Abstract

The aim of this work was to check the suitability of control materials to normalize intermethod results for the measurement of free triiodothyronine in patient sera. In the main experiment, 108 patient sera and 11 commercially available control materials were assayed by a pair of methods. In a confirmatory experiment, two of the control materials and 142 patient sera were assayed with an alternative pair of methods. In the main experiment, the intermethod variability of 6/11 control materials differed significantly from that of patient sera (i. e. control materials were non-commutable). Recalculation of patient results using control materials as calibrators lowered the intermethod difference only if commutable materials were used. The confirmatory experiment demonstrated that the pattern of commutability changed if a different pair of methods was used. We conclude that in the case of free triiodothyronine the commutability of control materials should be tested if they are to be used to normalize patient results obtained by different methods.