barcelona . spain - European Association for the Study of the Liver

BARCELONA . SPAIN
98 POSTGRADUATE COURSE SYLLABUS ALCOHOLIC LIVER DISEASE 99
APRIL 18 - 19/2012 THE INTERNATIONAL LIVER CONGRESS TM 2012
ALCOHOLIC HEPATITIS OR DECOMPENSATED CIRRHOSIS:
DIAGNOSTIC AND PROGNOSTIC ISSUE
Frederik Nevens
Leuven, Belgium
E-mail: frederik.nevens@uzleuven.be
KEY POINTS
• There are different types of chronic liver failure: chronic hepatic decompensation and acuteon-chronic
liver failure.
• Multiorgan failure occur before sepsis in ACLF and therefore scores as the SOFA score predict
outcome better than MELD.
• One of the most frequent precipitating events of ACLF is severe alcoholic hepatitis.
• Prognostic scores used for patients with ASH are disease-specific: modified Maldrey’s
discrimination function, the Lille score, the Glasgow alcoholic hepatitis score and the ABIC
and the non disease-specific score: MELD.
• Early control of ASH by corticosteroids prevents evolution to ACLF.
• During corticosteroid treatment a Lille score at day 7 ≤ 0.16 select responders.
Alcohol can provoke different types of liver injury. The most critical ill patients are patients suffering
from severe alcoholic hepatitis (ASH) (mostly superimposed on alcoholic cirrhosis) and patients with a
decompensated cirrhosis. Recently a new clinical syndrome has been described : acute-on-chronic liver
failure (ACLF).
GENERAL ASPECTS OF LIVER FAILURE AND PROGNOSIS
There are different types of liver failure. It is important to recognize them since they all have a different
prognosis and require a different therapeutic approach. Traditionally liver failure is divided in acute liver
failure (ALF) which occurs suddenly in a patient without a previous liver disease and chronic liver failure
due to chronic hepatic decompensation (CHD) which is found in patients with end-stage cirrhosis as a
result of a slowly progression of their underlying liver disease. Since the introduction of albumin dialysis
another subtype of chronic liver failure is recognized : ACLF. This is an acute and rapid deterioration within
weeks of a patient with a well-compensated chronic liver disease and is clinically characterized by deep
jaundice, renal impairment, hepatic encephalopathy and rapidly evolving multiorgan failure (1). In patients
with alcoholic cirrhosis admitted to the hospital for a complication, ACLF has a 3 month mortality of 60 %
versus 20 % in case of end-stage CHD (2) (Fig1). Spontaneous reactivation of hepatitis B presenting as
ACLF in the absence of liver transplantation has a 3 month mortality of 80% illustrating the severity of this
syndrome (3).
The most important clinical complication in all these conditions of liver failure which directly affects outcome,
is sepsis and multiorgan failure. A low threshold to start antibiotics has significantly improved the outcome
of patients with ALF (4). In case of CHD sepsis occur late in the progression of the disease at the moment of
severe impairment of liver function (5). In this situation prognosis at admission of the hospital is especially
related to the degree of liver dysfunction and therefore MELD score is accurate to predict the mortality in
this condition. In ACLF multiorgan failure occur before sepsis at an early stage of the clinical syndrome.
In this regard we found in a recent prospective study that in ACLF the SOFA score was the best score to
predict outcome at admission in the hospital (2).
The exact pathofysiology of ACLF is still unclear. In most cases of ACLF there is a discordance between the
severity of jaundice on the one hand and the degree of other features of liver impairment on the other hand.
This resulted to the toxin accumulation hypothesis. Since albumin dialysis removes both water soluble and
non soluble toxins a beneficial effect in ACLF was expected and indeed an improvement in haemodynamic
instability and hepatic encephalopathy was observed. (6, 7, 8) None of these devices however could
improve yet overall survival. Currently an excessive pro-inflammatory response to bacterial components is
thought to play an important role linking the gut, liver and systemic circulation in this syndrome (9).
A crucial role in the onset of ACLF are the precipitating events. This is illustrated by the fact that early
and aggressive control of these triggers allow reversal of ACLF. A common precipitating event of ACLF
is variceal haemorrhage. Prophylactic use of antibiotics has significantly improved the outcome of these
patients and in this regard it has been recently demonstrated that in high risk patients early control of
variceal haemorrhage by TIPS prevents ACLF (10). A similar observation came recently from India. Early
suppression of HBV viremia with tenofovir prevented the development of multiorgan failure in patients with
spontaneous reactivation of Hepatitis B presenting as ACLF (3). The most common trigger in the Western
countries of ACLF is severe alcoholic hepatitis and this will be discussed more in detail below.
DIFFERENT TYPES OF LIVER FAILURE IN PATIENTS WITH ALCOHOLIC LIVER DISEASE
a.Severe ASH as a trigger for ACLF
One of the most important cause of chronic liver failure is alcohol. CHD is the most frequent subtype and is
characterized in the initial stage by the complications of portal hypertension and mild to moderate jaundice.
The one year mortality in case of the appearance of ascites is 29% and if hepatic encephalopathy occurs
this is 64 % (11). The prognosis is predominantly affected by abstinence. ACLF is less frequent but in
tertiary hospitals ACLF can account for more than 40 % of the emergency hospitalization due to alcoholic
cirrhosis (2). In Western countries the percentage of alcoholic cirrhosis included in the studies with albumin
dialysis (MARS of Prometheus) for the treatment of ACLF range from 39 to 92% (8, 12, 13). In patients with
alcoholic cirrhosis ACLF can be induced by several precipitation events but the most frequent one is severe
ASH, which occurred in around 25 % of our patients with ACLF (2).
b.Prognostic scores used for patients with ASH
Severe ASH is characterized by a rapid onset of jaundice, elevated AST (< 300 IU/ml), AST/ALT ratio > 2,
bilirubin > 5 mg/dl, elevated INR and neutrophilia (14). As has been demonstrated for variceal haemorrhage
and exacerbation of HBV, the best way to reverse ACLF is a rapid control of ASH in patients of which
spontaneous recuperation is doubtful. Therefore prognostic scores are of importance.
There are different scores which assess the severity of ASH (Table 1). There are the disease-specific
prognostic models (modified Maddry’s discrimination function, the Glasgow alcoholic hepatitis score and
the ABIC score) and there is the non disease-specific prognostic model: MELD. In 2 retrospective studies it
was found that a MELD score of > 21 was a useful clinical tool to assess mortality of these patients (15,16).
In one of the studies also a decrease in MELD by 2 points at day 7 was associated with a good prognosis.
Still one of the most commonly used disease-specific score to predict outcome of ASH is the modified
Maddry’s discriminant function (DF) (17). Patients with a DF < 32 have a spontaneous survival at 28 days
around 90 %; if the DF is ≥ 32 the spontaneous 2 month survival is only 50 % - 65 %. These patients mostly
progress to the systemic inflammatory response syndrome and multisystem organ failure which is also
seen in other types of ACLF.
In patients with severe ASH without contraindication (gastro-intestinal bleeding, hepatorenal syndrome,
sepsis and HBV) corticosteroids seems to improve outcome. A recent meta-analysis of individual data from
5 randomized trials demonstrated that 28-day survival was higher in corticosteroid-treated patients (80%
vs 66%) (18). Two disease-specific scores assessed which patients might benefit from corticosteroids: for
the Maddrey score the cut-off is ≥ 32.and for the Lille score this is ≥0.45 (19). Another score which allow to
select the best candidates for treatment is the Glasgow score (20). The Glasgow alcoholic hepatitis score