Improving survival in alcohol-related hepatitis: what's new?

Abstract

Alcohol-related hepatitis (AH) is the most florid presentation of alcohol-related liver disease and carries a high short-term and long-term mortality rate. Specific treatment options remain inadequate. The current management approach for AH focuses on early identification, careful screening and treatment of infection, as well as identification of those patients who may benefit from corticosteroid therapy based on validated prognostic scoring systems.

In recent years, there has been growing interest in exploring novel therapies for AH, which may offer alternative treatment options beyond the traditional approaches. Additionally, early liver transplantation (LT) has emerged as a promising option in selected cases with growing evidence supporting its role. In this review, we will discuss the current evidence base for the assessment and treatment of AH, and how these advances are shaping practice to improve outcomes in the UK.

Key points

• Alcohol-related hepatitis (AH) should be diagnosed based on clinical criteria; if patients do not fulfil all criteria, a liver biopsy may be considered.

• Screening for infection should be performed in patients presenting with AH. When infection is identified, it must be treated aggressively.

• Prognosis of patients with AH should be assessed using a validated scoring system.

• Corticosteroid treatment should be considered in patients with indicators of likely benefit, such as Glasgow Alcoholic Hepatitis Score≥9, neutrophil-to-lymphocyte ratio 5–8, Model for End-Stage Liver Disease 21–51.

• Response to steroids should be assessed at day 4 or day 7 to determine the effectiveness of treatment.

• Early liver transplantation for AH has shown favourable outcomes but has not yet been adopted in the UK.

• Abstinence from alcohol is key to long-term survival after AH.

Introduction

Alcohol-related hepatitis (AH) is the most florid presentation of alcohol-related liver disease (ARLD) characterised by recent-onset jaundice and a history of prolonged alcohol excess. A clinical diagnosis of AH can be made based on agreed consensus criteria. Onset of jaundice occurs within the previous 8 weeks on a background of alcohol excess, (typically >5 UK Units (40 g) per day for women or ˃7.5 UK Units (60 g) per day for men) with less than 60 days of abstinence before the onset of jaundice.¹ Biochemically, AH can be diagnosed by an elevated aspartate aminotransferase (AST), typically >50 IU/mL, with an AST/alanine transaminase ratio ˃1.5 and both values ˂400 IU/mL. The acuity of jaundice and biochemical features help to differentiate AH from decompensated cirrhosis. In cases where there is diagnostic uncertainty a liver biopsy may be done, but it rarely changes the diagnosis in individuals who meet clinical criteria.²

Severe forms of AH carry a dismal short-term mortality rate, with 20% of individuals dying within 28 days of admission,³ higher than most acute medical emergencies. Even less severe forms of AH can still have a poor outcome with relatively high risk of mortality.⁴ In recent years, there have been numerous studies aimed at improving prognostic scores, guiding timing of therapeutic interventions and identifying patients who are most likely to benefit from treatment. Significant developments have been made in understanding inflammatory pathways, biomarkers and novel therapeutics targets, leading to several clinical trials aiming to improve patient outcomes. A role for early liver transplantation (LT) in severe AH is emerging. This review will describe important advances in the management of AH in UK practice.

Who to treat?

Specific treatment options in AH remain inadequate and controversial. First, it is crucial to make a correct diagnosis and differentiate AH from decompensated ARLD cirrhosis. Corticosteroids are the only pharmacological treatment currently recommended. Results of the STOPAH trial showed that corticosteroids were associated with a modest improvement in mortality at 28 days, though the improvement did not reach conventional statistical significance and any benefit was not sustained at 90 days.⁵ Subsequent meta-analysis also supports the use of corticosteroids in selected patients.⁶ However, corticosteroids have side effects and are associated with an increased risk of infection,⁷ highlighting the importance of careful patient selection for treatment.

