Meld is superior to king's college and clichy's criteria to assess prognosis in fulminant hepatic failure

MELD Is Superior to King’s College and Clichy’s
Criteria to Assess Prognosis in Fulminant
Hepatic Failure
Silvina E. Yantorno,1 Walter K. Kremers,2 Andre
´s E. Ruf,1 Julio J. Trentadue,1 Luis G. Podesta
Federico G. Villamil1
1Liver Unit, Fundacio´n Favaloro, Buenos Aires, Argentina; 2William J. von Liebig Transplant Center,
Mayo Clinic, Rochester, MN
Assessment of prognosis in fulminant hepatic failure (FHF) is essential for the need and appropriate timing of orthotopic livertransplantation (OLT). In this study we investigated the prognostic efficacy of King’s College criteria, Clichy’s criteria, Model forEnd-Stage Liver Disease (MELD), and Pediatric End-Stage Liver Disease (PELD) in 120 consecutive patients with FHF.
Survival with medical therapy (18%), death without OLT (15%), and receipt of a liver transplant were similar in adults (n ϭ 64)and children (n ϭ 56). MELD scores were significantly higher in patients who died compared to those who survived withoutOLT, both in adults (38 Ϯ 7 vs. 26 Ϯ 7, P ϭ 0.0003) and children (39 Ϯ 7 vs. 23 Ϯ 6, P ϭ 0.0004). Using logistic regressionanalysis in this cohort of patients, concordance statistics were significantly higher for MELD (0.95) and PELD (0.99) whencompared to King’s College (0.74) and Clichy’s criteria (0.68). When data was analyzed in a Cox model including patientsreceiving transplants and censoring the time from admission, the concordance statistic for MELD (0.77) and PELD (0.79)remained significantly higher than that of King’s College criteria but not higher than that of Clichy’s criteria. In conclusion, thisstudy is the first to show that MELD and PELD are superior to King’s College and Clichy’s criteria to assess prognosis in FHF.
However, because data was generated from a single center and included a rather low number of patients who survived or diedwithout OLT, further confirmation of our findings is required. Liver Transpl 13:822-828, 2007. 2007 AASLD.
Received August 22, 2006; accepted December 19, 2006.
cations, 4 died on the waiting list, and only 28 (57%) received transplants. Similarly, applicability of OLT was66% in a multicenter study of 308 consecutive patients Fulminant hepatic failure (FHF) is the most severe and with FHF reported by Ostapowicz et al.4 Accurate as- dramatic of all liver diseases. Reported mortality rates sessment of prognosis early after referral is a key factor with supportive medical therapy range from 60 to90%.1,2 The advent of orthotopic liver transplantation for the appropriate timing of OLT and the outcome of (OLT) significantly improved outcome for adults and FHF. Effective prognostic markers should allow the dif- children with FHF. However, major benefits provided by ferentiation of patients likely to survive with medical OLT are limited by its relatively low applicability, either therapy, and thus with no need for OLT, from those due to development of contraindications such as irre- with poor prognosis in whom OLT should not be de- versible brain damage or multiorgan failure or the un- layed. At present, the King’s College criteria reported by availability of an organ donor in a timely fashion.3,4 O’Grady et al.5 and the Clichy’s criteria reported by Castells et al.3 showed that among 49 patients with Bernuau and Benhamou6 and Bernuau et al.7 are con- FHF meeting criteria for OLT, 17 developed contraindi- sidered to be the most valuable tools to assess progno- Abbreviations: FHF, fulminant hepatic failure; OLT, orthotopic liver transplantation; MELD, Model for End-Stage Liver Disease; PELD,Pediatric End-Stage Liver Disease.
Supported in part by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (DK 34238) and fromFundacio´n para la Docencia e Investigacio´n de las Enfermedades del Hı´gado (FUNDIEH).
Address reprint requests to Professor Federico G. Villamil, Liver Unit, Fundacion Favaloro, Avenida Belgrano 1782 (C1093AAS), Piso 5, BuenosAires, Argentina. Telephone: 54 11 4378 1366; FAX: 54 11 4378 1392; E-mail: fvillamil@ffavaloro.org DOI 10.1002/lt.21104Published online in Wiley InterScience (www.interscience.wiley.com).
