Calcolo dell'incertezza nella determinazione dell'attività catalitica della γ-glutamiltransferasi: Confronto fra tre approcci

The calculation of the uncertainty of measurement is a mandatory requirement for a reference laboratory, but it would also be requested for any clinical laboratory seeking for accreditation according to ISO 15189:2007 standard. This study compares three different approaches for the calculation of the uncertainty related to the measurement of the catalytic activity of γ-glutamyltransferase, performed with the IFCC reference procedure. The first approach was based on the "Guide to the expression of uncertainty in measurement" (GUM), whereas the other two approaches derived from the "Handbook for calculation of measurement uncertainty in environmental laboratories". Using GUM, all the sources of uncertainty were identified and their contribution to the final measurement uncertainty was calculated using literature data or ad hoc experiments. Wavelength setting resulted the major uncertainty contributor, while the effect of pH was relevant only when using quality control materials. The calculated expanded uncertainty was 4.34%. In the second approach, we estimated the uncertainty by using the imprecision data from the long term IQC and the data obtained from an EQAS. Using the proposed calculation, the expanded uncertainty resulted 7.06%. The third approach relied on the data obtained with repeated measurements of the certified reference material ERM-AD452/IFCC. Using these data, the expanded uncertainty resulted 3.92%. In conclusion, approaches one and three gave similar figures, with estimated uncertainties of ∼4.0%, while the data derived with the second approach were biased by the results of EQAS that likely utilized non commutable materials, thus amplifying a bias source (pH) that is negligible when real patient samples are used.