ORIGINAL ARTICLE

MELD Era: Is This Time to Replace The Original Child-Pugh Score in

Patients with Decompensated Cirrhosis of Liver
Samiullah Shaikh1, Hanif Ghani1, Sadik Memon2, Ghulam Hussain Baloch1, Mukhtiar Jaffery1 and Khalid Shaikh1

ABSTRACT
Objective: To compare the predictive value of MELD (Model of end stage liver disease) and Child-Pugh (CP) scores in
patients with decompensated cirrhosis of liver.
Study Design: Descriptive study.
Place and Duration of Study: Medical Department, Liaquat University of Medical and Health Sciences, Jamshoro/
Hyderabad, from August 2006 to October 2007.
Methodology: This study included 110 consecutive patients with decompensated cirrhosis of liver diagnosed either
clinically or radiologically were followed-up during hospital stay. Studied variables included demographic data, cirrhosis
related complications and investigations. Patients were classified according to original CP classification into A, B and C.
MELD score was estimated from serum bilirubin, serum creatinine and INR (International normalized ratio) of the patients.
Duration of hospitalization and in-hospital mortality were made as the end points of the study. T-test and Chi-square test
were done for continuous and categorical data. Original CP and MELD score were compared by the ROC curve. 0.05 was
kept as the level of significance.
Results: There were 110 patients with decompensated cirrhosis of liver. Mean age was 46.76+12.93 years. There were
72 (65%) male and 38 (35%) females patients. Hepatitis C was the most prevalent cause of cirrhosis of liver present in
60/110 (60%) cases. Ascites was present in 93/110 (83%) patients. The mean MELD scores were 2.23+0.712 (95% CI
2.09 – 2.36) and for CTP 2.52+0.586 (95%; CI 2.41-2.63). The outcome of the patients were 12 deaths (11%); 54 (49%)
remained hospitalized for up to 14 days and 44 (40%) for > 14 days. The majority of deaths and prolong hospitalization
were found in patients with MELD score > 15 as well as with Child-Pugh grade C. The c-statistic was 0.726 (p=0.001) for
CP score, and 0.642 for MELD score (p=0.021).
Conclusion: The MELD score was not found to be superior to CTP score for short-term prognostication of patients with
cirrhosis in this study.

Key words: MELD. Child pugh score. Cirrhosis.

INTRODUCTION patients are grouped into three classes. Patients with

Patients with cirrhosis of liver not only remain
hospitalized for prolong period of time but also die due
to cirrhosis related complications. Because of this
increase in morbidity and mortality, it is very important to
know about the good prognostic markers in patients
with decompensated cirrhosis of liver.1,2 Since the last
40 years, Child-Turcotte-Pugh (CP) score has been
used successfully for assessing the prognosis of
patients with cirrhosis.3 The Child-Pugh gain the
popularity because of its simplicity as it can be used at
the bed side.4,5 The CP score included the presence and
severity of ascites and encephalopathy, prolongation of
prothrombin time and the levels of albumin and bilirubin.
It ranges from 5 to 15. On the basis of their score
1 Department of Medicine, Liaquat University of Medical and
Health Sciences, Jamshoro/Hyderabad.
2 Department of Medicine, Isra University Hospital, Hyderabad.
Correspondence: Dr. Samiullah Shaikh, House No.55, Green
Homes, Qasimbad. Hyderabad.
E-mail: shaikh135@hotmail.com,
Received May 29, 2008; accepted April 19, 2010.
score 5-6 were named as CP class A, with 7-9 as class
B and > 9 as class C.6 Although the predictive value of
CP score has been proved over the years in many
studies,2,7 recently it has been challenged because of
inclusion of ascites and encephalopathy being
subjective variables with interobserver variability as well
as the ceiling effect of the CP score.8 This led to the
search for another scoring system which should be
more objective and had better prognostic markers. In
1999 the Model for end stage liver disease (MELD) was
introduced for assessing the risk of mortality and
morbidity in patients with cirrhosis of liver.9 MELD score
calculation is based on the etiology of cirrhosis and
three simple and objective laboratory variables, viz.
serum bilirubin, serum creatinine and prothrombin time
expressed as international normalized ratio (INR).The
MELD score is not without pitfalls as it includes
logarithmic transformation and multipilication by several
factors which require a software.6,16 The MELD score
ranges from 8 to 40 and is preferred over CTP because
it includes objective variables, lack of ceiling effect and
similarity of its results in various centers and in particular
inclusion of serum creatinine which is regarded as an

