Harmonization and standardization of thyroid function tests.

This issue of Clinical Chemistry contains 3 reports from the IFCC Working Group for Standardization of Thyroid Function Tests. The first report relates to thyroidstimulating hormone (TSH) (1 ), the second addresses free thyroxine (FT4) and free triiodothyronine (FT3) (2 ), and the third discusses total thyroxine (T4) and total triiodothyronine (T3) (3 ). Each of these reports is authored by a group of experts who relate not only their own experience but also the experience of various professional organizations. This broad representation is important for the potential harmonization role that the authors hope to promote with these reports. The chair of this group, Linda Thienpont, was also project leader for the measurements of the total and free thyroid hormones in the prior groundwork project initiated by the European Commission, “Feasibility Studies for the Development of Reference Measurement Systems for Thyrotropin (TSH) and for Free Thyroxine (FT4), and Validation of Reference Measurement Systems (Procedure and Material) for Thyroxine (T4) and Triiodothyronine (T3) in Human Serum (2001–2005)” (4 ). This working group also included Catherine Ronin (Laboratories de Neurobiologic, Marseille, France), who was the project leader for the TSH measurements in this European Commission project. International leaders in standardization from both academia and industry participated in this undertaking. The major goal for standardization of thyroid function tests is the harmonization of patient test results. Standardization and harmonization are 2 separate but closely linked concepts. Standardization of assays does not always lead to harmonized patient test results, and harmonization of test results does not always require standardized reference methods. When analytes are well defined and standard reference procedures are available, calibration of production assays to match the reference methods is a good method for working toward harmonization. Harmonization of patient test results still may not be achieved, however, especially if there are problems with commutability of the reference standards and/or reference procedures (5 ). Many analytes, such as TSH, are not well defined, and reference procedures are not available. TSH is a glycoprotein that has multiple forms in blood. Donadio and colleagues have shown that circulating TSH molecules have various glycosylation patterns that may differ with different thyroid disease states (6 ). Differences in the relative concentrations of these circulating forms and differences in immunoreactivity with the various reagent antibodies make TSH standardization very difficult. The use of targeted glycosylated forms of TSH as assay calibrators might help with the standardization, but assay differences probably would continue to exist. Further analyses of the circulating forms of TSH and their relationships to thyroid diseases may help target the preferred protein fragments and/or protein forms for use in standard reference procedures. Lopez and colleagues recently published mass spectrometric data regarding the circulating forms of parathyroid hormone and showed how they related to immunoassay reactivity (7 ). Similar analyses of TSH may help in the development of reference methods for TSH. Even when reference standards are not available, major steps toward harmonization can be achieved via intramethod comparisons, such as those provided in the TSH part of this series (1 ). Weighted Deming regression analyses that compare the mean of 6 replicates of the measurements for each sample contained in the panel to the overall mean for all assays is an effective way to identify the offsets for each assay. These offsets then could be used, at least theoretically, to “correct” each assay to the “consensus” mean. Normalization factors will harmonize only the systematic differences among the assays. That is, if the regression slope is different from 1.0, the results could be factored to generate a slope of 1.0; however, if individual samples show different offsets compared with the overall regression slopes, global factors will not harmonize these results. Analyses of the data cross-plots, CVs, and correlation coefficients often help to identify the discordant subgroups. The standardization of the smaller molecules, such as T4 and T3, should be easier because these molecules 1 Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN. * Address correspondence to the author at: 200 First St. S.W., Rochester, MN 55905. Fax 507-266-5193; e-mail [email protected]. Received March 29, 2010; accepted March 30, 2010. Previously published online at DOI: 10.1373/clinchem.2010.145540 2 Nonstandard abbreviations: TSH, thyroid-stimulating hormone; FT4, free thyroxine; FT3, free triiodothyronine; T4, thyroxine; T3, triiodothyronine. Clinical Chemistry 56:6 879–880 (2010) Editorials