Steroids show modest 28-day benefit in severe alcohol-associated hepatitis

Corticosteroids modestly reduce 28-day mortality in patients with severe alcohol-associated hepatitis, but the overall certainty of evidence is low and durable survival benefits remain unproven, according to a large systematic review and meta-analysis published in Clinical Gastroenterology and Hepatology.

For the work, corresponding author Juan Pablo Arab, MD, FRCPC, director of alcohol sciences at the Stravitz-Sanyal Institute for Liver Disease and Metabolic Health at Virginia Commonwealth University, Richmond, and colleagues pooled data from 52 randomized controlled trials including 5,121 adults hospitalized with severe alcohol-associated hepatitis, defined as Maddrey’s Discriminant Function of at least 32 or Model for End-stage Liver Disease score greater than 20. The primary outcome was all-cause mortality at 28 days; secondary outcomes included 90-day mortality, infections, acute kidney injury, hepatorenal syndrome and liver transplantation.

In a sensitivity analysis excluding one outlier trial, corticosteroids were associated with a significant reduction in 28-day mortality compared with placebo (risk ratio [RR], 0.62; nine studies; 890 participants; low-certainty evidence). Nineteen percent of patients treated with corticosteroids (84 of 449) died by 28 days vs 26% (116 of 441) receiving placebo. Heterogeneity was moderate (I² = 63%).

In the primary pooled analysis including all 10 trials (918 participants), the difference did not reach statistical significance (RR, 0.69; I² = 61%). Only one small study showed a 90-day mortality benefit (RR, 0.28; 61 participants).

Rates of serious adverse events, infections, acute kidney injury and progression to hepatorenal syndrome did not differ significantly between corticosteroids and placebo, with GRADE ratings ranging from low to moderate certainty.

In other findings, pentoxifylline did not improve survival compared with corticosteroids (RR for 28-day mortality, 0.99; seven studies; 994 participants) or placebo (RR, 0.73; three studies; 686 participants). However, in one study, pentoxifylline reduced progression to hepatorenal syndrome compared with placebo (RR, 0.24; 101 participants; low-certainty evidence).

Several combination strategies showed larger relative effect sizes, although each was supported by one or two small trials.

Adding N-acetylcysteine to corticosteroids reduced 28-day mortality compared with corticosteroids alone (8% [7 of 85] vs 24% [21 of 89]; RR, 0.35; 174 participants; low-certainty evidence). The 90-day difference was not statistically significant. Combination therapy also reduced infections (RR, 0.45; moderate-certainty evidence) and progression to hepatorenal syndrome (RR, 0.48).

Data cited in the review found that granulocyte colony-stimulating factor plus pentoxifylline lowered 28-day mortality compared with pentoxifylline alone (17% [7 of 41] vs 63% [27 of 43]; RR, 0.27; two studies; 84 participants; low-certainty evidence) and reduced 90-day mortality (RR, 0.32; three studies; 129 participants; moderate-certainty evidence). Progression to hepatorenal syndrome was also less frequent (RR, 0.13).

Metadoxine combined with corticosteroids reduced 28-day mortality compared with corticosteroids alone (26% vs 54%; RR, 0.47; 70 participants) and reduced 90-day mortality (RR, 0.39). Certainty was low to moderate.

Fecal microbiota transplantation did not significantly affect 28-day mortality but was associated with lower 90-day mortality compared with corticosteroids (RR, 0.58; 120 participants; moderate-certainty evidence).

On the other hand, anakinra plus zinc was associated with higher mortality than corticosteroids at 28 days (RR, 6.91) and 90 days (RR, 2.68).

A post hoc Bayesian network meta-analysis supported reductions in 28-day mortality with corticosteroids, pentoxifylline and corticosteroids plus N-acetylcysteine compared with placebo, and reductions in 90-day mortality with corticosteroids, granulocyte colony-stimulating factor plus corticosteroids and metadoxine plus corticosteroids.

Of the 52 trials, 39 had an unclear risk of bias in at least one area, and 23 were judged to have a high risk of bias because blinding was not done well or not reported clearly. Many of the results were also considered less certain because the studies were small, with several comparisons including fewer than 100 participants.

The authors noted that for clinicians, the findings reinforce corticosteroids as first-line therapy in eligible patients with severe alcohol-associated hepatitis, consistent with current guideline thresholds. However, the lack of consistent 90-day benefit and the low certainty of most adjunctive data underscore the need for careful patient selection, close monitoring for infection and kidney injury, and enrollment in clinical trials when available.

Three of the study authors reported having received consulting/advisory board fees and/or grants from several pharmaceutical companies. Dr. Arab had no disclosures to report.

GI & Hepatology News invited corresponding author John Pablo Arab, MD, to share his perspective on the study’s implications.

When you had all the data in front of you, was there a finding, or perhaps more than one, that surprised you?

Dr. Arab: Two things stood out. First, while corticosteroids continue to show the most consistent short-term signal, the overall certainty remains limited, and the effect is primarily concentrated early (28-day outcomes). In our primary analyses, heterogeneity was substantial in the steroids vs placebo comparison, and the signal became clearer only after sensitivity analyses (eg, excluding an outlier study).

Second, some “combination” approaches showed potentially meaningful early benefits (eg, corticosteroids plus N-acetylcysteine; and metadoxine plus corticosteroids), but these signals are generally based on a small number of trials and participants, so the overall certainty ranges from low to moderate. That gap between “promising point estimates” and “limited confidence due to sparse data” is both striking and important.

What do these findings mean for real-world clinical practice?

Dr. Arab: In practice, our results support a pragmatic approach: corticosteroids remain the first-line pharmacologic therapy for steroid-eligible patients with severe alcohol-associated hepatitis, with the clearest evidence for short-term (around 28-day) benefit, but limited certainty overall.

For adjunctive or alternative agents, the data suggest some regimens may be beneficial, but most are not yet supported strongly enough to recommend as routine standard of care across settings, because findings often come from single trials or small study clusters and vary in certainty.

So, the clinical takeaway is: we need better treatments!, in the meantime, continue to prioritize careful patient selection for steroids, provide optimal supportive care (including infection vigilance and management of complications), and view promising combinations as hypothesis-generating or for use in specific contexts where local expertise/protocols and patient factors justify them — while we await larger, multicenter trials to confirm benefit and generalizability.

Key Takeaways

• Steroids still lead the pack. Across 52 randomized trials, corticosteroids remained the only therapy consistently associated with improved short-term survival in severe alcohol-associated hepatitis.

• Combination therapy shows promise. Adding N-acetylcysteine to corticosteroids may provide additional survival benefit in some studies, though evidence remains mixed.

• Alternatives lack strong support. Other pharmacologic strategies — including pentoxifylline-based regimens — have not demonstrated consistent survival benefit.

• Mortality remains high. Severe alcohol-associated hepatitis still carries up to 50% mortality within 90 days, highlighting the need for better therapies and earlier intervention.

• More high-quality trials needed. The review underscores significant gaps in evidence and calls for better-designed studies to guide treatment decisions.