432 Journal of the College of Physicians and Surgeons Pakistan 2010, Vol. 20 (7): 432-435
MELD score in decompensated cirrhosis of liver
independent predictor of survival during the course of
the disease.10-12 MELD score was initially designed for
patients undergoing transjugular intrahepatic porta-
systemic shunts (TIPS) in order to set out the priority for
liver transplant allocation in patients with decom-
pensated cirrhosis of liver.9,13 With the passage of time
MELD is now being used as a predictor of survival in all
patients irrespective whether patient is undergoing for
transplant or not.9 Although many studies have
compared the predictive value of MELD and original CP
in assessing the survival of patients undergoing for TIPS
for allograft allocation, only a few have examined the
usefulness of these models in patients hospitalized with
decompensated cirrhosis not undergoing the TIPS
procedure. Patients with more severe disease remain in
the hospital for a longer period of time which is also as
important as survival of the patients.
The aim of this study was to compare the accuracy of
MELD and CP score for predicting prolong hospi-
talization as well as in-hospital mortality in patients with
decompensated cirrhosis of liver.
METHODOLOGY
This descriptive study included 110 consecutive patients
with cirrhosis of liver admitted in the Department of
Medicine between August 2006 to October 2007. Primary
objective of the study was in-hospital mortality or prolong
hospitalization for more or less than 14 days. Patients with
hepatocellular carcinoma, severe primary cardiopulmonary
failure or intrinsic kidney disease were excluded. The
diagnosis of decompensated cirrhosis was based on
clinical, laboratory and radiological signs of cirrhosis with at
least one sign of liver decompensation (ascites, variceal
bleeding, hepatic encephalopathy and non-obstructive
jaundice). The patient was considered to be a case of
hepatitis B if found positive for HBsAg and hepatitis C if
positive for anti-HCV antibodies. Patients’ biodata, detailed
medical history, complete physical examination and
laboratory test were enrolled in a well designed proforma.
The age, gender, cause of cirrhosis, reason of admission,
first and previous complications of decompensated
cirrhosis including history of encephalopathy, H/O
haematemesis was taken as well as complete blood count
including platelet count, prothrombin time and INR, serum
urea and creatinine, total bilirubin, alanine amino-
transferase (ALT), alkaline phosphatase, serum albumin
and globulins and ascitic fluid characteristics were recorded
for all patients. The CP score (range 5-15) was calculated
for every patient included in the study and each patient
was classified according to the suggestion by Pugh et al.
into grade A (up to 6), grade B (7-9) and grade C (> 9).6
MELD score which ranged from 6-40 was calculated
according to the formula proposed by Kamath et al.9 i.e.
[9.57 x Log creatinine (mg/dL) + 3.78 x Log bilirubin
(mg/dL) + 11.20 x Log INR + 6.43] which was a slight
modification of the risk score used in the original TIPS
Journal of the College of Physicians and Surgeons Pakistan 2010, Vol. 20 (7): 432-435
model.13 The patients were categorized into A, B, and C
with MELD score < 10, 10-15 and >15 respectively.
According to outcome, the patients were grouped into three
with A assigned for in-hospital death, B for hospitalization
< 14 days and C for those patients who remained
hospitalized for > 14 days.
Continuous variables such as age, bilirubin, hemoglobin,
platelet count, serum albumin, serum creatinine and serum
sodium were expressed as mean with standard deviation.
Categorical variables such as gender, cause of cirrhosis,
clinical features, Child-Pugh class, MELD score and
outcome were expressed in frequency and percentage.
Univariate analysis of variance were used for comparing
the MELD and Child pugh score for the outcome of the
patients. Receiver-operating characteristic (ROC) curves
were used to determine the cut off values of CTP score and
MELD and modified CTP score with the best sensitivity and
specificity in discriminating between patients who remained
hospitalized for prolonged period and those who died. The
validity of the models was measured by means of the
concordance (c) statistic (equivalent to the area under the
ROC curve), and the c-statistic of models were compared
using chi-square test as described by Hanley and McNeil.14
A p-value of 0.05 or less was considered as statistically
significant. A model was considered to have diagnostic
accuracy if the c-statistic was 0.5 and excellent diagnostic
accuracy if the c-statistic was > 0.5. All calculations were
done using SPSS version 16 (Chicago, IL, USA).
RESULTS
There were 110 consecutive patients with
decompensated liver cirrhosis. The mean age of the
patients was 46.76+12.93 years. There were 72 (65. %)
male and 38 (35%) female patients. Regarding the
etiology of cirrhosis, 60 (55%) were suffering from
hepatitis C, 38 (35%) from hepatitis B and 12 (11%) had
a combination of hepatitis C and B. Ascites was the
most prevalent feature on clinical examination being
present in 93 (84%) cases. History of encephalopathy
was found in 19 (17%), hematemesis in 27 (24%) and
diuretic therapy in 67 (61%). The outcome of the
patients was such that there were 12 deaths (11%); 54
(49%) remained hospitalized for up to 14 days and 44
(40%) for > 14 days as shown in Table I. The table also
shows the relationship of MELD and CTP with the
outcome of the patients. With univariant analysis of
variance a non-siginificat difference was found
comparing the MELD and CP for the outcome of the
patient (p=0.46). The ROC curves of CP score, and
MELD score are shown in Figure 1. The c-statistic for
CP was 0.726 ranging from 0.633 to 0.82 (p=0.001) and
0.642 for MELD score ranging from 0.53 to 0.745
(p=0.021). Table II shows the comparison of sesitivity,
specificity, positive and negative predictive values
between CP and MELD.
433
Samiullah Shaikh, Hanif Ghani, Sadik Memon, Ghulam Hussain Baloch, Mukhtiar Jaffery and Khalid Shaikh

