MELD Score: Definition and Calculation
The MELD (Model for End-Stage Liver Disease) score is a prognostic scoring system based on three objective laboratory values—serum bilirubin, serum creatinine, and INR (international normalized ratio for prothrombin time)—that predicts 3-month mortality risk in patients with end-stage liver disease. 1
What MELD Measures and Why It Matters
The MELD score was originally developed to predict short-term survival in patients undergoing TIPS (transjugular intrahepatic portosystemic shunt) procedures for gastrointestinal bleeding 1, 2. Since 2002, it has been adopted worldwide as the primary tool for:
- Prioritizing patients on liver transplant waiting lists based on mortality risk rather than waiting time 1
- Predicting 3-month mortality in patients with cirrhosis 1, 2
- Determining transplant candidacy, with MELD ≥15 recommended for listing patients with end-stage liver disease 1
The score provides a continuous scale from 6 to 40, with higher scores indicating greater mortality risk (90% survival at MELD 6 vs. 7% survival at MELD 40) 2.
The Standard MELD Formula
The original MELD calculation uses:
- Serum bilirubin (mg/dL)
- Serum creatinine (mg/dL)
- INR (international normalized ratio)
These three objective laboratory parameters are weighted in a mathematical formula to generate the score 1, 3, 4.
Evolution: MELD-Na and MELD 3.0
MELD-Na adds serum sodium to improve predictive accuracy, particularly for patients with hyponatremia 1, 5.
MELD 3.0 (the most recent refinement) incorporates:
- Female sex (credits an extra 1.3 points to women)
- Serum albumin
- Interactions between bilirubin-sodium and albumin-creatinine
- Upper creatinine bound at 3.0 mg/dL
MELD 3.0 demonstrates superior discrimination compared to MELD-Na (C-statistic 0.869 vs. 0.862) and addresses sex disparities in transplant access, correctly reclassifying 8.8% of patients to higher priority tiers 6. This newer model resulted in fewer waitlist deaths in simulation studies 6.
Clinical Thresholds and Decision Points
MELD ≥15: Recommended threshold for listing patients for liver transplantation 1
MELD <15: In patients with MELD ≤14,1-year survival is actually lower with transplantation than without, making transplant inappropriate at this stage 1
MELD >30: Signals very high perioperative mortality and morbidity risk that must be carefully weighed 1
MELD >35: One of the few scenarios where MELD predicts post-transplant mortality 1
Critical Limitations and MELD Exceptions
MELD fails to capture mortality risk in 15-20% of patients 4, 7. Several conditions require "exception points" or expert adjudication because standard MELD underestimates their severity 1:
Manifestations of Cirrhosis Requiring Exceptions:
- Refractory ascites
- Recurrent or chronic hepatic encephalopathy
- Hepatopulmonary syndrome
- Portopulmonary hypertension
- Recurrent gastrointestinal bleeding
- Intractable pruritus
Specific Disease States:
- Hepatocellular carcinoma (HCC): Receives time-dependent exception points based on tumor size, number of nodules, AFP levels, and response to downstaging 1
- Cholangiocarcinoma
- Budd-Chiari syndrome
- Primary hyperoxaluria
- Familial amyloidotic polyneuropathy
- Polycystic liver disease
Other Predictive Models:
Delta MELD (ΔMELD)—the change in MELD over time—may better predict mortality than a single static score 1.
Common Pitfalls to Avoid
Don't use MELD alone for DCD (donation after circulatory death) grafts: Match controlled DCD livers to recipients with MELD <25 due to higher reperfusion injury risk 8
MELD doesn't predict post-transplant outcomes except at very high scores (>35) 1
Inter-laboratory variability in creatinine and INR measurements can affect score accuracy 5
Creatinine underestimates renal dysfunction in women, contributing to sex disparities in transplant access—a key reason MELD 3.0 was developed 9, 6
Recognize when MELD doesn't apply: Patients with complications like hepatorenal syndrome, severe encephalopathy, or HCC need expert multidisciplinary evaluation beyond the numerical score 1
Practical Application
MELD is good for predicting short-term pre-transplant mortality risk (Grade II-1 recommendation) 1. It should drive initial allocation decisions but final transplant decisions must incorporate multiple parameters including donor-recipient matching, local priorities, and clinical factors not captured by laboratory values 1.
For patients with liver diseases whose severity isn't described by MELD, expert panels should assign different priority through standardized exception criteria 1.