DAY-4 LILLE SCORE AS PREDICTOR OF RESPONSE TO CORTICOSTEROIDS AND SHORT TERM MORTALITY IN SEVERE ALCOHOLIC HEPATITIS.

Background: Alcoholic hepatitis (AH) is a clinical condition with high mortality rates. Maddrey's Discriminant function (MDF) is a commonly used prognostic score for patients with AH with a DF score ≥ 32 is classied as severe alcoholic hepatitis (sAH). Corticosteroids is still the mainstay of treatment for Patients with sAH, demonstrating modest improvement in survival at 28 days corticosteroid therapy has been associated with increased risk of infections in patients with sAH. We evaluated whether the use of the Lille Model at day 4 (LM4) is useful to predict response to corticosteroids and short term mortality at 28 days as compared with conventional Lille score at day 7 (LM7). A Materials And Method Prospective observational study was performed, and all consecutive patients with AH admitted over a one year period were observed. Relevant lab data was utilised to assess baseline severity assessment scores (MDF,MELD,ABIC.GASH) . Patients with sAH as evidenced by MDF > 32 without contraindications to corticosteroids were treated as per standard protocol with corticosteroids . Patients receiving corticosteroid therapy were assessed with LM 4 scores at day 4 of treatment and LM7 Scores at day 7 of treatment , all patients receiving steroids where continued on therapy till day 7 before a decision on withdrawal if LM7 >0.45,28-day mortality was assessed in all patients . Prediction of response to corticosteroid therapy was assessed by LM4 and LM7 scores and compared. A total of 62 patients w Results ere admitted with AH during the study period of I year, 48 (77.4%) of these patients had SAH as evidenced by MDF >32, of the patients with sAH 34/48 (70.8 %) were eligible for steroid therapy. The short term mortality at 28 days was observed in 6 /62 (0.96%) patients, all of whom had sAH and 2 were non responders to corticosteroid therapy and 4 were ineligible to steroid therapy. There was no difference between the proportion of patients with a responder LM4 versus LM7 (28% vs. 35%, p = 0.30). 92.0% of patients were correctly identied as responders or not by LM4 compared with LM7. LM4 has comparable accuracy Conclusion with LM7 in predicting response to corticosteroids, and 28- mortality in patients with sAH. Assessing the efcacy of corticosteroids at an earlier time point using LM4 scores may help avoid a more prolonged futile use of steroid therapy and help reduce complications associated with prolonged steroid use in sAH.