Model For End Stage Liver Disease (MELD) and Child-TurcottePugh (CTP) PDF
Model For End Stage Liver Disease (MELD) and Child-TurcottePugh (CTP) PDF
Model For End Stage Liver Disease (MELD) and Child-TurcottePugh (CTP) PDF
Original Article
Abstract
Background: There are many patients awaiting liver transplantation with only few donors providing the organ. The Child-Turcotte-Pugh
score (CTP) and the Model for End stage Liver Disease (MELD) are the most common scores for prioritizing patients on waiting lists. In this
study, we compared the ability of these scores to predict mortality or removal from the waiting list due to poor medical conditions.
Methods: A total of 257 patients were included in our study and we observed their status in the waiting list over a 9-month period. MELD
and CTP of the patients at the time of listing were calculated. We used both ROC-curve and Area Under the Curve (AUC) to calculate the
predictive ability of these scores.
Results: During follow up, 22 patients died and 9 patients were removed from the waiting list due to poor medical conditions. Comparing
the predictive ability of CTP and MELD, the AUC for CTP was larger than that of MELD (0.75 versus 0.69; P-value = 0.065). The best cutoff
point for discriminating mortality or removal from the waiting list due to severe deterioration is 8 for CTP and 13.67 for MELD. The sensitivity
DQGVSHFL¿FLW\ZDVDQGUHVSHFWLYHO\IRU&73DQGDQGUHVSHFWLYHO\IRU0(/'
Conclusion: The CTP score can predict mortality or removal from the liver transplantation waiting list better than the MELD over a
9-month period. However, better improved models need to be developed for prioritization of patients in the waiting list.
Cite this article as: Rahimi-Dehkordi N, Nourijelyani K, Nasiri-Tousi M, Ghodssi-Ghassemabadi R, Azmoudeh-Ardalan F, Nedjat S. Model for End stage Liver
Disease (MELD) and Child-Turcotte-Pugh (CTP) scores: Ability to Predict Mortality and Removal From Liver Transplantation Waiting List due to Poor Medical
Conditions. Arch Iran Med. 2014; 17(2): 118 – 121
center of Imam Khomeini Hospital, Tehran, Iran, which is one of R-package version 2.14.2.14
the two liver transplantation centers in Iran. We began the study
in September 2010 and followed the patients up to June 2011. Results
A diagnosis of cirrhosis was made by two physicians, based on
clinical, biochemical, ultrasonographic or computed tomography We studied the sickness behavior of 118 women and 139 men
¿QGLQJV DQG OLYHU ELRSV\ 7KH SDWLHQWV XQGHUZHQW DQ HODERUDWH over 9 months. The mean age of patients was 40.8 years (range:
evaluation for the etiology of liver disease. We did not include 18–69). BMI was available for only 149 patients. The mean BMI
patients younger than 18, those with a history of previous liver was 23.95 (range: 15.43–43.28). Among our patients, 34.2% suf-
transplantation or transjugular intrahepatic portosystemic shunt fered from cryptogenic liver disease and 19.5% from autoimmune
placement,orpatients with incomplete biochemical data. The hepatitis. Hepatitis B, hepatitis C, primary sclerosing cholangitis,
study was approved by the ethics committee of the hospital. Wilson’s disease, primary biliary cirrhosis and alcohol accounted
for 18.7%, 14.8%, 8.9%, 2.3%, 1.2%, and 0.4% of the patients,
MELD and CTP score respectively (Table 1). Ascites, hepatic encephalopathy, spontane-
,Q WKLV VWXG\ ZH XVHG WKH PRGL¿HG IRUP RI 0(/' VFRUH GH- ous bacterial peritonitis and hepatorenal syndrome were found in
scribed by Kamath and his team.8 The formulation is: 3.78 × Ln 47.1%, 20.2%, 0.8%, and 0.4% of patients, respectively. Other
(bilirubin mg/dL) + 11.2 × Ln (INR) + 9.57 × Ln (creatinine mg/ comorbidities not related to liver disease were diabetes (8.95%),
dL) + 6.43. To avoid negative scores, laboratory values less than 1 hypertension (2.3%), coronary artery disease (0.4%), and ulcer-
mg/dL were rounded off to 1, and the maximum allowed measure ative colitis (2.3%).
for creatinine was 4 mg/dLin the MELD score model. The mean waiting time on liver transplantation list was 814.67
7KHPRGL¿HG&73VFRUHLQFOXGHV¿YHSDUDPHWHUVVHUXPELOLUX- days (range 14–3667). At the end of our study, 12 patients were
bin and albumin levels, prothrombin time, and the presence and transplanted and 22 patients died. 9 patients were removed from
severity of as cites and encephalopathy.2 Some more sensitive im- the waiting list due to severe deterioration; one did not require
aging studies such as ultrasonography, in addition to physical ex- the liver transplantation anymore because of improvement, and
amination, were used to detect ascites. Encephalopathy was con- another patient voluntarily relinquished waiting (Table 2).
