The C-STFT proposed a harmonization approach for measurement of TSH. This glycoprotein circulates in the bloodstream typically as mixture of components comprising disease-specific glycoforms. Current TSH immunoassays all are traceable to the same WHO IRP TSH 80/558 (10), and express their measurements in mIU/L (defined by that IRP). The latter is from human cadaver pituitary and may comprise glycosylation forms that differ from those typically present in serum TSH. Therefore, the above issues of non commutability of the IRP apply, and hence, IRP-traceable TSH measurements do not give equivalent or interchangeable results. In principle, the harmonization proposal by the C-STFT still takes the basic metrological concepts into consideration, because, after all, also for TSH the measurement paradigm applies, i.e., procedures that claim the same measurand and are traceable to the same standard should give equivalent results. While this seems to be “the logic itself”, this has not been realized yet in laboratory medicine for proteins in general, nor for TSH in particular.
Definition of the measurand TSH
Important considerations that had to be made before proposing a harmonization concept was that protein/TSH measurement is mixture analysis. This implies that each component in the mixture must contribute significantly to the diagnostic application of the assay used for mixture analysis, and that the latter should measure the components in the mixture in an “equimolar” way, but to the diagnostic relevant extent. This means that the important changes in overall concentration of the mixture in health and disease should outweigh the minor changes in composition of the mixture in individual samples. This requirement of equimolar measurement is of utmost relevance for TSH, since glycoforms are different in euthyroidism versus subclinical and overt hypothyroidism (11), and because there is evidence that immunoassays may be sensitive to the differences (12), (13). Another consequence of mixture analysis is that the definition of the measurand should be commensurate with its realization by a measurement standard (and vice versa). This may require the definition of a “quasi surrogate component-mixture”, which is related to the epitopes at invariate peptide sequences that immunoassays must recognize. The C-STFT defined the measurand as “TSH, intact, total, glycosylation encountered in diagnostic applications which should be specified”. This definition requires in first instance that TSH assays examined for harmonization deliver a measure for “total TSH”. It also implies that harmonization should be based on the glycoforms present in native samples.
The “All Procedure Trimmed Mean, APTM” as surrogate RMP for TSH
Another consideration before harmonizing TSH assays is that the measurement results are expressed in IUs (concentrations in mIU/L). However, it is to expect that one day it will be possible to establish the relation between the IU and mass units (SI) for TSH. From this point of view, the reference measurement system for TSH and its realization should be viewed as continuum from discovery of the analyte to final translation into the SI, with different needs at different stages of the continuum. Therefore, the C-STFT decided to preserve the traceability of TSH assays to the WHO IRP in their proposed harmonization process, and recommends the “All Procedure Trimmed Mean” (APTM, expressed in IU) as surrogate RMP. It is proposed to statistically derive the APTM from data of a dedicated method comparison study with native human samples, in which as many TSH assays as possible participate (14). Note such a method comparison is also referred to as “split sample multiple method comparison study”. According to this proposal, the APTM targeted panel becomes the key tool in the harmonization process. It has the big advantage that it is composed of samples that contain the typical glycoforms of TSH in the blood circulation, thus are as commutable as possible. Note that after being assigned with the APTM (expressed in mIU/L) derived from the measurement values of the WHO traceable assays that participated in the method comparison study, the panel is intended to serve as a new measurement standard, in other words, at that point in the harmonization process, the IU of the WHO IRP is transferred to the panel. Note that for the process to be successful, proof of sufficient correlation of the concerned assays to the APTM is a prerequisite. Also the new standard should be sustainable. This requires that the IU is transferred from the first panel to the follow-up panels via analogous method comparison studies. The very first panel will – so to speak – serve as a predicate panel.