Maddrey Discriminant Function

The Maddrey Discriminant Function (DF) is a clinical scoring tool used to stratify the severity of alcoholic hepatitis (AH) and identify patients who are likely to benefit from corticosteroid therapy. First described by Maddrey and colleagues in 1978, the function was developed from a randomised controlled trial of prednisolone in severe alcoholic hepatitis, and it uses prothrombin time prolongation and serum bilirubin to produce a single composite score. A discriminant function score of 32 or above defines severe alcoholic hepatitis, a threshold associated with a 30-day spontaneous mortality of approximately 30–50% without treatment. Patients meeting this threshold are candidates for prednisolone 40 mg daily for 28 days, assuming there are no contraindications such as active infection, gastrointestinal bleeding, or hepatitis B infection. The biological rationale for the DF rests on two key pathophysiological features of severe alcoholic hepatitis: impaired hepatic synthesis of coagulation factors (reflected by prothrombin time prolongation) and severe hepatocellular dysfunction and cholestasis (reflected by markedly elevated bilirubin). The original formula used a control prothrombin time from a reference laboratory, but this is often simplified by using a fixed control of 12 seconds in clinical practice. The Maddrey DF remains one of the most widely studied and applied scores in alcoholic liver disease, typically used alongside the MELD score and, following initiation of corticosteroids, the Lille score to assess early treatment response at day 7.

PrimeCalcPro provides professional-grade tools trusted by businesses and academics.

1. 1Measure the patient's prothrombin time (PT) in seconds and obtain the laboratory's control PT (typically 12–13 seconds in most labs). Calculate the PT prolongation: patient PT minus control PT.
2. 2Multiply the PT prolongation by 4.6 to obtain the coagulation component. This weighting reflects the importance of synthetic liver failure in prognosis.
3. 3Measure the serum total bilirubin in mg/dL. If the bilirubin is reported in µmol/L, convert by dividing by 17.1 (1 mg/dL = 17.1 µmol/L).
4. 4Add the bilirubin (mg/dL) directly to the weighted PT prolongation to obtain the DF score.
5. 5Interpret the result: DF ≥ 32 defines severe alcoholic hepatitis with approximately 30–50% 30-day spontaneous mortality. DF < 32 indicates non-severe alcoholic hepatitis with better short-term prognosis.
6. 6If DF ≥ 32 and no contraindications exist, initiate prednisolone 40 mg orally daily. Evaluate Lille score at day 7 to assess steroid response.
7. 7If the Lille score at day 7 is ≥0.45 (non-responder), discontinue prednisolone — continued steroid therapy does not improve survival and may increase infection risk.

• !Using bilirubin in µmol/L instead of mg/dL — this produces DF scores ~17x too high and a completely erroneous clinical interpretation.
• !Omitting the control PT and using the patient's absolute PT value instead of the prolongation — the formula requires the difference (patient PT minus control PT), not the raw patient PT.
• !Not screening for hepatitis B before initiating prednisolone, risking catastrophic HBV reactivation in co-infected patients.
• !Continuing prednisolone beyond day 7 in patients identified as Lille score non-responders (Lille ≥0.45) — this provides no mortality benefit and increases serious infection risk.
• !Applying the Maddrey DF to alcoholic liver disease other than acute alcoholic hepatitis — it was specifically validated for the acute presentation, not stable compensated alcohol-related cirrhosis.
• !Forgetting nutritional supplementation — it is part of the standard of care for all patients with alcoholic hepatitis irrespective of DF score, yet is frequently overlooked when focus centres on steroid therapy.