sidered in case of appearance of signs of occasional forgetfulness, In our univariate analysis, serum albumin and total bilirubin,
insomnia or distorted sleep pattern. INR, CTP, MELD, and presence of ascites and encephalopathy
We used MELD and CTP scores for all patients at the time of ZHUHVLJQL¿FDQWO\GLIIHUHQWEHWZHHQWKRVHVWLOOZDLWLQJRQWKHOLVW
listing for numerical analysis. and those who died or were removed from the waiting list due to
poor medical condition (P value < 0.05). Serum creatinine, however,
Statistical methods ZDVQRWVLJQL¿FDQWO\GLIIHUHQWEHWZHHQWKHWZRJURXSV7DEOH
Categorical variables were expressed as count with percentage, Statistical analysis for prediction of mortality or removal from
and continuous variables as mean with standard deviation. The the waiting list due to severe deterioration, showed AUC of 0.75
Mann-Whitney test and X2-test were used for continuous and cat- for CTP and 0.69 for MELD. Comparing the validity of these
egorical variables, respectively. In order to assess the validity of scores yielded a P value of 0.065 (Figure 1). The best cutoff for
MELD and CTP score for prediction of death or withdrawal from &73 ZDV ZLWK VHQVLWLYLW\ DQG VSHFL¿FLW\ ZKLOH IRU
the waiting list due to severe deterioration, we used the area un- 0(/' LW ZDV ZLWK VHQVLWLYLW\ DQG VSHFL¿FLW\
der the receiver operating characteristic (ROC) curve. In order to Nevertheless, 5 patients (16.1%) of those who died or were re-
compare the area under the two ROC curves, we used a non-para- moved from the waiting list due to poor medical condition had
metric test introduced by DeLong which exploits the properties of CTP < 8 and MELD < 13.67 (Table 4). The sensitivity and speci-
the Mann-Whitney statistic.13 All computations were done using ¿FLW\IRURWKHUFXWRIIVDUHOLVWHGLQ7DEOH
Table 1. Baseline characteristics at the onset of the study.
Mean (SD) Range
Age (year) 40.77 (13.45) 18–69
Body mass index* (kg/m2) 23.95 (4.53) 15.43–43.28
MELD† 14.06 (4.6) 6.43–30.41
CTP‡ 7.34 (1.96) 5–13
Number Percentage%
Female/Male 118/139 45.9/54.1
Etiology
Cryptogenic 88 34.2
Autoimmune hepatitis 50 19.5
Hepatitis B 48 18.7
Hepatitis C 38 14.8
Primary Sclerosing Cholangitis 23 8.9
Wilson’s disease 6 2.3
Primary Biliary Cirrhosis 3 1.2
Alcohol 1 0.4
Comorbidity
Chronic kidney disease 0 0
Coronary artery disease 1 0.4
Hypertension 6 2.3
Diabetes 23 8.95
Ulcerative colitis 6 2.3
Spontaneous bacterial peritonitis 2 0.8
Hepatorenal syndrome 1 0.4
Ascites 121 47.1
Hepatic encephalopathy 52 20.2
*Body mass index was available for 149 patients. †Model for End stage Liver Disease; ‡Child-Turcotte-Pugh score
Table 3. Univariate analysis comparing the group still on waiting list and the group expired or removed from the waiting list due to poor medical condition.
In list, (n=212) Died or worsen, (n=31)
P-value
Mean ± SD (95% CI), number Mean ± SD (95% CI), number
Serum total bilirubin (mg/dL) 2.32 ± 1.92 (2.06, 2.58) 4.52 ± 6.47 (2.15, 6.89) 0.024
Creatinine (mg/dL) 0.88 ± 0.24 (0.85, 0.91) 0.89 ± 0.26 (0.79, 0.98) 0.751
International Normalized Ratio 1.49 ± 0.43 (1.43, 1.55) 1.6 ± 0.47 (1.51, 1.86) 0.012
Albumin (g/dL) 3.71 ± 0.61 (3.63, 3.79) 3.09 ± 0.37 (2.95, 3.22) < 0.001
Presence of ascites 88 23 0.001
Presence of hepatic encephalopathy 33 13 < 0.001
CTP 7.01 ± 1.79 (6.77, 7.25) 8.68 ± 1.78 (8.03, 9.33) < 0.001
MELD 13.37 ± 4.2 (12.80, 13.94) 16.26 ± 4.36 (14.66, 17.86) 0.001
Table 4. CTP and MELD of the patients expired or removed from the list due to severe deterioration.