CP system was further confirmed by Teh-Ia Huo et al.

who found CP score as a better predictor of mortality
and morbidity than MELD.16 Durand in a meta-analysis
of comparison between CP and MELD in different study
populations, focused on TIPS, transplantation or
Abbreviation: cirrhosis in general suggested that the accuracy for
ROC= Reciver operating predicting outcome of Child-Pugh score is superior to
characteristic
MELD= Model of end stage that of MELD.17
liver disease
CP= Child-Pugh
Receiving-operating-characteristic (ROC) curve and the
derived c-statistic provide a global and standardized
appreciation of the accuracy of a marker or a composite
score for predicting an event. This statistic allows a
simple comparison of the accuracy of different
prognostic scores within the same population. In this
study, c-statistics for prediction of in-hospital mortality or
Figure 1: ROC curve of MELD and Child-Pugh.. prolong hospitalization by the MELD score ranged from
0.53 to 0.74 (p=0.021) being compatible with previous
Table I: Baseline characteristics of 110 patients. findings in other retrospectively evaluated cohorts of
Parameters Median (range) Mean + SD 95% confidence patients with decompensated cirrhosis.18 The c-statistics
Intervel
of Child-Pugh ranged from 0.63 to 0.82 (p=0.001).
Age (in years) 45 (15-80) 46.76 + 12.9 44.32-49.21
Hemoglobin-g/dl 9.5 (3.8-14.2) 9.190 + 1.9 8.82-9.55 According to Durand the accuracy of Child-Pugh score
Platelet count 157 (55-310) 155.2 + 65.1 154-189 is not always inferior to that of MELD.17 In some
Bilirubin mg/dl 1.8 (0.3-7.9) 2.04 + 1.25 1.6-2.5 instances, it can be equivalent or even superior. As a
INR 1.4 (1.0-4.2) 1.7 + 0.79 1.58-1.88 result, there is no clear evidence, that applying MELD
S.Albumin-g/dl 2.6 (1.2-5.7) 2.7 + 0.88 2.60-2.94
S.Sodium-mmol/l 137 (120-156) 136 + 6.1 135.2-137.53
score to a single patient provides a prognostic
S.Creatinine-mg/dl 0.9 (0.6-3.4) 1.07 + 0.44 0.99-1.63 information superior to that of Child–Pugh score.
Gender n (%) MELD Score n (%)
The MELD score has shown to be superior to CP score
Male 72 (65 %) < 10 18 (16.4%)
Female 38 (35%) 10-15 49 (44.5%)
in having only objective variables that are statistically
> 15 43 (39.1%) weighted and a continuous scale with no ceiling or floor
Cause of cirrhosis: effects.8 It has been shown to be equivalent to or better
HBV 38 (35%) than CP score in predicting short, intermediate or long-
HCV 60 (54%) Child Pugh –
HBV+HCV 12 (11%) Grade
term survival in patients with decompensated liver
Jaundice 54 (49.1%) A 5 (4.5%) disease.18,19 On the other hand, MELD score cannot be
Odema 67 (60.9%) B 43 (39.1%) calculated at the bedside and is much more complex
Anemia 70 (63.6) C 62 (56.4%) than the easy to calculate CP score, since it includes
Ascites 93 (84.5%) Outcome
logarithmic transformations and multiplication by several
Splenomegaly 82 (74.5%) Death 5 (4.5%)
H/o hematemesis 27 (24.5%) upto 14 days 53 (48.2%)
factors.18
H/o diueretic 67 (60.9%) > 14 days 49 (44.5%) The etiology of cirrhosis of liver was found to be either
H/o encephalopathy 19 (17%)
hepatitis C or hepatitis B in accordance to the Western
and far Eastern studies where hepatitis C and hepatitis
Table II: Results of AUCROC (Area under curve receiver operating
characteristic) MELD and CP with the outcome of the patients. B respectively, are the most common causes of
Variables Sensitivity Specificity PPV NPV p-value cirrhosis.20,21 Ascites is a common form of decom-
(%) (%) pensation in patients with cirrhosis as evident from this
MELD 44 75 16.48 92.3 0.021 study population, 84% of whom had ascites. This is in
CP 77.7 63 18.9 96..2 0.001
MELD = Model of end stage liver disease; CP = Child-Pugh; PPV = Positive predictive value;
concurrence with other reported studies where the
NPV = Negative predictive value. incidence ranges from 57% to 86%,22,23 thus suggesting
that CP score may be a more suitable prognostic
DISCUSSION indicator in the local population. Mishra found ascites to
In this study CP was found to be a better prognosis be present in 64% of patient he studied.22 Nazish in a
indicator than MELD for predicting in-hospital mortality small study comprising 50 patients, found ascites in
as well as prolonged hospitalization in patients with 52% of cases.23 At the time of presentation, 4.5%
decompensated cirrhosis of liver. According to Said et al. patients were in CP class A, 30% in CP-B and 65.5% in
MELD score seemed to be slightly inferior to CP score CP-C. Sarin et al. stated that 26.3%, 30.2% and 43.4%
in predicting short-term mortality in patients with liver of their patients were in CP A, CP B, and CP C classes,
cirrhosis and HCC.15 The prognostic significance of the respectively.24

