Lille score at Day 7 (LM7) helps to predict the outcome of patients with severe alcoholic hepatitis (sAH) undergoing corticotherapy. Several scores such as Maddrey's discriminant function (MDF), MELD, ABIC, and GAHS are used for a 28-day mortality prognosis. Our study aimed to evaluate if the assessment of the Lille score at 4 days (LM4) is as useful as the Lille score at Day 7 (LM7) to predict response to corticosteroids and 28-day mortality and evaluate the utility of severity scores at admission for predicting the prognosis of patients with liver cirrhosis (LC) and severe alcoholic hepatitis (sAH). A retrospective study was performed, and all consecutive patients with AH and MDF > 32 without contraindications to corticosteroids were included. Prognostic scores were evaluated at admission, and 28-day mortality was assessed. Response to corticotherapy was assessed by LM4 and LM7. Results: A total of 55/103 patients with sAH (51.5%) had MDF > 32 and received corticosteroids. There was no difference between the proportion of patients with a responder LM4 versus LM7 (27% vs. 36%, p = 0.31). The mean value for LM4 was 0.64 ± 0.3 versus 0.60 ± 0.3 for LM7 (p = 0.48). Precisely 90.3% of patients were correctly identified as responders or not by LM4 compared with LM7. The best model for predicting 28-day mortality was composed of MELD and LM4/LM7, with an accuracy of 0.90 for both combinations. Conclusion: LM4 could be used instead of LM7 for predicting response to corticosteroid therapy in patients with sAH and LC, as well as 28-day mortality. Using LM4, we could avoid prolonged use of this therapy and its complications.
Keywords: Lille score; alcoholic hepatitis; alcoholic liver cirrhosis; corticosteroid therapy; prognostic scores.