Pathogenesis and management of alcoholic hepatitis

Authors


  • The terms alcohol and ethanol are used interchangeably in the present paper.

Dr Paul S Haber, Royal Prince Alfred Hospital, Camperdown, NSW 2050, Australia. Email: phaber@mail.usyd.edu.au

Abstract

Abstract  Alcoholic hepatitis is a potentially life-threatening complication of alcoholic abuse, typically presenting with symptoms and signs of hepatitis in the presence of an alcohol use disorder. The definitive diagnosis requires liver biopsy, but this is not generally required. The pathogenesis is uncertain, but relevant factors include metabolism of alcohol to toxic products, oxidant stress, acetaldehyde adducts, the action of endotoxin on Kupffer cells, and impaired hepatic regeneration. Mild alcoholic hepatitis recovers with abstinence and the long-term prognosis is determined by the underlying disorder of alcohol use. Severe alcoholic hepatitis is recognized by a Maddrey discriminant function >32 and is associated with a short-term mortality rate of almost 50%. Primary therapy is abstinence from alcohol and supportive care. Corticosteroids have been shown to be beneficial in a subset of severely ill patients with concomitant hepatic encephalopathy, but their use remains controversial. Pentoxifylline has been shown in one study to improve short-term survival rates. Other pharmacological interventions, including colchicine, propylthiouracil, calcium channel antagonists, and insulin with glucagon infusions, have not been proven to be beneficial. Nutritional supplementation with available high-calorie, high-protein diets is beneficial, but does not improve mortality. Orthotopic liver transplantation is not indicated for patients presenting with alcoholic hepatitis who have been drinking until the time of admission, but may be considered in those who achieve stable abstinence if liver function fails to recover.

INTRODUCTION

In Australia, alcoholic liver disease (ALD) accounts for 1000 deaths and 64 000 hospital admissions per year.1 Alcoholic liver disease comprises an overlapping spectrum of pathological processes, including steatosis (alcoholic fatty liver), alcoholic hepatitis, and alcoholic cirrhosis. The purpose of the present review is to describe the current understanding of the pathogenesis and treatment of alcoholic hepatitis.

PATHOGENESIS

The pathogenesis of alcoholic hepatitis is multifactorial (Fig. 1). Metabolism of ethanol to toxic metabolites, Kupffer cell stimulation by endotoxin, and nutritional impairment lead to hepatic injury, inflammation, and fibrosis. Ethanol is also associated with impaired hepatic regeneration. Each of these pathological processes presents a potential therapeutic target. The disease process may be accelerated by several cofactors. Detailed reviews of pathogenesis may be found elsewhere.2

Figure 1.

Overview of the pathogenesis of alcoholic hepatitis. ADH, alcohol dehydrogenase; TGF-β, transforming growth factor beta; TNF-α, tumor necrosis factor alpha.

Ethanol metabolism

Ethanol is the preferred fuel for the liver, displacing other substrates when present. Metabolism leads to the elimination of ethanol at the risk of toxicity. A variable proportion of ethanol undergoes first-pass metabolism mainly by the gastric isoenzymes of alcohol dehydrogenase (ADH).3 Ethanol alters intracellular signaling pathways4,5 and the physical properties of cell membranes.6 The toxicity of ethanol is thought to result chiefly from its metabolism rather than from ethanol itself.

Ethanol metabolism involves oxidative and non-oxidative pathways. In the first step of oxidation, ethanol is converted to acetaldehyde.7 Alcohol dehydrogenase is the major enzyme. The microsomal ethanol-oxidizing system (MEOS) involves several cytochrome P450 isoenzymes, of which cytochrome P450 2E1 (CYP2E1) is the major constituent.8 Expression of CYP2E1 is increased by chronic ethanol consumption and this is thought to account for increased hepatic ethanol oxidation observed in this setting. This pathway generates reactive oxygen species (ROS) that contribute to tissue injury. A peroxisomal system involving catalase is capable of substantial ethanol oxidation, but its activity in vivo is limited by the availability of the cofactor H2O2.9 The products of these enzyme systems are all thought to contribute significantly to liver injury. Non-oxidative ethanol metabolism involves formation of fatty acid ethyl esters (FAEE) from free fatty acids and ethanol. The potential for FAEE to contribute to liver injury has received little attention.

