MELD Score and Child-Pugh Score Explained: What the Numbers on Your Cirrhosis Report Mean

A MELD score estimates how likely a person with cirrhosis is to die within three months, on a scale of 6 to 40, using three blood tests: bilirubin, creatinine and INR. Among 3,437 patients awaiting transplant, three-month mortality was 1.9% below MELD 9 and 71.3% at MELD 40 or above. Child-Pugh grades the same liver as class A, B or C from five bedside measures.
— Dr Chetan Kalal, DM Hepatology, Gleneagles Hospital Mumbai

Picture a composite of a family I see most weeks at Gleneagles Hospital: a 48-year-old schoolteacher from Thane with cirrhosis from hepatitis B, and his wife holding a transplant work-up report with “MELD-Na 22, Child-Pugh C (11 points)” circled in red. Nobody has told them what either number means. He feels reasonably well. She has read online that “MELD above 15 means transplant” and has not slept in two nights.

That conversation happens often enough to deserve a written answer. The Child-Pugh score and the MELD score are the two numbers every hepatologist uses to describe how sick a cirrhotic liver is — they decide who is referred for transplant, who is listed and, in many systems, who receives an organ first — yet they are rarely explained to the person they describe. This article explains what each score is built from, what your number means for the months ahead, when it should trigger a transplant discussion, and what neither can tell you. If you have been told you have cirrhosis, read it with the report in front of you.

What is the Child-Pugh score?

The Child-Pugh score is a bedside grading of cirrhosis severity built from five items. Pugh and colleagues described it in 1973 in a study of surgery for bleeding oesophageal varices, to predict who would survive the operation. It has outlived the operation.

Each of the five items — serum bilirubin, serum albumin, clotting (prothrombin time, now usually expressed as INR), ascites, and hepatic encephalopathy — earns 1, 2 or 3 points, for a total of 5 to 15. A total of 5 or 6 is class A, 7 to 9 is class B, and 10 to 15 is class C. Class A is a liver still coping; class C is one failing on several fronts.

Any doctor with a blood report and a stethoscope can calculate it in a minute. Its weakness is that two of the five items depend on the examiner’s eye — how much ascites is “moderate”? — and it has a ceiling: a bilirubin of 4 and a bilirubin of 30 both score 3 points, though their outlook is very different. MELD was built to fix exactly these limits.

The class still matters because it tracks the shape of the disease. A systematic review of 118 studies describes cirrhosis as running through a compensated phase, in which life expectancy is long, and a decompensated phase in which survival falls sharply. Child-Pugh A usually sits in the first; B and C mark the second, and the complications that move you across are described on my portal hypertension page.

What is the MELD score and what does it use?

MELD — the Model for End-stage Liver Disease — uses three laboratory values, bilirubin, creatinine and INR, fed through a formula to produce a number from 6 to 40. Nothing in it depends on an examiner’s opinion: the same blood results give the same score in Mumbai or Rochester.

It was never meant to be a transplant score. The Mayo Clinic group built it to predict survival after a TIPS procedure (a shunt placed inside the liver to lower portal pressure). In 2001 Kamath and colleagues tested it across four independent patient groups, from decompensated inpatients to stable outpatients. It predicted death within three months with a concordance statistic of 0.87 in the hospitalised group and 0.80 in outpatients, and adding ascites, variceal bleeding or spontaneous bacterial peritonitis improved the prediction by less than 0.01 to 0.03.

The decisive test came in 2003. Wiesner and colleagues applied MELD prospectively to 3,437 adults placed on the US transplant waiting list for chronic liver disease. Twelve per cent died within three months, and mortality climbed in proportion to the listing score. Head to head, MELD ranked the patients better than Child-Pugh — an area under the curve of 0.83 against 0.76. The United States moved allocation to MELD, and most of the world followed. (Why a kidney test inside a liver score? In advanced cirrhosis the kidneys buckle first, and a rising creatinine is one of the surest signs that time is short.)

What does my MELD score mean?

Your MELD score is a three-month weather forecast, not a life sentence. In the 2003 waiting-list study, patients with a MELD below 9 had a three-month mortality of 1.9%; patients with a score of 40 or above had a mortality of 71.3%. Between those poles the curve is not a straight line; risk accelerates through the twenties.

In clinic language: single digits means a damaged but stable liver, and the work is prevention. The mid-teens to low twenties is where a transplant evaluation belongs. The high twenties and thirties usually describe a patient in hospital, often in intensive care, where the question is no longer whether to transplant but whether it can be done in time.

MELD-Na. The original score had a blind spot: serum sodium, which predicts death on its own. In 2008, Kim and colleagues analysed 6,769 adults registered for transplant in the United States in 2005: each MELD point raised the hazard of death 1.21-fold, and each 1 mmol/L fall in sodium between 125 and 140 raised it 1.05-fold. The sodium effect was strongest at low MELD scores. Applied to 2006 data, 32 of the 477 patients who died on the list (7%) would have been ranked meaningfully higher. MELD-Na is what most Indian transplant centres report today.