2007 American Association for the Study of Liver Diseases.
sis in patients with FHF. The Model for End-Stage Liver only from patients with FHF of viral etiology. MELD and Disease (MELD) and Pediatric End-Stage Liver Disease PELD scores were calculated according to United Net- (PELD) have been found to be excellent predictors of work for Organ Sharing.14 One patient with chronic 3-month mortality in adults and children with chronic renal failure on hemodialysis developed fulminant hep- liver disease listed for OLT.8-10 However, experience atitis B and died on the waiting list. Otherwise, no with MELD and PELD in FHF is limited. Kremers et al.11 patient with FHF required renal replacement therapy recently investigated the ability of MELD to predict pre- for acute renal dysfunction. FHF was considered of and post-OLT survival in 720 patients listed as status 1 indeterminate etiology in patients with no previous ex- in the Organ Procurement and Transplantation Net- posure to hepatotoxic drugs and with negative immu- work/United Network for Organ Sharing. This study noglobulin M antibodies to hepatitis A virus, immuno- showed that patients with nonacetaminophen FHF had globulin M antibodies to hepatitis B core antigen, statistically significant lower survival rates while await- hepatitis B surface antigen, hepatitis C virus ribonu- ing OLT than those with primary nonfunction or he- cleic acid by qualitative polymerase chain reaction, au- patic artery thrombosis and that the risk of death cor- toantibodies, and metabolic markers. Fulminant auto- related significantly with MELD scores. In addition, the immune hepatitis was diagnosed in patients with no group with nonacetaminophen FHF had the greatest history of chronic liver disease, acute onset with coagu- survival benefit with OLT. The goal of the present study lopathy and encephalopathy, detectable autoantibod- was to investigate the prognostic accuracy of the King’s ies, and massive or submassive hepatic necrosis in the College criteria, Clichy’s criteria, MELD, and PELD in explant or liver biopsy. In this series, there were no cases of FHF due to acetaminophen toxicity. All pa-tients received standard medical therapy in the inten- sive care unit. Intracranial pressure monitoring wasindicated in those who progressed to stage 3-4 hepatic The study included 120 consecutive patients with FHF encephalopathy. Liver support devices were not utilized who were referred to our institution between June 1995 in this study. OLT was indicated in patients with stage and August 2004. Of these, 64 (53%) were adults and 4 hepatic coma and in those with progression or lack of 56 (47%) were children. Among the pediatric group, improvement of encephalopathy and/or coagulopathy only 5 patients (8.7%) were aged 11 to 16 yr. Due to the during hospitalization. Medical care and criteria for list- low number of cases, prognosis of FHF in adolescents ing and OLT remained mostly unchanged throughout was not analyzed either as a separate subgroup or in the 9-yr study period. The study was approved by the combination with adults. FHF was defined as the acuteonset of coagulopathy and hepatic encephalopathy within 8 weeks of initial symptoms in patients with no previous history of liver disease.12 Clinical variants of Data were summarized using means Ϯ standard devi- FHF were defined according to the criteria reported by ation (range) for numeric variables, and counts and Bernuau and Benhamou6 and O’Grady et al.13 King’s percents for categorical variables. Group comparisons College criteria, Clichy’s criteria, and MELD score were for numerical variables are based upon a t-test, bino- calculated based on the results of blood tests obtained mial variables based upon a Fisher’s exact test, and on hospital admission and compared to each other in other categorical variables based upon a chi-squared 40 of 120 patients (33%) who either survived or died test. Positive and negative predictive values, diagnostic without OLT. Patients who underwent OLT were ex-cluded from the analysis of prognosis, except for the accuracy, and concordance statistic are used to de- Cox model which included the entire