Acetaldehyde

Acetaldehyde is highly reactive and is thought to play a central role in hepatic damage.10 Acetaldehyde directly affects many aspects of cellular functioning and may bind covalently to various compounds, particularly proteins, to form acetaldehyde adducts. This process contributes to liver injury as some adducts are immunogenic and some protein adducts do not function normally.11 Several acetaldehyde adducts, including malondialdehyde and 4-hydroxynonenal (HNE) adducts, have been observed at sites of liver damage in alcoholic subjects.12

Genetically determined polymorphisms of acetaldehyde dehydrogenase (ALDH) with low enzyme activity may lead to accumulation of higher levels of acetaldehyde that promote liver injury. Japanese persons with ALDH2 deficiency are susceptible to liver disease.13 In Caucasians populations, the ALDH polymorphisms found to date are uncommon and do not explain individual variations in ALDH activity or susceptibility to ALD. Levels of acetaldehyde in blood and liver tissue are increased in patients with ALD.14 This has been attributed to mitochondrial damage, which is an early feature of ALD. Mitochondrial reoxidation of the cofactor nicotine adenine dinucleotide (NADH) to NAD+ is reduced, leading to impaired acetaldehyde elimination.10

Oxidant stress

An important factor in ethanol toxicity is the generation of oxidative stress via both increased formation of ROS and decreased defences against oxidant stress. The induction of CYP2E1 by alcohol intake stimulates the formation of ROS during ethanol oxidation.8 Several xenobiotics that are metabolized by CYP2E1 have a synergistic effect with ethanol in the generation of liver damage, including carbon tetrachloride and paracetamol.10 Total hepatic levels of glutathione (GSH) are low in alcoholics, but GSH is selectively depleted in mitochondria.15 Carotenoids are major natural protective agents against free-radical-mediated liver damage and are depleted.16 Vitamin E depletion in the liver may also be associated with alcohol consumption and predispose to damage by ROS.17 Non-transferrin bound iron contributes to oxidant stress and tissue injury.18

Reactive oxygen species interact with nucleic acid, proteins, and lipids. Mitochondrial DNA is particularly susceptible to ROS because of selective GSH depletion.19 Mitochondria from alcohol-treated hepatocytes are susceptible to tumor necrosis factor (TNF)-α toxicity, resulting in cell necrosis.20 Lipid peroxidation leads to impaired membrane function and has been implicated in steatosis and stellate cell activation.

Endotoxin and the inflammatory response

Inflammation is a hallmark of alcoholic hepatitis and is characterized by intrahepatic infiltration of neutrophils or mononuclear cells and by altered local and systemic expression of inflammatory cytokines. Recent studies have highlighted the role of endotoxin in the initiation of liver injury and inflammation.

Endotoxin is a toxic lipopolysaccharide (LPS) present in the cell wall of all Gram-negative bacteria21 that may be absorbed and thus enter the portal and systemic circulation. Circulating endotoxin binds to a LPS-binding protein (LBP) and this complex binds CD14, a receptor on the surface of Kupffer cells. CD14 interacts with the membrane spanning Toll-like receptor type 4 (TLR4).22 The TLR signals the upregulation of pro-inflammatory cytokines (notably TNF-α).23 Tumor necrosis factor-α binds to cognate receptors on hepatocytes and other cells, leading to tissue damage via induction of oxidant stress, stimulation of other cytokines and induction of apoptosis.24 A number of other mediators are released from Kupffer cells, including nitrous oxide (NO), TxA2 and interleukin (IL)-6.21 Circulating levels of endotoxin are increased in alcoholics25 as a result of increased gut permeability to large molecules, increased numbers of bacteria within the gut, and reduced capacity for endotoxin scavenging. CD14 gene expression is upregulated in alcohol-fed rats.26 Experimental alcohol-induced liver injury is potentiated by coadministration of endotoxin27 and is reduced by coadministration of antibiotics28 and in animals with TLR4 mutations.29

A major consequence of endotoxin stimulation of Kupffer cells is release of TNF-α, a key mediator of ALD. Increased levels of TNF-α30 and TNF-α receptors31 have been observed in the blood of patients with this disease. Levels parallel the severity of alcoholic hepatitis30 and correlate with prognosis.32 Tumor necrosis factor-α knockout mice do not develop liver disease after alcohol exposure.33 Altered levels of TNF-α receptors persist even after abstinence from alcohol is achieved, and may contribute to perpetuation of disease activity occasionally observed in this setting.31

Fibrosis

The hepatic scar is derived mainly from activated hepatic stellate cells (HSC), also called lipocytes, Ito cells, or fat-storing cells.34 Similar cells are found in many other tissues, including the pancreas, another major target of alcohol-induced fibrosis.35 Hepatic stellate cells are activated to become myofibroblast-like, secretory cells resulting in both an increase in fibrinogenesis and in cell proliferation.36 Transforming growth factor (TGF)-β is a key pro-fibrotic cytokine and is mitogenic in human HSC. Tissue markers of stellate cell activation have also been demonstrated in the alcoholic fatty liver in the absence of hepatitis,37 suggesting that fibrogenesis may be stimulated by factors that are independent of necrosis or histologically evident inflammation. In experimental studies, HSC are directly stimulated by acetaldehyde, by products of lipid peroxidation, and by decreases in antioxidant GSH.

