A 40 year old was admitted to hospital with his first presentation of alcoholic liver disease with symptoms of jaundice (bilirubin 248), poor mobility, hallucinations and passing of black stool. On admission to hospital, he was lethargic with features of Grade II encephalopathy, was coagulopathic (INR 3.1), had deranged electrolytes (sodium 114, potassium 2.9), but a normal creatinine (54) and a raised white cell count (15.9). He was haemodynamically stable and had a haemoglobin of 119g/L with no signs of active bleeding. His abdomen was distended (ascites), he was visibly jaundiced and had spider naevi on his chest. An abdominal ultrasound was performed that showed liver cirrhosis, borderline splenomegaly, small volume ascites and normal kidneys. A full liver screen revealed no infective cause and his AST:ALT ratio suggested alcoholic liver disease. His prognostic indicator scores were all suggestive of severe alcoholic liver hepatitis (Maddrey score: 131; Childs: C; Lille Score: 1; GAHS: 10; MELD: 29). His serum ammonia level was 170. He was commenced on terlipressin, prednisolone and pentoxyphylline and thiamine. Despite this, his encephalopathy progressed to grade 4 and he required intubation and ventilation for airway protection and a presumed aspiration pneumonia. His liver function and coagulopathy continued to worsen, and he developed an acute kidney injury necessitating commencement of renal replacement therapy. He required noradrenaline to support his blood pressure. Ascitic tap ruled out spontaneous bacterial peritonitis. He was discussed with regional liver centres, but was not felt to be a transplant candidate. His liver and renal function continued to deteriorate and eventually treatment was withdrawn nearly 3 weeks into his admission.

Describe the scoring systems for assessing the severity of acute hepatic dysfunction.

Emma Fitzgerald

Alcoholic liver disease represents an important condition amongst critically ill patients admitted to intensive care units. In the last 15 years, alcohol-related deaths have doubled [1] and the number of admissions to ICU with decompensated alcoholic liver disease has significantly increased. This is important since they usually young patients and they experience a particularly high morbidity and mortality secondary to multiple organ failure with an estimated mortality of between 40% and 90% (if ≧3 failing organ systems on admission to ICU) [2].

Accurate prognostic indicators for patient survival in ICU are important, especially in these patients where mortality remains high despite intensive support. They can also be used to compare the effectiveness of treatments. Many scoring systems have been used to predict mortality in patients with alcoholic liver cirrhosis. Established scoring systems include liver specific (Childs Pugh, Maddrey and MELD) and more general ICU scores (Apache II and SOFA). These predictive scores provide information that can risk stratify patients, help in the discussions with families, making comparisons between units and provide thresholds around which treatments can be initiated, for example, the American college of gastroenterologists recommended that corticosteroids be considered with mDF>32.

SCORING SYSTEMS

Discriminant Function
Maddrey et al first described this in 1978 in a placebo controlled corticosteroid trial [3]. He concluded that a DF>93 predicted a 25% 28 day survival, whilst a score <93 predicted 100% survival.

Modified DF
This was described in 1989 and used in a placebo-controlled corticosteroid trial [4]. It showed that a mDF score ≧32 and/or presence of encephalopathy in placebo-treated patients had a 28 day survival of 68% and a 28 day survival of 93% with an mDF score <32.

Model for End-stage Liver Disease Score
This score was developed in non-emergency patients and is mainly used for assessing for patients with alcoholic hepatitis for planned liver transplantation. A study showed that with a MELD score of >11 there was a 45% 30-day survival, while a score <11 had a 30-day survival of 96% [5]. The score has been applied to patients with acute illness alcoholic hepatitis and huge differences were seen compared to their pre-admission MELD scores (outcomes of critically ill patients with cirrhosis admitted to intensive care). It has never been validated for its accuracy of assessing cirrhotic patients admitted to ITU. Compared to Maddrey scoring, MELD has been seen to be a better indicator of mortality (it not only uses PT and bilirubin, but also the creatinine as a measure of renal dysfunction) [6].

