In the current issue of Transplantation, Verma et al2 analyze the impact of laboratory variation in model for end-stage liver disease (MELD) scores focusing, in particular, on differences in creatinine based on the interfering substance of bilirubin. The investigators took 30 individual patient samples and sent them to 8 different laboratories at 7 transplant centers within a single united network for organ sharing region. The study was able to measure only creatinine, sodium, and bilirubin given that the remaining component of the MELD-Na calculation, the international normalized ratio (INR), is not able to be measured on a stored sample. Among the 8 different labs, 6 different analyzer platforms were used for processing. Depending on the lab, either the kinetic alkaline picrate (Jaffe) or the enzymatic method was used to measure creatinine, and also depending on the lab, either the ion selective electrode or potentiometric methods for obtaining serum sodium were used. All sites used the same method for determining total serum bilirubin. None of the laboratories used the same combination of analyzer platforms and assays.
The main finding of the current analysis is that there was difference in the calculated MELD-Na which ranged from 1 to 6 for individual samples depending on which laboratory performed the analysis, with a mean difference of 3. The difference was more pronounced in the setting of higher bilirubin, related to inference of bilirubin with creatinine assay. Overall, the mean results of the calculated scores ranged from 25.9 to 27.9 among the different labs. The values used to rank order patients on the waitlist are rounded to the nearest whole number, and thus, 3 labs had a mean score of 26, while 3 others had a mean score of 27, and 1 had a mean score of 28. Whether individual patients would have been able to access transplant based on the different scores during the study period could not be determined by this analysis, though a supplemental figure highlights the probability of patients being transplanted within 30 days according the MELD score.
The impact of laboratory variation across different sites on the calculation of MELD highlights an important problem which also has been demonstrated in previous analyses. Trotter et al3,4 demonstrated substantial variation in the calculated MELD score depending on the laboratory, first in 2004 at a US single-center level, and then in 2007 at the national level in the United States. While those differences were primarily based on variability in INR, the current report demonstrates similar variability exists in creatinine measurements. Lisman et al5 sent samples from patients listed for liver transplantation to 7 different European laboratories and found similar variability to the current study, with a mean difference of 4.8. Schouten et al6 also found variability among 6 centers and similar to the current analysis, they assessed the impact on the match MELD on the probability of receiving a transplant. Beyond a mandate for a uniform testing platform there are few solutions proposed by previous authors, though 1 Italian group proposed creating a normalized scale to adjust for the variation noted between 2 different laboratories using a specific formula (“corrected value = measured value × Vmax lab 1/Vmax lab 2”), though it seems for this to be possible patients would need to have their blood tests performed at the same place each time.7 A similar analysis to the current 1 was also recently performed in Germany, demonstrating the impact of nonrandom variability of calculated MELD score based on the laboratory techniques used within a specified sharing area.8
One of the more remarkable findings of the current analysis was that as a quality control measure, 2 standardized samples were purchased from the National Institute for Standards and Technology (NIST) to determine the reproducibility of creatinine at each of the labs. Only 1 out of the 8 laboratories met the quality control standard for variability within the allowable range for both NIST samples on all tests. At 3 of 8 sites, all results were outside the allowable range for NIST sample 1 while 2 of 8 laboratories, all results were outside the allowable range for NIST sample 2. This is independent of inference by bilirubin or any other substance, and if reproducible, identifies a potential need for quality control and sharing of best practices to ensure the best outcomes for patients, which could extend far beyond liver allocation. It seems that 1 laboratory is performing better than the others in the reliable measurement of creatinine, and presuming this high rate of variability is not a byproduct of the methods used in the current study (such as an extra freeze-thaw cycle or pooling of the samples), working to get laboratories to adopt the technology that allows for the most reliable measurement may ultimately have an significant impact on patient outcomes given that creatinine is used not only to identify patients who may require diagnosis and treatment of renal issues, but more importantly, it is used to determine dose adjustments required to avoid toxicity from a vast number of medications.
Ideally, as suggested by the title of the current analysis, the platforms and methods used to perform these very common laboratory tests can be standardized to the best practice method which produces the most consistent result, and thus eliminating the technology used to perform the test as a potential source of bias and inaccuracy in the result. This type of oversight is outside the scope of the organ allocation system, but efforts to reduce inter laboratory variation and encourage standardization toward a unified best practice would hopefully be within the purview of the oversight agencies responsible for laboratory oversight. Centers that participate in transplantation in the United States are held to a very rigorous set of unified standards both through the Organ Procurement and Transplantation Network (OPTN) and the Center for Medicare Services (CMS) which consists of an exhaustive list of “Conditions of Participation” that centers must meet to perform transplantation. A similar adoption of best practices should be expected from the other aspects of the healthcare system to ensure, as identified by the by the authors, “that the MELD score itself can be measured both accurately and reproducibly across regions and transplant centers.”
1. Shakespeare W. Hamlet. Act 1scene 3line 564
2. Verma EC, Connelly C, Dove LM, et al. Center-Related Bias in MELD scores within a liver transplant UNOS region: a call for standardization. Transplantation. In press
3. Trotter JF, Brimhall B, Arjal R, et al. Specific laboratory methodologies achieve higher model for endstage liver disease (MELD) scores for patients listed for liver transplantation. Liver Transpl. 2004; 10:995–1000
4. Trotter JF, Olson J, Lefkowitz J, et al. Changes in international normalized ratio (INR) and model for endstage liver disease (MELD) based on selection of clinical laboratory. Am J Transplant. 2007; 7:1624–1628
5. Lisman T, van Leeuwen Y, Adelmeijer J, et al. Interlaboratory variability in assessment of the model of end-stage liver disease score. Liver Int. 2008; 28:1344–1351
6. Schouten JN, Francque S, Van Vlierberghe H, et al. The influence of laboratory-induced MELD score differences on liver allocation: more reality than myth. Clin Transplant. 2012; 26:E62–E70
7. Ravaioli M, Masetti M, Ridolfi L, et al. Laboratory test variability and model for end-stage liver disease score calculation: effect on liver allocation and proposal for adjustment. Transplantation. 2007; 83:919–924
8. Al-Saeedi M, Yassein T, Schultze D, et al. Impact of inter-laboratory variability on model of end-stage liver disease (MELD) score calculation. Ann Transplant. 2016; 21:675–682