Recent studies have shown that fibrogenesis induced by HSC may be reversible. During resolution of injury, HSC participate in matrix degradation and undergo apoptosis.38 Thus, prevention of HSC activation or collagen synthesis and/or enhancement of collagen breakdown or HSC apoptosis represent another therapeutic strategy.

Hepatic regeneration

The normal liver has an extensive capacity to regenerate after injury. Hepatocytes and other liver cells quickly leave the resting Go state and enter the cell cycle via a process that requires TNF-α, IL-6 and other factors.39 Hepatic stem cells are also involved.40

Chronic ethanol consumption is known to inhibit hepatic regeneration41 and has been shown to arrest the regenerative process during the prereplicative period.42 Several factors are thought to contribute. The action of TNF-α on nuclear factor-kappa light polypeptide B (NF-κB) that leads to DNA synthesis is impaired.42 Intracellular calcium release by epidermal growth factor is known to stimulate mitosis43 and is also impaired after chronic alcohol administration.44 Transmission of growth  signals  through  insulin  receptor  substrate (IRS)-1 mediated signal transduction cascade is also impaired.45 Impaired nutrition observed in many alcoholics is also likely to inhibit regeneration.

Genetic predisposition and cofactors

The risk of liver disease is related to the amount of alcohol consumed, but genetic factors, female sex, obesity, chronic viral hepatitis, hepatotoxins, and nutritional impairment all accelerate the disease process. A classic twin study showed that the concordance rate for ALD is three-fold higher in monozygotic twins compared with dizygotic twins.46 Subsequent studies are only beginning to define the specific genes contributing to this risk: TNF-α polymorphisms47 and ALDH213 deficiency contribute to a small degree.48

The increased susceptibility of women to ALD is well documented.49 Female susceptibility may be explained by increased effective consumption of alcohol as a result of sex differences in body composition, or lower first-pass ethanol metabolism leading to higher systemic absorption for a given dose.50 Despite a smaller average body weight, liver volume and alcohol clearance are similar in women compared to men.51 At least in the rat model, the sex difference has been ascribed to vulnerability to endotoxin-related toxicity.52

Several acquired factors are associated with accelerated liver disease. Hepatitis C is more common in alcoholics than in the general community,16 an observation attributed to increased prevalence of injecting drug use. Alcohol abuse and chronic hepatitis C interact, leading to accelerated progression to cirrhosis.53 This interaction is less clear for moderate levels of alcohol consumption.54 A similar interaction has been postulated between chronic hepatitis B infection and alcohol,55 but the evidence is unclear. Long-standing obesity is an independent risk factor for liver disease in alcoholics.56

Excessive drinkers appear to be at increased risk of liver injury from therapeutic drug administration. For example, paracetamol in apparently therapeutic doses has led to acute liver failure in alcoholics57 and methotrexate induces hepatic fibrosis more rapidly in this context.58

DETECTION OF ALCOHOL ABUSE

It is recommended that a quantitative alcohol history is obtained for all patients. Collateral reports may provide important additional information. Alcohol abuse may be recognized via other typical medical or psychosocial problems such as depression, neurological toxicity, abdominal pain, relationship difficulties, poor work performance, trauma, and violence.

Several brief questionnaires have been validated for the detection of alcohol abuse. The CAGE questionnaire is short and still widely used (Table 1).59 Two or more positive responses suggest the presence of alcohol-related problems and the need for further assessment.60 The CAGE questionnaire is sensitive and specific for alcohol dependence, but less useful for less severe alcohol abuse. The Alcohol Use Disorders Identification Test (AUDIT) questionnaire is more sensitive, but has 10 questions61 and is not in routine clinical use.

Table 1.  The CAGE screening questions for alcohol abuse59
C – Have you ever felt the need to cut down you drinking?
A – Have you ever felt annoyed by criticism of your drinking?
G – Have you ever felt guilty about your drinking?
E – Have you ever taken a drink (eye-opener) first thing in the morning?

Laboratory markers such as gamma-glutamyl transferase (γ-GT) and mean corpuscular volume (MCV) are only moderately sensitive and specific. However, the γ-GT level is almost invariably elevated in ALD.62 False positive values may be seen in other liver diseases and with some drugs. The MCV is generally less sensitive and less specific than γ-GT.63 Combined assessment of MCV and γ-GT detects 70% of the alcohol-dependent population.

