Maddrey's Discriminant Function

Maddrey's Discriminant Function (DF), also known as the Modified Discriminant Function (mDF), is the original and most widely used clinical tool for assessing the severity of alcoholic hepatitis and determining which patients should receive corticosteroid therapy. Introduced by Maddrey et al. in 1978 and subsequently modified by Carithers et al. in 1989, the DF uses a simple formula incorporating prothrombin time prolongation and serum bilirubin to identify patients with severe disease who face high short-term mortality without treatment.

The formula is: DF = 4.6 x (Patient PT - Control PT) + Total Bilirubin (mg/dL). The coefficient of 4.6 for the PT prolongation was derived from the original study data and reflects the strong prognostic importance of coagulopathy in alcoholic hepatitis. The control PT represents the laboratory's mean normal prothrombin time, typically 11-13 seconds. The resulting DF value provides a continuous measure of disease severity, with the critical threshold set at 32.

A DF of 32 or greater defines severe alcoholic hepatitis and identifies patients with approximately 35-50% 30-day mortality without treatment. These patients are candidates for corticosteroid therapy with prednisolone 40 mg daily for 28 days, which has been shown to improve 28-day survival in multiple clinical trials. A DF below 32 identifies patients with less severe disease who have lower short-term mortality and can generally be managed with supportive care, alcohol cessation, and nutritional supplementation without corticosteroids.

The biological rationale for the DF components is straightforward. Prothrombin time reflects the liver's ability to synthesize coagulation factors (II, V, VII, IX, X), which have relatively short half-lives (6 hours to 5 days). Prolonged PT indicates acute severe impairment of hepatic synthetic function. Elevated bilirubin reflects both hepatocellular damage (reduced bilirubin uptake, conjugation, and excretion) and cholestasis. Together, these parameters capture the two cardinal manifestations of severe alcoholic hepatitis: synthetic failure and impaired bilirubin metabolism.

Despite its widespread use and historical importance, the DF has limitations. It uses PT rather than INR, creating potential interlaboratory variability depending on the thromboplastin reagent used. The control PT must be obtained from the same laboratory, adding a step. The DF does not account for renal function, which is a strong independent predictor of mortality in alcoholic hepatitis. These limitations led to the development of the MELD score as an alternative severity assessment, with MELD of 21 or greater having similar prognostic performance to DF of 32 or greater for identifying severe alcoholic hepatitis.

In clinical practice, the DF remains the standard initial assessment tool for alcoholic hepatitis severity. Guidelines from AASLD and EASL recommend calculating the DF (or MELD) in all patients with suspected alcoholic hepatitis to guide treatment decisions. Patients meeting severity criteria should be evaluated for corticosteroid contraindications (active infection, GI bleeding, hepatorenal syndrome), treated with prednisolone if eligible, and reassessed at day 7 using the Lille Score to determine whether to continue or discontinue steroids.