Prognostic scoring systems help define disease severity and identify those patients who are most likely to benefit from treatment. These scores can be divided into static scores which use values taken at baseline, and dynamic scores, which assess changes in biochemical variables after initiation of treatment (table 1). The most well-established static score is the discriminant function (DF), which combines prothrombin time (PT) and total bilirubin, with a score of ≥32 typically defining severe AH and suggesting a poor prognosis.⁸ In recent years, other static scores have been developed, including the Glasgow Alcoholic Hepatitis Score (GAHS)⁹ and the Age, serum Bilirubin, INR and Creatinine (ABIC) score,¹⁰ both of which risk stratify patients using age and biochemical parameters. The Model for End-Stage Liver Disease (MELD) score has also been used in AH, although thresholds indicating a poor prognosis vary between studies.¹¹

Table 1.

Comparison of prognostic scores in AH

			Age	WCC	Urea	Bili	PT/INR	Cr	Alb	Lymph count	Neut count	Thresholds
Static	Prognosis	ABIC	+			+	+	+				≥9=25% 90 day survival 6.71 to <9=70%, <6.71=100%.
	Prognosis and CS indication	MELD				+	+	+				>20=severe 25–39 CS treatment
		DF				+	+					≥32 Severe AH. CS treatment.
		GAHS	+	+	+	+	+					≥9 D1 or day 6–9=poor prognosis If DF≥32 and GAHS≥9 CS treatment
		NLR								+	+	Baseline NLR 5–8 CS treatment
Dynamic	Assess response	Lille score	+		+	+ D0+D7	+	+	+			≥0.45=25% 6 month survival
		ECBL				+ D0+D7						If ECBL present=good prognosis. If no ECBL, then 77% 6-month mortality
	Outcome prognosis	%Δ bilirubin				+ D0+D6-9						Non-responder <25% Δ bilirubin. 57.9% 56-day mortality rate vs 11.1% of responders

ABIC, Age, serum Bilirubin, INR and Creatinine ; AH, alcohol-related hepatis; Alb, serum albumin; Bili, serum bilirubin; lymph count, serum lymphocyte count; Cr, serum creatinine; CS, corticosteroid; DF, discriminant function; ECBL, Early Change in Bilirubin Levels; GAHS, Glasgow Alcoholic Hepatitis Score; INR, international normalised ratio; MELD, Model for End-Stage Liver Disease; Neut, serum neutrophil count; NLR, neutrophil-to-lymphocyte ratio; PT, prothrombin time; WCC, white cell count.

Using STOPAH trial data, all three newer static scores (GAHS, ABIC and MELD) were found to have a higher predictive value for 28-day and 90-day mortality than DF.¹² Specifically, patients with consistently high static scores had a higher risk of mortality at 28 days and were more likely to benefit from corticosteroids. Another multicentre study found that MELD score had superior prognostic ability compared with GAHS and ABIC, while DF had the lowest predictive value.¹¹ These findings suggest that clinicians should rely less on DF and instead use newer scores to predict short-term mortality.

Dynamic scores, such as the Lille score,¹³ the Early Change in Bilirubin Levels¹⁴ and the percentage change in bilirubin¹⁵ predict mortality based on a reduction in bilirubin level in patients who have received corticosteroid therapy. The Lille score is the most widely used dynamic score and combines several variables, including age, renal function, albumin, PT and bilirubin at baseline and 7 days. A non-response to corticosteroids is defined as a Lille score ≥0.45, and this is associated with a much poorer survival at 6 months compared with responders (25% vs 85%).¹³ To limit unnecessary steroid exposure and reduce risk of infection, two small studies have shown that calculating the Lille score at 4 days is similarly accurate in predicting response and mortality^{16 17} though validation in larger studies would be helpful.

Recent data demonstrates that combining both static and dynamic prognostic scores can create joint-effect models that better predict survival in AH. A combination of the MELD and Lille scores has been found to provide the greatest predictive value at both 2 and 6 months.¹⁸ By incorporating both baseline and post-treatment variables, this approach enables clinicians to make more accurate and informed decisions regarding timing of therapeutic interventions, as well as identifying patients who are most likely to benefit from treatment.

