Maddrey Discriminant Function Calculator for Alcohol-Associated Hepatitis
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.
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 Maddrey Discriminant Function calculator reproduces the modified mDF for an adult already in an appropriately suspected or clinician-established alcohol-associated-hepatitis assessment context. Alcohol-associated hepatitis was historically called alcoholic hepatitis. It uses actual patient PT, the corresponding laboratory control PT, and total bilirubin; it does not diagnose the condition. [2, 3]
The public threshold is the current conventional historical marker mDF ≥32. It is not a corticosteroid order or a complete current severity assessment; current ACG guidance principally uses MELD >20 for severe-AH context and requires clinical eligibility assessment. [3, 4]
Interpretation
| Exact raw mDF | Public status | Meaning on this page |
|---|---|---|
| <32 | Historical threshold not reached | Does not establish mild disease, safety, disposition, or treatment. |
| ≥32 | Historical threshold reached | Historical severity-model context; not a diagnosis, individual mortality estimate, or steroid instruction. |
Carithers 1989 enrolled patients with DF >32 or spontaneous hepatic encephalopathy. Later literature and current ACG context commonly operationalize mDF ≥32; exactly 32 counts as reached here. Encephalopathy was a separate trial-entry condition, not an mDF input or score override. [2, 3]
Original Maddrey DF vs modified Maddrey DF
The 1978 trial identified PT prolongation and bilirubin as important prognostic variables. The 1989 multicenter trial used the now-common modified expression with patient PT minus laboratory control PT, supporting comparison across laboratories with different PT standards. Today, “Maddrey DF” calculators generally mean this modified control-adjusted form. [1, 2, 6]
Why the laboratory control PT matters
Use the corresponding control/reference time associated with the laboratory PT method and reagent. Do not substitute a generic 12-second value, a reference-range midpoint or upper limit, another laboratory's control, or the patient's INR. If the correct control PT is unavailable, this calculator will not invent one. Because mDF uses the raw PT difference, changing laboratory methods can change the result. [2, 3]
Patient PT and control PT should belong to the corresponding laboratory method; bilirubin should represent the same presentation or an explicitly appropriate paired assessment. Do not mix unrelated institutions, reagents, admissions, or pre/post-treatment records without a defined clinical reason.
Can Maddrey DF be calculated from INR?
No direct substitution is source-defined. INR depends on patient PT, reference PT, and the International Sensitivity Index, so reconstructing PT from INR adds assay assumptions. QuickMedCalc requires the actual patient PT and laboratory control PT and does not implement INR, ISI, or estimated-PT mode. [2, 3]
Maddrey DF vs MELD in alcohol-associated hepatitis
| Model | Inputs and role | Important distinction |
|---|---|---|
| Modified Maddrey DF | Patient PT, laboratory control PT, and bilirubin; simple historical severity context tied to classic corticosteroid trials. | Laboratory-dependent PT and no creatinine; not the current sole severity definition. |
| MELD | INR, bilirubin, creatinine, and model-version-specific variables. | Current ACG uses MELD >20 in its AH evidence context and considers MELD more accurate; model versions are not interchangeable. |
In contemporary comparison data summarized by ACG, mDF discrimination was lower than MELD, GAHS, and ABIC. mDF remains recognizable and historically important rather than “obsolete.” The separate MELD Score Calculator supports multiple model versions; do not automatically map the AH literature's MELD >20 threshold onto current OPTN MELD 3.0 or another version without version-specific evidence. [3]
Maddrey and Lille answer different questions
mDF is a baseline/presentation severity model. Lille is a separate dynamic response model for patients after corticosteroid treatment, using treatment-course information including bilirubin change and additional variables at day 4 or day 7. This page does not calculate Lille and does not add treatment-day inputs. [3, 5, 6]
Historical mortality and treatment limits
ACG summarizes mDF ≥32 as historically associated with roughly 20%–50% short-term, approximately 30-day mortality across older evidence. That is population-level context, not an individual probability or a continuous risk curve. A result of 33 and a result of 60 do not receive invented patient-specific percentages here. [3, 4]
Current corticosteroid decisions require the diagnosis and severity context, eligibility and contraindication assessment, infection and bleeding considerations, renal and multiorgan status, and clinician judgment. This calculator does not select a medicine, dose, duration, discontinuation rule, admission, transplant pathway, or other treatment. [3]
Worked arithmetic examples
- 20 s patient PT, 12 s control PT, and 10 mg/dL bilirubin: 4.6 × (20 − 12) + 10 = 46.8.