Glasgow Hepatitis Score
Developed in 2005 for use specifically in alcoholic hepatitis [7]. This score, gives a similar cut-off point that is “equivalent” to that used by MELD, scores <9 suggest prognosis is better (93% 28-day survival), and much worse if the score is ≧9 (47% 28 day survival). In 2007 it was validated against the mDF with reference to the implementation of corticosteroid therapy [8].

Lille Model
This scoring system was put forward to improve the prediction of mortality in severe AH (i.e. with DF>32) [9]. The prospective cohort study demonstrated a high sensitivity and specificity for high risk of death at 6 months. Its aim is to not to compete with other models, but to predict poor mortality in patients with severe AH treated with corticosteroids. Above its cut-off of 0.45, it predicts 76% of the observed deaths, below this level there was observed an 85% survival.

Child-Pugh
This was created in 1973 as a modification of the Child and Turcotte classification. This widely used liver-specific prognostic score was originally developed in acutely unwell patients, but has since developed a profile in staging stable out-patient disease. However, there is a clear association between disease stage and mortality:

Others

Non-liver specific, general ICU scoring systems include SOFA and APACHE II. Although these are widely used, the predictive accuracy has only been evaluated in a few studies. There is some evidence to suggest that SOFA is better able to predict mortality in their cohort of acutely ill cirrhotic patients than APACHE II or CP [10].

---

Lessons Learnt:

Scores are not always transferrable to the critically unwell patient and they must be interpreted carefully. Some are organ specific and some more general. They should each be used in the context that they were designed for. Some aspects are very subjective, such as the encephalopathy score. Some only use a score at one point in time, whilst others consider progression of disease (Lille). Some use INR, measures of renal dysfunction, or both and it is therefore important to interpret the scores with with care as one abnormal variable could significantly skew the score.

Some of the scores are organ specific and do not consider any pulmonary, renal and cardiovascular dysfunction (as this this patient had) which is highly relevant in a critical care situation. This can be overcome by also using other scores such as APACHE II, but these are very non-specific.

---

References:
1. Jackson P. Alcoholic Liver Disease. BJA Contin Educ Anaesth Crit Care Pain (2010) 10 (3): 66-71
2. Cholongitas 2006 Review article: scoring systems for assessing prognosis in critically ill adult cirrhotics. Alimentary Pharmacology & Therapeutics. (2006) 24(3):453-464
3. Maddrey WC, Boitnott JK, Bedine MS, et al. Corticosteroid therapy of alcoholic hepatitis. Gastroenterology 1978;75:193–9.
4. Mathurin P, Mendenhall CL, Carithers J, et al. Corticosteroids improve short- term survival in patients with severe alcoholic hepatitis (AH): individual data analysis of the last three randomized placebo controlled double blind trials of corticosteroids in severe AH. J Hepatol 2002;36:480–7.
5. 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 Gastroenterol 2002;2:2
6. Kamath PS et al. The model for end-stage liver disease (MELD), Hepatology. 2007 Mar;45(3):797-805.
7. Forrest EH et al. Analysis of factors predictive of mortality in alcoholic hepatitis and derivation and validation of the Glasgow alcoholic hepatitis score. Gut. 2005 Aug;54(8):1174-9.
8. Forrest EH et al. The Glasgow alcoholic hepatitis score identifies patients who may benefit from corticosteroids. Gut. 2007;56:1743–1746
9. Louvet A et al. The Lille Model: A New Tool for Therapeutic Strategy in Patients with Severe Alcoholic Hepatitis Treated with Steroids. Hepatology 2007;45:1348-1354.
10. Cholongitas E et al. Review article: scoring systems for assessing prognosis in critically ill adult cirrhotics Alimentary Pharmacology & Therapeutics (2006) 24 (3);453–464.

Appendix:
DF = (4.6 x PT)+serum bilirubin (mg/dl)
mDF = 4.6 (PTpatient – PT control)+ serum bilirubin (mmol/l)/17.1
MELD score = 3.8 x loge(bilirubin (mg/dl))+1.2 x loge(INR)+ 9.6 x loge(creatinine (mg/dl))