Scoring Systems for Patients with Hepatic Dysfunction:
           A lecture for continuous medical education

Mohammad  Ezzat  Moemen

Introduction: 
The liver dysfunction problem represents a difficult international medical equation due to the dreadful spread of chronic viral hepatitis, chronic bilharzial periportal fibrosis and alcohol consumption in some countries.
It is well known that patients with cirrhosis have reduced life expectancy.  Gines et al (1) reported a median survival time of 8.9 years for patients with newly diagnosed cirrhosis, which decreased to 1.6 years after the onset of the first major complication as jaundice, ascites, encephalopathy or gastrointestinal haemorrhage.
It has been estimated that 10% of all patients with liver disease undergo operative procedures during the final two years of their lives (2).  Anaesthesia and surgery are known to have decompensatory effects on patients with cirrhosis with a consequent increase in morbidity and mortality rates. So, preoperative recognition of compensated or decompensated liver disease is important to the patients and their treating doctors to improve patient outcome.

(I): Scoring systems and liver cirrhosis
As early as the middle of the last century, the ultimate outcome in a cirrhotic patient was not easily determined prior to a surgical procedure (2). Since then, many authors attempted to predict perioperative mortality in cirrhotic patients by assessing different variables (3-11).
Cayer and Sohmer (3) reported that in surgical cirrhotic patients, ascites, hypo-albuminemia, prolonged prothrombin time (PT) and anaemia were significant between survivors and non-survivors.  
Wirthlin et al (4) developed a 5-parameters scoring system to assess surgical risk: raised bilirubin, varices with past history of ascites, prolonged PT, hypo-albuminemia and elevated ammonia. They proved the preoperative bilirubin level as the most reliable indicator for poor surgical risk in cirrhotic patients.
Maddrey et al (5) in 1978 introduced a modified discriminant function (DF) score as a predictor of mortality based on a retrospective analysis of a severely-ill subgroup of patients with alcoholic hepatitis over two decades. It was calculated based on two laboratory values; prolonged PT and total bilirubin level.
DF= 4.6 [patient PT-control PT] + total bilirubin (mg/dl)
This score could recently identify those persons with alcoholic cirrhosis who had greater than 50% mortality at one month and who might benefit from corticosteroid or pentoxifylline therapy. However, the combination of a total bilirubin greater than 8mg/dl and presence of ascites was shown to be a more highly predictor of mortality than the DF score (6).
Doberneck et al (7) found mortality to be associated with hypo-albuminemia, elevated alkaline phosphatase, prolonged PT, an emergency procedure, an alimentary tract operation, ascites, operative blood loss and post-operative complications. One factor present correlated with 5% mortality, while more than six factors correlated with 67% mortality.
Garrison et al (8) tested 54 perioperative variables in a retrospective study of 100 survivors and non-survivors after abdominal surgery. They utilized a computer-generated multivariate discriminant analysis of the variables for each patient. Ten preoperative variables proved to be significant in a descending order of importance: Child’s class, ascites, abdominal infection or contramination, emergency procedure, poor nutrition, raised serum bilirubin, hypo-albuminemia, prolonged PT, prolonged partial thromboplastin time (PTT) and increased white blood count (wbc).  Out of these preoperative variables, albumin concentration, presence of infection and prolonged PTT (given appropriate weight by coefficients) were utilized to predict patient outcome through a deduced equation for each patient.  Each patient score deduced from the application of the equation was compared to means of survivors and non-survivors generated by the computer program. Accumulative overall accuracy of the entire equations in predicting survival or non-survival was 83%.  Blood transfusion was the only significant intra-operative variable contributing to survival. Postoperative application of the equation showed four significant variables (pulmonary failure, blood requirements, urinary tract infection, ascites development) with an accumulative overall accuracy of 98%.
Child and Turcotte (9) in 1964 designed their classification (table 1) to predict mortality after portosystemic shunts in cirrhotic patients.  Patients were classified according to increasing severity from class A to class C, on the basis of 5 parameters: serum bilirubin, serum albumin, ascites control, grade of encephalopathy and nutrition.  The problem of the Child-Turcotte classification was that 3 of the elements (ascites, encephalopathy and nutrition) were very subjective.

