Editorial 

May 9, 2023

The Role of Prophylactic Antibiotics for Patients With Severe Alcohol-Related Hepatitis

Ewan Forrest, William Bernal

JAMA. 2023;329(18):1552-1553. doi:10.1001/jama.2023.1729

Few effective pharmacologic therapies are available for alcohol-related hepatitis. In patients diagnosed with alcohol-related hepatitis, greater severity of liver injury and systemic inflammation are associated independently with early mortality.¹ Three months after a diagnosis of alcohol-related hepatitis, patients who continue to use alcohol had poorer prognosis than those who abstained (no alcohol relapse, 66 deaths in 8451 patient-months observed; relapse of ≥30 g/d, 55 deaths in 2654 patient-months; hazard ratio, 2.89; P < .001).²

Corticosteroids have been used to treat alcohol-related hepatitis for more than 50 years, but their role in the treatment of alcohol-related hepatitis remains controversial. In a meta-analysis of 5 randomized clinical trials of patients with alcohol-related hepatitis, compared with placebo, corticosteroids were associated with lower mortality at 28-day follow-up (81/490 [16.5%)] vs 111/466 [23.8%]; hazard ratio, 0.64; P = .003), but in a meta-analysis that included 4 randomized clinical trials, compared with placebo, corticosteroids were not significantly associated with lower mortality at 6-month follow-up (170/455 [37.3%] vs 157/435 [36.0%]; hazard ratio, 0.99; P = .93).³ In randomized clinical trials, interleukin 1 antagonists such as anakinra and granulocyte colony-stimulating factor were not associated with improved outcomes.^4,5 Other than corticosteroids, treatment of alcohol-related hepatitis consists primarily of supportive care, managing alcohol use disorder, and consideration of liver transplant when appropriate.

Alcohol-related hepatitis is a complex inflammatory state characterized by high-grade systemic inflammation, which is associated with liver tissue injury, portal hypertension, and extrahepatic organ dysfunction, especially acute kidney injury. Alcohol-related hepatitis is also associated with immunosuppression. Patients with alcohol-related hepatitis have impaired monocyte and neutrophil phagocytic and oxidative burst capacity, defined as rapid release of reactive oxygen species from different cell types, and these immune system abnormalities are associated with increased rates of infection.⁶ Patients with alcohol-related hepatitis have a higher rate of bacterial infection compared with those with decompensated alcohol-related cirrhosis who do not have alcohol-related hepatitis.⁷ Elevated biomarkers of infection such as serum lipopolysaccharide, bacterial DNA, and procalcitonin are associated with increased risk of multiorgan failure and higher mortality in alcohol-related hepatitis, suggesting that infection may contribute to increased rates of mortality.^1,8 Patients with alcohol-related hepatitis who are treated with corticosteroids have higher rates of infection compared with those not treated with corticosteroids, especially during the first week of treatment.⁹ Among 547 patients with alcohol-related hepatitis treated with corticosteroids, the presence of infection during corticosteroid treatment was associated with a higher rate of mortality compared with those without infection (39% vs 22%; odds ratio, 2.27; 95% CI, 1.52-3.38), but this association was not observed in patients with alcohol-related hepatitis who were not treated with corticosteroids.⁸ Therefore, treating patients who have alcohol-related hepatitis with antibiotics in combination with corticosteroids has the potential to prevent infection and associated adverse outcomes, while allowing the anti-inflammatory actions of the corticosteroids to take effect.

