MELD Score
MELD estimates 3-month mortality risk in patients with end-stage liver disease using serum creatinine, bilirubin, and INR. It is the primary score used for organ allocation in liver transplantation (UNOS/Eurotransplant).
inputs
✓ when to use
Use in patients with chronic liver disease/cirrhosis to estimate short-term mortality, guide transplant listing priority, assess candidacy for TIPS, and inform prognosis discussions. Calculate at every clinical encounter in advanced liver disease. MELD is preferred over Child-Pugh for transplant allocation because it uses only objective laboratory variables.
✗ when not to use
MELD may overestimate mortality in patients with renal impairment from causes other than hepatorenal syndrome (e.g., diabetic nephropathy, CKD). Does not account for portal hypertension complications (variceal bleeding, ascites) or hepatocellular carcinoma — exception points are added separately in transplant allocation. Not validated for acute liver failure (use KCH criteria).
clinical pearls
- MELD uses only objective lab values (creatinine, bilirubin, INR), eliminating the subjectivity of Child-Pugh's ascites and encephalopathy grading. This is why it replaced Child-Pugh for transplant allocation.
- Lab values <1.0 are set to 1.0 (prevents negative log values). Creatinine is capped at 4.0 mg/dL. These boundary rules are specific to the UNOS implementation and must be applied for the score to be valid.
- MELD does not capture the impact of hyponatraemia, which is an independent predictor of mortality in cirrhosis. MELD-Na addresses this limitation and is now used by UNOS.
- In patients on warfarin, INR is elevated by the drug rather than by hepatic dysfunction, artificially inflating MELD. Some transplant centres adjust for this when interpreting MELD.
- MELD is a mortality predictor, not a comprehensive disease severity tool. A patient with MELD 12 can still have refractory ascites, recurrent encephalopathy, and severely impaired quality of life — clinical assessment must supplement the score.