cohort. PELD was scribe the predictive and discriminative value of the evaluated only in the pediatric population. Follow-up predictors of survival. Concordance between mortality MELD and PELD scores were not analyzed because for the 30-day period since FHF onset and predictors of several patients received transfusions of fresh frozen mortality were derived both from the logistic model to plasma before placement of intracranial pressure mon- allow comparison with earlier studies of survival in pa- itors or other invasive procedures. All data was col- tients with end-stage liver disease,9,15 as well as for the lected prospectively and analyzed retrospectively. Indi- Cox model,16,17 which accounts for the variable fol- cators of poor outcome of the King’s College criteria for low-up due to transplantation. For the logistic model, patients with nonacetaminophen FHF are either an in- concordance only took into consideration those pa- ternational normalized ratio of prothrombin Ͼ6.5 or tients who survived 30 days or who died within 30 days presence of at least 3 of 5 variables including age (Ͻ10 of FHF onset (n ϭ 41) and does not use information on or Ͼ40 yr), interval from jaundice to encephalopathy those individuals who were transplanted within 30 Ͼ7 days, indeterminate or drug-induced etiologies, in- days. One patient with sub-FHF underwent OLT be- ternational normalized ratio of prothrombin Ͼ3.5, or yond 30 days of admission and was therefore included serum bilirubin Ͼ300 ␮mol/L.5 Clichy’s criteria indi- in the logistic model. For the Cox model, concordance is cate a poor prognosis when hepatic encephalopathy is essentially the fraction of patient pairs in which the associated with factor V concentrations Ͻ20% for pa- model correctly identifies which patient died first. Cox tients aged Ͻ30 yr or Ͻ30% for those older than 30 analysis included the entire cohort of patients with yr.6,7 Of note, these prognostic variables were derived FHF. Differences in concordance for the prognostic LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases TABLE 1. Characteristics on Admission of Adults and Children With Fulminant Hepatic Failure Abbreviations: NA, not analyzed; INR, international normalized ratio.
TABLE 2. Outcome of Adults and Children With Fulminant Hepatic Failure Abbreviation: OLT, orthotopic liver transplantation.
*Multiorgan failure ϭ 6, brain death ϭ 3.
scores were assessed using the jackknife method. A P OLT (15%), and receipt of an OLT (67%) were similar in value Ͻ0.05 was considered statistically significant.
adults and children with FHF (Table 2). Among the 22patients who survived with medical therapy, only 4 (2/11 adults and 2/11 children) were listed for OLT.
The remaining 18 patients significantly improved or Demographics, etiology, clinical variants, and severity resolved hepatic encephalopathy within 48-72 hours of of liver failure in adults and children with FHF are hospitalization and therefore were not listed. Of the 18 described in Table 1. Hepatitis A was the most frequent patients who died without OLT, 9 had contraindications identifiable etiology in the pediatric group (48%) and for the procedure that were present on admission (5/11 autoimmune hepatitis (19%) and drug-induced hepato- adults and 4/7 children) and the other 9 died while toxicity (17%) among adults. FHF was classified as of awaiting an organ donor (Table 2). The diagnostic ac- indeterminate etiology in approximately one-third of curacy of King’s College criteria and Clichy’s criteria in both adults and children. Clinical variants of FHF, asdefined by the interval between jaundice and encepha- patients who either survived or died without OLT was lopathy,6,13 serum bilirubin, and MELD scores, did not 73% and 71%, respectively, as shown in Table 3. Cli- differ between groups. Although the proportion of chil- chy’s criteria had a higher positive predictive value dren with stage 3-4 hepatic coma was lower than in (87% vs. 65%) than King’s College criteria but a lower adults, international normalized ratio of prothrombin negative predictive value (67% vs. 83%). King’s College was significantly higher in the pediatric population.
