Staging and prognosis of chronic liver disease in ICU


This topic is neglected in the fellowship exam. Of the historical SAQs available to this author, only Question 1d from the second paper of 2001 has ever touched on this topic. The question is one of the older style long-form questions, the answer to which required a degree of thinking. The topic of the question was a 42 year old alcoholic who presented in haemorrhagic shock due to a variceal bleed, and went on to develop hepatic encephalopathy. In Part 4 of this 4-part question, the wife asks you “what are his chances of survival”. The college model answer is somewhat evasive, sounding a bit like a gastroenterology consultant on the phone at night. Something is muttered about reversible factors, and ultimately the patient is described as “not unsalvageable”, but no firm prognostic statements are made. If I were the wife, I would not be satisfied that my question was well answered.

So: the objective of this chapter is to arm the candidates for future questions about prognosticating in severe liver disease. In contrast to many topics in the CICM fellowship exam, a part of this is useful in the real life setting – it offers an arsenal of arguments for and against the admission of critically ill liver disease patients to the ICU. The real skill is in determining who will benefit from aggressive intensive care therapy, and distinguishing this population from those for whom it is futile.

 Scoring systems for prognostication in chronic liver disease

The Child-Pugh scoring system

This is a long-lived scoring system devised originally by Child and Turcotte in 1964, in order to stratify patients undergoing shunt surgery. It combined objective measurements (albumin and bilirubin) with subjective assessments (ascites, encephalopathy and malnutrition) to predict outcomes following portosystemic shunting. The Pugh contribution in 1972 was to replace nutritional status with prothrombin time. Since then, everyone refers to the score as the Child-Pugh score, completely forgetting  the contribution of Dr Jeremiah G. Turcotte.

The score is calculated as follows:


1 point

2 points

3 points

Bilirubin (μmol/L)

Under 34


Over 50

Albumin (g/L)

Over 35


Under 28


Under 1.7


Over 2.3



Mild (responds to diuretics)

Severe (refractory to diuretics)



Grade I-II

Grade II-IV

A critique of the Child-Pugh score

This was all fine and good when it was purely used for crystal-gazing. Unfortunately, the system is somewhat defective when it comes to making decisions regarding survival. For instance, in transplant listing, it would be unfair to hang somebody’s only chance of survival on the presence or absence of subjective characteristics such as the “mildness” of  ascites, or the scoring of encephalopathy. An excellent critique is offered by Cholongitas et al (2005), whose article has acted as the source for most the ensuing summary.

In short, the following drawbacks have been noted in the Child-Pugh scoring system:

  • Use of subjective assessments (ascites and encephalopathy) is interpreter-dependent
  • It is unclear whether one should use clinical ascites,  ultrasonographic  or CT radiological evidence of ascites (which are more sensitive than the original methods of assessment from 1964)
  • Ascites and encephalopathy respond to treatment. It is unclear whether you score the patient  at their best  or worst numbers. In ICU a Child-Pugh class C will turn to A with only a little effort.
  • Even though the objective measures (eg. Bilirubin) are supposedly more scientific, the cut-off values are entirely arbitrary.
  • The cut-offs have a certain “ceiling effect”,  with absurd results. Mr Bloggs has a bilirubin of 60, and statistically has the same chance of survival as Mr Cloggs whose bilirubin is 300, because both belong in Child-Pugh class C.
  • The scoring system completely ignores other influences on mortality, such as renal failure or dilated cardiomyopathy – these have implications for transplant eligibility as well as for survival, duration of mechanical ventilation, length of ICU stay, and so forth.

Influence of Child-Pugh class on ICU outcomes

So, with all of these aforementioned problems, how much does Child-Pugh class matter when it comes to ICU stay? It seems, not much. Among 135 Taiwanese patients who mainly had cirrhosis due to Hep B, Ho et al (2005) found that the usually useless APACHE II score was actually better than the Child-Pugh system in discriminating survivors from non-survivors. In the Western world where alcoholism is the dominant cuase of cirrhosis,  Levesque et al (2012) prospectively followed 377 cirrhosis patients through the ICU and also found that the SOFA and SAPS II scores were much better at predicting mortality than the Child-Pugh class. A more recent retrospective audit of 85 patients by Yeoh et al (2016) again confirmed that ICU-specific organ dysfunction scores are superior.