434 Journal of the College of Physicians and Surgeons Pakistan 2010, Vol. 20 (7): 432-435
MELD score in decompensated cirrhosis of liver
Child-Pugh classification has been a reference for more
than 30 years for assessing the prognosis of cirrhosis.
MELD score comes as the most serious challenger for
replacing Child-Pugh score and overcoming its
limitations. The principal advantages of MELD score are
that (a) it is based on variables selected by statistical
analysis rather than clinical judgment, (b) the variables
are objective and unlikely to be influenced by external
factors, (c) each variable is weighted according its
proper influence on prognosis and (d) the score is
continuous which helps scoring individuals more
precisely among large populations. However, MELD
score also has limitations including the need for
computation, the absence of clearly defined cut-off
values for categorizing cirrhotic patients and the
absence of validation in some clinical trails.
CONCLUSION
The MELD score was not found to be superior to CTP
score for short-term prognostication of patients with
cirrhosis in this study.
REFERENCES
1. D'Amico G, Morabito A, Pagliaro L, Marubini E. Survival and
prognostic indicators in compensated and decompensated
cirrhosis. Dig Dis Sci 1986; 31:468-75.
2. Forman LM, Lucey MR. Predicting the prognosis of chronic liver
disease: an evolution from Child to MELD. Mayo end-stage liver
disease. Hepatology 2001; 33:473-5. Comment in: Hepatology
2001; 34:215-6.
3. Oellerich M, Burdelski M, Lautz HU, Rodeck B, Duewel J, Schulz
M, et al. Assessment of pre-transplant prognosis in patients with
cirrhosis. Transplantation 1991; 51:801-6.
Child CG, Turcotte JG. Surgery and portal hypertension. Major
4.
Probl Clin Surg 1964; 1:1-85.
5. Conn HO. A peak at the Child-Turcotte classification. Hepatology
1981; 1: 673-6.
Reuben A. Child comes of age. Hepatology 2002; 35:244-5.
6.
7. Infante-Rivard C, Esnaola S, Villeneuve JP. Clinical and
statistical validity of conventional prognostic factors in predicting
short-term survival among cirrhotics. Hepatology 1987; 7:660-4.
8. Freeman RB Jr, Edwards EB. Liver transplant waiting time does
not correlate with waiting list mortality: implications for liver
allocation policy. Liver Transpl 2000; 6:543-52.
9. Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau
TM, Kosberg CL, et al. A model to predict survival in patients with
end-stage liver disease. Hepatology 2001; 33:464-70.
10. Pagliaro L. MELD: the end of Child-Pugh classification? J Hepatol
2002; 36:141-2.
11. Christensen E, Krintel JJ, Hansen SM, Johansen JK, Juhl E.
Prognosis after the first episode of gastrointestinal bleeding or
● ● ● ● ●
Journal of the College of Physicians and Surgeons Pakistan 2010, Vol. 20 (7): 432-435
