GastroAGI Logo
OverviewBlogsAbout
Trending TopicsDaily BriefConference

117 calculators match

GastroAGI flagship

1
MASLD–MASH NITIntegrated non-invasive assessment of MASLD fibrosis and at-risk MASH — FIB-4, APRI, NFS, FAST, Agile 3+, Agile 4, ELF and ADAPT in one pass

Most used

21
MELD-NaAssesses the severity of chronic liver diseaseChild-Pugh ScoreAssesses the prognosis of chronic liver disease, mainly cirrhosisFIB-4 IndexLiver fibrosis scoring indexAPRIAST to platelet ratio — liver fibrosisMaddrey's DFAlcoholic hepatitis severityGlasgow-BlatchfordUpper GI bleed risk stratificationGAHSGlasgow alcoholic hepatitis scoreMontreal IBDIBD classification — CD & UCMayo ScoreUlcerative colitis activityBISAP ScoreBedside index for severity of pancreatitisCLIF-SOFAOrgan failure scoring in cirrhosisAlcohol ContentStandard drinks & alcohol grams calculatorAARC-ACLFAcute-on-chronic liver failure gradePELD / CR ScorePediatric end-stage liver diseaseHarvey-BradshawCrohn's disease activity indexCTSICT severity index — pancreatitisRockall ScoreGI bleed rebleeding & mortality riskCAGEAlcohol use disorder screening (4 questions)MELD 3.0Updated MELD — sex-inclusive formulaAUDIT ScoreAlcohol use disorders identification testVOCAL-Penn ScorePost-operative mortality risk in cirrhosis surgery

Liver & Cirrhosis

17
ALBI GradeAlbumin-bilirubin liver function grade in HCCUKELD ScoreUK model for end-stage liver diseaseMELD-XIMELD excluding INR — for anticoagulated patientsWest Haven CriteriaHepatic encephalopathy gradingMilan CriteriaLiver transplant eligibility in hepatocellular carcinomaLI-RADS v2018 (CT/MRI)Liver observation category from size, APHE and major featuresBCLC StagingHepatocellular carcinoma stage and treatment allocationSimplified AIH CriteriaSimplified criteria for autoimmune hepatitisRevised Original AIH ScoreIAIHG 1999 comprehensive autoimmune hepatitis scoreSAAGSerum-ascites albumin gradient — cause of ascitesR FactorHepatocellular vs cholestatic pattern in liver injuryCLIF-C ACLFMortality prediction in acute-on-chronic liver failureKing's College CriteriaTransplant criteria in acute liver failureGALAD ScoreHCC detection from gender, age, AFP-L3, AFP and DCPMetroticket 2.0AFP-adjusted up-to-seven for HCC transplant eligibilityRUCAMCausality in drug- and herb-induced liver injuryBaveno VII CriteriacACLD, CSPH and sparing screening endoscopy

Fibrosis & MASLD

8
NAFLD Fibrosis ScoreAdvanced fibrosis probability in MASLD/NAFLDBARD ScoreBMI, AST/ALT ratio, diabetes — MASLD fibrosisFatty Liver IndexPredicts hepatic steatosis from routine labsFibrotic NASH Index (FNI)At-risk NASH probability from AST, HbA1c and HDLNAFLD Activity Score (NAS)Histologic activity grade — steatosis, inflammation, ballooningMEFIB IndexMRE + FIB-4 rule for significant fibrosis (≥F2) in MASLDFAST ScoreFibroScan-AST — at-risk NASH from LSM, CAP and ASTSAFE ScoreSteatosis-Associated Fibrosis Estimator for MASLD in primary care

Pancreas & Biliary

8
Ranson's CriteriaAcute pancreatitis severity at 48 hoursGlasgow-Imrie CriteriaAcute pancreatitis severity — the PANCREAS criteriaHAPSHarmless acute pancreatitis scoreTokyo Guidelines — CholangitisTG18 diagnosis and severity grade for acute cholangitisTokyo Guidelines — CholecystitisTG18 diagnosis and severity grade for acute cholecystitisBiliary Pain (Rome IV)Rome IV — defining biliary-type pain before interventionFunctional Pancreatic SODRome IV — pancreatic sphincter of Oddi disorderRevised Atlanta ClassificationAcute pancreatitis severity — mild, moderately severe, severe