When and how to treat with steroids?

In AH, the rate of infection is significantly higher in those treated with steroids.^{7 19} Therefore, steroids should be used judiciously after careful screening and aggressive treatment of any infection at baseline. Moreover, timing of infection relative to steroid treatment is also important. If a patient has a pre-existing infection that is effectively treated before initiating steroid therapy, the risk of mortality is reduced. However, if an infection develops while on steroids the prognosis can be poor, and the risk of complications increases significantly. It is essential to monitor closely for signs of infection in patients receiving steroids, and appropriate treatment should be initiated promptly when infection is suspected or confirmed.

It is important to identify the optimal cohort who may benefit from steroid treatment, thereby avoiding exposing patients with milder disease to unnecessary steroid therapy, while ensuring that those who are likely to benefit most receive them (figure 1). In patients with a DF≥32, a GAHS≥9 can help to identity patients who may benefit from steroid therapy.²⁰ Furthermore, retrospective analysis of 789 patients from the STOPAH trial⁵ found that a baseline neutrophil-to-lymphocyte ratio (NLR) of between 5 and 8 can help identify those patients likely to benefit from corticosteroid treatment²¹ while also predicting 90-day mortality. The authors propose incorporating NLR into GAHS, creating a modified GAHS which may improve its predictive value, although further validation of this modified score is needed.

Figure 1.

Diagnosis and management algorithm for alcohol-related hepatitis. AH, alcohol-related hepatis; ALT, alanine transaminase; AST, aspartate aminotransferase; DF,discriminant function; GAHS, Glasgow Alcoholic Hepatitis Score; LT, liver transplant; MELD, Model for End-Stage Liver Disease; NLR, neutrophil-to-lymphocyte ratio.

The MELD score has also been used for prognostication in AH. Patients with an MELD score between 25 and 39 derived the greatest survival benefit from corticosteroid treatment in a large multicentre cohort study.²² However, patients with an MELD score >51 did not benefit from steroid treatment. In addition, the trajectory of serum bilirubin appears to be an important determinant of survival. Recent analysis has shown that patients can be categorised into three groups based on bilirubin trajectory. ‘Fast fallers’, those with a rapidly falling bilirubin level, had better 90-day survival compared with patients with rapidly rising or static bilirubin.²³ This information can help to identify those patients who are likely to spontaneously recover without steroid therapy and thus minimise exposure to unnecessary treatment.

What else to treat with?

Supportive therapy

Given that infections are a common and serious complication in patients with AH, universal antibiotic prophylaxis has been proposed for all with AH or at least those being treated with steroids. The AntibioCor study investigated co-amoxiclav and prednisolone, compared with prednisolone and placebo in 284 patients with biopsy proven severe AH. Although the incidence of infection was significantly lower in the intervention group compared with placebo (29.7% vs 41.5%, p=0.015) this did not translate into a significant improvement in overall survival at 2 months (82.7% in intervention group vs 78.1% in placebo group, p=0.3).²⁴ Therefore, current evidence does not support the use of prophylactic antibiotics for all patients with AH when steroid treatment is planned.

Malnutrition is also common in severe AH and associated with reduced survival.²⁵ Enteral nutrition with nasogastric feeding is often used for patients who are unable to meet their nutritional requirements. A multicentre randomised controlled trial (RCT) involving 136 patients with biopsy-proven AH compared intensive enteral or conventional nutrition alongside steroid therapy. While there was evidence of histological and biochemical improvement, there was no significant difference in 6-month mortality between the two groups, with authors concluding that oral nutrition should be the preferred first-line intervention.²⁶ However, low daily calorie intake was associated with higher mortality, highlighting the importance of achieving adequate nutritional intake as a main goal of treatment.

