About the Model for End-Stage Liver Disease with Sodium (MELD-Na)
MELD-Na estimates three-month mortality in chronic liver disease from four laboratory values — total bilirubin, INR, creatinine and serum sodium — and returns a score from 6 to 40. It is the MELD score with a correction for hyponatraemia, added because patients with a low sodium die on the transplant waiting list faster than their MELD score alone predicts. Roughly, a score of 9 or below carries about 1.9% three-month mortality, 10–19 about 6%, 20–29 about 19.6%, 30–39 about 52.6%, and 40 about 71.3%.
Formula
MELD(i) = 10 × [ 0.957 × ln(creatinine) + 0.378 × ln(bilirubin) + 1.120 × ln(INR) + 0.643 ]
MELD-Na = MELD(i) + 1.32 × (137 − Na) − [ 0.033 × MELD(i) × (137 − Na) ]- ln
- Natural logarithm — the reason every input has a floor of 1.0.
- creatinine
- mg/dL, floored at 1.0 and capped at 4.0. Set to 4.0 if the patient has had dialysis twice in the past week.
- bilirubin
- Total bilirubin in mg/dL, floored at 1.0.
- INR
- Floored at 1.0.
- Na
- Serum sodium in mEq/L, bounded to 125–137 before use.
- MELD(i) is rounded to the nearest integer and bounded to 6–40 before the sodium term is applied.
- The sodium correction is only applied when MELD(i) is above 11, per OPTN policy. Below that, MELD-Na equals MELD.
- The final result is rounded and bounded to 6–40 again.
- For SI units, divide bilirubin in µmol/L by 17.1 and creatinine in µmol/L by 88.4 before entering — or use the unit toggles on this page, which do it exactly.
Interpreting the result
Read the score as a three-month mortality estimate for a population of similar patients, then look at the trend. A single MELD-Na is much less informative than three taken over six weeks: a stable 22 and a 22 that was 14 a month ago are different clinical situations, and only the second is an emergency. The bands below come from waiting-list cohorts, so they describe patients under transplant assessment; a hospitalised patient at the same score carries a higher risk than the table suggests, and the original validation work found exactly that.
| Score | Band | What it means | Action |
|---|---|---|---|
| ≤ 9 | Low | About 1.9% three-month mortality | Treat the underlying disease; no transplant-specific action from the score |
| 10–19 | Moderate | About 6% three-month mortality | Refer for transplant evaluation; look for and treat precipitants |
| 20–29 | High | About 19.6% three-month mortality | Active transplant work-up; escalate level of care |
| 30–39 | Very high | About 52.6% three-month mortality | Urgent transplant assessment; intensive monitoring |
| 40 | Maximum | About 71.3% three-month mortality | Critical care; the score is capped, so the patient may be sicker than 40 suggests |
Scroll the table sideways for every column.
What the MELD-Na needs (5 inputs)
- Total bilirubin (mg/dL)
- Serum total bilirubin. Values below 1.0 are treated as 1.0, because the formula takes a natural logarithm and would otherwise return a negative contribution.
- INR
- International normalised ratio of prothrombin time. Also floored at 1.0. Note that INR is affected by warfarin and other anticoagulants, which is a known distortion rather than a feature — MELD-XI exists for that situation.
- Serum creatinine (mg/dL)
- Floored at 1.0 and capped at 4.0. The cap matters: renal failure beyond that point does not increase the score further, which is a deliberate ceiling rather than a claim that it stops mattering.
- Serum sodium (mEq/L)
- Bounded to 125–137 mEq/L. Values outside that window are clamped to the boundary rather than extrapolated, so a sodium of 118 is scored as 125.
- Dialysis in the past week
- Two or more sessions of haemodialysis, or 24 hours of continuous veno-venous haemodialysis, within the previous seven days. If yes, creatinine is set to 4.0 regardless of the measured value.
Units. Enter bilirubin and creatinine in mg/dL, or switch the unit toggles to µmol/L and the calculator converts exactly (bilirubin ÷ 17.1, creatinine ÷ 88.4). Sodium is in mEq/L, which is numerically identical to mmol/L. Doing the conversion by hand before entering mg/dL gives the same answer; the toggles just remove a rounding step.