IBD

9
Truelove & Witts CriteriaAcute severe ulcerative colitis — admission decisionUCEISUlcerative colitis endoscopic index of severitySCCAISimple clinical colitis activity index — symptoms onlyCDAICrohn's disease activity index — the trial standardSES-CDEndoscopic severity in Crohn's diseasePUCAIPaediatric ulcerative colitis activity indexTravis (Oxford) CriteriaDay 3 colectomy risk in acute severe ulcerative colitisHo IndexDay 3 steroid failure risk in acute severe ulcerative colitisRutgeerts ScorePostoperative Crohn's recurrence at ileocolonoscopy

GI Bleeding

7
EVendo ScorePredicts oesophageal varices needing treatmentForrest ClassificationPeptic ulcer bleeding — rebleeding risk at endoscopyAIMS65 ScoreUpper GI bleed mortality — five bedside criteriaOakland ScoreSafe-discharge risk for acute lower GI bleedingABC ScoreAge, blood tests, comorbidities — GI bleed mortalitySarin ClassificationEndoscopic classification of gastric varicesEGUS (Gastric Ulcer)Malignancy risk in a gastric ulcer, and who needs repeat endoscopy

Alcohol

2
ABIC ScoreAge, bilirubin, INR, creatinine — alcoholic hepatitisLille ModelSteroid response at day 7 in alcoholic hepatitis

Upper GI

4
Chicago Classification v4.0Oesophageal motility pattern from high-resolution manometryLA Classification (Oesophagitis)Los Angeles grade A–D for erosive oesophagitisPrague C & M CriteriaCircumferential and maximal extent of Barrett's oesophagusEREFS (Eosinophilic Oesophagitis)Endoscopic reference score — oedema, rings, exudates, furrows, stricture

Colorectal

3
Boston Bowel Prep ScaleColonoscopy preparation adequacy by segmentStool Osmotic GapOsmotic vs secretory diarrhoea from stool electrolytesATLAS Score (C. difficile)Predicted response to therapy in Clostridioides difficile infection

Functional GI

37
Bristol Stool ScaleStool form types 1–7 and colonic transitRome IV Criteria for IBSIrritable bowel syndrome diagnosis and subtypeFunctional ConstipationRome IV — two of six items, IBS excludedOpioid-Induced ConstipationRome IV — constipation tied to opioid therapyFunctional DiarrhoeaRome IV — loose stools without predominant painFunctional Bloating / DistensionRome IV — bloating without other bowel disorder criteriaUnspecified Functional Bowel DisorderRome IV — bowel symptoms fitting no other categoryCentrally Mediated Abdominal Pain (CAPS)Rome IV — continuous pain unrelated to gut eventsNarcotic Bowel SyndromeRome IV — opioid-induced hyperalgesia of the gutFaecal Incontinence (Rome IV)Rome IV — the criteria, and why nobody is askedFunctional Anorectal PainLevator ani, unspecified pain and proctalgia fugaxFunctional Defecation DisordersRome IV — dyssynergia and inadequate propulsionInfant RegurgitationRome IV — the happy spitter, and the alarm features that rule it outInfant ColicRome IV — recurrent unexplained crying in a well infant under 5 monthsInfant DyscheziaRome IV — straining before a soft stool, and why not to intervenePaediatric Functional ConstipationRome IV — two of six over one month, with overflow soiling as a criterionToddler's DiarrhoeaRome IV functional diarrhoea of childhood — painless, thriving childPaediatric Cyclic Vomiting SyndromeRome IV — both age bands, with different criteria for eachPaediatric Rumination SyndromeRome IV — infant and child/adolescent criteriaFunctional Nausea & Vomiting (Children)Rome IV — two separate disorders that can be met togetherAerophagiaRome IV — distension that increases through the dayPaediatric Functional DyspepsiaRome IV — four times a month, with PDS and EPS subtypingPaediatric Irritable Bowel SyndromeRome IV — plus the constipation clause clinicians missAbdominal MigraineRome IV — stereotypical incapacitating episodes weeks apartFunctional Abdominal Pain — NOSRome IV — the residual category, reached after the other threeNonretentive Faecal IncontinenceRome IV — soiling without retention, where laxatives make it worseFunctional DyspepsiaRome IV — with PDS and EPS subtypingRumination SyndromeRome IV — effortless regurgitation without retchingCyclic Vomiting SyndromeRome IV — stereotypical episodic vomitingCannabinoid HyperemesisRome IV — CVS pattern relieved by cannabis cessationChronic Nausea & VomitingRome IV — chronic nausea and vomiting syndromeBelching DisordersRome IV — supragastric vs gastric belchingFunctional HeartburnRome IV — heartburn with normal acid exposureReflux HypersensitivityRome IV — normal acid exposure, positive symptom associationFunctional Chest PainRome IV — non-cardiac, non-reflux chest painGlobusRome IV — painless lump-in-throat sensationFunctional DysphagiaRome IV — dysphagia with normal endoscopy and manometry
  1. Calculators
  2. /
  3. MELD 3.0
Liver & CirrhosisMost used

