Re-weighting the model for end-stage liver disease score components

Area of Science:

• Transplant medicine
• Hepatology
• Nephrology

Background:

• Liver transplant candidates with mild liver dysfunction and severe kidney problems can have high Model for End-Stage Liver Disease (MELD) scores.
• Existing MELD coefficients may not accurately reflect current waiting list mortality risks.

Purpose of the Study:

• To re-estimate MELD coefficients using recent national data.
• To evaluate the impact of updated MELD scores on liver transplant waiting list rankings and mortality prediction.

Main Methods:

• Analysis of 38,899 adult liver transplant candidates wait-listed from September 2001 to December 2006.
• Utilized a time-dependent Cox regression model to estimate updated MELD component coefficients.
• Computed rank correlation between existing and updated MELD scores.

Main Results:

• Updated MELD coefficients showed significant differences compared to existing ones, with lower weight for creatinine and international normalized ratio, and higher weight for bilirubin.
• Updated MELD demonstrated improved prediction of overall and 90-day waiting list mortality (Index of Concordance 0.68 vs. 0.64).
• High rank correlation (0.95) between existing and updated MELD scores was observed, but significant differences emerged for candidates with MELD >= 20.

Conclusions:

• Existing MELD coefficients are significantly different from those derived from current national waiting list data.
• The updated MELD formula provides a more accurate prediction of waiting list mortality.
• Implementing the updated MELD for liver allocation could change the ranking of candidates on the waiting list.