CTP < 8 &73
MELD < 13.67 5 3
0(/' 3 20
Table 5. 6HQVLWLYLW\DQGVSHFL¿FLW\IRUGLIIHUHQWFXWRIIVRIWKH0(/'DQG&73VFRUHV
MELD Sensitivity 6SHFL¿FLW\ CTP Sensitivity 6SHFL¿FLW\
13 0.77 0.50 6 0.93 0.23
13.5 0.74 0.55 7 0.90 0.46
13.6 0.74 0.57 8 0.74 0.67
13.67 0.74 0.58 9 0.61 0.79
13.7 0.70 0.58 10 0.25 0.88
13.8 0.67 0.59 11 0.16 0.95
13.9 0.67 0.60 12 0.06 0.98
14 0.64 0.60 13 0.00 0.99
CTP is superior in predicting mortality or removal from the wait- model for end-stage liver disease score to serum sodium ratio index as
ing list due to severe deterioration after 9 months; nevertheless, a prognostic predictor in patients with cirrhosis. Gastroenterol Hepa-
tol. 2009; 24: 1547 – 1553.
the P YDOXHLVERUGHUOLQH1HLWKHU0(/'VFRUH 2. Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R.
QRU&73VFRUHFRXOGSUHGLFWWKHVLWXDWLRQRIRISDWLHQWV Transection of the oesophagus for bleeding oesophagealvarices. BrJ
who had the prognosis of mortality or removal from the waiting Surg. 1973; 60: 646 – 649.
list due to severe deterioration in 9 months. 3. Christensen E. Prognostic models including the Child-Pugh, MELD
and Mayo risk scores--where are we and where should we go? J Hepa-
Some previous studies have shown that MELD performs better tol. 2004; 41: 344 – 350.
than CTP inpredicting mortality on liver transplantation waiting 4. Robert A, Chazouilleres O. Prothrombin time in liver failure: time,
list in 3 months.9,21,22 A prospective study by Wiesner et al., on ratio, activity percentage, or international normalized ratio? Hepatol-
ogy. 1996; 24: 1392 – 1394.
3,437 patients on liver transplantation waiting list in 2003 showed 5. Durand F, Valla D. Assessment of the prognosis of cirrhosis: Child-
WKDW0(/'LVVLJQL¿FDQWO\VXSHULRUWR&73IRUSUHGLFWLQJPRUWDO- Pugh versus MELD. J Hepatol. 2005; 42: S100 – S1007.
ity in 3 months (AUC 0.83 versus 0.76). The patients in that study 6. Malinchoc M, Kamath PS, Gordon FD, Peine CJ, Rank J, Ter Borgs
were listed as status 2A and 2B, but not status 3 (which usually re- PCJ. A Model to predict poor survival in patients undergoing tran-
sjugular intrahepatic portosystemic shunts. Hepatology. 2000; 31: 864
fers to patients with lower CTP values).10 Another study by Heu- – 871.
man et al., on 6,958 patients listed as status 2A, 2B and 3, showed 7. Freeman RB, Wiesner RH, Harper A, McDiarmid SV, Lake J, Edwards
WKDW &73 VOLJKWO\ EXW QRW VLJQL¿FDQWO\ RXWSHUIRUPV 0(/' LQ E, et al. The new liver allocation system: moving toward evidence-
predicting 3-month mortality (AUC 0.766 versus 0.759).12 Heu- based transplantation policy. Liver Transplant. 2002; 8: 851 – 858.
8. Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Thereneau TM,
man claimed that Wiesner’s conclusion, superiority of the MELD Kosberg CL, et al. A model to predict survival in patients with end-
score, is probably the result of a selection bias. Some other studies stage liver disease. Hepatology. 2001; 33: 464 – 470.
KDYHIDLOHGWRFRQ¿UPWKHVXSHULRULW\RI0(/'WR&7311,20 9. Wiesner RH, Mc Diarmid SV, Kamath PS, Edwards EB, Malinchoc
Although our results are in line with Heuman’s study, there are M, Kremers WK, et al. MELD and PELD: application of survival
models to liver allocation. Liver Transplant. 2001; 7: 567 – 580.
differences between the study populations. Cryptogenic disease 10. Wiesner R, Edwards E, Freeman R, Harper A, Kim R, Kamath P, et
and alcohol were the most and least common etiologies, respec- al. Model for end-stage liver disease (MELD) and allocation of donor
tively, for liver disease on our waiting list. Also, hepatitis B and livers. Gastroenterology. 2003; 124: 91 – 96.