MELD 3.0

Updated MELD — sex-inclusive formula

Bounded to 1.5–3.5 g/dL in the formula.

Capped at 3.0 mg/dL in MELD 3.0 — lower than MELD-Na's 4.0.

Female sex adds 1.33 points. This is deliberate: women were shown to be disadvantaged by earlier MELD versions relative to their actual mortality.

MELD 3.0 adds 1.33 points for female sex to correct the under-prioritisation of women, includes albumin, and caps creatinine at 3.0 rather than 4.0.

When to use
Use it for liver transplant waiting-list priority in the United States, where it is the operative allocation formula, and for prognosis in chronic liver disease generally. It applies to candidates aged 12 and over; PELD covers younger children. Use MELD-Na instead only when you need comparability with older literature or are working in a jurisdiction that has not adopted MELD 3.0. As with every MELD variant, it is not validated in acute liver failure and it is not a surgical risk model.
Why use it
Because MELD-Na was measurably unfair to women, and MELD 3.0 was built to fix that. Women were listed and transplanted at lower rates than their actual mortality warranted, for reasons that include creatinine systematically overstating renal function in patients with lower muscle mass. MELD 3.0 addresses this directly with a sex term and by adding albumin, and simulation showed the change would avert waiting-list deaths — 7,788 against 7,850 for MELD-Na. The discrimination gain over MELD-Na is small in absolute terms; the equity correction is the substantive reason it replaced it.
Formula, evidence and interpretation

About the MELD 3.0 (Model for End-Stage Liver Disease, 2021 revision)

MELD 3.0 is the current US liver transplant allocation score, adopted by UNOS in 2023. It keeps MELD-Na's four variables — bilirubin, INR, creatinine and sodium — and adds serum albumin plus 1.33 points for female sex, while lowering the creatinine cap from 4.0 to 3.0 mg/dL. It discriminates 90-day waiting-list mortality slightly better than MELD-Na (C-statistic 0.869 versus 0.862) and, more importantly, corrects a sex disparity: it reclassified a net 8.8% of patients who died to a higher MELD tier, particularly women.

On this page

  • Formula
  • Interpreting the result
  • Inputs
  • What it returns
  • How it is calculated
  • Facts & figures
  • Evidence
  • How it compares
  • Pearls & pitfalls
  • Critical actions
  • Why it exists
  • About the creator
  • Limitations
  • If you are the patient
  • FAQ
  • Related calculators
  • References

Formula

MELD 3.0 = 1.33 (if female) + 4.56 × ln(bilirubin) + 0.82 × (137 − Na) − 0.24 × (137 − Na) × ln(bilirubin) + 9.09 × ln(INR) + 11.14 × ln(creatinine) + 1.85 × (3.5 − albumin) − 1.83 × (3.5 − albumin) × ln(creatinine) + 6
ln
Natural logarithm, which is why bilirubin, INR and creatinine are all floored at 1.0.
bilirubin
mg/dL, floored at 1.0. Also appears in an interaction with sodium.
Na
Serum sodium in mEq/L, bounded to 125–137.
INR
Floored at 1.0.
creatinine
mg/dL, floored at 1.0 and capped at 3.0. Set to 3.0 after two or more dialysis sessions in the past week.
albumin
g/dL, bounded to 1.5–3.5. Also appears in an interaction with creatinine.
1.33 (female)
Added for female sex, correcting the under-prioritisation of women under earlier MELD versions.
  • Two interaction terms make this genuinely different in shape from MELD-Na: sodium with bilirubin, and albumin with creatinine. The effect of each variable therefore depends on the others, which is why MELD 3.0 cannot be derived from a MELD-Na score.
  • The result is rounded and bounded to 6–40.
  • The creatinine cap is 3.0 mg/dL, not 4.0. This is one of the easiest differences from MELD-Na to overlook.
  • For SI units, divide bilirubin in µmol/L by 17.1, creatinine in µmol/L by 88.4, and albumin in g/L by 10 — or use the unit toggles on this page.

