Journal Editors’ requirements for full descriptions of laboratory methods need to include information on metrological traceability and analytical specificity

With respect to the consensus report from Journal Editors on laboratory methods, ¹ I believe that information on the traceability chain and analytical specificity of methods must also be provided, so that the accuracy of results and appropriateness of the units may be assessed. This should help prevent the problem of laboratory tests coming into premature clinical use with firm guidelines before the measurand is properly characterized or standardized.

For Type 1 Joint Committee for Traceability in Laboratory Medicine (JCTLM List 1 ² ) analytes, (unique molecular entities which are identical wherever found), where a complete traceability chain back to an SI (Systeme Internationale) definition of the analyte is possible, ‘analytical assays’ may be constructed which give method-independent results with SI mass or molar units. ³ To assess their accuracy panels of unprocessed single donations of clinical material containing an appropriate range of analyte concentrations are required, whose values have been assigned by an established reference method (catalogued in the JCTLM database ⁴ ) in a reference laboratory which, ideally, participates in the International Federation of Clinical Chemistry RELA (IFCC External Quality Assessment Scheme for Reference (Calibration) Laboratories in Laboratory Medicine) scheme, ⁵ as has been done for cortisol. ⁶ Manufacturers should be required to publish regression and difference plots of results of split sample comparisons for such panels in their instructions for use and on their websites, and these data should be refreshed whenever significant changes in the method are made. These plots should demonstrate that the method is properly calibrated and analytically specific for the measurand to support ongoing vigilance of methods, external quality assessment (EQA) services should regularly distribute patient-like materials whose values have been assigned by the same reference methods as above.^7,8

For Type 2 (JCTLM List 2 ² ) analytes, where uncertainty over their structure and/or heterogeneity of forms make it impossible for a complete traceability chain back to an SI definition to be made, only ‘comparative assays’ can be constructed, which give method-dependent results in arbitrary units. ³ International reference preparations may be used as measurement standards, but uncertainty over the specificity of the method for the measurand molecule and its isoforms or fragments, render the exact nature of what is being measured open to question. A practical solution to this problem, as exemplified by human chorionic gonadotropin, is to characterize and make reference preparations for each clinically significant molecular form and then establish, through epitope mapping, which commercial antibodies react to these in the different assays. ⁹ Manufacturers should collaborate with reference institutions in clarifying the analytical specificity of their methods, so that differences may be explored and set in clinical context. The lack of correlation between analytical recovery of spiked reference materials and bias in EQA schemes for many peptide hormones strongly supports such an approach. ¹⁰

Care must be taken not to indicate a false sense of trueness by assigning SI unitage to biomarkers when these are Type 2 analytes and/or no recognized standards exist. Whereas it may be possible to define the mass or molar content of a recombinant material, traceability of a patient's result back to an SI definition of the measurand (required by the use of SI units) is not possible, as the sample may contain molecular forms which give a signal in the assay but which are not the same molecular entity as the working standard. An example of this problem is in the same issue of the Annals, where SI mass units are used for urinary neutrophil gelatinase-associated lipocalin, ¹¹ and it also applies to troponins, given as an example in the main report. ¹²

I believe we have a moral and professional duty to meet patients’ requirements for universally accurate and properly interpreted test results through the rigorous development of traceability chains and analytical and clinical quality specifications. Investigations must be evidence-based, and underpinned by scientific metrological principles. Journal Editors should engage with professional organizations, national metrology institutions, reference laboratories and diagnostic manufacturers, to develop strict requirements for clear statements to be made about the metrological characteristics of assays for existing and new biomarkers intended for clinical use.

Declarations

Competing interests: JGM is a member of the IFCC Committee on Traceability in Laboratory Medicine but this letter is a personal view and does not represent the views of C-TLM or the IFCC.

Funding: None.

Ethical approval: None required.

Guarantor: JGM.

Contributorship: JGM is the sole contributor.