Elevated levels of carbohydrate-deficient transferrin (CDT) may be more sensitive in evaluating drinking status than clinical measurement of γ-GT.64 Carbohydrate-deficient transferrin has fewer than the normal 4–5 terminal sialic acid residues. However, application of CDT to clinical detection of alcohol abuse has limitations. The relative value (% of total) of CDT is a more useful marker than total CDT because of variations in total serum transferrin.65 Iron overload and iron deficiency decrease the accuracy of CDT.66 The methodology used to measure CDT has been shown to influence the accuracy of the results and the widely available kit method may be no more useful than γ-GT.67 Carbohydrate-deficient transferrin has been reported to be less useful in the presence of alcoholic68 and non-alcoholic liver disease.69 The only report of CDT in alcoholic hepatitis found that the test was not useful, but only a small number of cases were studied.70

Other laboratory parameters associated with alcohol abuse, including uric acid, triglycerides, and high-density lipoprotein cholesterol may be helpful if elevated, but are not adequately sensitive or specific for use as a screening test. Newer markers are under investigation, including acetaldehyde adducts, serum beta-hexosaminidase, and the ratio of urinary serotonin metabolites.

NATURAL HISTORY

In the short term, the mortality of alcoholic hepatitis is closely related to the severity of illness on presentation. Overall, there is a 30-day mortality of approximately 15% and 39% mortality by 1 year. For severe cases, the short-term mortality remains approximately 50%.71 Hepatic and extra-hepatic factors contribute to the causes of death. Established hepatorenal syndrome carries a high mortality. Extrahepatic mortality of alcoholics is related to trauma and the common comorbidities of depression and tobacco dependence, leading to suicide, stroke, and cancers of the lung, upper alimentary tract, and breast. In the longer term, survival is 58% for those with alcoholic hepatitis alone, but only 35% for those with cirrhosis.72

Continuing alcohol consumption is a major factor that influences the survival of patients with alcoholic hepatitis.72 Alcoholic hepatitis persists in those that continue to drink, with approximately 38% progressing to cirrhosis within 18 months.73 The benefit of abstinence is less clear for decompensated liver disease, as one study of patients presenting with esophageal varices showed no benefit,74 whereas a study of patients presenting with ascites found a marked survival benefit.75 Once the disease is advanced, the outcome may be determined by irreversible liver damage rather than by continuing alcohol consumption.

The prognosis of alcoholic hepatitis is worse for women than for men.76 In addition, women suffer from higher levels of psychological comorbidity, primarily depression. A study has shown that a higher proportion of women progressed to cirrhosis, with 30 g or more of ethanol/day correlating with a significant chance of developing cirrhosis.77

Other liver diseases influence the outcome of patients with alcoholic hepatitis, most notably chronic hepatitis C. As many as 35–40% of patients that drink excessively test positive for hepatitis C virus (HCV) antibodies.78,79 Hepatitis C virus infection in patients with alcohol-induced liver disease is associated with more severe liver disease.80 Past hepatitis B infection is common among patients with ALD, but seropositivity is not related to the severity of liver disease.78,80

MANAGEMENT OF ALCOHOLIC HEPATITIS

Assessment of alcohol use disorder

The alcohol use disorder is assessed according to recognized diagnoses (e.g. International Classification of Diseases, ICD-10). Harmful consumption refers to continuing consumption despite harms recognized by the patient. Dependence refers to the presence of three or more of the following symptoms: drinking despite harm, a strong desire to drink, loss of control over drinking, withdrawal symptoms, tolerance, and salience (giving a higher priority to drinking than to other activities). Psychological comorbidities such as depression and anxiety are very common and treatment of these may increase the chance of controlling alcohol consumption.

Assessment of alcoholic hepatitis

In cases of mild alcoholic hepatitis, there may be no symptoms or signs, with abnormal serum biochemistry the only manifestation of the disease. Others have fatigue, anorexia, weight loss, jaundice, fever, and tender hepatomegaly, with physical examination revealing signs of chronic liver disease such as spider nevi. In severe cases, signs of hepatic decompensation are evident, including ascites, jaundice, and encephalopathy.

The severity of alcoholic hepatitis can be assessed using quantitative indices. Mendenhall compared the Maddrey Discriminant Function (MDF), Combined Clinical and Laboratory Index and the Child's score and concluded that the MDF gave the best indication of short-term survival and was also the simplest to use.81 The MDF = 4.6 × (prolongation of prothrombin time in seconds) + bilirubin (µmol/L)/17. Thirty-day survival rates of 80–100% have been reported for mild to moderate disease (MDF < 32), decreasing to 50% for severe disease (MDF > 32).