coma in cirrhosis: survival and prognostic factors. Scand J
Gastroenterol 1989; 24:999-1006.
12. Fernández-Esparrach G, Sánchez-Fueyo A, Ginès P, Uriz J,
Quintó L, Ventura PJ, et al. A prognostic model for predicting
survival in cirrhosis with ascites. J Hepatol 2001; 34:46-52.
13. Malinchoc M, Kamath PS, Gordon FD, Peine CJ, Rank J, ter
Borg PCJ. A model to predict poor survival in patients
undergoing transjugular intrahepatic portosystemic shunts.
Hepatology 2000; 31:864-71. Comment in: Hepatology 2000;
32:881.
14. Hanley JA, McNeil BJ. The meaning and use of area under a
receiver operating characteristic (ROC) curve. Radiology 1982;
143:29-36.
15. Said A, Williams J, Holden J, Remington P, Gangnon R, Musat
A, et al. Model for end stage liver disease score predicts mortality
across a broad spectrum of liver disease. J Hepatol 2004; 40:
897-903.
16. Huo T, Lin H, Wu J, Lee F, Hou M, Lee P, et al. Proposal of a
modified Child-Turcotte-Pugh scoring system and comparison
with the model for end-stage liver disease for outcome prediction
in patients with cirrhosis. Liver Transpl 2006; 12:65-71.
17. Durand F, Valla D. Assessment of the prognosis of cirrhosis:
child-pugh versus MELD. J Hepatol 2005; 42:S100-7. Epub 2004
Dec 24.
18. Botta F, Giannini E, Romagnoli P, Fasoli A, Malfatti F,
Chiarbonello B, et al. MELD scoring system is useful for
predicting prognosis in patients with liver cirrhosis and is
correlated with residual liver function: a European study. Gut
2003; 52:134-9.
19. Schepke M, Roth F, Fimmers R, Brensing KA, Sudhop T, Schild
HH, et al. Comparison of MELD, Child-Pugh and emory model for
the prediction of survival in patients undergoing transjugular
intrahepatic portosystemic shunting. Am J Gastroenterol 2003;
98:1167-74.
20. Stroffolini T, Sagnelli E, Almasio P, Ferrigno L, Craxi A, Mele A,
et al. Characteristics of liver cirrhosis in Italy: results from a
multicenter national study. Dig Liver Dis 2004; 36:56-60.
21. Sakugawa H, Nakasone H, Nakayoshi T, Kawakami Y,
Yamashiro T, Maeshiro T, et al. Clinical characteristics of patients
with cryptogenic liver cirrhosis in Okinawa, Japan.
Hepatogastroenterology 2003; 50:2005-8.
22. Mishra P, Desai N, Alexander J, Singh DP, Sawant P.
Applicability of MELD as a short-term prognostic indicator in
patients with chronic liver disease: An Indian experience.
J Gastroenterol Hepatol 2007; 27:1232-5.
23. Zahra N, Inayatullah M, Nasir SA, Arshad M, Salma T, Naqvi AB.
Liver cirrhosis: clinical presentation. Professional Med J 2002;
9:207-12.
24. Sarin SK, Chari S, Sundaram KR, Ahuja RK, Anand BS, Broor
SL. Young versus adult cirrhotics: a prospective, comparative
analysis of the clinical profile, natural course and survival. Gut
1988; 29:101-7.
✯
● ● ● ● ●
435