What it returns
- MELD-Na score (6–40)
- An integer. The floor of 6 and ceiling of 40 are allocation conventions, not the natural limits of the equation — a patient at 40 may be considerably sicker than one at 40.
- Estimated three-month mortality
- The band the score falls into, with the mortality observed for that band in waiting-list cohorts. This is a population estimate, not an individual prediction.
How it is calculated
MELD is a Cox proportional-hazards model, and the coefficients above are the published hazard weights for each variable. Because each term is a logarithm, the score responds steeply to changes at low values and much more gently at high ones: a bilirubin moving from 1 to 2 mg/dL adds more to the score than a move from 10 to 11. The sodium term is a later addition fitted on waiting-list data, and it is interactive rather than additive — the `0.033 × MELD × (137 − Na)` term means the penalty for hyponatraemia shrinks as the underlying MELD rises, because at a high MELD the sodium is telling you less that the other variables have not already said.
Facts & figures
| Variant | Inputs | What it added | Where it is used |
|---|---|---|---|
| Original MELD (2001) | Bilirubin, INR, creatinine, plus an aetiology term | The model itself — built to predict survival after TIPS, then repurposed for allocation | Superseded; the aetiology term was dropped early in allocation use |
| MELD (allocation form) | Bilirubin, INR, creatinine | Removal of the aetiology term, which was subjective and disputed | The pre-sodium allocation standard; still the score much older literature reports |
| MELD-Na | Adds serum sodium | Corrects for hyponatraemia, which predicts waiting-list death independently of the other three | Became the US allocation standard in January 2016 |
| MELD 3.0 | Adds female sex and albumin; recalibrated | Corrects the disadvantage MELD-Na imposed on women, largely through creatinine | Adopted for US allocation in 2023 |
| MELD-XI | Bilirubin and creatinine only | Removes INR entirely | Patients on anticoagulation, where INR reflects the drug rather than the liver |
Scroll the table sideways for every column.
All are the same construct at different vintages, which is why they share this page's calculator rather than each having their own. Pick the variant your allocation system currently uses, and note which one a published cohort reported before comparing scores across studies.
Evidence
Derivation — MELD, Mayo Clinic
2001MELD was originally built to predict three-month survival after transjugular intrahepatic portosystemic shunt (TIPS), then repurposed for waiting-list allocation. It was validated across four independent datasets: patients hospitalised with hepatic decompensation, ambulatory non-cholestatic cirrhosis, primary biliary cirrhosis, and an unselected 1980s cirrhosis cohort.
Concordance statistic 0.78 (95% CI 0.74–0.81) for three-month survival in the derivation cohort, and 0.87 in hospitalised patients, 0.80 in ambulatory non-cholestatic cirrhosis, 0.87 in primary biliary cirrhosis and 0.78 in the historical cohort.
Sodium correction — OPTN waiting-list cohorts
2008 · n = 6,769Adult candidates for primary liver transplantation registered with the Organ Procurement and Transplantation Network in 2005 (derivation) and 2006 (validation). The 2005 cohort held 6,769 registrants, of whom 1,781 were transplanted and 422 died within 90 days of listing; 477 died within three months in the 2006 validation cohort.
At any given MELD score, waiting-list mortality rose as serum sodium fell, and the effect was most pronounced below 125 mEq/L. This is the finding the MELD-Na correction encodes.
How it compares
MELD-Na vs MELD 3.0
MELD 3.0 has superseded MELD-Na for US allocation — if you are asking about listing priority today, MELD 3.0 is the current formula.
MELD 3.0 keeps the same four variables, adds serum albumin and a 1.33-point correction for female sex, lowers the creatinine cap from 4.0 to 3.0, and reweights the sodium interaction. The female-sex term exists because women were shown to be systematically disadvantaged by earlier MELD versions relative to their actual mortality. MELD-Na remains useful for comparison with older literature and is still used outside the United States.
MELD-Na vs Child-Turcotte-Pugh
Use MELD-Na for prognosis and transplant priority, and Child-Turcotte-Pugh for anything written in terms of its classes — drug dosing, trial eligibility, and procedural guidance are still expressed as class A/B/C.