Interpreting the result

Read it exactly as you would MELD-Na — as a 90-day mortality estimate driving waiting-list priority — but do not compare a MELD 3.0 with a MELD-Na as though they were the same measurement. The scales coincide at 6 and 40 and the interaction terms mean the two diverge unpredictably in between, so a patient whose score 'went up' when a centre switched formulas has not necessarily deteriorated. As with all MELD variants, the trend across repeated measurements carries more information than any single value, and a low score does not exclude significant frailty, sarcopenia or refractory ascites, none of which appear in the equation.

ScoreBandWhat it meansAction
≤ 9Low severityLow 90-day waiting-list mortalityTreat the underlying liver disease; no transplant-specific action from the score alone
10–19Moderate severityIntermediate 90-day mortality; rising waiting-list priorityTransplant evaluation; identify and treat precipitants
20–29High severityHigh 90-day mortalityActive transplant work-up; escalate level of care
30–40Very high severityVery high 90-day mortality without transplantUrgent transplant assessment and critical care input

Scroll the table sideways for every column.

What the MELD 3.0 needs (7 inputs)

Total bilirubin (mg/dL)
Floored at 1.0, as in every MELD variant. It also appears in an interaction term with sodium, so its effect on the score depends on how hyponatraemic the patient is.
INR
Floored at 1.0. Still distorted by warfarin and direct oral anticoagulants — MELD 3.0 did not solve that, and MELD-XI remains the alternative for anticoagulated patients.
Serum creatinine (mg/dL)
Floored at 1.0 and capped at 3.0 — lower than MELD-Na's 4.0. The reduced cap limits how much renal dysfunction can dominate the score, part of the same correction that addresses the sex disparity.
Serum sodium (mEq/L)
Bounded to 125–137. It appears both on its own and in an interaction with bilirubin, so the penalty for hyponatraemia is not a fixed number of points.
Serum albumin (g/dL)
New in MELD 3.0, bounded to 1.5–3.5. It enters both directly and in an interaction with creatinine. Note that recent albumin infusion — after large-volume paracentesis, for instance — will raise the measured value without any change in synthetic function.
Sex
Female sex adds 1.33 points. This is a deliberate equity correction, not a biological claim about disease severity: it compensates for the systematic under-prioritisation of women under earlier MELD versions.
Dialysis in the past week
Two or more sessions of haemodialysis, or 24 hours of continuous veno-venous haemodialysis, in the previous seven days. Sets creatinine to the 3.0 cap.

Units. Enter bilirubin and creatinine in mg/dL and albumin in g/dL, or switch the unit toggles to µmol/L and g/L and the calculator converts exactly (bilirubin ÷ 17.1, creatinine ÷ 88.4, albumin ÷ 10). Sodium is in mEq/L, numerically identical to mmol/L. Remember the creatinine cap here is 3.0 mg/dL, which is 265 µmol/L — lower than MELD-Na's 4.0 mg/dL.

What it returns

MELD 3.0 score (6–40)
An integer on the same 6–40 scale as MELD-Na, so the numbers look comparable — but they are not interconvertible, and the same patient can score differently under each.
Estimated 90-day mortality band
The severity band the score falls into. MELD 3.0 was fitted against 90-day waiting-list mortality specifically.

