OpiCalc Logo

OpiCalc

989 Clinical Tools

Logo
OpiCalc
Child-PughFIB-4 IndexFibroScan E-RatioLille ModelMELD ScoreMaddrey's DF
OpiCalc Logo

OpiCalc

Open-access clinical infrastructure. Built to the standard every clinician deserves — fast, private, and free.

Zero data stored
Always free
Our mission & transparency

Get in Touch

Tool request, clinical feedback, or partnership inquiry — we read everything.

WhatsApp feedback
Email us
Partnership inquiry

© 2026 OpiCalc • Calculated Care

ProtocolsAboutPrivacyTerms

Maddrey DF

Maddrey's Discriminant Function: Severity score for Alcoholic Hepatitis. DF ≥ 32 defines severe disease and guides corticosteroid therapy decision. Sensitivity ~79%.
Guidelines & Evidence

Clinical Details

Section 1

When to Use

When to Use

Active alcoholic hepatitis with jaundice to assess severity
Determine eligibility for corticosteroid therapy
Identify patients requiring ICU-level monitoring
Prognosticate 30-day mortality in active drinkers presenting with liver failure

Required Clinical Context

Must have confirmed or highly suspected alcoholic hepatitis (jaundice + recent heavy alcohol use + AST:ALT > 2:1 + excluding other causes). DF is not meaningful in non-alcoholic liver disease.
Section 2

Formula & Logic

Formula

DF = 4.6 × (PT patient − PT control) + Serum Bilirubin (mg/dL)

Interpretation

01
DF < 32: Non-severe AH. Supportive management. No steroids.
02
DF ≥ 32: Severe AH. 30-day mortality 35–45%. Consider prednisolone 40mg/day × 28 days if no contraindications.

Bilirubin Units

The original Maddrey formula uses bilirubin in mg/dL. If your lab reports in µmol/L, divide by 17.1 before entering the value.
Section 3

Pearls/Pitfalls

Lille Model — Day 7 Assessment

After 7 days of prednisolone, calculate the Lille Model score
Lille score < 0.45: Steroid responder — complete 28-day course
Lille score ≥ 0.45: Non-responder — discontinue steroids (no benefit, continued infection risk)

Contraindications to Steroids

Active GI bleeding
Active untreated bacterial infection (treat infection first, then reassess)
Renal failure (creatinine > 2.5 mg/dL — consider pentoxifylline)
Uncontrolled hepatic encephalopathy or coma
HBV or HCV co-infection (risk viral reactivation)

MELD vs DF

Both Maddrey DF and MELD score predict 30-day mortality in AH. MELD > 20 correlates with DF ≥ 32. Some centres use Glasgow AH Score (GAHS) or ABIC score as additional severity tools.
Section 4

Next Steps

Severe AH Management (DF ≥ 32)

01
Prednisolone 40mg orally daily × 28 days (preferred over methylprednisolone)
02
Nutritional support: EN via NGT if oral intake inadequate (target 35–40 kcal/kg/day, 1.2–1.5g protein/kg/day)
03
Thiamine 300mg IV daily × 3 days, then oral B-vitamins
04
Lille Model on Day 7: if ≥ 0.45, stop steroids
05
Prophylactic PPI for GI bleeding risk
06
Transplant evaluation in selected non-responders

Complementary Tools

MELD Score
Child-Pugh Score
Glasgow-Blatchford Score
Section 5

Evidence Appraisal

Original Study

Corticosteroid therapy of alcoholic hepatitis.

Maddrey WC et al. • Gastroenterology.. 1978;75(2):193–199. DF ≥ 93 originally predicted benefit; revised to ≥ 32 in subsequent validation studies.

Section 6

Literature

Dr. Willis C. Maddrey

A hepatologist who published the discriminant function in 1978 as part of the first RCT of corticosteroids in alcoholic hepatitis. His formula — combining PT prolongation with bilirubin — elegantly captured both synthetic dysfunction and cholestasis in a single score, remaining the entry criterion for steroid trials in AH for over 45 years.

Last Comprehensive Review: 2026

Guidelines & Evidence

Clinical Details

Section 1

When to Use

When to Use

Active alcoholic hepatitis with jaundice to assess severity
Determine eligibility for corticosteroid therapy
Identify patients requiring ICU-level monitoring
Prognosticate 30-day mortality in active drinkers presenting with liver failure

Required Clinical Context

Must have confirmed or highly suspected alcoholic hepatitis (jaundice + recent heavy alcohol use + AST:ALT > 2:1 + excluding other causes). DF is not meaningful in non-alcoholic liver disease.
Section 2

Formula & Logic

Formula

DF = 4.6 × (PT patient − PT control) + Serum Bilirubin (mg/dL)

Interpretation

01
DF < 32: Non-severe AH. Supportive management. No steroids.
02
DF ≥ 32: Severe AH. 30-day mortality 35–45%. Consider prednisolone 40mg/day × 28 days if no contraindications.

Bilirubin Units

The original Maddrey formula uses bilirubin in mg/dL. If your lab reports in µmol/L, divide by 17.1 before entering the value.
Section 3

Pearls/Pitfalls

Lille Model — Day 7 Assessment

After 7 days of prednisolone, calculate the Lille Model score
Lille score < 0.45: Steroid responder — complete 28-day course
Lille score ≥ 0.45: Non-responder — discontinue steroids (no benefit, continued infection risk)

Contraindications to Steroids

Active GI bleeding
Active untreated bacterial infection (treat infection first, then reassess)
Renal failure (creatinine > 2.5 mg/dL — consider pentoxifylline)
Uncontrolled hepatic encephalopathy or coma
HBV or HCV co-infection (risk viral reactivation)

MELD vs DF

Both Maddrey DF and MELD score predict 30-day mortality in AH. MELD > 20 correlates with DF ≥ 32. Some centres use Glasgow AH Score (GAHS) or ABIC score as additional severity tools.
Section 4

Next Steps

Severe AH Management (DF ≥ 32)

01
Prednisolone 40mg orally daily × 28 days (preferred over methylprednisolone)
02
Nutritional support: EN via NGT if oral intake inadequate (target 35–40 kcal/kg/day, 1.2–1.5g protein/kg/day)
03
Thiamine 300mg IV daily × 3 days, then oral B-vitamins
04
Lille Model on Day 7: if ≥ 0.45, stop steroids
05
Prophylactic PPI for GI bleeding risk
06
Transplant evaluation in selected non-responders

Complementary Tools

MELD Score
Child-Pugh Score
Glasgow-Blatchford Score
Section 5

Evidence Appraisal

Original Study

Corticosteroid therapy of alcoholic hepatitis.

Maddrey WC et al. • Gastroenterology.. 1978;75(2):193–199. DF ≥ 93 originally predicted benefit; revised to ≥ 32 in subsequent validation studies.

Section 6

Literature

Dr. Willis C. Maddrey

A hepatologist who published the discriminant function in 1978 as part of the first RCT of corticosteroids in alcoholic hepatitis. His formula — combining PT prolongation with bilirubin — elegantly captured both synthetic dysfunction and cholestasis in a single score, remaining the entry criterion for steroid trials in AH for over 45 years.

Last Comprehensive Review: 2026

Related Hepatology Tools

MELD / MELD-Na
Child-Pugh Score
FIB-4 Index
NAFLD Fibrosis Score
Lille Model
Have feedback about this calculator?Let us know.