Iacob, S., Gheorghe, C. - Predictors of death on the waiting list for liver transplantation characterized by a long waiting time. Transpl. Int., 2005, 18:572.
6. Gheorghe, L., Iacob, S., Iacob, R., Gheorghe, C., Popescu, I. - Variation of the MELD score as a predictor of death on the waiting list for liver transplantation. J. Gastrointestin. Liver Dis., 2007, 16:267.
7. Broelsch, C.E., Frilling, A., Malago, M. - Should we expand the criteria for liver transplantation for hepatocellular carcinoma – Yes, of course! J. Hepatol., 2005, 43:569.
8. Kamath, P.S., Wiesner, R.H., Malinchoc, M., Kremers W, Therneau TM, Kosberg CL, D'Amico G, Dickson ER, Kim WR. - A model to predict survival in patients with end-stage liver disease. Hepatology, 2001, 33:464.
9. Wiesner, R., Edwards, E., Freeman, R., Harper, A., Kim, R., Kamath, P., Kremers, W., Lake, J., Howard, T., Merion, R.M., Wolfe, R.A., Krom, R. - United Network for Organ Sharing Liver Disease Severity Score Committee. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology, 2003, 124:91.
10. Bass, N.M., Yao, F.Y. - Portal hypertension and variceal
bleeding. În “Gastrointestinal and Liver Disease. Pathophysiology /Diagnosis/Management” sub redacţia Sleisenger & Fordtran’s, W.B. Saunders Company (Philadelphia), 2002, pag. 1487-1515.
11. Runyon, B.A. - Ascites and spontaneous bacterial peritonitis. În “Gastrointestinal and Liver Disease. Pathophysiology/ Diagnosis/Management.” sub redacţia Sleisenger & Fordtran’s. WB Saunders Company (Philadelphia), 2002, pag. 1517-1542.
12. Fitz, J.G. - Hepatic encephalopathy, hepatopulmonary
syndromes, hepatorenal syndrome, coagulopathy and endocrine complications of liver disease. În “Gastrointestinal and Liver Disease. Pathophysiology/ Diagnosis/Management.” sub redacţia Sleisenger & Fordtran’s. WB Saunders Company (Philadelphia), 2002, pag. 1543-1566.
13. Sanyal, A., Bosch, J., Blei, A., Arroyo, V. - Portal Hypertension and Its Complications. Gastroenterology, 2008, 134:1715.
14. Ginès, P., Berl, T., Bernardi, M., Bichet DG, Hamon, G., Jimenez, W., Liard, J.F., Martin, P.Y., Schrier, R.W. - Hyponatremia in cirrhosis: from pathogenesis to treatment. Hepatology, 1998, 28:851.
15. D’Amico, G., Garcia-Tsao, G., Pagliaro, L. - Natural history and prognostic indicators of survival in cirrhosis: A
systematic review of 118 studies. J. Hepatol., 2006, 44:217.
16. Freeman, R.B. - MELD: the holy grail of organ allocation? J. Hepatol., 2005, 42:16.
17. Cholongitas, E., Marelli, L., Shusang, V., Senzolo, M., Rolles, K., Patch, D., Burroughs, A.K. - A systematic review of the performance of the model for end-stage liver disease (MELD) in the setting of liver transplantation. Liver Transpl., 2006, 12:1049.
18. Forman, L.M., Lucey, M.R. - Predicting the prognosis of chronic liver disease: an evolution from Child to MELD. Hepatology, 2001, 33:473.
19. Durand, F., Valla, D. - Assessment of the prognosis of
cirrhosis: Child-Pugh versus MELD. J. Hepatol., 2005, 42 (Suppl):S100.
20. Freeman, R.B., Rohrer, R.J., Katz, E., Lewis, W.D., Jenkins, R., Cosimi, A.B., Delmonico, F., Friedman, A., Lorber, M., O’Connor, K., Bradley, J. - Preliminary Results of a Liver Allocation Plan Using a Continuous Medical Severity Score That De-Emphasizes Waiting Time. Liver Transpl., 2001, 7:173.
21. Huo, T.I., Lin, H.C., Wu, J.C., Hou, M.C., Lee, F.Y., Lee, P.C., Chang, F.Y., Lee, S.D. - Limitation of the model for end-stage liver disease for outcome prediction in patients with cirrhosis-related complications. Clin. Transplant, 2006, 20:188.
22. Gheorghe, C., Gheorghe, L., Iacob, S., Iacob, R., Popescu, I. - Primary prophylaxis of variceal bleeding in cirrhotics awaiting liver transplantation. Hepatogastro-enterology, 2006, 53:552.
23. Yoo, H.Y., Edwin, D., Thuluvath, P. - Relationship of the Model for End-Stage Liver Disease (MELD) Scale to Hepatic Encephalopathy, as Defined by Electroencephalography and Neuropsychometric Testing, and Ascites. Am. J. Gastroenterol., 2003, 98:1395.
24. Wadhawan, M., Dubey, S., Sharma, B.C., Sarin, S.K. - Hepatic venous pressure gradient in cirrhosis: correlation with the size of varices, bleeding, ascites, and Child's
status. Dig. Dis. Sci., 2006, 51:2264.
25. Ripoll, C., Banares, R., Rincon, D., Catalina, M.V., Lo, Iacono, O., Salcedo M, Clemente G, Núñez O, Matilla A, Molinero LM. - Influence of hepatic venous pressure gradient on the prediction of
survival of patients with cirrhosis in the MELD Era. Hepatology, 2005, 42:793.
26. Guardiola, J., Baliellas, C., Xiol, X., Fernandez Esparrach, G., Ginès, P., Ventura, P., Vazquez, S. - External Validation of a Prognostic Model for Predicting Survival of Cirrhotic Patients With Refractory Ascites. Am. J. Gastroenterol., 2002, 97:2374.
27. Martin-Llahi, M., Pepin, M.N., Guevara, M., Diaz, F., Torre, A., Monescillo, A., Soriano, G., Terra, C., Fabrega, E., Arroyo, V., Rodes, J., Gines, P. - Terlipressin and Albumin vs Albumin in Patients With Cirrhosis and Hepatorenal Syndrome: A Randomized Study. Gastroenterology, 2008, 134:1352.
28. Biggins, S.W., Rodriguez, H.J., Bacchetti, P., Bass, N.M., Roberts, J.P., Terrault, N.A. - Serum sodium predicts mortality in patients listed for liver transplantation. Hepatology, 2005, 41:32.
29. Ruf, A.E., Kremers, W.K., Chavez, L.L., Descalzi, V.I., Podesta, L.G., Villamil, F.G. - Addition of serum sodium into the MELD score predicts waiting list mortality better than MELD alone. Liver Transpl., 2005, 11:336.
30. Londoño, M.C., Cardenas, A., Guevara, M., Quinto, L., de Las Heras, D., Navasa, M., Rimola, A., Garcia-Valdecasas, J.C., Arroyo, V., Gines, P. - MELD score and serum sodium in the
prediction of survival of patients with cirrhosis awaiting liver transplantation. Gut, 2007, 56:1283.