Table (1): Child-Turcotte classification of liver disease (9)

Pugh et al (10) in 1973, through studying transection of the oesophagus for bleeding oesophageal varices in 38 cirrhotic patients, published a modification of the original Child-Turcotte classification of liver disease (table 2). This Child-Pugh classification allowed better grading of ascites and encephalopathy and substituted PT in place of the nutritional status, therefore obviating the most subjective of the 3 elements; encephalopathy ascites, and nutritional status. To determine surgical risk, each parameter in class A, B and C was assigned 1, 2 and 3 scoring points respectively.  Total points of 5-6, 7-9, and 10-15 denoted good, moderate and poor surgical risk respectively.  The percentage of patients dying in A, B and C grades was 29%, 38% and 78% respectively (10).

Table (2): Child-Pugh modified classification 
of liver disease (10)

In another restrospective study on 189 patients undergoing portocaval shunt procedures using the Child-Pugh score, mortality was 7.7% in A, 25.9% in B an 53.1% in C. (11)
Although mortality increased through the grades in both studies, the ability of the Child-Pugh score to predict those likely to die was not satisfactory since many patients died in grades A and B.  Webster (12) holds the opinion that patients with cirrhosis may not solely present to the anaesthetists for hepatic surgery and that most anaesthetists would be interested in a predictive scoring system which had been validated in a prospective study of patients undergoing all forms of surgery.
Armonis et al (13) in 1997 revived the old hepatic venous catheterization test with hepatic venous pressure of more than 16 mmHg as a prognostic mortality maker in surgical cirrhotic patients.
Iwao et al (14) in 1997 used the Doppler ultrasound technique to verify the presence of varices with portal vein diameter less than 13mm and time averaged mean portal blood flow velocity less than 15cm/s.
These invasive and non-invasive techniques have been suggested to improve the prognostic information of the Child scoring systems (13, 14).
The wider re quirement for predicting prognosis in cirrhosis led some experts to advocate flow-dependent quantitative measurements of liver function(15).
The Bromsulphthalin and Indocyanine Green clearance tests signify that inert dyes are cleared from the plasma evaluating hepatic blood flow. The Aminopyrine Breath test signifies that labelled aminopyrine given by mouth is metabolized by the P450 cytochrome system and expired labelled carbon dioxide correlates with aminopyrine clearance from the plasma (16).  Lidocaine is metabolized by the P450 system to monoethyl glycine xylidide (MEGX) which correlates with the rate of lidocaine clearance and hepatic blood flow. The MEGX test is sensitive to predict graft functions in liver transplantation (17).  Nicholas et al (18) reported that MEGX formation test and gastric intramucosal PH tonometry were the most discriminatory regarding survival or death, due to mismatch between hepatic metabolic demand and hepate blood flow.
Moemen et al (19) have prospectively introduced a recently modified Child-Pugh scoring system (table 3) by the addition of four parameters: serum sodium (s.Na), serum creatinine, wbc and arterial/ alveolar oxygen tension ratio (Pa/AO2).  Each parameter in class A, B and C gets 1, 2 or 3 points respectively with minimum and maximum scoring points of 9 and 27.  The surgical risk is classified according to the scoring points into: mild (9-10 points), moderate (11-14 points) and severe (15-27 points).

