An international cooperative study evaluating serum thyroglobulin standards.

A number of research laboratories have reported great variability in the levels of serum thyroglobulin (Tg) in normal subjects, the reason for which is not immediately apparent. The present study was designed to determine how important these variations were by submitting three identical standards to all participating laboratories. Three lyophilized human sera (standards A, B, and C) with increasing concentrations of Tg (5.3, 30.6, and 80.6 ng/ml, respectively) were submitted to 37 laboratories (40 assays) in 18 different countries. Standard A gave detectable values in 19 assays. The mean serum Tg concentration was 6.3 +/- 1.4 (+/- SEM) ng/ml (n = 18). Standard B was detected in all but 3 assays. The mean serum Tg concentration in standard B was 15.7 +/- 1.4 ng/ml (n = 37). All laboratories were able to detect Tg in standard C, and reported a mean serum Tg concentration of 36.5 +/- 3.2 ng/ml (n = 40). Lyophilization affected the recovery of Tg in our assay. This was confirmed by a study in which lyophilized standards A, B, and C and frozen standards were analyzed in the same assays. The remarkable finding was that the variability in serum Tg values reported by the various assays was great despite the submission of an identical set of standards of each of the laboratories. Wide interassay variation raises problems with respect to the applicability of threshold levels proposed by certain studies. The latter is particularly germane to the follow-up of patients with differentiated thyroid cancer. It is concluded that the development of a world standard for Tg may be a first and important step toward standardization of Tg assays, and that other components of the assays may need standardization as well.