Implementation of standardization in clinical practice: not always an easy task

As soon as a new reference system (RS) is adopted and implemented, validation of the correctly calibrated commercial methods should take place. Tracing back the calibration of routine tests to a RS (i.e. implementing standardization) can actually modify the analyte results and this may invalidate some of the clinical decision-making criteria currently used. In order to maintain the clinical experience, the quantitative relationship to the previous calibration system should be established and, if necessary, the clinical decision-making criteria should be adjusted accordingly. Measurement of serum creatinine and its use for estimating glomerular filtration rate provide a good example. Following implementation of creatinine methods with calibration traceable to the isotope dilution-mass spectrometry, equations used to estimate kidney function will give values that are higher than the values obtained using traditionally calibrated creatinine assays. This calibration change may significantly affect interpretative criteria based on these estimates of kidney function and in order to avoid this risk a reexpression of equations with standardized creatinine results is required. A MDRD equation has indeed been reexpressed for standardized creatinine results with the best approximation. Conversely, this has never been done for Cockcroft-Gault equation resulting in a worsening of its clinical performance. The prostatespecific antigen (PSA) is another example. Two sources of calibration are in use, one based on the traditional calibration scheme that produces results consistent with the first PSA assay marketed by Hybritech, used to establish the clinically relevant PSA cutoff of 4.0 mg/L, and the second providing traceability to the WHO Reference Preparation 96/670. Recalibrating a PSA assay from an original ‘Hybritech’ calibration to the new WHO calibration results, however, in ~20% lower PSA values indicating that the 4.0 mg/L cutoff would not provide optimal clinical efficacy for the WHO standardized assays, but a newly recalculated cutoff of 3.1 mg/L must be employed. In the case of HbA1c, reliable linear relationships between results traceable to the IFCC RS and previous national recommended methods have been demonstrated allowing the conversion of analytical and clinical data from one system to another. In practice it is therefore possible to translate HbA1c target values generated in previous landmark clinical studies, using methods not traced to the RS. All in all, the implementation of standardization should take place in a concerted action of laboratorians, clinicians, manufacturers,EQAS organizers, and other stakeholders. Dedicated meetings with manufacturers should be organized to discuss changes, e.g. process of assay recalibration, and studies should be performed to obtain convincing evidence that the standardization works improving comparability in clinical setting. Another important issue relates to the post-market surveillance of the performance of standardized assays through the organization of appropriate EQAS, requiring the availability of commutable control materials with target values assigned by accredited reference laboratories. Last but not least, uncertainty of measurement that fits for purpose must be defined across the entire traceability chain, starting with the provider of reference materials, extending through the manufacturers and their processes for assignment of calibrator values, and ultimately to the final result reported to clinicians by laboratories.