BACKGROUND: Previous studies have described the clinical impact of infection in alcoholic hepatitis (AH) but none have comprehensively explored the aetiopathogenesis of infection in this setting. We examined the causes, consequences and treatment of infection in a cohort of patients with AH.

METHODS: We undertook a retrospective cohort study of patients with AH admitted between 2009 and 2014 to seven centres in Europe and the USA. Clinical and microbiological data were extracted from medical records. Survival was analysed with Kaplan-Meier analysis and Cox proportional hazards analysis to control the data for competing factors. Propensity score matching was used to examine the efficacy of prophylactic antibiotics administered in the absence of infection.

RESULTS: We identified 404 patients with AH. Of these, 199 (49%) showed clinical or culture evidence of infection. Gut commensal bacteria, particularly Escherichia coli and Enterobacter species, were most commonly isolated in culture. Fungal infection was rarely seen. Cultured organisms and antibiotic resistance differed markedly between centres. Infection was an independent risk factor for death (hazard ratio for death at 90 days 2.33, 95% confidence interval 1.63-3.35, p < 0.001). Initiation of antibiotic therapy on admission in the absence of infection did not reduce mortality or alter the incidence of subsequent infections. Corticosteroid use increased the incidence of infection but this did not impact on survival.

CONCLUSIONS: In this large real-world cohort of patients with AH, infection was common and was associated with reduced short-term survival. Gram-negative, gut commensal bacteria were the predominant infective organisms, consistent with increased translocation of gut bacteria in AH; however, the characteristics of infection differ between centres. Infection should be actively sought and treated, but we saw no benefits of prophylactic antibiotics.