Biopsy is the definitive diagnostic tool, but typically is not carried out, as the results are not likely to alter the diagnosis or therapy.82 Mathurin et al. reported that up to 30% of patients with clinical features of severe alcoholic hepatitis do not have alcoholic hepatitis on biopsy.71 In practice, liver biopsy is desirable in cases where the alcohol etiology cannot be clarified clinically and to confirm diagnosis before commencement of specific therapy. In most severe cases, coagulopathy precludes percutaneous biopsy, but the transjugular route can be safely used.

Histological features of alcoholic hepatitis include steatosis, ballooning necrosis, acidophil bodies, Mallory's hyaline with cellular infiltration and fibrosis (Fig. 2). Macro- and microvesicular steatosis results in part from the inhibition of fatty acid oxidation by the redox shift that follows ethanol oxidation. These effects are exacerbated by high dietary fat, protein deficiency, and diabetes. Ballooning necrosis of hepatocytes results from impaired secretion of water and protein as a result of acetaldehyde disruption of microtubule function. Acidophilic bodies represent apoptosis of hepatocytes. Mallory's hyaline/bodies are reddish-pink intracellular aggregations of disorganized intermediate filaments within the cytoplasm, involving cytokeratins 8 and 18. The cellular infiltrate comprises mainly polymorphs surrounding hepatocytes, and is associated with increased serum levels of IL-883 and intercellular adhesion molecule (ICAM)-1.84 Mononuclear cells may also be seen. Fibrosis commences in zone 3 of the hepatic lobule around the central vein as a result of relative hypoxia and is termed perivenular. ‘Chicken-wire’ fibrosis surrounds lobular hepatocytes. Cirrhosis is present in 90% of severe cases of alcoholic hepatitis.71

Figure 2.

Histological features of alcoholic hepatitis. (a) Low-power view demonstrating the cardinal features of steatosis, fibrosis, inflammation, and hepatocellular injury (HE ×10). (b) (Black arrows) Mallory bodies are irregular eosinophilic cytoplasmic structures with a rope-like appearance. Although typical of alcoholic liver disease, they may be found in other liver disorders including non-alcoholic steatohepatitis. Further details are in the text. (Open arrow) Ballooning degeneration of hepatocytes is also shown (HE ×40). (c) (Open arrow) Pericellular fibrosis, also termed ‘chicken-wire’ fibrosis, surrounds individual degenerate hepatocytes. (Black arrow) Perivenular fibrosis extends from the central vein (Masson's trichrome ×20). (d) The unit lesion (also termed satellitosis) comprises a degenerating hepatocyte with (arrow) a surrounding cuff of neutrophils. Marked steatosis is also evident.

TREATMENT (TABLE 2)

Table 2.  Management options for alcoholic hepatitis
Interventions to reduce alcohol consumption
 Brief intervention
 Counseling
 12-step including Alcoholics Anonymous
 Pharmacotherapy
  Acamprosate
  Naltrexone
Treatment of liver injury
 Nutritional supplementation
 Corticosteroids
 Propylthiouracil
 Pentoxifylline

Mild to moderately severe alcoholic hepatitis almost invariably recovers without specific treatment and management is directed to the alcohol use disorder. Severe cases are associated with poor outcomes despite abstinence and warrant specific treatment.

Tailored management of alcohol use disorder

Goal setting

For those with clinically significant liver disease, lifelong abstinence is the appropriate treatment goal. Those with minimal abnormalities and without alcohol dependence may eventually return safely to consumption within recommended levels with appropriate monitoring. It is important to assist the patient to set a clinically appropriate treatment goal.

How can abstinence be achieved?

Treatment for alcohol abuse is moderately effective.85 Management of harmful consumption without dependence is by brief intervention, which is effective and can be offered by non-specialized staff in about 5–10 min.86 The FLAGS acronym summarizes one approach (Table 3). Follow up is recommended with repeated intervention if necessary.

