From the AGA Journals

Updated MELD score adds serum albumin, female sex

View on the News

This could achieve equitable distribution

Introduction of the Model for End-Stage Liver Disease (MELD) score in 2002, consisting of objective measurements of creatinine, bilirubin, and international normalized ratio, revolutionized liver allocation in the United States. To minimize patient wait-list mortality and reduce geographic variability, further improvements to allocation system including the National Share for status 1 and Regional Share for MELD score greater than 35 in 2013, adoption of MELDNa score in 2016, and most recently the introduction of the Acuity Circles distribution system were implemented. Unfortunately, MELD tends to disadvantage women whose lower muscle mass translates to lower normal creatinine levels thereby underestimating the degree of renal dysfunction and wait-list mortality. MELD performance characteristics were also shown to be less accurate in patients with alcoholic and nonalcoholic fatty liver disease when compared with patients with hepatitis C, likely contributing to MELD’s decreasing accuracy in predicting mortality over the years with changing patient population.

Alexandra Shingina, MD, MSc, is an assistant professor of medicine in the division of gastroenterology, hepatology, and nutrition at Vanderbilt University Medical Center, Nashville, Tenn.

Dr. Alexandra Shingina

To address these deficiencies, the study by Kim and colleagues explores a new iteration of organ prioritization system – MELD 3.0 – which includes adjustments for gender, albumin level, and lowering the upper limit of creatinine to 3.0 mg/dL (from 4.0 mg/dL) with validation in a contemporary cohort of listed patients. Undoubtedly, this is a step in the right direction for gender equity in organ allocation as well more accurate assessment of renal dysfunction. The incorporation of albumin into the model is more controversial. The indications for albumin administration ranges from large volume paracentesis to volume expansion for many admitted patients and is more likely to occur in patients with worse liver disease. The risks and benefits of such a volatile component will need to be carefully weighed before implementation. MELD 3.0 holds promise in bringing equity to liver organ allocation as well as improving wait-list mortality and we are likely to see MELD 3.0 (or a variation thereof) dominate the field in the near future.

Alexandra Shingina, MD, MSc, is an assistant professor of medicine in the division of gastroenterology, hepatology, and nutrition at Vanderbilt University Medical Center, Nashville, Tenn. She has no conflicts.



A newly updated version of the Model for End-Stage Liver Disease (MELD) score was effective for predicting short-term mortality in patients with end-stage liver disease and addressed important determinants of wait list outcomes that haven’t been addressed in previous versions, according to findings from a recent study. The new model, termed MELD 3.0, includes new variables such as female sex, serum albumin, and updated creatinine cutoffs.

Thomas Northcut/Getty Images

“We believe that the new model represents an opportunity to lower wait list mortality in the United States and propose it to be considered to replace the current version of MELD in determining allocation priorities in liver transplantation,” wrote study authors W. Ray Kim, MD, of Stanford (Calif.) University and colleagues in Gastroenterology.

In patients with end-stage liver disease, the MELD score was shown to be a reliable predictor of short-term survival, according to the researchers. The original version of MELD consists of international normalized ratio of prothrombin time and serum concentrations of bilirubin and creatinine; MELDNa consists of the same with the addition of serum concentrations of total sodium. Since 2016, MELDNa has been utilized in the United States to allocate livers for transplant.

Despite the utility of the current MELD score, questions have been raised concerning the accuracy of the tool’s ability to predict mortality, including a study by Sumeet K. Asrani, MD, MSc, and colleagues. Changes in liver disease epidemiology, the introduction of newer therapies that alter prognosis, as well as increasing age and prevalence of comorbidities in transplant-eligible patients are several drivers for these concerns, according to Dr. Kim and colleagues. Also, there is an increasing concern regarding women and their potential disadvantages in the current system: At least one study has suggested that serum creatinine may overestimate renal function and consequently underestimate mortality risk in female patients, compared with men with the same creatinine level.

Dr. Kim and colleagues sought to further optimize the fit of the current MELD score by considering alternative interactions and including other variables relevant to predicting short-term mortality in patients awaiting liver transplant. The study included patients who are registered on the Organ Procurement and Transplantation Network Standard Transplant Analysis and Research files newly wait-listed from 2016 through 2018. The full cohort was divided 70:30 into a development set (n = 20,587) and a validation set (n = 8,823); there were no significant differences between the sets in respect to age, sex, race, or liver disease severity.

The investigators used univariable and multivariable regression models to predict 90-day survival following wait list registration. The 90-day Kaplan-Meier survival rate in the development set was 91.3%. Additionally, model fit was tested, and the investigators used the Liver Simulated Allocation Model to estimate the impact of replacing the current version of the MELD with MELD 3.0.

In the final MELD 3.0 model, the researchers included several additional variables such as female sex and serum albumin. Additionally, the final model was characterized by interactions between bilirubin and sodium as well as between albumin and creatinine. Also, an adjustment to the current version of MELD lowering the upper bound for creatinine from 4.0 mg/dL to 3.0 mg/dL.

The MELD 3.0 featured significantly better discrimination, compared with the MELDNa (C-statistic = 0.8693 vs. 0.8622, respectively; P < .01). In addition, the researchers wrote that the new MELD 3.0 score “correctly reclassified a net of 8.8% of decedents to a higher MELD tier, affording them a meaningfully higher chance of transplantation, particularly in women.” The MELD 3.0 score with albumin also led to fewer wait-list deaths, compared with the MELDNa, according to the Liver Simulated Allocation Model analysis (P = .02); the number for MELD 3.0 without albumin was not statistically significant.

According to the investigators, a cause of concern for the MELD 3.0 was the addition of albumin, as this variable may be vulnerable to manipulation. In addition, the researchers note that, while differences in wait list mortality and survival based on race/ethnicity were observed in the study, they were unable to describe the exact root causes of worse outcomes among patients belonging to minority groups. “Thus, inclusion in a risk prediction score without fully understanding the underlying reasons for the racial disparity may have unintended consequences,” the researchers wrote.

“Based on recent data consisting of liver transplant candidates in the United States, we identify additional variables that are meaningfully associated with short-term mortality, including female sex and serum albumin. We also found evidence to support lowering the serum creatinine ceiling to 3 mg/dL,” they wrote. “Based on these data, we created an updated version of the MELD score, which improves mortality prediction compared to the current MELDNa model, including the recognition of female sex as a risk factor for death.”

The researchers reported no conflicts of interest with the pharmaceutical industry. No funding was reported for the study.

Next Article: