A collaborative trial for the evaluation of blood cholesterol measurement in clinical laboratories in Italy

A collaborative trial for the evaluation of blood cholesterol measurement in Italy was carried out, with the use of two lyophilized controls, whose target values, respectively 4.42 and 6.21mmol/l were established by means of 'definitive' methodology (isotope dilution/mass spectrometry). Results from 480 participants showed a somewhat broad dispersion (CV 6.1% and 6.3% respectively), and a definite bias (-0.25mmol/l and -0.61mmol/l respectively) with respect to the target values. The different analytical systems were characterized by different combinations of inaccuracy and imprecision; however, the bias observed for the higher concentration sample was a constant finding. The behaviour of the control materials, in comparison with that exhibited by patients' sera, was assessed in a manual enzymatic procedure and in the Kodak Ektachem 700 and Technicon Chem 1 systems. The peculiar properly of one control material to behave differently from patients' sera in some analytical systems, i.e. the lack of commutability, was found to be partially responsible for the observed bias in the three methods studied. The importance of testing for commutability of the control materials to be used for the control of accuracy is stressed.