Table 3.  Brief intervention for alcohol abuse
Feedbackto patient the nature and extent of alcohol-related problems
Listento patient concerns
Advisepatient clearly to reduce consumption
Goalsnegotiate clinically appropriate goals acceptable to the patient
Strategiesspecific suggestions to modify drinking

Management options for alcohol dependence include counseling of various types, pharmacotherapy, and peer support.87 There is no clear evidence to favor one treatment modality over the others and they may be combined according to patient preference and availability. Empirically, pharmacotherapy is commenced once the patient is well enough to contemplate resumption of alcohol consumption and ideally before discharge from hospital. Naltrexone is an orally active opioid antagonist that reduces the psychological reward provided by alcohol, leading to reduced desire to drink.85 Naltrexone is uncommonly hepatotoxic, typically at higher doses than are used routinely.88 Hepatic naltrexone clearance is reduced in cirrhosis,89 but no increased risk of hepatotoxicity has been reported. Acamprosate inhibits central nervous system N-methyl-d-aspartate receptors and gamma-aminobutyric acid (GABA) transmission. Side-effects are uncommon and mild and include diarrhoea, dizziness, and pruritus.90 Acamprosate is not metabolized by the liver and its pharmacokinetics are not altered in liver failure.91 Disulfiram inhibits ALDH, leading to accumulation of acetaldehyde after alcohol consumption, which causes an aversive reaction with nausea and vomiting. Controlled studies provide little evidence of effectiveness, because of side-effects and poor compliance.85 Disulfiram is occasionally hepatotoxic and inhibits CYP2E1, leading to drug interactions. It is not recommended for patients with alcoholic hepatitis.

Alcoholic Anonymous (AA) is a fellowship that conceptualizes ‘alcoholism’ as a medical disease that can be controlled by adopting the 12-step approach to achieve abstinence. The spiritual component of AA deters some potential participants. There are no randomized controlled trials of effectiveness, but the duration of abstinence is associated with the number of AA meetings attended.92

Monitoring of alcohol intake

Monitoring of alcohol intake by history taking provides adequate information in most cases. Alcohol may be measured in blood, breath, or urine and positive results indicate consumption within the last 12 h, but not earlier consumption. Trace levels in blood (up to 0.01 g%) may reflect endogenous ethanol production.93 A urine marker, 5-hydroxytryptophol, has been reported as being an accurate marker in assessing alcohol intake in the past 24 h, but is not yet widely available.94 In abstinent patients, levels of γ-GT fall, with an apparent half-life of 2 weeks and are usually normal within 6 weeks.95 Levels that remain elevated suggest continuing alcohol consumption or persisting liver disease. The MCV falls over several months and is insensitive to drinking relapse.96 An increase of CDT to 25–30% above baseline values is considered diagnostic of relapse in the absence of severe liver disease. Elevated CDT has been reported to often precede self-report of relapse by at least 1 month.97

Corticosteroids

Corticosteroids are the most studied treatment for alcoholic hepatitis, but despite 13 controlled clinical studies and five meta-analyses, their role remains controversial.98 The variation in published results has been attributed to small study numbers, subject mismatching, differences between study populations, different treatment regimens, different concurrent treatments, and use of different outcome measures. When data from all published studies are pooled (momentarily ignoring all differences between studies), the mortality for controls was 112/330 (34%) and for steroid-treated patients was 76/329 (23%), with a relative risk reduction of 0.32 (95% confidence intervals, 0.13–0.47), and the number needed to treat was 9.2.98 However, this is not a valid meta-analysis and formal meta-analyses led to conflicting conclusions because of the difficulties of pooling small studies that differ so widely from each other. The most recent of these studies, showing a survival benefit in those with MDF > 32, was of higher quality. No benefit on other outcomes has been found. When used, the active form of prednisolone, rather than the inactive precursor prednisone, is preferred because of reduced activation of prednisone in active liver disease.99 In alcoholic hepatitis, corticosteroid therapy generally produces few complications, but gastrointestinal (GI) bleeding and serious infections have been reported.100

Corticosteroids may be beneficial in alcoholic hepatitis by inhibition of cytokine production101 and suppression of extracellular matrix collagen synthesis, leading to an anti-inflammatory and antifibrotic effect. Corticosteroids have also been shown to reduce blood levels of ICAM-1.84 This molecule may play an important role in cellular chemotaxis, as it is upregulated on sinusoidal cells in alcoholic hepatitis patients, and facilitates leukocyte migration across hepatic endothelia. Corticosteroids also increase production of albumin102 and may thus improve ascites. Finally, Helman et al. reported that prednisolone significantly improved caloric intake by improving appetite and overall well-being.103

The American Society of Gastroenterology has recommended the use of corticosteroids for severe alcoholic hepatitis.104 The ideal patient has MDF > 32 with spontaneous encephalopathy, failure to improve during the first few days of hospitalization, and no contraindications such as active GI bleeding, hepatorenal syndrome, sepsis, or viral hepatitis. These individuals are rare, approximately one per annum for the average hospital. From the calculated number needed to treat of 9.2, it can be seen that corticosteroids can make little overall impact on the outcomes of this disease.