How it is calculated

Kim and colleagues refitted the MELD model on all US liver transplant waiting-list candidates from January 2016 to December 2018, testing candidate variables against 90-day mortality. Female sex and serum albumin were retained as independent predictors beyond the MELD-Na components, and two interaction terms were introduced because the effect of sodium proved to depend on bilirubin and the effect of albumin on creatinine. Lowering the creatinine cap to 3.0 was part of the same reweighting: creatinine had been carrying more of the score than its true prognostic contribution justified, and that over-weighting was one of the mechanisms by which women — who tend to have lower creatinine at equivalent renal function because of lower muscle mass — were disadvantaged.

Facts & figures

What changed from MELD-Na
FeatureMELD-NaMELD 3.0
Bilirubin, INR, creatinine, sodiumYesYes
Serum albuminNoYes, bounded 1.5–3.5 g/dL
Female sex adjustmentNone+1.33 points
Creatinine cap4.0 mg/dL3.0 mg/dL
Interaction termsSodium × MELDSodium × bilirubin; albumin × creatinine
Score range6–406–40
US allocation statusSupersededCurrent since 2023

Scroll the table sideways for every column.

Derivation performance (Kim 2021)
MetricValue
Derivation cohortAll US waiting-list candidates, Jan 2016 – Dec 2018
C-statistic, MELD 3.00.869
C-statistic, MELD-Na0.862 (P < .01)
Decedents reclassified to a higher tierNet 8.8%, particularly women
Simulated waiting-list deaths7,788 with MELD 3.0 vs 7,850 with MELD-Na (P = .02)

Scroll the table sideways for every column.

Evidence

Derivation — US national transplant registry

2021

All adult liver transplant waiting-list candidates registered in the US national registry between January 2016 and December 2018, refitting the MELD model against 90-day mortality and testing additional candidate variables.

C-statistic 0.869 for MELD 3.0 against 0.862 for MELD-Na (P < .01). MELD 3.0 correctly reclassified a net 8.8% of decedents into a higher MELD tier, affording them a higher chance of transplantation, particularly among women.

Simulation of allocation impact

2021

Simulation modelling of the same registry population, comparing waiting-list deaths under each formula.

Fewer waiting-list deaths under MELD 3.0 than MELD-Na — 7,788 versus 7,850 (P = .02) — alongside the stated aim of addressing the sex disparity in access.

External validation

2025

MELD 3.0 has since been validated in independent cohorts outside the United States, including a German clinical cohort study assessing it alongside ReMELD-Na.

Broadly reproduces the discrimination reported in the derivation work; comparative performance against locally derived alternatives varies by cohort.

How it compares

MELD 3.0 vs MELD-Na

MELD 3.0 has superseded MELD-Na for US allocation and is the one to use for current listing decisions; MELD-Na remains relevant for comparability with older literature and in jurisdictions that have not adopted the update.

MELD 3.0 adds albumin and 1.33 points for female sex, lowers the creatinine cap from 4.0 to 3.0, and introduces interaction terms between sodium and bilirubin and between albumin and creatinine. Its discrimination advantage is modest (C-statistic 0.869 versus 0.862) but it reclassified a net 8.8% of decedents to a higher tier and reduced simulated waiting-list deaths. Because of the interaction terms the two scores are not interconvertible — you cannot adjust a MELD-Na to get a MELD 3.0.

Open the MELD-Na calculator →Kim WR, Mannalithara A, Heimbach JK, et al. MELD 3.0: The Model for End-Stage Liver Disease Updated for the Modern Era. Gastroenterology. 2021;161(6):1887-1895.

MELD 3.0 vs Child-Turcotte-Pugh

Use MELD 3.0 for transplant priority and prognosis, Child-Pugh whenever guidance is written in classes — the two are not substitutes and cannot be converted.

MELD 3.0 is a fitted survival model on six objective variables plus sex; Child-Pugh is an unweighted five-item score containing two subjectively graded items. MELD 3.0 discriminates 90-day mortality far better. But drug labels, trial eligibility and procedural guidance are expressed as Child-Pugh class A/B/C, and no MELD score answers that question, so both are usually recorded in patients under assessment.

Open the Child-Turcotte-Pugh calculator →

MELD 3.0 vs MELD-XI

Use MELD-XI instead when the patient is anticoagulated, since MELD 3.0 still includes INR and is therefore still inflated by warfarin or a direct oral anticoagulant.