In this issue of JAMA, Louvet et al¹⁰ report results of a multicenter randomized clinical trial that compared amoxicillin-clavulanate combined with prednisolone with placebo combined with prednisolone in 292 patients with biopsy-confirmed “definite” severe alcohol-related hepatitis. There was no significant difference in the primary outcome of mortality at 60-day follow-up between those randomized to amoxicillin-clavulanate combined with prednisolone and those randomized to placebo combined with prednisolone (17.3% in the amoxicillin-clavulanate group and 21.3% in the placebo group [P = .33]; between-group difference, −4.7% [95% CI, −14.0% to 4.7%]; hazard ratio, 0.77 [95% CI, 0.45-1.31]). Similarly, there was no significant effect of amoxicillin-clavulanate combined with prednisolone, compared with placebo combined with prednisolone, on the incidence of hepatorenal syndrome at 60-day follow-up (7.7% vs 8.5%; P = .81). However, the combination of antibiotics and prednisolone significantly reduced the incidence of infection during the first month of antibiotic prescription (13.4% vs 28.1%; P = .001). The finding that antibiotics significantly reduced infection rates without improving liver function or preventing complications of liver failure is consistent with the hypothesis that sepsis is a consequence of severe alcohol-related hepatitis and/or its treatment, rather than the cause of worsening liver function.

The study also clarified the significance of baseline infection treated prior to the initiation of corticosteroid therapy. In the current study, participants were required to have had any bacterial infection “cured” by antibiotics for at least 7 days before entering the trial. Earlier observations from Louvet and colleagues¹¹ suggested that treated baseline infection did not affect the predefined response to corticosteroids, measured by improvement in the Lille score (comprising patient age and laboratory parameters of plasma albumin, change in bilirubin over 7 days, creatinine, and prothrombin time), and did not affect survival, but rather suggested that new infections that occurred after beginning corticosteroids influenced treatment response and increased rates of adverse outcomes. However, data from the STOPAH trial, the largest randomized clinical trial in severe alcohol-related hepatitis, suggested an interaction between antibiotic use and subsequent outcome in corticosteroid-treated patients who had infection at baseline.⁹ Specifically, patients whose antibiotic treatment of baseline infection overlapped with the initiation of corticosteroids had significantly improved survival compared with those whose antibiotics were discontinued prior to corticosteroid use.^8,9

In the current trial, there was a suggestion of a role for antibiotics in those with baseline infection who are subsequently treated with corticosteroids.¹⁰ Although not statistically significant, the subgroup analysis of those with infection prior to randomization showed a higher 2-month survival in those treated with the combination of antibiotics and corticosteroids (83.9% vs 65.0%; hazard ratio, 0.40; 95% CI, 0.13-1.16). Considered together with the STOPAH observations, this could suggest a clinically relevant role for the overlap of antibiotic treatment for baseline infection with the initiation of corticosteroids for ongoing alcohol-related hepatitis.

The current clinical trial by Louvet et al suggests no role for prophylactic prescription of antibiotics when treating all patients who have alcohol-related hepatitis with corticosteroids.¹⁰ This finding was similar to observations from a post hoc analysis of hospitalized patients with decompensated cirrhosis in the ATTIRE clinical trial, many of whom had alcohol-related liver disease.¹² In this study, prescription of antibiotics in 777 patients hospitalized with decompensated cirrhosis without documented infection did not improve survival compared with those who did not receive antibiotics.¹² For patients with alcohol-related hepatitis, developing infection while taking corticosteroids is an indicator of poor prognosis and should be treated.¹³ This may be particularly relevant for patients without a Lille score biochemical response to corticosteroids, since these patients have an increased rate of infection compared with those with a biochemical response (42.5% vs 11.1% in responders; P < .001).¹¹ Guidelines recommend discontinuation of corticosteroids for patients who do not show biochemical response to treatment as reflected by the Lille score,¹³ but close monitoring for infection and early institution of antibiotic therapy for documented infection should continue, as the risk of infection, especially lung infection, remains high even after corticosteroid therapy is discontinued.¹⁴ In addition, clinicians should evaluate all patients with alcohol-related hepatitis for infection, and treat identified infections, prior to initiating corticosteroids as recommended in the American College of Gastroenterology guidelines for the management of alcohol-related liver disease.¹³ If there are features of active ongoing alcohol-related hepatitis after treatment of baseline infection, then overlapping the antibiotic treatment with subsequent corticosteroid therapy could be considered.