Nutritional supplementation

Alcohol provides 30 kJ of energy / gram, but is nutritionally ‘empty’, lacking in protein, vitamins, and minerals. In patients with ALD, alcohol may provide up to 50% of calories, so that some evidence of malnutrition may be found in almost all patients.105,106 Nutritional supplementation attempts to correct this deficiency and provide the substrates needed for hepatic regeneration and systemic nutrition. Maintaining a positive nitrogen balance correlates with an improved outcome,107 but is difficult to achieve.

The evidence that supplementary nutrition improves the outcome of alcoholic hepatitis remains limited.98 Two studies of enteral supplementation in patients with alcoholic hepatitis reported improvements in both liver function and nutritional status using 1 g protein/kg/day, but failed to show any beneficial effect on survival. Studies of parenteral supplement have showed varied results. Nasrallah and Galambos found an improvement in both liver function tests and survival.108 Improvement in biochemical, metabolic, and nutritional parameters has been reported; however, amino acid supplementation did not affect mortality at 1 month.105

A more recent study71 compared total enteral nutrition (TEN) with prednisolone therapy and found that, although steroids have been widely supported as the mainstay of therapy in alcoholic hepatitis, patients on TEN do no worse, and further studies are needed to evaluate the potential synergistic effect when combined.

In practise, it is essential that all patients with alcoholic hepatitis are adequately nourished and should receive the standard daily requirements of at least 30 kcal/day and 1 g protein/kg/day. Supplements should be given first via oral or enteral routes, but if difficulties arise, a parenteral route may be used.98

Pentoxifylline

Pentoxifylline is a dimethylxanthine derivative used as an adjunctive agent for peripheral vascular disease.109 The drug decreases blood viscosity (via increasing erythrocyte deformity), decreases red cell and platelet aggregation, and improves blood flow. Pentoxifylline is also thought to improve organ microcirculation and tissue oxygenation, and lower portal hypertension in experimental cirrhosis, an observation attributed to the rheological effects of the drug. Pentoxifylline has other actions that suggest a broader therapeutic role. It attenuates TNF-α release and action and exerts an antifibrinogenic action.110

Pentoxifylline has recently been shown to improve short-term (4-week) survival in a prospective randomized controlled trial of 96 patients with severe alcoholic hepatitis.32 Mortality was associated with rising serum TNF-α levels, which were reversed using pentoxifylline. This survival benefit was related to a significant decrease in the rate of development of hepatorenal syndrome. Replication of this result by other clinical units is required before its use can be confidently recommended. The only significant side-effects were epigastric pain, vomiting, and dyspepsia and these resolved soon after discontinuation of treatment. The drug is inexpensive and widely available. Patients with sepsis and active GI bleeding were excluded from the study and the use of this drug in such patients has not been evaluated.

Propylthiouracil

Alcohol and hypoxia are both known to induce centrilobular hepatic injury.98 These changes can be partially restored by antithyroid medication, such as propylthiouracil (PTU). As with corticosteroids, the use of PTU for alcoholic hepatitis remains controversial. Reduced mortality in both short-111 and long-term112 administration of 300 mg PTU/day have been reported. No clinically important side-effects were seen. In contrast to the above, Halle et al. reported that PTU did not reduce mortality, change the incidence of complications, or speed the improvement of liver tests.113 However, because of limited numbers of patients, they concluded a beneficial effect of PTU could not be excluded. Propylthiouracil is not generally used because of the conflicting evidence described above.

Insulin and glucagon

Hepatic regeneration is inhibited by alcohol consumption and several hormones have been assessed for their potential to stimulate regeneration in alcoholic hepatitis. Of these, insulin and glucagon are the best studied. There is mixed evidence for the use of insulin and glucagon in alcoholic hepatitis.98 Currently, insulin/glucagon therapy is not recommended. Moreover, continuing venous access and careful monitoring of blood glucose levels are required to maintain euglycemia throughout the 3-week course of treatment. It is hoped that investigation of mechanisms of hepatic regeneration will give rise to more specific and effective treatments.

Colchicine

Colchicine has several effects that may be beneficial in the treatment of alcoholic hepatitis. These include inhibition of granulocyte migration, inhibition of microtubule assembly, inhibition of collagen secretion, and increased collagenase production. A recent Cochrane review has found that there is no evidence of a therapeutic effect in liver disease.114 Two randomized controlled studies of colchicine in alcoholic hepatitis have been reported, both showing no benefits from treatment.115,116 Based on the above results, colchicine cannot be recommended for alcoholic hepatitis.

