Introduction Current assumption in the published literature is that AAH is not dissimilar to other forms of ACLF; however, the clinical syndrome of AAH is unique, characterised by profound jaundice and immune dysfunction. Therefore, the outcomes of patients requiring organ support in this setting may differ from other forms of ACLD.

Aim To determine whether the clinical outcomes of AAH patients with multi-organ failure admitted to ITU differ from those with other forms of ACLD.

Method Single-centre retrospective study of consecutive patients admitted to ITU with AAH (AAH ITU) between 10/2014 and 07/2017. Two comparator cohorts were identified - patients with AAH hospitalised but not requiring ITU (non-ITU AAH); and patients with non-AAH ACLD admitted to ITU (non-AAH ITU). The diagnosis of severe AAH was made prospectively adhering to the STOPAH trial criteria, and confirmed retrospectively by two independent Hepatologists; 37% of AAH patients had the diagnosis confirmed histologically.

Results 62 patients were diagnosed with severe AAH during the study period – at the time of hospital admission the median bilirubin was 319μmol/l and 63% had a Glasgow Alcoholic Hepatitis Score ≥9. 21/62 patients were admitted to ITU. AAH ITU patients were more likely to have CLIF-C ACLF (100% vs 80%, p=0.017), but had a similar SOFA score (p=0.064) and total number of organ supports (2 vs 2, p=0.447) to non-AAH ITU patients (n=70).

The 90-day survival was 29% for the AAH ITU patients, compared with 90% and 60% for non-ITU AAH and non-AAH ITU patients, respectively (p<0.001).

Overall, 15% of AAH ITU and 56% of non-AAH ITU patients who received any organ support survived to hospital discharge. Of the AAH ITU patients with a CLIF-C ACLF grade of 3, 1 patient (8%) survived to discharge, compared with 8/23 (35%) non-AAH ITU patients. Of the AAH ITU patients who required 3 organ support (n=8), none survived to hospital discharge, compared with 6/18 (33%) non-AAH ITU patients.

Conclusion Patients with AAH admitted to ITU do comparatively worse than non AAH patients receiving a similar level of organ support; and prognostic variables such as the CLIF-C ACLF score may not be as discriminatory in this cohort. As such, ceilings of care should be considered carefully and on an individual case basis.