MELD Score Calculator — Model for End-Stage Liver Disease

About

The MELD (Model for End-Stage Liver Disease) score was initially developed at the Mayo Clinic by Kamath et al. in 2001 to predict 3-month survival in patients undergoing transjugular intrahepatic portosystemic shunt (TIPS) procedure. It was subsequently validated to accurately predict 3-month mortality in patients with end-stage liver disease regardless of the planned intervention and became the basis for liver transplant organ allocation in the United States through the UNOS (United Network for Organ Sharing) system. The score ranges from 6 to 40, with higher scores indicating more severe disease and greater mortality risk. The original MELD uses three objective, reproducible laboratory variables: serum bilirubin (reflecting hepatic synthetic and excretory function), INR (reflecting hepatic clotting factor synthesis), and serum creatinine (reflecting renal function — an important independent predictor in cirrhosis due to the hepatorenal axis). The current MELD score incorporates a cap on creatinine at 4.0 mg/dL for patients on renal replacement therapy (hemodialysis ≥2 times in the past 7 days or 24 hours of continuous venovenous hemodialysis). All laboratory values are capped at 0.1 for minimum values to allow logarithmic transformation. The MELD-Na variant, adopted by UNOS in 2016, adds serum sodium to improve mortality prediction, since hyponatremia in cirrhosis reflects portal hypertension and is independently associated with waitlist mortality. MELD has been validated across diverse populations including cirrhosis of all etiologies, acute liver failure, alcoholic hepatitis, and non-alcoholic fatty liver disease. Evidence level: Grade A, supported by multiple large prospective cohort studies and systematic reviews.

Formula

MELD = 3.78 × ln(bilirubin) + 11.2 × ln(INR) + 9.57 × ln(creatinine) + 6.43

The MELD score is calculated using natural logarithm (ln) of three laboratory values, each weighted by a coefficient derived from a Cox proportional hazards regression model of survival in the original Mayo Clinic cohort. The equation: MELD = 3.78 × ln(bilirubin mg/dL) + 11.2 × ln(INR) + 9.57 × ln(creatinine mg/dL) + 6.43. Each term represents the relative prognostic contribution of that variable — INR has the highest weight. Steps: (1) Obtain serum bilirubin (total), INR, and serum creatinine. (2) If the patient has received hemodialysis ≥2 times in the past 7 days OR 24 hours of CVVHD, set creatinine to 4.0 mg/dL. (3) All values are floored at 0.1 — if any value is <0.1, set to 0.1 before taking the logarithm to avoid undefined ln(0). (4) Compute the natural logarithm for each variable. (5) Multiply each ln by its coefficient. (6) Sum the products and add the constant 6.43. (7) Multiply by 10 (some historical versions; the standard UNOS equation does not multiply by 10). (8) Round to the nearest integer. (9) Cap the score: if the result < 6, set to 6; if > 40, set to 40. For MELD-Na, a correction formula is used: MELD-Na = MELD + 1.32 × (137 − Na) − 0.33 × MELD × (137 − Na), where Na is capped between 120 and 137 mEq/L. Serum sodium values <120 are set to 120 and >137 are set to 137. MELD-Na is capped similarly at 6–40. The UNOS exception points system allows additional points for specific conditions (e.g., hepatocellular carcinoma within Milan criteria, hepatopulmonary syndrome) that may have lower MELD scores but high short-term mortality without transplant. Interpretation: Each 1-point increase in MELD is associated with approximately 1% increase in 3-month mortality risk, though this is non-linear — risk accelerates at higher MELD thresholds.

Score Interpretation

MELD has transformed liver transplant allocation from a time-based (waiting list time) to a disease-severity-based system, substantially reducing waitlist mortality. Before MELD implementation in 2002, donor livers were allocated primarily by waiting time, which incentivized early listing and did not prioritize the most urgent candidates. After MELD-based allocation was introduced by UNOS, median waitlist time decreased and organs were preferentially directed to patients with highest short-term mortality risk. The MELD score has been validated extensively: in a landmark study of over 3,000 patients on the liver transplant waiting list, the c-statistic for 3-month mortality prediction was 0.83, indicating excellent discriminative ability. The score is used in over 20 countries including the US (UNOS), Eurotransplant, and transplant centers in Asia, South America, and the Middle East. Key clinical applications: (1) Transplant listing and allocation — patients with MELD ≥15 typically qualify for listing, with allocation priority proportional to score. (2) Predicting surgical risk in cirrhotic patients undergoing non-transplant abdominal surgeries — MELD >14 is associated with increased 30-day mortality. (3) Prognosis in acute liver failure — MELD score at admission predicts mortality and need for urgent transplantation. (4) Alcoholic hepatitis — MELD (with or without the Lille model) predicts 90-day survival. (5) TIPS procedure candidacy — the original validation population. MELD score correlates with the severity of portal hypertension complications: ascites severity, variceal bleeding risk, hepatic encephalopathy stage, and spontaneous bacterial peritonitis risk. The MELD-Na variant improved the c-statistic to 0.85 by incorporating the negative prognostic effect of hyponatremia found in approximately 30% of patients with decompensated cirrhosis. Studies have shown that a MELD-Na ≥21 is the optimal threshold for identifying patients with 3-month mortality risk ≥15%. The score has limitations: exception points for HCC can create disparities, MELD does not capture quality of life or functional status, and the constant 6.43 floor can underestimate risk in patients with severe complications but preserved lab values.

Reference Ranges