New approaches to therapy

S-adenosylmethionine

S-adenosylmethionine is a source of cysteine for the production of GSH. In cirrhosis, SAMe synthase activity is reduced. S-adenosylmethionine supplementation of ethanol-fed rats restores the mitochondrial pool of GSH,117 suggesting a role for therapy with SAMe. Treatment with SAMe has been shown to lead to increased survival of patients with alcoholic cirrhosis.118 Patients in this study did not have alcoholic hepatitis and SAMe has not yet been evaluated in this setting.

Polyenylphosphatidylcholine

Studies in experimental animals have shown that polyenylphosphatidylcholine (PPC) reverses a number of the adverse effects of ethanol on the liver. Polyenylphosphatidylcholine acts as an antioxidant, downregulates CYP2E1 activity, restores SAMe synthase activity, and reduces stellate cell activation and collagen synthesis. The most active component is dilinoleoylphosphatidylcholine (DLPC), but pure preparations are not readily available. A large cooperative Veterans’ Affairs study of PPC in ambulatory patients with ALD has shown significant protection against cirrhosis.119 One study of alcoholic hepatitis showed a trend to long-term improved outcomes, but the efficacy of this treatment is unclear.120

N-acetyl cysteine

Glutathione depletion is common to the toxicity of both paracetamol and alcohol. Alcoholic hepatitis patients who had taken apparently non-toxic amounts of paracetamol were given N-acetyl cysteine, leading to reduced hospital stay compared with historical controls.121 A subsequent controlled study found no benefit.

Cytokines

In view of the central role of TNF-α in alcoholic hepatitis, inhibition of TNF-α might be useful clinically. Monoclonal antibodies are now available for clinical use (Infliximab). A recent case series suggested increased survival compared with historical controls122 and a controlled trial was initiated. This trial was stopped after deaths from sepsis in the TNF-α antibody treated group. However, TNF-α also plays a central role in hepatic regeneration, suggesting that TNF-α inhibition may have adverse effects as well.

Inhibition of TGF-β is a potential antifibrotic therapy. Interleukin-10 is an anti-inflammatory cytokine and preliminary data have been reported from a controlled trial in alcoholic hepatitis, suggesting a beneficial effect.123 A recent study has shown that the soluble receptor of TGF-β is effective in an animal model of non-alcoholic liver fibrosis,124 but there are no data in human ALD.

Others

Many other potential therapies have been evaluated for ALD with varied outcomes, including malotilate, silymarin, methionine/choline, betane, metadoxine, and paromomycin, but have not as yet been evaluated in alcoholic hepatitis.

Combination therapy

In view of the multistep process involved in the pathogenesis of alcoholic hepatitis, combination therapy targeting multiple aspects of this disease may improve clinical outcomes.125 For example, nutritional supplementation might be coupled with pharmacological agents that suppress inflammation and oxidant stress. Although this is an attractive concept, no trials of combination therapy have been reported.

Liver transplantation

For patients with ALD, clinical outcomes of orthotopic liver transplantation (OLT) are comparable to those of patients undergoing transplantation for other liver diseases.126 Alcoholic liver disease has become one of the most common indications for OLT.127 Only 6% of those with end-stage ALD are transplanted in the USA because of a combination of donor shortage, lack of patient suitability, and other factors. In Australia, alcohol  has  been  reported  to  be  the  primary  etiology  for 6% and a contributing factor for another 6% of transplants.128

There is a risk of recurrent alcohol abuse that may lead to recurrent ALD in the graft, non-compliance with immunosuppression, and other alcohol related problems. About one-third of survivors return to alcohol consumption and, in most cases, consumption remains at very moderate levels.129 Careful patient selection is crucial, but objective criteria to predict a successful outcome have not been defined.130 Most transplantation units require a minimum of 6 months of abstinence before consideration for transplantation. Many patients who achieve stable abstinence will recover sufficiently and not require transplantation. Consultation with specialists with relevant experience is required.

Patients with alcoholic hepatitis do not meet the above criteria for transplantation. They have been drinking to the time of admission and are typically difficult to assess from a psychosocial perspective as they are severely ill, and often encephalopathic. Consequently, few cases of OLT for alcoholic hepatitis have been reported and the results of these are not encouraging.131 Orthotopic liver transplantation for alcoholic hepatitis is not currently recommended.

ACKNOWLEDGMENTS

Michael Buckland from the Department of Anatomical Pathology at Royal Prince Alfred Hospital assisted with the photomicrography for Figure 1 and Sarah Hutchison assisted with the references.

Ancillary