autoimmune hepatitis were other major etiologies on our waiting 11. Cholongitas E, Marelli L, Shusang V, Senzolo M, RollesK,Patch D, et
al. A systematic review of the performance of the Model for End-stage
list. These are different from the above studies in which the most Liver Disease (MELD) in the setting of liver transplantation. Liver
common etiologies were hepatitis C and alcohol. Moreover, data Transplant. 2006; 12: 1049 – 1061.
in those studies were analyzed when transplantation was done 12. Heuman DM, Mihas AA. Utility of the MELD score for assessing
3-month survival in patients with liver cirrhosis: one more positive
based on the time the patients had waited. In our transplantation
answer. Gastroenterology. 2003; 125: 992 – 993. Author reply 994
center, however, MELD and CTP are used for prioritization and – 995.
this may have caused a selection bias, because patients transplant- 13. DeLong ER, DeLong DM, and Clarke-Pearson DL. Comparing the
ed were not included in the analysis. Therefore, we expect lower areas under two or more correlated receiver operating characteristic
curves: a nonparametric approach. Biometrics. 1988;44:837–845.
AUC for both CTP and MELD. Despite this bias, CTP and MELD 14. R Development Core Team (2012). R: A language and environment
still had AUC above 0.6 which shows that these scores still can be for statistical computing. R Foundation for Statistical Computing,
utilized for prioritizing patients on the waiting list. Vienna, Austria. ISBN 3-900051-07-0. Available from: URL: http://
A similar study by Gotthardt showed that the optimum cutoff www.R-project.org/.
15. Alavian SM, Fallahian F, Lankarani KB. The changing epidemiology
to predict mortality or the need for removal from the waiting list of viral hepatitis B in Iran. J Gastrointest Liver dis. 2007; 16: 403 –
due to severe deterioration in one year is 9 for the CTP and 14.4 406.
for the MELD.20 In our study, the MELD and CTP means and 16. Khademolhosseini F, Malekhosseini SA, Salahi H, Nikeghbalian S,
cutoffs were smaller than those in Gotthardt’s study. The reason is Bahadori A, Lankarani KB. Outcome and characteristics of patients
on the liver transplant waiting list: Shiraz experience. Middle East J
probably that we included patients some of whom had entered the Digest Dis. 2009; 11: 63 – 67.
waiting list before the onset ofour study and had better prognosis. 17. 6DEHUL¿URR]L06HUDWL$0DOHNKRVVHLQL6$6DODKL+%DKDGRU$OL
Therefore, it resulted in their survival during the time on list. As Lankarani KB. Analysis of patients listed for liver transplantation in
the previous studies have shown, lower scores of CTP and MELD Shiraz, Iran. Indian J Gastroenterol. 2006; 25: 11 – 13.
18. Murray KF, Carithers RL. AASLD practice guidelines: evaluation of
are associated with better prognosis.2,6 Thus, it seems reasonable the patient for liver transplantation. Hepatology. 2005; 41: 1 – 26.
that CTP and MELD means and cutoffs yielded lower values for 19. Fink MA, Berry SR, Gow PJ, Angus PW, Wang BZ, Muralidharan
our study population. V, et al. Risk factors for liver transplantation waiting list mortality. J
Gastroenterol Hepatol. 2007; 22: 119 – 124.
Our study does not contradict the use of the CTP or the MELD
20. Gotthardt D, Weiss KH, Baumgärtner M, Zahn A, Stremmel W,
scores for allocating patients in the liver transplantation waiting Schmidt J, et al. Limitations of the MELD score in predicting mor-
list in short term and shows that the CTP score outperforms the tality or need for removal from waiting list in patients awaiting liver
MELD’s in predicting mortality or removal from the waiting list transplantation. BMC Gastroenterol. 2009; 9: 72.
21. Huo TI, Wu JC, Lin HC, Lee FY, Hou MC, Lee PC, et al. Evaluation
due to bad medical conditions in 9 months. However, our overall of the increase in model for end-stage liver disease (DMELD) score
perception and conclusion is a need for more valid models of pri- over time as a prognostic predictor in patients with advanced cirrho-
oritizing patients on liver transplantation waiting lists. We suggest sis: risk factor analysis and comparison with initial MELD and Child-
future studies to seek variables that can improve such prioritiza- Turcotte-Pugh score. J Hepatol. 2005; 42: 826 – 832.
22. Huo TI, Lin HC, Wu JC, Lee FY, Hou MC, Lee PC, et al. Proposal of
tion algorithms. DPRGL¿HG&KLOG7XUFRWWH3XJKVFRULQJV\VWHPDQGFRPSDULVRQZLWK
the Mode for End-stage Liver Disease for outcome prediction in pa-
References tients with cirrhosis. Liver Transplant. 2006; 12: 65 – 71.
1. Lv XH, Liu HB, Wang Y, Wang BY, Song M, Sun MJ. Validation of