Table (3): Moemen modified classification of liver disease (19)

Of the changes in different electrolytes during liver disease, those concerning sNa level promised Moemen et al (19) to be of importance in predicting periopeative outcome.  Dilutional hypo-natraemia is a common finding in patients with liver cirrhosis, and sNa levels should be maintained above 130 meq/L, even by fluid restriction.  The mortality for patients with severe liver disease with sNa levels below 120 meq/L and central nervous symptoms exceeds 60%. (20)
Renal failure, mostly in the form of hepato-renal syndrome (HRS) may complicate patients with liver cirrhosis.  This HRS is diagnosed by oligurea with urine volume <500 ml/d, urine sodium <10mg/L, and sNa<130 meq/L.  Toblas (21) reported that the HRS has been associated with nearly 100% mortality, if the operation is any other than liver transplantation. Serum creatinine level has been considered a risk factor for predicting survival or death of patients after liver transplantation, and a preoperative serum creatinine value <1.7mg/dl predicted survival in 79% of patients (22).  So, the inclusion of serum creatinine promised Moemen et al (19) to be a valuable parameter for predicting perioperative outcome.
A retrospective study showed that infection is one of the most sensitive indicators of mortality or survival after abdominal surgery (23).  The role of infection for producing severe sepsis, septic shock and multiple organ failure has been overemphasized (8, 24, 25).
As many as 12%-28% of patients with cirrhosis demonstrated moderate hypoxaemia with PaO2 <70 mmHg and a smaller percentage of patients had severe hypoxaemia with PaO2<50 mmHg (26, 27).
Over the last few decades, oxygen transport and flow-related oxygen variables including cardiac index, oxygen delivery, oxygen consumption and oxygen extraction ratio have been better understood, and optimization of oxygenation became an aim to be achieved in the management of critically-ill patients (28-32).
The P(a/A)O2 is an oxygenation parameter that adjusts for fluctuations of PaCO2 and considers barometric pressure (33). It is a mathematical manipulation that eliminates the influence of FiO2 on the alveolar-arterial oxygen tension gradient P(A-a)O2 (34, 35) Gilbert and Kerghlery reported normal P(a/A)O2 as 0.77-0.74 and 0.80-0.82 on FiO2 of 0.21 and 1.0 respectively.  Men (36) reported that a ratio of 0.75 indicates alveolar hypoventilation and a ratio ranging from 0.55 to less than 0.75 indicates ventilation perfusion mismatch (V/Q), while a ratio below 0.55 suggests true blood shunting.
The P(a/A)O2 as an oxygenation parameter is measured through a minimally invasive technique through radial artery cannulation. If this parameter is difficult to measure, the total Moemen et al score will include 24 points for only 8 parameters and grades A, B and C will include 8-9, 10-12 and 13-24 points respectively (19).
The study of Moemen et al (19) was prospective and enrolled 210 surgical cirrhotic patients.  No patients in grade A showed morbidity or mortality.  Postoperative morbidity and mortality were more closely related to the scoring class.  By using this dynamic score, the study recommended the ethical preoperative preparation of surgical cirrhotic patients with optimization of potentially reversible variables as an obligation for good postoperative outcome.  This new scoring system proved to be a more accurate and specific predictor of patient outcome (12).

(II): Scoring systems and liver transplantation
The primary principal underlying organ allocation in liver transplantation is to prioritize recipients who have a substantial risk of death. This requires a disease severity index based on generalizable, verifiable and easily obtained variables. (37)
When the Child-Pugh scoring system was used as a disease severity index for organ allocation, it had a number of limitations:
a) limited discriminatory ability:
The Child-Pugh scoring system has only 8 points of difference between the least and most sick transplant candidates. All patients with markedly abnormal 10-15 scoring points are C-classed equally.

b) Variability in the measurement of laboratory parameters:
The three objective elements (bilirubin, albumin, PT) may have variable evaluations by different laboratories.
c) Variability in the measurement of subjective parameters:
The two subjective elements (ascites, encephalopathy) may have different evaluations by different persons. Although the degree of ascites is recently based on ultrasound examination, no standard definition has been given to refractory encephalopathy.
Before 1997, the initial allocation scheme of the United Network of Organ Sharing (UNOS) status stratified liver transplantation urgency according to hospital status and waiting time before transplantation. Because prolonged waiting time gave recipients an advantage, early listing of patients with well-compensated liver disease became a common practice (38).  In 1997, the UNOS adopted a minimal Child-Pugh score greater than 7 as essential for patient listing (39).  However, accumulated waiting time, not the disease severity, continued to determine organ allocation.  In 1998, the UNOS status was redefined as status 2A, 2B and 3, based on the Child-Pugh scoring system. This could eliminate hospitalization as a criterion for allocation of donor livers. In 1999, it became apparent to the Department of Health and Human Services (DHHS) in USA that donated livers should be allocated according to uniform medical criteria to be developed by the transplantation community.