- 171 µmol/L bilirubin normalizes exactly to 10 mg/dL under the source-defined ÷17.1 convention and gives the same 46.8 result.
- 17 s patient PT, 12 s control PT, and 9 mg/dL bilirubin gives exactly 32, so the public ≥32 threshold is reached.
These are arithmetic examples, not diagnoses, prognoses, or treatment recommendations. [2]
References
- Maddrey WC, Boitnott JK, Bedine MS, Weber FL Jr, Mezey E, White RI Jr. Corticosteroid therapy of alcoholic hepatitis. Gastroenterology. 1978;75(2):193–199. PMID 352788. DOI 10.1016/0016-5085(78)90401-8.
- Carithers RL Jr, Herlong HF, Diehl AM, et al. Methylprednisolone therapy in patients with severe alcoholic hepatitis. A randomized multicenter trial. Ann Intern Med. 1989;110(9):685–690. PMID 2648927. DOI 10.7326/0003-4819-110-9-685.
- 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.
- Mathurin P, Mendenhall CL, Carithers RL Jr, et al. Corticosteroids improve short-term survival in patients with severe alcoholic hepatitis (AH): individual data analysis of the last three randomized placebo controlled double blind trials of corticosteroids in severe AH. J Hepatol. 2002;36(4):480–487. PMID 11943418. DOI 10.1016/S0168-8278(01)00289-6.
- 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.
- 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 modified Maddrey Discriminant Function (mDF) is a historical severity model for an adult already in an appropriate alcohol-associated-hepatitis assessment context. Alcohol-associated hepatitis was historically called alcoholic hepatitis. It combines patient PT prolongation relative to the laboratory control PT with total bilirubin; it does not diagnose the condition.
mDF = 4.6 × (patient prothrombin time in seconds − corresponding laboratory control PT in seconds) + total bilirubin in mg/dL. For bilirubin reported in µmol/L, the 1989 formula convention divides by 17.1 to obtain the mg/dL term.
Sources: [2]
QuickMedCalc treats an exact unrounded mDF of 32 or greater as reaching the conventional historical threshold used in alcohol-associated-hepatitis severity literature. Reaching it does not independently diagnose severe disease, provide a personal mortality probability, or determine corticosteroid treatment; current ACG guidance principally uses MELD >20 for severe-AH context.
The modified formula uses patient PT minus the corresponding laboratory control/reference PT because PT depends on the laboratory method and reagent. The correct value is not a universal 12-second constant, a reference-range midpoint or upper limit, another institution's value, or the patient's INR. If it is unavailable, this calculator does not invent one.
The 1989 trial abstract used DF >32 or spontaneous hepatic encephalopathy as enrollment conditions. Later literature and current ACG context commonly use mDF ≥32. This page uses the current conventional ≥32 comparison, counts exactly 32 as reached, and preserves the original trial wording as provenance; encephalopathy is not a formula input or override.
Yes. Select µmol/L and the runtime divides the submitted value by the exact source-defined 17.1 conversion before using the formula. For example, 171 µmol/L becomes exactly 10 mg/dL. Switching between an already selected unit clears bilirubin so the same number is not silently reinterpreted.
Sources: [2]
mDF uses patient PT, laboratory control PT, and bilirubin. MELD uses INR, bilirubin, creatinine, and model-version-specific variables. Current ACG guidance considers MELD more accurate and uses MELD >20 in its alcohol-associated-hepatitis evidence context. The QuickMedCalc MELD page supports several versions, which must not be treated as interchangeable with one historical threshold.
Sources: [3]
No. The formula assumes an appropriate clinical assessment context but cannot establish the diagnosis, identify another cause of hepatitis or jaundice, review alcohol exposure, assess infection or bleeding, or evaluate renal and multiorgan status.
Sources: [3]
No. mDF has a historical relationship to corticosteroid trials, but a score alone cannot establish current severity, eligibility, contraindications, medicine, dose, duration, or treatment response. Current decisions require complete clinical assessment, and response after treatment is a separate Lille task.
The 1978 Maddrey trial established the prognostic importance of PT prolongation and bilirubin. The 1989 multicenter trial used the now-common modified expression with patient PT minus laboratory control PT. This page implements that 1989 laboratory-control-adjusted formula rather than claiming the final modern expression came directly from the 1978 paper.
Related Calculators
Lille Score
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.
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.
Child-Pugh
Calculate the standard five-domain Child-Turcotte-Pugh score and class for established cirrhosis using bilirubin, albumin, INR or PT prolongation, ascites, and encephalopathy.
Disclaimer
Educational and informational reference only. Not intended to replace professional medical advice, diagnosis, treatment, or independent verification.