Introduction There are limited data on outcome of AIH outside large centres. We have conducted an outcome audit of AIH in 28 UK centres of varying size. We report overall outcome results in this cohort.

Methods We attempted to collect all prevalent cases since 2000 and all incident cases since 2007 by searching electronic patient letters, histology databases and hospital coding. Validation was by 1999 IAIHG diagnostic criteria. Information, on diagnosis, initial severity, treatment and outcome was entered into a web-based data collection system.

Results There were 1267 patients (80% female, age at diagnosis (median (range) 55 years (8–86)). IAIHG criteria were not met in 79 (6%) but they were still treated as AIH. At diagnosis, cirrhosis (biopsy, varices, ascites or Fibroscan) was present in 317 (25%) patients and decompensation (ascites, oedema, encephalopathy, variceal bleed or MELD score >15) at diagnosis was present in 282 (22%). Follow-up (time from diagnosis to last seen) was 4 (0–14) years. AIH relapse occurred in 371 patients (141 had ≥2 and 57 ≥ 3 relapses). De-novo cirrhosis developed in 78 (6%) patients, decompensation in 42, and hepatocellular cancer in 7. Fifteen patients died of liver disease, 28 had a liver transplant and 31 died of liver-unrelated causes.

Five- and 10 year death/transplant rates were, respectively 6.4±0.1% and 10.0±0.1% (all-cause) and 3.6±0.1% and 5.8±0.1% (liver related). Factors associated with death/transplant rate by Cox regression analysis were (a) cirrhosis, and decompensation at diagnosis (p < 0.001 both all-cause and liver related) (b) age and cardiovascular/respiratory comorbidity (all-cause only, p = 0.004 and 0.014) (c) failure to normalise serum transaminases (p < 0.001); however because of “early” deaths/transplants, direction of cause and effect is unclear. In contrast, death/transplant rate showed no association with failure to achieve ALT normalisation at specific time points (over 1–12 months). Nor was it associated with gender, peak initial serum ALT, number of relapses or concurrent presence of PBC “overlap”, diabetes or cancer.

Both all-cause (p-0.05) and liver-related (p = 0.021) death/transplant were lower in patients attending District General Hospitals (DGHs), compared to University Hospitals; these differences persisted in multivariate analysis (p = 0.017 and 0.036), adjusting for the above baseline predictive factors.

Conclusion In this multicentre UK AIH Audit, disease outcome is similar to that in patients reported from large single centres. Main determinants are disease severity, age and comorbidity at presentation. Normalisation of serum ALT at specific time points may not be an accurate longer term outcome marker and better ones are needed. Outcome of AIH in UK patients attending DGHs seems at least as good as those attending University hospitals.

Disclosure of Interest None Declared