The MELD scoring system
In 2001, the article of Kamath et al (40) in Hepatology, established the value of a model for end-stage liver disease (MELD) in predicting survival in four independent adult populations with advanced liver disease.  In February 2002, the MELD score was instituted by UNOS for organ allocation. The MELD score ranges from 6 to 40 points, yielding a wide range of severity of illness, thus reducing the impact of waiting time on allocation.
The MELD scoring system includes bilirubin as a measure of liver function, creatinine as a measure of kidney function, international normalized ratio (INR) as a measure of the ability of the blood to clot and the etiology of liver disease (41).  The formula of the MELD score is: 3.8 log (bilirubin mg/dl) + 9.6 log (creatinine mg/dl) + 11.2 log (INR) + 6.4 etiology (0 if cholestatic or alcoholic, 1 if otherwise). Kamath et al (40) found that the inclusion of the etiology of liver disease as a parameter in the MELD score is not necessary.
The MELD score was originally designed to predict 3-month mortality after transjuglar intrahepatic portosystemic shunt surgery (TIPSS), done to control an episode of variceal bleeding (41).  The MELD score now replaces the previous UNOS status 2A, 2B and 3 for adult patients, but status 1 patients with acute liver failure and life expectancy less than 7 days remain in place and are not affected by MELD (40).
Merion et al (42) have recently shown that a change in MELD score over 30 days was a more significant determinant of mortality than a one-time measured score. For a patient with a stable MELD score for 30 days (MELD = 0), the mortality was lower than for a patient in whom the MELD score has increased (positive MELD), but higher than for a patient in whom the score have decreased (negative MELD).  A patient’s MELD score may go up or down over-time depending on the status of his/her liver disease, while being on the waiting list.
It is important to know the causes of deteriorating or improving MELD scores. An increase may be due to reversible factors related to worsening of liver disease as dehydration or sepsis.  On the other hand, a decrease over-time may indicate correction of such reversible factors. So long as transient and reversible factors are ruled out or corrected, an increase in MELD score over-time is related to worsening of liver function.  Therefore, the emphasis changed form waiting time to selecting the sickest patients for liver transplantation, keeping waiting time as a tie-breaker for patients with identical MELD scores.

MELD versus Child-Pugh scores
If ascites and encephalopathy as subjective criteria are eliminated from the Child-Pugh score, the remaining difference between MELD and Child-Pugh scores is creatinine in MELD versus albumin in Child-Pugh.  So, the magnitude of the superiority of MELD score over Child-Pugh score is limited to populations at high risk of renal failure (43).
Renal failure complicating chronic liver disease may be caused by hypovolaemia, HRS or intrinsic renal disease.  Volume loading to treat prerenal volume depletion mostly due to the use of diuretics, frequently treats transient increases in creatinine. In such group of patients, the accuracy of MELD and Child-Pugh scores to predict outcome should be similar with equivalent accuracy in the absence of renal failure (40).  In a study, allowing entry of renal failure patients, the predictive accuracy of MELD proved to be superior to Child-Pugh (44).  Thus, the main advantage of MELD over Child-Pugh score may be in waiting list patients who have the probability to develop renal failure.  While day-to-day fluctuations in bilirubin and PT are less dramatic, the dependence of MELD score on dramatic day-to-day fluctuating creatinine levels means that the MELD score also fluctuates.
The UNOS has mandated updates of the MELD score to reduce the risk of the effect of transient fluctuations in creatinine in determining organ allocation (43). The MELD score is recalculated every 7 days for patients with MELD score of 25 or more, every 30 days for those of 19-24, every 90 days for those of 11-18, and every year for those equal to or less than 10 scoring points.
It should be remembered that the MELD and Child-Pugh scores have 8 and 34 points range between the least and most sick transplant candidates respectively.  The MELD score increases as its three constituent parameters deteriorate, whereas individual scoring elements in the Child-Pugh score remain fixed once a defined threshold has been reached, i.e. all patients with 10-15 scoring points are C-classed equally.