MELD 3.0 addressed the sex disparity and added albumin but left the INR problem untouched. In an anticoagulated patient the INR reflects the prescription as much as the liver, and the score overstates severity. MELD-XI drops INR entirely and uses bilirubin and creatinine only; it discriminates less well in patients who are not anticoagulated, so it is a targeted substitution rather than a general improvement.

Open the MELD-XI calculator →

Pearls & pitfalls

  • The creatinine cap is 3.0, not 4.0. Carrying MELD-Na's cap over is the commonest error when moving between the two.
  • The 1.33 points for female sex is an equity correction for measurement bias, not a statement that female sex worsens liver disease. Expect to have to explain that.
  • MELD 3.0 and MELD-Na are not interconvertible. The interaction terms mean the difference between them varies by patient, so a score change on switching formulas is not clinical deterioration.
  • Recent albumin infusion inflates the albumin term. A patient given albumin after large-volume paracentesis will score better without being better.
  • Albumin is bounded to 1.5–3.5 g/dL, so a profoundly low albumin contributes no more than 1.5 does.
  • INR remains distorted by anticoagulation. MELD 3.0 did not fix this — use MELD-XI when the patient is anticoagulated.
  • The 6 and 40 bounds are unchanged, so the score still cannot distinguish among the sickest patients.
  • PELD, not MELD 3.0, applies below age 12.

Critical actions

  • Confirm which formula your centre or registry is using before quoting a score, since MELD-Na and MELD 3.0 give different numbers for the same patient.
  • Check whether albumin has been infused recently before trusting the score.
  • Check whether the patient is anticoagulated before trusting an INR-driven score.
  • Track the trend across repeated measurements rather than acting on a single value.
  • Look for a precipitant whenever the score rises — infection, variceal bleeding, hepatorenal syndrome, drug injury.
  • Assess frailty and sarcopenia separately; both affect transplant outcome and neither is in the equation.
  • Do not use MELD 3.0 to judge the risk of a specific operation — use a surgical risk model that accounts for the procedure.

Why this score exists

MELD 3.0 is unusual among clinical scores in that its principal purpose was fairness rather than accuracy. The discrimination gain over MELD-Na is real but small — a C-statistic of 0.869 against 0.862 — and on its own would not have justified changing a national allocation policy. What did justify it was evidence that women were being systematically under-prioritised relative to their actual mortality, partly because creatinine understates renal impairment in people with less muscle mass. The authors' response was twofold: add variables that capture severity independently of muscle mass (albumin), and add an explicit sex term to correct the residual gap. The 1.33 points for female sex is therefore not a claim that women's liver disease is worse at equivalent laboratory values — it is a correction for a known measurement bias in the other variables.

About the creator

  • W. Ray Kim

    First author, 2021 derivation study

    Led the development of MELD 3.0, adding sex and albumin and recalibrating the coefficients to correct the disadvantage women faced under MELD-Na.

  • Allison J. Kwong

    Senior author

    Co-led the analysis of national registry data underlying the revision.

Limitations

  • Derived in a US waiting-list population between 2016 and 2018, so it reflects that era's referral patterns, aetiologies and practice.
  • The discrimination gain over MELD-Na is small in absolute terms; the substantive advance is the equity correction, which is harder to observe in an individual patient.
  • The explicit sex term requires explanation and can be misread as a biological claim rather than a correction for measurement bias.
  • Albumin is confounded by infusion, sepsis and proteinuria, none of which reflect synthetic liver function.
  • INR remains distorted by anticoagulation.
  • The 6–40 bounds still compress the extremes, so the sickest candidates remain indistinguishable at 40.
  • It omits frailty, sarcopenia and quality of life, all of which affect transplant outcome.
  • It is not validated in acute liver failure and does not apply below age 12.

If you are the patient

MELD 3.0 is the score used in the United States to decide how urgently someone needs a liver transplant, and it replaced an older version in 2023. It is worked out from five blood tests — bilirubin, INR, creatinine, sodium and albumin — plus whether you are male or female. A higher number means more advanced liver disease and a higher priority on the transplant waiting list. The reason for the update is worth knowing: under the old version, women were being placed lower on the list than their actual risk justified, largely because one of the blood tests (creatinine) tends to read lower in people with less muscle, making kidney problems look milder than they are. MELD 3.0 corrects for that, which is why the calculation includes sex. Two other things: the number can go up and down, and the direction over time matters more than any single reading; and the score does not capture everything your team considers, such as your strength, nutrition and general fitness, which they assess separately.