Abstinence from alcohol is the single most important determinant which influences long-term survival after an episode of AH,^{27 28} regardless of current, emerging or transplant options for AH, which aim to reduce short-term mortality. There are limited studies evaluating the management of abstinence in those surviving an episode of AH. Integrated treatment programmes, including both specialist psychosocial and pharmacological therapies, have been shown to improve engagement, treatment retention and outcomes.²⁹ Therefore, management strategies must focus on providing education, support and access to resources to ensure and maintain abstinence in the longer term. The importance of abstinence from alcohol in AH cannot be overstated.

Anti-inflammatory agents

Alcohol consumption upregulates production of proinflammatory cytokines³⁰ and inflammasome activation is a potential therapeutic target in AH. Interleukin-1 (IL-1) is a cytokine which plays a key role in regulating the immune response and pathogenesis of inflammation, therefore, blocking activity of IL-1 has been proposed as therapy in AH.

Anakinra, an IL-1 inhibitor, is currently used to treat several inflammatory conditions. The Defeat Alcoholic Steatohepatitis study was a multicentre, randomised double-blind trial to evaluate the efficacy of anakinra in AH. Patients with severe AH (DF≥32 and MELD≥20) were randomised to receive either 28 days of methylprednisolone or the combination of 14 days of anakinra, plus 28 days of pentoxifylline and 180 days of zinc with the primary endpoint being 180-day mortality. The study found no significant difference in 6-month mortality between the combination group and the prednisolone group (67.9% vs 56%, respectively).³¹ The study was underpowered, with only 104 patients undergoing randomisation, and took 5 years to recruit, which was followed by an additional 4 years before results were published. The prolonged study duration rendered the role of pentoxifylline outdated. A higher number of fungal infections occurred in the prednisolone-treated group, and the authors suggested that IL-1 inhibitors may be a better option than prednisolone for patients at high risk of systemic fungal infection.

However, because pentoxifylline was included in the combination, a direct comparison between anakinra and steroids was not possible. To further evaluate this, a phase 2b double-blind multicentre RCT (AlcHepNet) was conducted, comparing the efficacy of anakinra (plus zinc) or prednisolone in patients with severe AH.³² The trial was prematurely halted after an interim analysis showed significantly poorer 90-day overall survival rates in those receiving Anakinra plus zinc.³³

The ISAIAH trial, a UK multicentre phase 2, randomised, double-blind, placebo-controlled study investigated the use of canakinumab in patients with biopsy-confirmed AH, a DF of ≥32 and an MELD≤27.³⁴ Canakinumab is an inhibitor of IL-1β, a proinflammatory cytokine of the peripheral immune response during infection and inflammation. The primary endpoint was histological improvement after 28 days following a single dose of canakinumab, with all patients receiving prophylactic antibiotics. Although histological improvement was observed, no clinical or biochemical improvement was noted in patients receiving canakinumab.³⁴ Notably, this was a small-scale study with a limited sample size of 48 patients, and some baseline clinical characteristics were unmatched (GAHS and performance status) between the two groups. Further research is needed to evaluate the role of canakinumab, potentially in combination with other therapies or in certain subgroups of AH.

Interleukin-22 (IL-22) has emerged as a promising therapeutic target in AH, owing to its potential beneficial effects on inflammation and impaired hepatic regeneration, which are key characteristics of AH.³⁵ A small open-label pilot study using an IL-22 agonist (F-652), a recombinant IL-22 fusion protein, demonstrated favourable outcomes and increased expression of markers of hepatic regeneration.³⁶ While it showed a good safety profile, larger RCTs will be needed to establish the efficacy of IL-22 agonists in treating AH.

Historically, other anti-inflammatory agents have been studied in the management of AH. Pentoxifylline, an inhibitor of tumour necrosis factor (TNF), had been thought to improve short-term survival³⁷ but the STOPAH trial showed that Pentoxifylline does not work.⁵ Similarly, controlled trials of infliximab and etanercept, which also inhibit TNF, showed no clinical efficacy in severe AH.^{38 39} These agents have fallen out of favour and are no longer recommended for use in AH.

