The Model for End-Stage Liver Disease, MELD, is a number built from blood tests that estimates how urgently a person with advanced liver disease needs a transplant. It was designed to replace waiting time with medical urgency: the sicker the patient, the higher the score, and the higher the priority for a donated liver. The original score used three labs, bilirubin, INR, and creatinine. The current US allocation score, MELD 3.0, adds sodium, albumin, and a sex term, and it is the score transplant centres use now. This guide explains all three versions, the clamps that keep any one lab from dominating, and what the score does and does not mean for one person.

Classic MELD: three labs, one logarithm

Classic MELD is 10 times (0.957 times the natural log of creatinine, plus 0.378 times the natural log of bilirubin, plus 1.120 times the natural log of INR, plus 0.643), rounded to the nearest whole number. The logarithms matter: they compress the scale, so doubling a lab value adds a fixed number of points rather than doubling the score, which keeps very high values from overwhelming the rest.

The worked example from the calculator: creatinine 1.2, bilirubin 2.0, INR 1.5. The logs are 0.182, 0.693, and 0.405. Weighted, they give 0.174 plus 0.262 plus 0.454, plus 0.643, or 1.534, times 10, or about 15. A healthy person with all values at 1.0 scores 6, the floor, because the logs of 1.0 are zero and only the constant remains.

The clamps are part of the formula, not an afterthought. Bilirubin and INR are floored at 1.0, so values below 1.0 count as 1.0. Creatinine is floored at 1.0 and capped at 4.0, and a patient on dialysis has creatinine set to 4.0 regardless of the measured value. Without the cap, kidney failure alone could drive the score to the top, swamping the liver measures the score exists to summarise.

MELD-Na: adding sodium

MELD-Na, adopted for US allocation in January 2016, adds serum sodium, because low sodium in cirrhosis independently predicts death. The adjustment is MELD plus 1.32 times (137 minus sodium), minus 0.033 times MELD times (137 minus sodium), applied only when MELD is above 11, with sodium clamped between 125 and 137. A sodium of 125 on a MELD of 20 adds about 11 points; a sodium of 137 adds nothing.

The interaction term, minus 0.033 times MELD times the sodium gap, keeps the sodium addition from growing without bound at high MELD scores. Sodium matters most in the middle range, where it separates patients whose liver labs look similar but whose prognosis differs. At very high MELD, the liver labs already dominate, and sodium adds little.

MELD-Na was a real improvement, but it carried a bias: creatinine reflects muscle mass, so women, with less muscle on average, scored lower than men at the same severity, and waited longer. That bias is what MELD 3.0 was built to correct.

MELD 3.0: the current US allocation score

MELD 3.0, adopted by OPTN for US allocation in July 2023, adds albumin and a female sex term, and refits every coefficient. The formula is 1.33 for female sex, plus 4.56 times log bilirubin, plus 0.82 times (137 minus sodium), minus 0.24 times (137 minus sodium) times log bilirubin, plus 9.09 times log INR, plus 11.14 times log creatinine, plus 1.85 times (3.5 minus albumin), minus 1.83 times (3.5 minus albumin) times log creatinine, plus 6, rounded, and clamped between 6 and 40.

The changes from classic MELD are deliberate. The creatinine cap falls from 4.0 to 3.0, reducing the weight of kidney function. Albumin enters because low albumin marks malnutrition and poor synthetic function. The 1.33 female term corrects the muscle-mass bias. The interaction terms let sodium matter more when bilirubin is low and albumin matter more when creatinine is low, reflecting how the risks combine.

The clamps in MELD 3.0 are: bilirubin and INR floored at 1.0, creatinine between 1.0 and 3.0 and set to 3.0 on qualifying dialysis, sodium between 125 and 137, and albumin between 1.5 and 3.5. The calculator applies all of them before evaluating the formula, which is why entering a creatinine of 8.0 and a creatinine of 3.0 give the same MELD 3.0.

The three versions, side by side

The same patient scores differently under each version, which is why the version must always be stated beside the number.

What the score means: 90-day mortality bands

MELD was built to predict death within 90 days without a transplant, and the classic bands, from the original Wiesner data, are widely quoted: a score under 10 carries roughly 2 percent 90-day mortality, 10 to 19 about 6 percent, 20 to 29 about 20 percent, 30 to 39 about 53 percent, and 40 or above about 71 percent. These are population averages from the era of the data, not predictions for one person, and modern care has improved outcomes at every band.

The band is the right way to read a score, because a single point means little and the bands mean a lot. A move from 18 to 22 crosses a band boundary and changes the urgency materially, while a move from 18 to 19 does not. Transplant teams read the score with the trend, the complications, and the patient in front of them, not as a standalone verdict.

MELD does not capture everything. It has no term for ascites, encephalopathy, or frailty, which is why Child-Pugh, a separate score that includes those clinical features, still appears in clinic discussions. MELD prioritises organs; Child-Pugh stages disease. The two answer different questions, and neither replaces clinical judgement.

Limits, and what the score cannot do

MELD is not suitable for children under 12, who are scored by PELD, or for acute liver failure, where the tempo is too fast for a chronic-disease score. It is not a general health score, and a MELD computed from labs drawn for another reason, without liver disease, means nothing. The calculator carries the same scope note: adult end-stage liver disease, for information, not for listing decisions.

Labs must be current and comparable. A creatinine measured after dialysis, a bilirubin after a transfusion, or an INR on anticoagulation all distort the score, and the calculator flags the dialysis rule explicitly because it changes the creatinine input entirely. Scores from different days should not be compared without checking what changed in the labs.

Finally, a MELD score is not a transplant guarantee at any level. Allocation also depends on blood type, body size, geography, and organ availability, and exception points exist for conditions MELD underestimates, such as hepatocellular carcinoma. The score sets medical urgency; the system sets the match.

Frequently Asked Questions

What is a MELD score?

A number from 6 to 40 estimating 90-day mortality in end-stage liver disease, computed from lab values. Higher means more urgent. The US allocation score since July 2023 is MELD 3.0.

How is the MELD score calculated?

Classic MELD is 10 times (0.957 log creatinine plus 0.378 log bilirubin plus 1.120 log INR plus 0.643), with clamps on every input. MELD 3.0 adds sodium, albumin, and a female sex term with refitted coefficients.

What is MELD-Na, and when does sodium count?

MELD-Na adds a sodium adjustment to classic MELD, applied when MELD is above 11, with sodium clamped between 125 and 137. Low sodium can add up to about 11 points. It was the US allocation score from 2016 to 2023.

Why is creatinine capped at 3.0 in MELD 3.0 but 4.0 in classic MELD?

To reduce the weight of kidney function and correct the bias against women, whose lower muscle mass gives lower creatinine at the same severity. MELD 3.0 also adds 1.33 points for female sex for the same reason.

What MELD score is needed for a liver transplant?

There is no fixed threshold: allocation depends on the score, blood type, body size, geography, and organ availability. Higher scores mean higher priority, and scores above 30 carry high 90-day mortality without transplant.

Is this the exact score my transplant centre uses?

The calculator implements the published OPTN formulas and clamps, but a centre score uses verified labs, the correct registration-age rules, and exception points where applicable. Treat the calculator result as information, not as a listing score.