criteria were more useful in adults (diagnostic accuracy Survival with medical therapy (18%), death without of 78% vs. 67%) and Clichy’s criteria in children (83% LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases MELD scores obtained on admission were signifi- TABLE 3. Positive Predictive Value, Negative cantly higher (P ϭ 0.03) in adults with FHF who died Predictive Value, and Diagnostic Accuracy of King’s after OLT (36 Ϯ 9, n ϭ 14) compared to those who College Criteria and Clichy’s Criteria in Patients Who survived (31 Ϯ 7, n ϭ 28). In contrast, no such differ- Survived or Died Without Liver Transplantation ence was observed in children (36 Ϯ 4 vs. 36 Ϯ 5) whodied (n ϭ 7) or survived (n ϭ 32) after liver transplan- A survey conducted in 2001 by the Argentina Society of Transplantation showed that among 212 adults with FHF referred for OLT there was not a single case of acetaminophen toxicity and that hepatitis A was the main cause of FHF in the pediatric group (127/219,58%) (F. Villamil, personal communication). Therefore, etiology of FHF in this study is representative of our negative predictive value; DA, diagnostic accuracy.
Over the last 2 decades, many static and dynamic vs. 61%). The diagnostic accuracy of the King’s College variables have been proposed to assess prognosis in criteria in adults with FHF of this study (78%) was patients with FHF. This rather long list includes, among similar to that of previously reported series (Table 4).
others: age,6,18,19 etiology,6,18,19 stage of encephalopa- MELD scores were significantly higher in patients who thy,6,18,19 biochemical tests, such as serum biliru- died compared to those who survived without OLT both bin,5,18 serum phosphate,20 alfa-fetoprotein,21 arterial in adults and children. Similar results were observed ketone body ratio,22 and vitamin D-binding protein,23 with PELD in the pediatric population (Table 5). Among coagulation parameters, such as prothrombin time,24 the 22 patients who survived with medical therapy, factor V,25 and factor VIII,26 and the extent of paren- MELD score was Յ30 in 20 (91%). Conversely, MELD chymal necrosis on biopsies obtained by the transjugu- was Ͼ30 in 17 of 18 (94%) patients who died without lar route.27 Although significant differences have been OLT (Fig. 1). PELD scores were Ͼ30 in the 7 children reported for some of these variables when comparing who died and Ͻ30 in 10 of 11 survivors (91%).
patients with FHF who survived or died, they are of little Using logistic regression analysis, all prognostic help to assess prognosis in an individual patient and, scores studied were significant predictors of death, with most importantly, to decide whether there is a need for concordance statistic values ranging from 0.68 to 0.99.
OLT. Since their original description in the late 1980s, However, concordance statistics were much higher for King’s College and Clichy’s criteria have been accepted MELD (0.95 in all patients) and PELD (0.99) when com- and validated as the most useful tools to establish the pared to Clichy’s criteria (0.68) and King’s College cri- risk of death and need for OLT among patients with teria (0.74), both in adults and children. (Table 6). Con- FHF.5,22,25,27-30 However, the major limitation of these sidering all patients, MELD score was significantly criteria is their low negative predictive value. As shown different from the King’s College criteria (P ϭ 0.0037) in Table 4, a significant proportion of patients with and Clichy’s criteria (P ϭ 0.0001). Comparison between negative criteria (23-70%) ultimately die or require OLT.
subgroups is described in Table 6. When the data was In addition, up to 21% of adults with FHF who fulfill analyzed in a Cox model including patients who sur- King’s College criteria will survive without OLT (Table vived or died without OLT and those who were trans- 4). These limitations mostly derive from the formula of planted censoring the time-interval from admission both the King’s College and Clichy criteria that allocate (n ϭ 120), the concordance score for MELD (0.77) and patients with FHF to only 2 categories, survival or PELD (0.79) remained higher than that of the Clichy’s death, which in clinical practice dictates the need for criteria (0.64). King’s College criteria were not a signif- OLT. The efficacy of a categorical score such as the icant predictor of death in this model (Table 7). Consid- King’s College criteria strongly relies on the accuracy of ering all patients, MELD score was significantly differ- its components to distinguish between these 2 major ent from the King’s College criteria (P ϭ 0.0001) and outcomes. As an example, when comparing 2 given marginally significantly different from the Clichy’s cri- patients, 1 with bilirubin of 40 mg/dL and international teria (P ϭ 0.064). Comparison between subgroups is normalized ratio of prothrombin of 6 and the other with 18 mg/dL and 3.6, respectively, no one will argue that Renal dysfunction, defined as serum creatinine con- the first case carries a higher risk of death. However, centrations Ͼ1.4 mg/dL, occurred in 5 of 11 (45%) according to the King’s College criteria, they both be- adults who died without OLT and in 3 of 11 (27%) who long to the same prognostic category. In contrast, out- survived with medical therapy. Mean serum creatinine come of 2 patients with FHF of the same etiology and was 2.8 Ϯ 3.0 mg/dL and 1.4 Ϯ 1.1 mg/dL, respec- bilirubin/international normalized ratio of prothrom- tively. All children had serum creatinine levels Ͻ1 mg/ bin of 16 mg/dL/3.4 and 18 mg/dL/3.6 should be similar, although according to the King’s College crite- LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases TABLE 4. Positive Predictive Value, Negative Predictive Value, and Diagnostic Accuracy of King’s College Criteria in Reported Series of Adults With Nonacetaminophen Fulminant Hepatic Failure Abbreviations: PPV, positive predictive value; NPV, negative predictive value; DA, diagnostic accuracy.
*Calculated from data described in the publication.
TABLE 5. MELD and PELD Scores in 40 Patients With Fulminant Hepatic Failure Who Survived or Died *P ϭ 0.0003 vs. survived and P ϭ 0.03 vs. transplanted.
†P ϭ 0.008 vs. survived.
‡P ϭ 0.0004 vs. survived.
Adults (n = 11) Children (n = 11)
Adults (n = 11) Children (n = 7)
Survived with medical therapy
Died with medical therapy
Figure 1.
MELD scores of 22 patients who survived with
medical therapy and 18 patients who died without liver trans-
plantation. Horizontal bars represent mean values.
ria they are allocated to the good and poor prognosticcategories, respectively.
MELD is a continuous score with no ceiling effect that were significantly higher among nonsurvivors (45 Ϯ 12) includes only 3 simple, readily available, objective, re- compared to survivors (34 Ϯ 13) and patients receiving producible, and quantitative variables. Validation stud- transplants (39 Ϯ 10).31,32 In our study, MELD score ies performed in the United States have shown that was Ն30 in 94% of patients who died without OLT and MELD is superior to a categorical score such as the Ͻ30 in 91% of those who survived with medical ther- Child-Turcotte-Pugh to assess the risk of death in pa- apy. Rather than proposing a value of MELD as a prog- tients with chronic liver disease.8,10 Our results sug- nostic dichotomous variable, our data suggest that gest that this is true also in FHF. The concordance MELD scores obtained upon admission may be of help statistic for MELD score in adults and children and for to establish the optimal timing for pre-OLT evaluation PELD score in the pediatric population with FHF, as and listing. However, the ideal cutoff value for MELD assessed by logistic regression, was Ͼ0.9 and signifi- requires further validation in larger and independent cantly higher than that of both King’s College and Cli- series of patients with FHF. Renal dysfunction occurred chy’s criteria. When patients receiving transplants were in 45% of adults who died with supportive medical included in the analysis using a Cox model, MELD and therapy. This represents an additional advantage of PELD remained as the most significant predictors of MELD over the King’s College and Clichy’s criteria, mortality within 30 days, with concordance statistics of whose formula does not include serum creatinine as a 0.77 and 0.79, respectively (Table 7). Of note, Clichy’s criteria were superior to King’s College criteria, espe- Worldwide, patients with FHF and those requiring cially in children. In agreement with our results, Aydin emergency re-OLT are listed in a special category with et al.31 recently showed that among 170 patients with priority for organ allocation designated as status 1 in FHF, MELD scores obtained on hospital admission the United States and emergency in other geographic LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases TABLE 6. Logistic Regression Analysis of King’s College Criteria, Clichy’s Criteria, MELD, and PELD in 41 Patients With Fulminant Hepatic Failure Who Survived or Died Without OLT Within 30 Days of Admission Abbreviations: NS, nonsignificant; CI, confidence interval; ϱ, infinity.
*P ϭ 0.15 (NS) vs. King’s College and 0.0026 vs. Clichy’s.
