Interlaboratory variability in assessment of the model of end-stage liver disease score

Background: The model of end-stage liver disease (MELD) score is nowadays widely used to prioritize patients for liver transplantation. Aims: To assess the contribution of the individual components of the MELD score in interlaboratory variability. Methods: We sent 15 samples from patients listed for liver transplantation to seven different European laboratories who were asked to measure all three variables. In addition, 10 samples from patients on oral anticoagulant treatment were sent to the same labs for the international normalised ratio (INR) measurement. Results and Conclusions: In all 15 samples, a substantial and clinically relevant variation in the calculated MELD score was observed between laboratories. The mean difference in the MELD score between the highest- and the lowest-scoring laboratory was 4.8. The variation in creatinine measurements resulted in differences of up to three MELD points in a single patient when comparing the highest and the lowest scoring lab. The variation in bilirubin measurements only accounted for a difference of one point between the highest- and the lowest-scoring laboratory, but the variation in INRs resulted in differences of 2 to 12 MELD points. MELD scores or INR values were not substantially different in laboratories that used the Owren instead of the more widely used Quick methodology for INR measurements. The variability in the INR in patients on oral anticoagulants was substantially less as compared with the variability in patients with liver disease. In conclusion, we observed a large interlaboratory variation in the MELD score. This variation in the MELD score is primarily caused by the INR.