The PELD scoring system
A pediatric end-stage live disease (PELD) score is currently used for candidates below 18 years of age (37).  It is similar to the MELD score, but uses a different statistical formula to recognize the needs for children growth and development. The used items are bilirubin (mg/dl), albumin (g/L), INR, growth failure (based on gender, height and weight) and age at listing (specially whether the child is less than one year old).
The PELD score now replaces the previous UNOS status 2B and 3 for pediatric patients, while status 1 patients with acute liver failure and life expectancy less than 7 days remain in place and are not affected by PELD.(37)

Allocation systems: an overview
The MELD and PELD formulas are objectable and verifiable yielding consistent results whenever the scores are calculated. They use specific programmable calculators with screens to collect the data of the scores.  However, added to the complexity of the used formulas, this has been considered, by some authors as a barrier for their bed-side clinical use (45-46).
The current MELD and PELD allocation systems do not identify patients who benefit the most from liver transplantation. They, rather, determine patients who should receive donor organs on priority. However, their effects must ideally include post-transplantation outcome, because allocation of donor livers to the sickest patients may yield significant increases in post-transplant mortality.  Nair et al (47), have examined the UNOS database and found that pretransplant renal failure was associated with diminished graft and patient survival. This is actually a complex issue because aiming to reduce mortality in sicker patients on the waiting list may be offset by increased mortality in less sick patients, with increased waiting times for donor organs (43).  This may also increase the frequency of hospitalization and morbid events and the total cost of medical care.  Answers to these concerns will await analysis of the impact of MELD and PELD scores on overall mortality and morbidity in liver transplanted patients.
Recently, the MELD score has been utilized in non-transplant patients, particularly with less severe chronic liver disease (48).  The predictive abilities of the MELD and the Child-Pugh scores for intermediate (1 year) and long-term (5 years) mortality has been compared in 1611 patients with chronic liver disease. The MELD score proved to be a valid prognostic score for intermediate mortality and to be equivalent to the Child-Pugh score in predicting survival.  In this study, the inclusion of hepatic encephalopathy proved to add a prognostic value to the MELD score (48).  On the other hand, another study has recently questioned the reason for replacing the Child-Pugh score by the MELD score for prognostic assessment of patients with live cirrhosis (49).
Goals for future research are to improve the accuracy and precision of clinical models and to develop quantitative methods for assessing hepatic function that can predict survival in the individual patient (43).