Gut microbiome

Chronic alcohol consumption disrupts the gut microbiome, leading to dysbiosis, intestinal bacteria overgrowth and increased bacterial translocation. These changes cause immune activation and inflammation, which play an important role in AH pathogenesis.⁴⁰ In murine models, AH is transmissible using human intestinal microbiota from patients with ARLD.⁴¹ The presence of an exotoxin (Cytolysin) secreted by certain bacterial species was found to correlate with poor outcomes in AH and bacteriophages which targeted this virulence factor were able to attenuate AH in mice.⁴² Phage treatment could be a potentially importantly novel therapeutic target for severe AH. Recent studies in AH have targeted these alterations in the gut microbiome as a potential therapeutic approach.

The RIFA-AH study was an open-label, single-arm trial investigating whether rifaximin could reduce 90-day mortality compared with a matched historic cohort who received standard therapy. Although the study was underpowered, the rifaximin group had a lower mean number of infections (0.29 vs 0.62, p=0.049), indicating that rifaximin may inhibit bacterial translocation in this patient population. Moreover, the treatment group had a lower incidence of acute-on-chronic liver failure (ACLF) relating to infection. Notably, there was a significantly lower rate of liver-related complications (such as de novo ACLF, ascites, gastrointestinal bleeding or hepatic encephalopathy) within the rifaximin group (0.43 vs 1.26 complications per patient, respectively, p=0.01). Although the treatment group had a lower 90-day mortality than the control group, this difference did not reach statistical significance (14.2% vs 30.9%, p=0.15).⁴³

The potential role of faecal microbiota transplantation (FMT) in AH management has been explored with encouraging results, further highlighting the important reciprocal relationship between the gastrointestinal tract and its microbiota with the liver. In a recent open-label trial, 33 patients with severe AH and features of ACLF received either FMT or standard of care (SOC). Survival at 28 and 90 days was improved in patients receiving FMT, with no significant adverse events.⁴⁴ In addition, a long-term retrospective analysis has found that patients with AH who received FMT had a lower incidence of complications such as ascites, hepatic encephalopathy, infection, hospitalisation and alcohol relapse in comparison to those who received SOC, with an improvement in 3-year survival.⁴⁵ FMT has also been shown to be associated with short-term reduction in alcohol craving and consumption in patients with ARLD cirrhosis and alcohol use disorder, in a small phase 1 RCT trial.⁴⁶ However, it should be noted that many of the FMT studies are small scale, not all are randomised and the participant characteristics of those included may not necessarily be generalisable internationally.

Other therapeutic options

Granulocyte colony-stimulating factor (GCSF) has been studied as a potential therapy for AH due to its ability to induce liver regeneration.⁴⁷ However, results from several trials have been conflicting. A recent multicentre RCT found no significant beneficial effect on patients’ survival with ACLF, including alcohol-precipitated ACLF which was thought to be predominantly AH.⁴⁸ A recent meta-analysis concluded that while GCSF may reduce mortality in AH, there was a high level of heterogeneity between Asian and European studies, with some reporting improved survival and others not showing any significant benefit.⁴⁹ As a result, the role of GCSF in AH is not currently recommended and further studies are needed.

N-acetylcysteine (NAC) has also been studied in AH due to its antioxidant and anti-inflammatory properties, which may help mitigate liver injury caused by oxidative stress and inflammation. The results thus far have been underwhelming. One study randomly assigned 174 patients to receive prednisolone plus NAC or prednisolone alone. While survival was improved at 1 month in those treated with NAC, the difference in mortality between the two groups was not significant at 3 and 6 months.⁵⁰ Rates of infection and hepatorenal syndrome were significantly lower in those treated with NAC. A subsequent study evaluated the combination of NAC and GCSF and found the addition of NAC did not provide any additional survival benefit at 90 days.⁵¹

What is the role for LT?