In short, stop calculating the Child-Pugh grade in the ICU; it is pointless to do so.

Model for Endstage Liver Disease (MELD) score

In answer to the drawbacks of the Child-Pugh classification, Malinchoc et al (2000) published their own scoring system (MELD) which is thought to be better grounded in objective data.  Their model was designed for patients reciving a TIPS procedure,  and was supposed to predict survival and identify those patients whose liver-related mortality would be 3 months or less. The MELD score is calculated by the following formula:

MELD = 3.78×ln[serum bilirubin (mg/dL)] + 11.2×ln[INR] + 9.57×ln[serum creatinine (mg/dL)] + 6.43

In this formula, the variables were selected not arbitrarily but by Cox proportional hazard regression analysis, thus identifying those biochemical variables which appear to be the best predictors of survival. These were bilirubin, creatinine and INR. The original MELD (where “M” stood for “Mayo Clinic”) had also included the aetiology of cirrhosis as a variable, but this turned out to be irrelevant (and was criticised as a subjective value judegement).

The MELD score ends up giving the following 3- month  mortality predictions in patients awaiting a liver transplant:

  • 40 or more — 71.3% mortality
  • 30–39 — 52.6% mortality
  • 20–29 — 19.6% mortality
  • 10–19 — 6.0% mortality
  • <9 — 1.9% mortality

A comparison of MELD and Child-Pugh scoring systems (Peng et al, 2016). One might imagine that the MELD score performed better, but in fact both were about equally useless.

  • In patients with acute-on-chronic liver failure, Child–Pugh score had a higher sensitivity and a lower specificity than MELD score.
  • In patients admitted to ICU, MELD score had a smaller negative likelihood ratio and a higher sensitivity than Child–Pugh score.
  • In patients undergoing surgery, Child–Pugh score had a higher specificity than MELD score.
  •  In other subgroup analyses, Child–Pugh and MELD scores had statistically similar discriminative abilities.

In ICU, the liver scores have little influence because many other organ systems play are role in death or survival. Well-rounded models (SOFA, APACHE) are better at predicting ICU outcomes than liver-specific scoring systems (Levesque et al, 2012)

LiFe score (Liver Injury and Failure Evaluation)

A more recent entry into the scene of scoring systems is the LiFe score, the subject of a paper by Edmark et al (2016). The score, consisting of measured admission values for lactate bilirubin and INR, was "constructed based on the opinions of 157 intensivists within the European Society for Intensive Care Medicine", which sounds terrible (imagine all those people in the same room, voices raised in disagreement). In the study, hospital mortality was as follows:

  • low-risk group (0 risk score points) = 17% mortality
  • intermediate-risk group (1–3 points) = 28% mortality
  • high-risk group (4–8 points) = 47% mortality
  • very high-risk group (>8 points) = 77% mortality

Influences on survival of liver disease patients in ICU

So, if liver-specific scoring systems are poor at predicting outcome, what influences the in-ICU and post-ICU survival of a patient with liver disease?

The following points can be made:

  • Childs-Pugh at admission is worth calculating: it is associated with survival at 12 months (Warren et al, 2015). Specifically:
    • Class A = 100% survival at 12 months
    • Class B = 50% survival at 12 months
    • Class C = 25% survival at 12 months
  • The following are negative influences on survival (Levesque et al, 2012):
    • Mechanical ventilation
    • Vasopressor theraot
    • High bilirubin levels at admission
    • Sepsis

The outlook for severe cirrhosis is poor. In 2004, Gildea et al published an analysis of a 1993-1998 cohort of cirrhosis patients who were admitted to ICU and had some combination of liver failure, cirrhosis, chronic liver disease, variceal bleeding, hepatic encephalopathy, or hepatorenal syndrome. Those who underwent transplantation were excluded. The 1-year mortality rate was 69%, and the 5-year mortality rate was 77%. Most of the mortality was concentrated in the first admission: the median survival time was 1 month. This may be a dated study, and perhaps modern ICU is so super-awesome that the survival may be better. However, all the techniques of  extracorporeal "liver support" therapy have never been shown to improve mortality, and these have been the major changes since the nineties. If one were to repeat this study today, one would likely end up with the same results.