References

1. Gines P, Quintero E, Arroya V, et al. Compensated cirrhosis: natural history and prognostic factors. Hepatology 1987; 7: 122-128.
2. Jackson FC, Christophersen EB, Peternel W, et al. Preoperative management of patients with liver disease. Surg Clin North Am 1968; 48: 907-930.
3. Cayer d, Sohmer MF: Surgery in patients with cirrhosis. Arch Surg 1955; 71: 828-838.
4. Wirthlin LS, Urk Hv, Malt RB, et al. Predictors of surgical mortality in patients with cirrhosis and non-variceal gastroduodenal bleeding. Surg Gynecol Obstet 1974; 139: 65-68.
5. Maddrey WC, Boitnott JK, Bedine MS, et al.  Corticosteroid therapy for alcoholic hepatitis.  Gastroenterology 1978; 75: 193-199.
6. Sheth M, Riggs M, Patel T. Utility of the Mayo End-Stage Liver Disease (MELD) score in assessing prognosis of patients with alcoholic hepatitis. BMC Gastroenterology 2002; 2: 2-13.
7. Doberneck RC, Sterling WA, Laaison DC. Morbidity and mortality after operations in non-bleeding cirrhotic patients.  Am J Surg; 1983, 146: 306-309.
8. Garrison RN, Cryer HM, Howard DA, et al. Clarification of risk factors for abdominal operations in patients with hepatic cirrhosis. Ann Surg 1984; 199: 648-655.
9. Child CG, Turcotte JG. Surgery and portal hypertension. In: Child CG, (editor); The liver and portal hypertension. Philadelphia, WB Saunders 1964.
10.  Pugh RNH, Murray LIM, Dawson JL, et al. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973; 60: 646-649.
11.  Turcotte JG, Lambert MJ. Variceal hemorrhage, hepatic fibrosis and portocaval shunts. Surgery 1973; 73: 810-817.
12. Webster NR. Scoring systems and surgery in cirrhotic patients. Editorial. EgJ Anaesth 2004; 20: 1-3.
13. Armonis A, Patch D, Burroughs A. Hepatic venous pressure measurement, an old test as a new prognostic marker in cirrhosis Hepatology 1997; 25: 245-250.
14. Iwao T, Toyonaga A, Oho K, et al. Value of Doppler ultrasound parameters of portal vein and hepatic artery in the diagnosis of cirrhosis and portal hypertension. Am J Gastroenterol 1997; 92: 1012-1018.
15. Abd-El-Fattah S. Prognosis of liver disease: do we need Egyptian predictors? (Review article). Afro-Arab Liver J 2004; 3: 47-52.
16. Sherlock S, Dooley J. In: Diseases of the liver and biliary system. Sherlock S, Dooley J (editors).  11th ed. Blackwell Scientific Publication 2002.  Chapter 2, p19.
17. Oellerich M, Raude E, Burdelski M, et al. Monoethyl glycine xylidide formation kinetics: a novel approach to assessment of liver function. J Clin Biochem 1987; 25: 853-854.
18. Nicholas DM, David JB, Neil RD, et al.  Liver function and splanchnic ischemia in critically-ill patients. Chest 1997; 111: 180-187.
19.  Moemen ME, Gaafar TY, Farag WA, et al. A. Prognostic categorization in cirrhotic patients undergoing abdominal surgery: a randomized trial. Eg J Anaesth 2004; 20: 7-14.
20. Arief A, Papadakis MA. Hyponatremia and hypernatremia in liver disease. In: Epstein M, (editor), The kidney in liver disease. Baltimore 1988, Williams & Wilkins.
21.  Toblas MD. Evaluation of patients with hepatic disease. In: Principles and practice of anesthesiology. 2nd ed, chap 19, Longnecker DE, Tinker JH, Morgan Jr GE, (editors), Saint Louis (1998), Mosby.
22.  Curevas MV, Millan I, Gavaler JS, et al. Prognostic value of preoperatively obtained clinical and laboratory data in predicting survival following orthotopic liver transplantation. Hepatology 1986; 6: 922-930.
23.  Bloch RS, Allaben RD, Walt AJ. Cholecystectomy in patients with cirrhosis: a surgical challenge. Arch Surg 1985;120: 669-672.
24.  Aranha GV, Sontag SJ, Greenlee HB. Cholecystectomy in cirrhotic patients: a formidable operation.  Am J Surg 1982; 143: 55-59.
25.  Schwartz SI. Biliary tract surgery and cirrhosis: a critical combination. Surgery 1981; 90: 577-583.
26. Agusti AGN, Roca J, Bosch J, et al. The lung in patients with cirrhosis. J Hepatol 1990; 10: 251-257.
27.  Krowka M J, Cortese DA. Hepato – pulmonary syndrome: an evolving prospect in the era of liver transplantation. Hepatology 1990; 11: 138-143.