MELD 3.0. In 2021 the same group updated the model on US registry data from 2016 to 2018. MELD 3.0 adds female sex and serum albumin and caps creatinine at 3.0 mg/dL. Discrimination improved only modestly (0.869 against 0.862), but the model reclassified a net 8.8% of patients who died into a higher priority tier — mostly women, previously under-scored — and in simulation meant fewer waiting-list deaths, 7,788 against 7,850. It is being adopted in the United States; the same patient carries a different number under each version, so check which one your report uses.

When does a MELD score mean I need a liver transplant?

A transplant is worth doing when the risk of the operation is lower than the risk of living without it, and MELD is the best single measure of that second risk. The clearest evidence comes from Merion and colleagues, who followed 12,996 adults on the US waiting list between 2001 and 2003 and compared those transplanted with comparable candidates who were not. Overall, recipients of a deceased-donor liver had a 79% lower risk of death. But the benefit was not evenly spread. At MELD 18–20, mortality was 38% lower among recipients; at the maximum score of 40, 96% lower. At low scores the picture reversed: in the first year after transplant, patients transplanted at MELD 6–11 had 3.64 times the mortality of those who waited, and at MELD 12–14, 2.35 times.

That is the origin of the “MELD 15” you may have read about. It is not a guideline rule but a reading of this evidence: below about 15 a transplant is more likely to shorten life than lengthen it in the first year, the benefit is not demonstrable until the high teens, and from around 18 it is real and grows with every point. A transplant at MELD 10 is not an early transplant. It is a harmful one.

Two things follow. Referral and listing are different steps: referral to a transplant hepatologist should happen at the first decompensation — the first ascites, bleed or confusion — whatever the score, because the work-up takes weeks and the score rarely waits. And the score is not the only door. Liver cancer, recurrent encephalopathy and uncontrollable bleeding carry a risk the blood tests miss, and transplant programmes account for them separately.

How organs are allocated once you are listed is set by the state transplant authority; the rules vary by state, and your transplant team will explain how the list works where you are treated. In India a large share of transplants come from living donors — a relative who gives part of their liver — and there the question becomes “is this the right time, and is the donor safe”. My pages on liver transplant surgery and what a transplant specialist does cover that decision.

What the scores cannot tell you

The scores are snapshots, and the liver is a moving picture. A MELD of 24 during an infection may be 17 three weeks later. A patient with alcohol-related cirrhosis who stops drinking completely can watch the score fall over months — some fall out of the transplant range altogether, the best outcome of all; see my alcohol-related liver disease page. The reverse is also true: a MELD of 12 can become 28 after one variceal bleed. No single reading should reassure or alarm you.

The components can mislead. Creatinine rises with dehydration and water tablets and falls when muscle is lost, so a thin, wasted patient with sarcopenia can carry a deceptively low MELD while the kidneys quietly fail — one more reason nutrition in cirrhosis is not a side issue. INR is distorted by blood thinners and vitamin K deficiency; bilirubin by bile-duct problems unrelated to the cirrhosis.

And neither score measures how you live. Disabling tiredness, itching, repeated admissions for encephalopathy, ascites tapped every fortnight — these wreck a life without moving the MELD by a point. Nor does either apply to sudden liver failure in a previously healthy person, a different disease covered on my acute liver failure page. If your report carries a number you do not understand, consult your doctor rather than a search engine, and ask three questions: which version of MELD is this, what was it last time, and what are we doing to bring it down.

What I would say to that family

MELD-Na 22 and Child-Pugh C put the schoolteacher in the range where transplant lengthens life. The right response is neither panic nor delay: complete the evaluation now, treat the hepatitis B, correct the nutrition, screen for varices and liver cancer, and find out whether a living donor exists. If the score falls with treatment, the transplant can wait. If it climbs, the preparation is already done. The number is where the conversation starts, not where it ends.

References

1. Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg. 1973;60(8):646–649. PMID 4541913. doi:10.1002/bjs.1800600817
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. PMID 11172350. doi:10.1053/jhep.2001.22172
3. Wiesner R, Edwards E, Freeman R, et al. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology. 2003;124(1):91–96. PMID 12512033. doi:10.1053/gast.2003.50016
4. 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. PMID 18768945. doi:10.1056/NEJMoa0801209
5. 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.e4. PMID 34481845. doi:10.1053/j.gastro.2021.08.050
6. Merion RM, Schaubel DE, Dykstra DM, Freeman RB, Port FK, Wolfe RA. The survival benefit of liver transplantation. Am J Transplant. 2005;5(2):307–313. PMID 15643990. doi:10.1111/j.1600-6143.2004.00703.x
7. D’Amico G, Garcia-Tsao G, Pagliaro L. Natural history and prognostic indicators of survival in cirrhosis: a systematic review of 118 studies. J Hepatol. 2006;44(1):217–231. PMID 16298014. doi:10.1016/j.jhep.2005.10.013