Lille Score Calculator for Alcohol-Associated Hepatitis
Calculate the adult Day-4 or Day-7 Lille score after corticosteroid therapy from baseline variables and bilirubin evolution, with exact formula provenance and a threshold-safe 0.45 audit.
Content updated: View sources
QuickMedCalc is developed and maintained by an independent developer. Medical content is not independently reviewed by a physician.
Formula and medical content are based on the references listed on this page. See sources, About, and Sources and Review Process.
Result
About
This Lille Score Calculator assesses an on-treatment response pattern in an adult with severe alcohol-associated hepatitis (historically called alcoholic hepatitis) who is already receiving corticosteroid therapy. Select Day 7 for the original Louvet model or Day 4 for the separately validated early assessment. It is not a baseline diagnostic or steroid-eligibility calculator. [1, 2, 5]
The primary source grouping is Lille <0.45 versus ≥0.45. The score does not independently establish diagnosis, treatment benefit, a patient-specific survival probability, or a medication decision. [1, 5]
Formula
Interpretation
| Exact raw Lille score | Source group | Boundary |
|---|---|---|
| <0.45 | Responder range | On-treatment cohort context; not an instruction to continue, dose, or taper corticosteroids. |
| ≥0.45 | Nonresponder range | Used in current reassessment guidance; confirm timing, inputs, diagnosis, contraindications, complications, and the complete clinical record before treatment decisions. |
The original full paper grouped scores below 0.45 separately from scores at or above 0.45. Current ACG prose states “greater than 0.45”; this calculator preserves the original model's exact </≥ boundary and explains rather than hides the wording difference. [1, 5]
Day 7 original Lille model and Day 4 validation
Louvet and colleagues developed the model from an exploratory cohort of 320 corticosteroid-treated patients and prospectively validated it in 118 patients; the original dynamic measurement is Day 7 bilirubin. The 2017 multinational study evaluated Day 4 using the same formula with Day-4 bilirubin and found similar discrimination to Day 7; its published corrigendum remains linked. Day 4 and Day 7 are therefore distinct provenance choices, not interchangeable labels for an arbitrary follow-up date. [1, 2, 3]
Bilirubin evolution direction
The evolution term is day-0 bilirubin minus Day-4 or Day-7 bilirubin after both values are normalized to µmol/L. A fall is positive; a rise remains negative. The calculator does not reverse the subtraction, take an absolute value, or round either normalized operand before computing R. [1, 2]
Renal-insufficiency flag
The source model uses a binary flag—not a continuous creatinine coefficient. In creatinine-derived mode, the submitted reporting system is compared directly: above 1.3 mg/dL or above 115 µmol/L gives 1; exactly either threshold gives 0. These two published thresholds are not presented as exact unit conversions. Documented mode is for a source-compatible record that already establishes the criterion, including the original creatinine-clearance pathway; this page does not calculate CrCl or substitute eGFR. [1]
PT and INR are source variants, not conversions
The original article published a PT-seconds formula and a separate INR formula using the same −0.0096 coefficient, reporting similar validation discrimination. Choose the value actually used by the record. The calculator never converts PT to INR or INR to PT, and control PT is not an input to Lille. [1]
Lille, Maddrey mDF, and MELD answer different questions
| Model | Timing | Responsibility |
|---|---|---|
| Modified Maddrey DF | Baseline/presentation | Historical PT-control and bilirubin severity context. |
| MELD | Baseline/current model-specific assessment | Version-specific liver severity/prognostic or allocation context. |
| Lille | Day 4 or Day 7 after corticosteroids begin | Dynamic early response assessment using bilirubin change and baseline variables. |
Use the separate Maddrey Discriminant Function Calculator or MELD Score Calculator for those different tasks. Lille does not diagnose alcohol-associated hepatitis or determine who should start corticosteroids. [5, 6, 7]
Secondary historical response strata
| Historical Lille range | Published label | 28-day survival in that meta-analysis |
|---|---|---|
| ≤0.16 | Complete response | 91.1% |
| >0.16 and <0.56 | Partial response | 79.4% |
| ≥0.56 | Null response | 53.3% |
These are secondary historical cohort strata from an individual-patient meta-analysis, not the primary 0.45 grouping, a current universal prognosis, or a patient-specific probability. They are not applied dynamically in the result. [4]
Current-guidance and treatment boundary
ACG 2024 recognizes Day 4 or Day 7 Lille assessment and discusses discontinuation for nonresponse, but an online calculation cannot verify the diagnosis, treatment course, data timing, eligibility, contraindications, infection or bleeding status, organ failures, goals, or local protocol. QuickMedCalc does not choose, start, continue, taper, dose, or stop a medicine and does not determine transplant referral or eligibility. [5, 6, 7]
Historical outcome context
In the original treated validation evidence, six-month survival was approximately 85% for Lille <0.45 and 25% for Lille ≥0.45. In the Day-4 study, 28-day survival was 90% versus 66% and 90-day survival was 76% versus 40% for responder and nonresponder groups. These estimates belong to those cohorts and periods; they are not individualized forecasts and are not reused across Day 4 and Day 7. [1, 2]
References
- Louvet A, Naveau S, Abdelnour M, et al. The Lille model: a new tool for therapeutic strategy in patients with severe alcoholic hepatitis treated with steroids. Hepatology. 2007;45(6):1348–1354. PMID 17518367. DOI 10.1002/hep.21607.