Early LT for AH improves survival in highly selected patients who do not respond to medical therapy. The first pilot study, conducted in France and Belgium in 2011, reported encouraging survival rates of 77%.⁵² This study led to a significant shift in practice in France, where surveys show that most clinicians are now open to considering early LT for AH with a more flexible approach to duration of abstinence.⁵³

Following this landmark study, other countries including the USA have also reported efficacy of early LT in treating severe AH cohorts. In a single-centre study published in 2016, 94 patients with severe AH who did not respond to steroid therapy were evaluated. Of these, nine patients underwent early LT with excellent 6-month survival rates compared with a matched control group (89% vs 11%, p<0.001).⁵⁴ In 2018, a consortium of 12 centres in the USA (ACCELERATE-AH) collaborated to publish their experience of early LT for AH. In the study, 147 patients underwent LT, with a median duration of 55 days abstinence prior to transplantation. Cumulative survival after LT was 94% at 1 year and 84% at 3 years. Sustained alcohol use was found in 11% of transplant recipients with a median follow-up time of 1.6 years.⁵⁵ Subsequent meta-analysis including 11 studies reported no significant difference between estimated risk of alcohol relapse in those undergoing early LT for AH or elective LT for ARLD cirrhosis.⁵⁶

In the first ever prospective controlled study of early LT for severe AH, conducted in France and Belgium, the risk of relapse within 2 years of transplantation was compared in those receiving early LT for AH and those transplanted via standard pathways for ARLD after 6 months of abstinence. The study did not establish non-inferiority between the two groups based on the prespecified margin of 10% (34% early LT vs 25% standard transplant group; absolute difference 9.1%, p=0.45).⁵⁷ Additionally, a higher rate of high alcohol intake (>30 g/day for women/>40 g/day for men) was observed in the patients who underwent early LT (22%) than in the standard pathway group (5%). These findings underscore the need for comprehensive and prospective data that encompass information about the pattern and impact of alcohol relapse over the long term. The only other European data available on early LT has recently emerged from a multicentre Italian cohort study. Using strict selection criteria similar to the Franco-Belgian team, 16 patients underwent early LT with 100% 2-year survival.⁵⁸

These favourable results have informed practice guidance from international societies including the European Association for the Study of Liver and the American Association for the Study of Liver Disease. They recommend considering early LT in carefully selected patients with favourable psychosocial profiles who have failed medical therapy.^{59 60} However, there is a need for further standardisation before this practice is more widely adopted. In the UK, a pilot programme was initiated between 2014 and 2017 for early LT in AH, aiming to transplant 20 steroid non-responders. This ultimately closed without any patients undergoing transplantation during this period. The criteria for listing were extremely strict, given concerns expressed about the wider impact on the organ donation process and the transplant waitlist.⁶¹ Consequently, early LT for AH has not yet been adopted, in part due to a lack of robust, objective and validated criteria to select individuals who might benefit most. Nevertheless, recent UK Liver Advisory Group guidelines recommend that patients with severe AH should be considered for transplantation after 3 months of abstinence following an episode of AH, provided they have a favourable psychosocial assessment.⁶²

Conclusion

Beyond LT, improvements in survival for AH are more likely to be achieved through small but significant improvements in aspects of medical care rather than a single intervention, such as corticosteroids. While early LT has been shown to further improve survival in other countries, the UK is still determining whether to adopt this approach.

As our understanding of AH continues to evolve, incorporating novel therapies and considering early LT in selected cases may hold promise for improving outcomes in this challenging condition. However, sustained abstinence from alcohol remains a crucial goal in AH management, and efforts should be directed towards supporting abstinence and reducing return to heavy drinking to achieve long-term treatment success in patients with AH.

Ethics statements

Patient consent for publication

Not applicable.