MELD was implemented as an objective measure to prioritize patients for liver transplant (LT). Accuracy and reproducibility of MELD is an essential assumption to ensure fairness in access to LT. Creatinine (Cr) assays were standardized in the US to ensure consistency in reporting eGFR. We assessed lab-based variability in MELD, and thus organ access, in all LT centers within UNOS region 9 post-standardization. Aliquots of 30 patient serum samples were sent to all 7 LT centers for analysis of Cr, bilirubin, and sodium. An additional 2 serum samples with Cr levels traceable to isotope dilution mass spectrometry (IDMS) were also tested. Descriptive statistics and unbiased hierarchical clustering were performed with R. The reproducibility of Cr measurement on 2 IDMS standards was acceptable at all sites [CV=1.0-1.8%]. MELD scores in this cohort ranged from 14 to 38. Despite standardization, and the mean CV for MELD between centers was 3.8% and unbiased hierarchical clustering suggested specific centers have values that result in higher MELDs. A single center had the highest MELD score on 15/30 samples, and one had the lowest MELD score on 13/30 samples, suggesting systematic bias. The mean range of MELD scores within each sample was clinically significant (mean 3 points, range 1-6) and increased with increasing bilirubin (mean of 4 points for bilirubin>20 mg/dL, and 6 for bilirubin>35 mg/dl). Bland-Altman plots confirmed that bilirubin interfered with Cr to varying degrees by platform, with lower Cr values in samples with higher bilirubin. Thus, despite implementation of Cr standardization, laboratories within this UNOS region report clinically significant differences in MELD on identical samples, with differences of up to 6 MELD points in the high MELD patients. Variable interference of bilirubin on Cr measurements among assay platforms caused some of this bias. The effect of common interfering substances present in patient samples is not taken into account in standardization methods. This systematic bias in MELD scores based upon laboratory methods and therefore center choice within a region must be addressed in the current efforts to eliminate disparities in LT access.

CITATION INFORMATION: Verna E, Connelly C, Dove L, Emond J, Adem P, Babic N, Corsetti J, Faix J, Hayden J, Lifshitz M, Chiles M, Mohan S, Eldad H, Kratz A. Life Threatening Disparities in Calculated MELD Due to Variation in Laboratory Methods within a UNOS Region. Am J Transplant. 2017;17 (suppl 3).

« Back to 2017 American Transplant Congress