†P ϭ 0.004 vs. King’s College and 0.067 (NS) vs. Clichy’s.
‡P ϭ 0.0037 vs. King’s College and 0.0001 vs. Clichy’s.
§P ϭ 0.046 vs. King’s College and 0.12 (NS) vs. Clichy’s.
TABLE 7. Cox Analysis of King’s College Criteria, Clichy’s Criteria, MELD, and PELD in 120 Patients With Fulminant Hepatic Failure Who Survived or Died With Medical Therapy or Underwent Liver Transplantation Abbreviations: NS, non significant; CI, confidence interval; ϱ, infinity.
*P ϭ 0.001 vs. King’s College and 0.013 vs. Clichy’s.
†P ϭ 0.055 (NS) vs. King’s College and 0.09 (NS) vs. Clichy’s.
‡P ϭ 0.0001 vs. King’s College and 0.064 (NS) vs. Clichy’s.
§P ϭ 0.02 vs. King’s College and 0.69 (NS) vs. Clichy’s.
areas such as Argentina. Within this category, organs dren with FHF. We acknowledge that our study has a are allocated according to waitlist time. Kremers et al.11 number of limitations. First, the number of patients recently showed that among patients listed as status 1 who survived or died without OLT was rather small and in the United States, the risk of death was significantly the analysis included both adults and children, whose higher in FHF when compared to those requiring re- prognosis may differ. In addition, hepatitis A is an in- OLT for primary nonfunction or hepatic artery throm- frequent etiology of FHF in children from most geogra- phies. Second, conclusions generated from single-cen- In conclusion, this study is the first to show that ter data may not be confirmed when assessed in larger MELD and PELD are superior to the King’s College and studies or different patient populations. Last, the ab- Clichy’s criteria to assess prognosis in adults and chil- sence of acetaminophen toxicity limits the generaliz- LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases ability of our findings. Additional studies are therefore 16. Harrell FE, Lee KL, Mark DB. Multivariable prognostic required to further assess the prognostic accuracy of models: issues in developing models, evaluating assump- MELD and PELD in FHF. However, if our results are tions and adequacy, and measuring and reducing errors.
Stat Med 1996;15:361-387.
confirmed in larger series, the benefits of stratifying 17. Johnson RE, Quade D, Langstron RD. The MATPAR pro- patients with FHF within the status 1 category accord- cedure. In: SUGI Supplemental Library User’s Guide, ver- ing to their MELD scores should be explored.
sion 5. Cary, NC: SAS Institute, Inc., 1986. pp. 295-306.
18. Hoofnagle JH, Carithers RL, Shapiro C, Ascher N. Fulmi- nant hepatic failure: summary of a workshop. Hepatology 19. Dhiman RK, Seth AK, Jain S, Chawla YK, Dilawari JB.
1. Lee WM. Acute liver failure. N Engl J Med 1993;329:1862- Prognostic evaluation of early indicators in fulminant he- patic failure by multivariate analysis. Dig Dis Sci 1998;43: 2. Mas A, Rodes J. Fulminant hepatic failure. Lancet 1997; 20. Chung PY, Sitrin MD, Te HS. Serum phosphorus levels 3. Castells A, Salmero´n JM, Navasa M, Rimola A, Salo´ J, predict clinical outcome in fulminant hepatic failure. Liver Andreu H, et al. Liver transplantation for acute liver fail- ure: Analysis of applicability. Gastroenterology 1993;105: 21. Murray-Lyon IM, Orr AH, Gazzard B, Kohn J, Williams R.
Prognostic value of serum alpha-fetoprotein in fulminant 4. Ostapowicz G, Fontana RJ, Schiodt FV, Larson A, Davern hepatic failure including patients treated by charcoal hae- TJ, Han SHB, et al. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States.
22. Saibara T, Onishi S, Sone J, Yamamoto N, Shimahara Y, Mori K, et al. Arterial ketone body ratio as a possible 5. O’Grady JG, Graeme JM, Hayllar KM, Williams R. Early indicator for liver transplantation in fulminant hepatic indicators of prognosis in fulminant hepatic failure. Gas- failure. Transplantation 1991;51:782-786.
23. Lee WM, Galbraith RM, Watt GH, Hughes RD, McIntire 6. Bernuau J, Benhamou JP. Fulminant and subfulminant DD, Hoffman B, Williams R. Predicting survival in fulmi- liver failure. In: McIntyre N, Benhamou JP, Bircher J, nant hepatic failure using serum Gc protein concentra- Rizzeto M, Rodes J, editors. Oxford Textbook of Clinical Hepatology. Oxford: Oxford Medical Publications, 1991: 24. Harrison PM, O’Grady JG, Keays RT, Alexander GJ, Wil- liams R. Serial prothrombin time as prognostic indicator 7. Bernuau J, Goudeau A, Poynard T, Dubois F, Lesage F, in paracetamol induced fulminant hepatic failure. Br Yvonnet B, et al. Multivariate analysis of prognostic factors in fulminant hepatitis B. Hepatology 1986;6:648- 25. Izumi S, Langley PG, Wendon J, Ellis AJ, Pernambuco RB, Hughes RD, Williams R. Coagulation factor V levels as a 8. Wiesner RH, Edwards EB, Freeman RB, Harper AM, Kim prognostic indicator in fulminant hepatic failure. Hepatol- R, Kamath PS, et al. Model for end stage liver disease (MELD) and allocation of donor livers. Gastroenterology 26. Pereira LM, Langley PG, Hayllar KM, Tredger JM, Williams R. Coagulation factor V and VIII/V ratio as predictors of 9. Kamath PS, Wiesner RH, Malinchoc M, Kremers WK, Th- outcome in paracetamol induced fulminant hepatic fail- erneau TM, Kosberg CL, et al. A model to predict survival ure: relation to other prognostic indicators. Gut 1992;33: in patients with end stage liver disease. Hepatology 2001; 27. Donaldson BW, Gopinath R, Wanless IR, Phillips MJ, 10. Wiesner RH, McDiarmid SV, Kamath PS, Edwars EB, Ma- Cameron R, Roberts EA, Greig PD. The role of tran- linchoc M, Kremers WK, et al. MELD and PELD: applica- sjugular liver biopsy in fulminant liver failure: relation tion of survival models to liver allocation. Liver Transpl to other prognostic indicators. Hepatology 1993;18: 11. Kremers WK, Van Ijperen M, Kim WR, Freeman RB, 28. Pauwels A, Mostefa-Kara N, Florent C, Le`vy VG. Emergency Harper AM, Kamath PS, Wiesner RH. MELD score as a liver transplantation for acute liver failure: evaluation of predictor of pretransplant and posttransplant survival in London and Clichy criteria. J Hepatol 1993;17:124-127.
OPTN/UNOS status 1 patients. Hepatology 2004;39:764- 29. Anand AC, Nightingale P, Neuberger JM. Early indica- tors of prognosis in fulminant hepatic failure: an as- 12. Trey C, Lipworth L, Chalmers TC, Davidson CS, Gottlieb sessment of the King’s criteria. J Hepatol 1997;26:62- LS, Popper H, et al. Fulminant hepatic failure: presumable contribution to halothane. N Engl J Med 1968;279:798- 30. Shakil AO, Kramer D, Mazariegos GV, Fung JJ, Rakela J.
Acute liver failure: clinical features, outcome analysis, 13. O’Grady JG, Schalm SW, Willliams R. Acute liver failure: and applicability of prognostic criteria. Liver Transpl redefining the syndromes. Lancet 1993;342:273-275.
14. United Network for Organ Sharing. Home Page. Avail- 31. Aydin C, Berk BS, Fung JJ, Shakil AO. Applicability of able at http:/www.unos.org. Last accessed February MELD scoring system to predict prognosis in patients with acute liver failure [Abstract]. Hepatology 2003;38(Suppl 15. Hanley JA, Mc Neil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Ra- 32. Sass DA, Shakil O. Fulminant hepatic failure. Liver LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases

Source: http://fundieh.org.ar/media/MELD-en-hepatitis-fulminante.pdf