Alcoholic Hepatitis Lille Score

Lille model explained

The Lille model stratifies mortality risk in alcoholic hepatitis patients who don’t respond to steroid treatment.

It is a model based on parameters tested at admission and on subsequent bilirubin determinations. The variables used in the model are explained in the table below:

The formula used in the Lille model accounts for the above variables by awarding them different weights, depending on the individual predictor value.

Lille Model Score = Exp(-R) / (1 + Exp(-R))

Where:

R = [3.19 - (0.101 x Age in years)] + (1.47 x Albumin in g/dL) + [0.28215 x (Bilirubin initial - Bilirubin day 7 in mg/dL)] - (0.206 x Creatinine in mg/dL) - (0.11115 x Bilirubin initial in mg/dL) - (0.0096 x Prothrombin time in seconds)

Alcoholic hepatitis prognosis

The Lille model provides an outcome prognostic in patients with alcoholic hepatitis that is unresponsive to corticosteroids at around 7 days since the therapy has begun.

The model is often used in association with the Maddrey score (for scores above 32), when the administration of treatment is urgent.

The subsequent bilirubin determinations can indicate whether poor prognosis is likely. The Lille model provides a cut off at 0.45:

■ Scores below 0.45 are consistent with a 6-month survival rate of 85%;

■ Scores above 0.45 indicate a 6-month survival rate of 25%.

If the patient scores above 0.45, alternative therapies should be considered, as well as referring the patient for a transplant waiting list.

There are other prognosis models for hepatic disease, such as the Model for End-Stage Liver Disease (MELD) or the Child-Pugh severity score.

About the study

The Lille model was created by Louvet et al. in 2007 and was aimed at early identification of patients with severe alcoholic hepatitis, not responding to corticosteroids.

The evolution of 320 patients in the development cohort and of 118 patients in the validation program, was supervised. The model was generated by logistic regression and ended up combining six reproducible variables, that were found statistically relevant for 6 month prognosis.

The area under the receiver operating characteristic (AUROC) curve of the model was found to be higher than that of other models (Maddrey and Child-Pugh) in the derivation cohort whilst in the validation cohort, Lille AUROC was higher than that of MELD.

The application of the Lille model can improve management of patients with severe alcoholic hepatitis.

Original source

Louvet A, Naveau S, Abdelnour M, Ramond MJ et al. The Lille model: a new tool for therapeutic strategy in patients with severe alcoholic hepatitis treated with steroids. Hepatology. 2007; 45(6):1348-54.

Validation

Sandahl TD, Jepsen P, Ott P, Vilstrup H. Validation of prognostic scores for clinical use in patients with alcoholic hepatitis. Scand J Gastroenterol. 2011; 46(9):1127-32.

Other references

1. Soultati AS, Dourakis SP, Alexopoulou A, Deutsch M, Vasilieva L, Archimandritis AJ. Predicting utility of a model for end stage liver disease in alcoholic liver disease. World J Gastroenterol. 2006; 12(25):4020-5.

2. Louvet A et al. Combining Data From Liver Disease Scoring Systems Better Predicts Outcomes of Patients With Alcoholic Hepatitis. Gastroenterology. 2015; 149(2):398-406.e8;

3. Papastergiou V, Burroughs AK, Tsochatzis EA. Prognosis and treatment of patients with acute alcoholic hepatitis. Expert Rev Gastroenterol Hepatol. 2014; 8(5):471-86.