- Garcia-Saenz-de-Sicilia M, Duvoor C, Altamirano J, et al. A Day-4 Lille Model Predicts Response to Corticosteroids and Mortality in Severe Alcoholic Hepatitis. Am J Gastroenterol. 2017;112(2):306–315. PMID 27922027. DOI 10.1038/ajg.2016.539.
- Garcia-Saenz-de-Sicilia M, Duvoor C, Altamirano J, et al. Corrigendum: A Day-4 Lille Model Predicts Response to Corticosteroids and Mortality in Severe Alcoholic Hepatitis. Am J Gastroenterol. 2017;112(4):666. PMID 28381859. DOI 10.1038/ajg.2017.76.
- Mathurin P, O'Grady J, Carithers RL, et al. Corticosteroids improve short-term survival in patients with severe alcoholic hepatitis: meta-analysis of individual patient data. Gut. 2011;60(2):255–260. PMID 20940288. DOI 10.1136/gut.2010.224097.
- Jophlin LL, Singal AK, Bataller R, et al. ACG Clinical Guideline: Alcohol-Associated Liver Disease. Am J Gastroenterol. 2024;119(1):30–54. PMID 38174913. PMCID PMC11040545. DOI 10.14309/ajg.0000000000002572.
- Crabb DW, Im GY, Szabo G, Mellinger JL, Lucey MR. Diagnosis and Treatment of Alcohol-Associated Liver Diseases: 2019 Practice Guidance From the American Association for the Study of Liver Diseases. Hepatology. 2020;71(1):306–333. PMID 31314133. DOI 10.1002/hep.30866.
- American Association for the Study of Liver Diseases. Why do we use steroids, Maddrey's Discriminant Function, and the Lille score in Alcohol-Associated Hepatitis? Liver Fellow Network. Reviewed September 8, 2026.
FAQ
The Lille model is an on-treatment response and prognostic model for adults with severe alcohol-associated hepatitis who are already receiving corticosteroid therapy. It combines baseline variables with bilirubin evolution at Day 7 in the original model or Day 4 in a separately validated early assessment; it does not diagnose alcohol-associated hepatitis or determine who should start treatment.
The source linear predictor is R = 3.19 − 0.101×age + 0.147×day-0 albumin in g/L + 0.0165×(day-0 minus follow-up bilirubin in µmol/L) − 0.206×binary renal insufficiency − 0.0065×day-0 bilirubin in µmol/L − 0.0096×day-0 PT in seconds or INR. Lille = exp(−R)/[1+exp(−R)].
Sources: [1]
Day 7 is the original Louvet model timing. Day 4 was evaluated and validated as an earlier assessment using the same formula with Day-4 bilirubin replacing Day-7 bilirubin. Select the actual intended time point from the same corticosteroid course; the calculator does not accept an arbitrary follow-up day.
An exact raw score below 0.45 falls in the original responder range. This is source-defined cohort context, not proof of benefit for one person and not an instruction to continue, dose, or taper corticosteroids. Diagnosis, timing, complications, contraindications, goals, and the complete clinical record still matter.
The original full paper grouped an exact raw score of 0.45 or greater as nonresponse. Current ACG guidance uses Lille reassessment in treatment decisions, but the calculator cannot verify the diagnosis, course, contraindications, infections, bleeding, organ failures, goals, or local protocol and does not tell a clinician to stop a medicine.
The original model uses a binary 0/1 renal-insufficiency variable. Serum creatinine above 1.3 mg/dL or above 115 µmol/L gives 1; exactly either threshold does not. The source also recognized creatinine clearance below 40 mL/min. QuickMedCalc can use an already documented source-compatible criterion but does not calculate CrCl, infer eGFR, or use creatinine as a continuous formula term.
Sources: [1]
Yes, as a distinct source-published variant. Louvet 2007 provided an INR formula using the same −0.0096 coefficient and reported similar discrimination to the PT-seconds formula. The two inputs are never converted into one another, and Lille does not use a laboratory control PT.
Sources: [1]
Modified Maddrey DF and MELD are baseline severity-model tasks with different inputs and model identities. Lille is a dynamic response model after corticosteroid therapy and includes paired bilirubin change at Day 4 or Day 7. The three scores are not interchangeable, and none independently establishes the diagnosis or a medication decision.
No. The 2007 Day-7 and 2017 Day-4 studies reported outcomes for particular treated cohorts, and later work described additional historical response strata. Those estimates are not personal forecasts. This calculator does not choose corticosteroid eligibility, medicine, dose, duration, discontinuation, transplant evaluation, or other treatment.
Related Calculators
Maddrey mDF
Calculate the modified Maddrey Discriminant Function from patient PT, the corresponding laboratory control PT, and total bilirubin, with exact threshold and current MELD/Lille context.
MELD 3.0
Calculate current OPTN MELD 3.0 for age 12+ or reproduce historical MELD-Na and classic MELD with version-specific inputs and limits.
Disclaimer
Educational and informational reference only. Not intended to replace professional medical advice, diagnosis, treatment, or independent verification.