How is the Lille score calculated on day 7 of prednisolone therapy for severe alcoholic hepatitis, and what cutoff determines whether to continue or stop steroids?

Lille Score: Calculation and Clinical Application in Severe Alcoholic Hepatitis

The Lille score must be calculated on day 7 of prednisolone therapy using a specific formula incorporating six variables, and steroids should be discontinued if the score is ≥0.56, as these null responders have only 53.3% 28-day survival compared to 91.1% in complete responders. 1, 2

Formula and Calculation

The Lille score is calculated using the following formula 2:

R = 3.19 – 0.101 × age (years) + 0.147 × albumin day 0 (g/L) + 0.0165 × (bilirubin day 0 – bilirubin day 7 in μmol/L) – 0.206 × renal insufficiency – 0.0065 × bilirubin day 0 (μmol/L) – 0.0096 × prothrombin time (seconds)

Then: Lille score = exp(–R) / [1 + exp(–R)] 2

Required Variables:

• Age in years (coefficient: -0.101) 2
• Albumin at day 0 in g/L (coefficient: +0.147) 2
• Bilirubin at day 0 in μmol/L (coefficient: -0.0065) 2
• Bilirubin at day 7 in μmol/L (used to calculate evolution: day 0 minus day 7, coefficient: +0.0165) 2
• Prothrombin time in seconds (coefficient: -0.0096) 2
• Renal insufficiency (coded as 0 if creatinine <1.3 mg/dL or 1 if ≥1.3 mg/dL, coefficient: -0.206) 2

Critical Cutoff Values and Response Categories

The score ranges from 0 to 1 and stratifies patients into three distinct prognostic groups 1, 2:

• Complete responders: Lille score ≤0.16 → 91.1% 28-day survival 1, 2
• Partial responders: Lille score 0.16-0.56 → 79.4% 28-day survival 1, 2
• Null responders: Lille score ≥0.56 → 53.3% 28-day survival 1, 2

An older threshold of 0.45 has also been validated, showing 6-month survival of 85% for scores <0.45 versus only 25% for scores >0.45 1.

Clinical Decision Algorithm

For Lille Score ≥0.56 (Null Responders):

Stop prednisolone immediately. 1, 3 These patients have significantly worse outcomes with continued steroids, and infection occurs more frequently in non-responders. 1 Consider alternative rescue therapies such as liver transplantation, as survival may be significantly higher with transplantation than continuing medical treatment alone. 1

For Lille Score 0.45-0.56 (Borderline Zone):

The decision requires individualized assessment based on 3:

• Presence of infection
• Trajectory of clinical deterioration
• Availability of liver transplantation

However, the guideline-recommended cutoff of ≥0.56 provides clearer decision-making, and most guidelines support stopping steroids at this threshold. 1

For Lille Score <0.45 (Responders):

Continue prednisolone 40 mg daily for an additional 21 days with or without tapering. 1 These patients demonstrate clear treatment benefit with 6-month survival of 85%. 1

Timing Considerations

Standard assessment is at day 7 of corticosteroid treatment. 1, 2, 3 However, emerging evidence suggests the Lille score can be calculated at day 4 with similar prognostic performance, potentially allowing earlier identification of non-responders and reducing complications from prolonged steroid exposure. 2, 4 One study showed 90.3% of patients were correctly identified as responders or non-responders by day 4 compared to day 7. 4

Integration with Other Prognostic Tools

Combining the Lille score with MELD provides optimal risk stratification. 2, 3 The MELD score is used for baseline severity assessment and longer-term prognostication, while the Lille score guides immediate treatment decisions at day 7. 3 The best model for predicting 28-day mortality combines MELD with Lille score (either day 4 or day 7), achieving accuracy of 0.90. 4

Common Pitfalls and Caveats

• Do not use the Lille score at baseline – it requires day 7 bilirubin to assess treatment response. 2
• Ensure units are correct: bilirubin must be in μmol/L (multiply mg/dL by 17.1 to convert), albumin in g/L, and prothrombin time in seconds. 2
• The score has excellent sensitivity (95%) but poor specificity (19%) for predicting 30-day survival, meaning it is better at identifying responders than non-responders. 5
• Baseline prothrombin time may predict steroid response – higher baseline PT is associated with non-response (OR = 2.954), potentially allowing earlier identification of futile therapy. 6