28.  Shoemaker WC, Appeal PL, Kram HB et al. Prospective trial of supranormal value of survivors as therapeutic goals in high risk surgical patients. Chest 1998; 94: 1176-1181.
29.  Fleming A, Bishop M, Shoemaker WC, et al. Prospective trial of supranormal values as goals of resuscitation in severe trauma. Arch Surg 1992; 127: 1175-1181.
30. Tuchschmidt J, Fried J, Astiz M, et al. Elevation of cardiac output and oxygen delivery improves outcome in septic shock. Chest 1992; 102: 216-220.
31. Bishop MH, Shoemaker WC, Appeal PL, et al. Prospective randomized trial of survivor values of cardiac index, oxygen delivery and oxygen consumption as resuscitation end points in severe trauma. J Trauma, Injury, Infection and Critical Care 1995; 38: 780-787.
32.  Boyd O, Grounds RM, Bennett ED. A randomized clinical trial of the effect of deliberate perioperative increase of oxygen delivery on mortality in high risk surgical patients. JAMA 1993; 270: 2699-2707.
33. Toblin MJ. Respiratory diseases in ICU. Am Rev Resp Dis 1988; 138: 1625-1640.
34. Gilbert R, Kerghlery JF. The arterial/alveolar oxygen tension ratio: an index of gas exchange applicable to varying inspired oxygen concentrations, Am Rev Resp Dis 1974; 109: 142-145.
35.  Gilbert R, Auchiclass H, Kuppinger MT. Stability of arterial/alveolar partial oxygen pressure ratio. Brit J Anaesth 1979; 37: 57-61.
36. Men SJ: Evaluation of arterial blood gas and acid-base homeostasis. Page26. In: Pulmonary disease. 4th ed, Litre, Brown Co. 1996 Boston.
37. Wiesner RH, Mediarmid SV, Kamath PS, et al. MELD and PELD: application of survival models to liver allocation. Liver Transpl 2001; 7: 567-580.
38.  Forman LM, Lucey MR. Predicting the prognosis of chronic liver disease: an evolution from Child to MELD. Hepatology 2001; 33: 473-475.
39. Lucey MR, Brown KA, Everson GT, et al. Minimal criteria for placement of adults on the liver transplant waiting list: a report of a national conference organized by the American Society of Transplant Physicians and the American Association for the Study of Live Diseases.  Liver Tranplant Surg 1997; 3: 628-637.
40. Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients with end-stage liver disease. Hepatology 2001; 33: 464-470.
41. Malinchoc M, Kamath PS, Gordon FD, et al. A model to predict poor survival in patients undergoing transjuglar intrahepatic protosystemic shunts. Hepatology 2000; 31: 864-871.
42. Merion RM, Wolfe RA, Dykstra DM, et al. Longitudinal assessment of mortality risk among candidates for liver transplantation. Liver Transpl 2003; 9: 12-18.
43. Everson GT. MELD: the answer to just more questions? Editorial. Gastroenterology 2003; 124: 1-6.
44. Weisner R, Edwards E, Freeman R, et al, and the United Network for Organ Sharing.  Liver Disease Severity Score Committee. The model for end-stage liver disease (MELD) and allocation for donor livers. Gastroenterology 2003; 124: 91-96.
45. Robert S, Brown JR, Shiva K, et al. Model for end-stage liver disease and Child-Turcotte-Pugh score as predictors of pre-transplantation disease severity: post-transplantation outcome and resource utilization in United Network for Organ Sharing status 2A patients. Liver Transpl 2002; 8: 278-286.
46. Sheth M, Riggs M, Patel T. Utility of the Mayo End-Stage Liver Disease (MELD) score in assessing prognosis of patients with alcoholic hepatitis.  BMC Gastroenterology 2002; 2: 2-13.
47. Nair S, Verma S, Thuluvath PJ. Pretransplant renal function predicts survival in patients undergoing orthotopic liver transplantation. Hepatology 2002; 35: 1179-1185.
48. Said A, Wiliams J, Hoden J, et al. Model for end-stage liver disease score predicts mortality across a broad spectrum of liver disease. Hepatology 2004; 40: 896-905.
49. Lee DH, Son JH, Kim TW. New scoring systems for severity outcome of liver cirrhosis and hepatocellular carcinoma: current issues concerning the Child-Turcotte-Pugh score and the Model of End-Stage Liver Disease (MELD) score. Available at: eastlee @hanyang.ac.kr.