Frequently asked questions

What is MELD 3.0?#

MELD 3.0 is the current US liver transplant allocation score, adopted by UNOS in 2023. It builds on MELD-Na by adding serum albumin and 1.33 points for female sex, and lowering the creatinine cap from 4.0 to 3.0 mg/dL. It was fitted against 90-day waiting-list mortality in US candidates from 2016 to 2018.

What is the difference between MELD 3.0 and MELD-Na?#

MELD 3.0 adds albumin and a 1.33-point female-sex adjustment, lowers the creatinine cap from 4.0 to 3.0 mg/dL, and introduces interaction terms between sodium and bilirubin and between albumin and creatinine. Because of those interactions the two are not interconvertible — the same patient can score differently under each, and the difference varies by patient.

Why does MELD 3.0 add points for female sex?#

To correct a documented disparity in access to transplantation. Women were being under-prioritised relative to their actual mortality, partly because creatinine understates renal impairment in people with lower muscle mass. The 1.33 points is a correction for that measurement bias, not a claim that female sex makes liver disease worse at equivalent laboratory values.

Is MELD 3.0 more accurate than MELD-Na?#

Slightly, and meaningfully fairer. The C-statistic for 90-day mortality was 0.869 against 0.862 for MELD-Na (P < .01) — a small absolute gain. The more consequential finding was that MELD 3.0 correctly reclassified a net 8.8% of patients who died into a higher tier, particularly women, and simulation showed fewer waiting-list deaths (7,788 versus 7,850).

What is the creatinine cap in MELD 3.0?#

3.0 mg/dL, lower than MELD-Na's 4.0 mg/dL. Carrying the old 4.0 cap across is one of the most common errors when switching between the two formulas. Two or more dialysis sessions in the previous week set creatinine to the cap.

Does MELD 3.0 include albumin?#

Yes — albumin is one of the two variables added over MELD-Na, bounded to 1.5–3.5 g/dL, and it also appears in an interaction term with creatinine. Be aware that recent albumin infusion, for example after large-volume paracentesis, raises the measured value without any change in synthetic liver function.

When did UNOS start using MELD 3.0?#

UNOS implemented MELD 3.0 for liver allocation in 2023, following its publication in Gastroenterology in 2021. MELD-Na had been in use since 2016, and the original MELD since 2002.

Can MELD 3.0 be used in children?#

No. MELD variants apply from age 12 upward; PELD, the Pediatric End-Stage Liver Disease score, is used for younger candidates and includes growth failure and age under one year rather than sodium and sex.

Related calculators

  • MELD-Na — Assesses the severity of chronic liver disease
  • Child-Pugh Score — Assesses the prognosis of chronic liver disease, mainly cirrhosis
  • MELD-XI — MELD excluding INR — for anticoagulated patients
  • PELD / CR Score — Pediatric end-stage liver disease
  • ALBI Grade — Albumin-bilirubin liver function grade in HCC
  • Milan Criteria — Liver transplant eligibility in hepatocellular carcinoma

References

Original / primary reference

  1. Kim WR, Mannalithara A, Heimbach JK, et al. MELD 3.0: The Model for End-Stage Liver Disease Updated for the Modern Era. Gastroenterology. 2021;161(6):1887-1895.

Preceding formulas

  1. Kim WR, Biggins SW, Kremers WK, et al. Hyponatremia and mortality among patients on the liver-transplant waiting list. N Engl J Med. 2008;359(10):1018-1026.
  2. Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients with end-stage liver disease. Hepatology. 2001;33(2):464-470.

External validation

  1. Validation of MELD 3.0 and ReMELD-Na scoring systems: a German clinical cohort study. 2025.

Last updated July 29, 2026. Clinical knowledge base written and curated by GastroAGI Team from primary medical literature.

Written from primary literature and not yet independently clinically reviewed.

For use by qualified healthcare professionals. This calculator supports clinical judgement and does not replace it.