MELD-Na is continuous, entirely objective, and validated for three-month mortality. Child-Turcotte-Pugh puts two subjectively graded variables (ascites, encephalopathy) into a five-item score with coarse bands, which is why it was replaced for allocation. But decades of literature, drug labels and eligibility criteria are written in Child-Pugh classes, so it has not gone away and converting between the two is not possible.
MELD-Na vs MELD-XI
Switch to MELD-XI when the patient is anticoagulated — it excludes INR, which is the variable that anticoagulation corrupts.
A patient on warfarin or a direct oral anticoagulant can have an INR that reflects their prescription rather than their liver, inflating MELD-Na without any change in hepatic function. MELD-XI drops INR and uses bilirubin and creatinine only. It is less discriminating in patients who are not anticoagulated, so it is a substitution for a specific situation rather than a general improvement.
Pearls & pitfalls
- Every lab is floored at 1.0. Entering a bilirubin of 0.4 changes nothing, which surprises people who expect a very well patient to score below 6.
- Creatinine is capped at 4.0, so a patient with a creatinine of 8 scores identically to one at 4. Answer the dialysis question rather than entering a very high creatinine — it produces the same cap by the intended route.
- Sodium is clamped to 125–137. A sodium of 118 is scored as 125, so the score understates the risk in profound hyponatraemia. That is a known limitation of the correction, not a bug in the arithmetic.
- The sodium adjustment does nothing at or below MELD 11. A hyponatraemic patient with a low MELD gets no uplift, by policy.
- INR is distorted by warfarin and direct oral anticoagulants, which inflates the score without any change in liver function. MELD-XI was built for exactly this case.
- The score is capped at 40, and a meaningful number of patients are sicker than 40. Treat a 40 as a floor on severity rather than a description of it.
- MELD-Na and MELD 3.0 are not interchangeable. MELD 3.0 is the current US allocation formula; it adds albumin and a female-sex correction and lowers the creatinine cap to 3.0.
Critical actions
- Check the trend, not just the value. Repeat the score rather than acting on one snapshot.
- Look for a precipitant every time the score jumps — infection, variceal bleeding, hepatorenal syndrome, alcohol, drug injury.
- Refer to a transplant centre once the score is consistently in double figures rather than waiting for a threshold; assessment takes time the patient may not have.
- Check whether the patient is anticoagulated before trusting an INR-driven rise.
- Do not use MELD-Na to decide whether a specific operation is safe. Use a surgical risk model built for that question.
- Remember the score says nothing about quality of life, sarcopenia or frailty, all of which affect transplant outcome and none of which appear in the equation.
Why this score exists
MELD was not designed for organ allocation. It was built at the Mayo Clinic to predict survival after TIPS, and it was adopted for the waiting list in 2002 because it happened to solve a different problem well: it ranked patients by urgency using only laboratory values, replacing a system in which subjective assessment of ascites and encephalopathy, and time already spent waiting, drove priority. That origin explains its shape — it was never intended to answer questions about specific operations, or about acute liver failure, and its authors have been consistent that a score built to rank a waiting list is being asked to do something else when it is used to make an individual prognostic promise.
About the creator
First author, 2001 original MELD study
Published the model for end-stage liver disease that this score extends.
Senior author of MELD; first author of the 2008 sodium study
Established that hyponatraemia independently predicts waiting-list mortality, which is the finding that added sodium to the model.
Limitations
- Derived and validated in waiting-list and referral populations. Applying it to unselected inpatients or to primary care overestimates how well it discriminates.
- Not validated in acute liver failure, where the King's College criteria and other tools are appropriate.
- The 6 and 40 bounds compress the extremes: the sickest patients are indistinguishable from one another at 40.
- It omits everything that is not a laboratory value — sarcopenia, frailty, ascites refractory to treatment, encephalopathy, quality of life, and the patient's trajectory.
- Creatinine is a poor marker of renal function in cirrhosis, where low muscle mass depresses it, so renal impairment tends to be under-represented.
- Laboratory-to-laboratory variation in INR and bilirubin assays produces real score differences between centres for the same patient.
