How to use it 01 Enter the chemistry once. Bilirubin, creatinine, INR, sodium and albumin feed the MELD cards and the fibrosis indices. 02 Tick dialysis if it applies. Dialysis at least twice in a week puts creatinine at the model's ceiling, not the measured value. 03 Read MELD-Na and MELD 3.0 side by side. The panel prints the pre-sodium MELD(i), the corrected score and the difference between the two models. 04 Set ascites and encephalopathy for Child-Pugh. These are clinical grades, so the score inherits their subjectivity. 05 Use FIB-4 and NFS as triage. Both print an indeterminate band; falling inside it is a result, not a failure.
Worked readouts Bilirubin 3.0 mg/dL, creatinine 1.5 mg/dL, INR 1.8, sodium 130 mmol/L, albumin 2.8 g/dL, male, no ascites or encephalopathy; FIB-4 inputs age 55, AST 60, ALT 40, platelets 150:
Card
The panel reads
MELD(i), before sodium
21
MELD-Na
25 — sodium correction applied (≈ 19.6 % 3-month mortality band)
MELD 3.0
26 (raw 25.54 , +1 against MELD-Na; female variant 27 )
Child-Pugh
8 / 15 — Class B (1-year 81 % , 2-year 57 % , perioperative 30 % )
FIB-4
3.48 — above the 3.25 cut-off (PPV 65 % )
NAFLD fibrosis score
2.185 — above 0.676 (PPV 90 % ), AST/ALT 1.50
Marking dialysis with a creatinine of 6 mg/dL gives MELD(i) 30 and MELD-Na 32 , because the model caps creatinine at 4.0.
Three generations, one patient MELD began as a log-linear model: MELD(i) = round(0.957·ln(Cr) + 0.378·ln(bili) + 1.120·ln(INR) + 0.643) × 10, with all three inputs floored at 1.0 and creatinine capped at 4.0. Because the logarithm compresses, the difference between a creatinine of 1.5 and 4.0 is large in points but not proportional in risk. Sodium entered in 2016 as a correction applied only when MELD(i) exceeds 11: MELD = MELD(i) + 1.32·(137 − Na) − 0.033·MELD(i)·(137 − Na). Here 21 becomes 25 , and the sodium is clamped to the 125–137 band first.
MELD 3.0 (2023) rewrote the model to remove the sex disadvantage: MELD 3.0 = 1.33·[female] + 4.56·ln(bili) + 0.82·(137 − Na) − 0.24·(137 − Na)·ln(bili) + 9.09·ln(INR) + 11.14·ln(Cr) + 1.85·(3.5 − Alb) − 1.83·(3.5 − Alb)·ln(Cr) + 6, with creatinine capped at 3.0 instead of 4.0 and albumin clamped to 1.5–3.5 g/dL. The same patient scores 26 — one point higher than MELD-Na and two higher as a woman. Child-Pugh predates all of this: five items worth 1–3 points each, summed to 5–15 and banded A/B/C, with survival figures attached to each class. The fibrosis indices differ again: FIB-4 is age × AST ÷ (platelets × √ALT), and NFS a weighted sum of age, BMI, glucose status, AST/ALT, platelets and albumin.
Assumptions and limits
Populations travel with each score. MELD-Na and MELD 3.0 apply to patients aged 12 and over with end-stage liver disease; acute liver failure, transplant exception points and non-hepatic causes of abnormal creatinine or INR sit outside them, and anticoagulants raise INR.
Child-Pugh contains judgement. Ascites and encephalopathy are graded clinically with limited inter-rater agreement, and albumin can fall for renal or nutritional reasons. FIB-4 is unreliable below 35 and needs a raised low cut-off at 65, and both fibrosis indices leave roughly a third of patients in an indeterminate band. What it does not do: it does not diagnose, decide listing or advise treatment, and it does not read a chart. For the cardiovascular scores use the Cardio panel ; for kidney-function arithmetic use the Renal panel ; for the metabolic indices use the Metabolic panel .
Caps and corrections change the answer The scores rearrange the same labs; the traps are where they differ. First, creatinine caps: MELD-Na takes 4.0 (or 4.0 for dialysis), MELD 3.0 takes 3.0 — a patient with a creatinine of 6 is not scored the same way in both. Second, the sodium correction is conditional: below MELD(i) 11 it is skipped, so a low sodium in a low-scoring patient moves nothing. Third, Child-Pugh boundaries are sharp: bilirubin exactly 3.0 scores 2, one decimal higher scores 3, which can flip the class on a single assay. Fourth, the FIB-4 threshold moves with age: at 70 the low cut-off is raised from 1.45 to 2.0 , and the panel prints that warning. Finally, both fibrosis scores have indeterminate zones — the NFS warning says about a third of patients land between −1.455 and 0.676, where the score alone cannot separate fibrosis from its absence.
Where it helps Comparing model generations The difference card makes the 2023 change concrete: the same patient reads 25 under MELD-Na and 26 under MELD 3.0, and the gap grows for women, whose under-scoring the sex term addresses.
Teaching score arithmetic Every item's points are exposed in the Child-Pugh table — bilirubin 2, albumin 2, INR 2, ascites 1, encephalopathy 1 — so the 8/15 total can be checked by hand.
Privacy Laboratory values and clinical grades are processed in your browser; nothing is uploaded and no result is stored beyond this device.
References
Kim WR et al., Hyponatremia and mortality among patients on the liver-transplant waiting list , pubmed.ncbi.nlm.nih.gov (访问日期:2026-10-07)— the sodium correction that turned MELD into MELD-Na.
Kim WR et al., MELD 3.0: the model for end-stage liver disease updated for the modern era , pubmed.ncbi.nlm.nih.gov (访问日期:2026-10-07)— the sex and albumin terms and the 3.0 creatinine cap.
Pugh RNH et al., Transection of the oesophagus for bleeding oesophageal varices , pubmed.ncbi.nlm.nih.gov (访问日期:2026-10-07)— the five-item Child-Pugh classification.
Sterling RK et al., Development of a simple noninvasive index to predict significant fibrosis , pubmed.ncbi.nlm.nih.gov (访问日期:2026-10-07)— the FIB-4 index and its 1.45 / 3.25 cut-offs.
Angulo P et al., The NAFLD fibrosis score , pubmed.ncbi.nlm.nih.gov (访问日期:2026-10-07)— the NFS formula and its indeterminate band.
Calculators in this hub
Hand-picked tools, one click away. The mini versions compute live and carry your values into the full calculator.
Sources & review
Reviewed by CalcX Editorial Team
Updated 2026-10-07 Kim WR et al., Hyponatremia and mortality among patients on the liver-transplant waiting list, N Engl J Med 359:1018–1026 (2008) (accessed 2026-10-07) Kim WR et al., MELD 3.0 — the model for end-stage liver disease updated for the modern era, Gastroenterology 161:1887–1895 (2021) (accessed 2026-10-07) Pugh RNH et al., Transection of the oesophagus for bleeding oesophageal varices, Br J Surg 60:646–649 (1973) (accessed 2026-10-07) Sterling RK et al., Development of a simple noninvasive index to predict significant fibrosis, Hepatology 43:1317–1325 (2006) (accessed 2026-10-07) Angulo P et al., The NAFLD fibrosis score, Hepatology 45:846–854 (2007) (accessed 2026-10-07)