Fichet, Jérôme, et al. "Prognosis and 1-year mortality of intensive care unit patients with severe hepatic encephalopathy." Journal of critical care 24.3 (2009): 364-370.

García-Martínez, Rita, Macarena Simón-Talero, and Juan Córdoba. "Prognostic assessment in patients with hepatic encephalopathy." Disease markers 31.3 (2011): 171-179.

Wendon, Julia, et al. "Critical care and cirrhosis: outcome and benefit." Current opinion in critical care 17.5 (2011): 533-537.

Pugh, R. N. H., et al. "Transection of the oesophagus for bleeding oesophageal varices." British Journal of Surgery 60.8 (1973): 646-649.

Forman, Lisa M., and Michael R. Lucey. "Predicting the prognosis of chronic liver disease: an evolution from Child to MELD." Hepatology 33.2 (2001): 473-475.

Child, CG_, and J. G. Turcotte. "Surgery and portal hypertension." Major problems in clinical surgery 1 (1964): 1.

Malinchoc, Michael, et al. "A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts." Hepatology 31.4 (2000): 864-871.

Zimmerman, Jack E., et al. "Intensive care unit admissions with cirrhosis: Risk‐stratifying patient groups and predicting individual survival." Hepatology 23.6 (1996): 1393-1401.

Cholongitas, E., et al. "Systematic review: the model for end‐stage liver disease–should it replace Child‐Pugh's classification for assessing prognosis in cirrhosis?." Alimentary pharmacology & therapeutics 22.11‐12 (2005): 1079-1089.

Ho, Yu-Pin, et al. "Outcome prediction for critically ill cirrhotic patients: a comparison of APACHE II and Child-Pugh scoring systems." Journal of intensive care medicine 19.2 (2004): 105-110.

Cholongitas, E., et al. "Risk factors, Sequential Organ Failure Assessment and Model for End‐stage Liver Disease scores for predicting short term mortality in cirrhotic patients admitted to intensive care unit." Alimentary pharmacology & therapeutics 23.7 (2006): 883-893.

Levesque, Eric, et al. "Prospective evaluation of the prognostic scores for cirrhotic patients admitted to an intensive care unit." Journal of hepatology 56.1 (2012): 95-102.

Yeoh, Sern Wei, et al. "Cirrhotics Treated In Intensive Care Unit Have High Short Term Survival in the Absence of Extrahepatic Organ Dysfunction." Journal of Gastroenterology and Hepatology Research 5.2 (2016): 1984-1988.

Wiesner, Russell, et al. "Model for end-stage liver disease (MELD) and allocation of donor livers." Gastroenterology 124.1 (2003): 91-96.

Saliba, Faouzi, et al. "Cirrhotic patients in the ICU: prognostic markers and outcome." Current opinion in critical care 19.2 (2013): 154-160.

Peng, Ying, Xingshun Qi, and Xiaozhong Guo. "Child–Pugh Versus MELD Score for the Assessment of Prognosis in Liver Cirrhosis: A Systematic Review and Meta-Analysis of Observational Studies." Medicine 95.8 (2016).

Edmark, Christin, et al. "LiFe: a liver injury score to predict outcome in critically ill patients." Intensive care medicine 42.3 (2016): 361-369.

Gildea, Thomas R., et al. "Predictors of long-term mortality in patients with cirrhosis of the liver admitted to a medical ICU." CHEST Journal 126.5 (2004): 1598-1603.