If you are the patient
MELD-Na is a number between 6 and 40 that your liver team calculates from four blood tests. A higher number means your liver disease is more advanced, and it is used to decide how urgently someone needs a liver transplant — the higher your score, the higher your priority on the waiting list. It is worked out purely from blood results, so it cannot be influenced by how well you seem on the day, which is deliberate: it makes the waiting list fairer. Two things are worth knowing. First, the direction matters more than any single number — your team will watch whether it is rising, falling or steady. Second, the score does not capture everything: it says nothing about how you feel, your strength, or your quality of life, all of which your team considers separately. If your score has changed, ask what caused it, because a sudden rise usually has a treatable trigger such as an infection.
Frequently asked questions
What is the difference between MELD and MELD-Na, and what changed in 2016?#
MELD-Na is MELD with a correction for serum sodium. Hyponatraemia predicts death on the waiting list independently of bilirubin, INR and creatinine, so patients with a low sodium were dying at a rate their plain MELD score understated. Adding sodium corrected that, and MELD-Na became the US allocation standard in January 2016. A cohort published before then almost certainly reports plain MELD, which matters when comparing scores across studies.
What was in the original 2001 MELD score?#
Bilirubin, INR and creatinine — as now — plus a term for the aetiology of the liver disease. The original model was built to predict survival after transjugular intrahepatic portosystemic shunt and only later repurposed for waiting-list allocation. The aetiology term was dropped early in that transition, because assigning a cause is a judgement rather than a measurement and it was open to dispute in a setting where the score determines priority.
What is the MELD-Na score?#
MELD-Na is the Model for End-Stage Liver Disease with a sodium correction. It estimates three-month mortality in chronic liver disease from total bilirubin, INR, creatinine and serum sodium, returning a score from 6 to 40, and it is used to rank patients by urgency on the liver transplant waiting list.
How is the MELD-Na score calculated?#
First MELD(i) = 10 × [0.957 × ln(creatinine) + 0.378 × ln(bilirubin) + 1.120 × ln(INR) + 0.643], with each lab floored at 1.0 and creatinine capped at 4.0. Then, if that result is above 11, MELD-Na = MELD(i) + 1.32 × (137 − sodium) − [0.033 × MELD(i) × (137 − sodium)], with sodium bounded to 125–137. The result is rounded and bounded to 6–40.
What MELD score is needed for a liver transplant?#
There is no single threshold. Patients are usually referred for transplant assessment once the score reaches the mid-teens, and in most US regions the score at which organs are actually allocated is considerably higher and varies by region and over time. The score determines priority among listed candidates rather than eligibility, and listing decisions also weigh frailty, comorbidity, substance use and social support — none of which appear in the score.
What does a MELD score of 20 mean?#
A MELD-Na of 20 falls in the 20–29 band, which carried roughly 19.6% three-month mortality in waiting-list cohorts. That figure describes a population of similar patients rather than an individual, and it is derived from patients under transplant assessment — a hospitalised patient with the same score has a higher risk than the band suggests.
Is MELD-Na the same as MELD 3.0?#
No. MELD 3.0 is the current US allocation formula and has superseded MELD-Na. It adds serum albumin and 1.33 points for female sex, and lowers the creatinine cap from 4.0 to 3.0. MELD-Na remains in use elsewhere and in older literature.
Why does serum sodium matter in MELD?#
Because at any given MELD score, patients with a lower serum sodium died faster on the waiting list. Analysis of OPTN registrants from 2005 and 2006 showed the effect clearly, and it was most marked below 125 mEq/L. MELD-Na adds a correction so that hyponatraemia raises the score rather than being invisible to it.
Can MELD-Na be used in a patient on warfarin?#
Not reliably. Anticoagulation raises the INR independently of liver function, which inflates MELD-Na. MELD-XI was developed for this situation: it drops INR and uses bilirubin and creatinine only.
Does a MELD score of 40 mean the maximum severity?#
It means the maximum the score can express, not the maximum severity. The 6–40 bounds are allocation conventions, so patients considerably sicker than 40 all score 40. Treat a 40 as a floor on severity rather than a full description of it.
What is the lowest possible MELD score?#
Six. Because every laboratory value is floored at 1.0, a patient with entirely normal results scores 6.43 before rounding, which is bounded to 6. There is no MELD below 6.
Further reading
- MELD-Na Score: How to Interpret It and When It Changes the Transplant ConversationA case-led piece on what happens when the sodium moves and the MELD does not — the narrative counterpart to the reference material on this page.
References
Original / primary reference
- 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.
- 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.