Establishment of reference distributions and decision values for thyroid antibodies against thyroid peroxidase (TPOAb), thyroglobulin (TgAb) and the thyrotropin receptor (TRAb).

BACKGROUND: The National Academy of Clinical Biochemistry (NACB) stresses that the reference intervals for thyroid peroxidase antibodies (TPOAb), thyroglobulin antibodies (TgAb) and thyroid stimulating hormone (TSH)-receptor antibodies (TRAb) should be based on young men who lack certain risk factors and have serum TSH between 0.5 and 2.0 mIU/L. However, some young men without any of the risk factors have autoantibodies, and cannot be identified by the present tools. A model for reference intervals and cut-off values should not be influenced by the prevalence of risk factors.

METHODS: We developed a model of "composite logarithmic Gaussian distributions" and tested it in 1441 well-characterised subjects without clinically overt thyroid disease.

RESULTS: TPOAb and TgAb could be measured in all individuals. The 97.5% upper limits 1) on a traditional non-parametric scale, 2) according to the NACB criteria, and 3) for our model were 284, 24 and 9.8 kIU/L for TPOAb, and 84, 22 and 19 kIU/L for TgAb, respectively. The decision value (defined as the concentration corresponding to 0.1% false positives) was 15 kIU/L for TPOAb and 31 kIU/L for TgAb. Concentrations above our reference intervals affected the corresponding distribution of TSH values. For TRAb the upper reference limits were 1) 0.75 and 2) 0.75 IU/L, while our model was not applicable to TRAb because only 2-3% of the results were above the functional assay sensitivity.

CONCLUSIONS: In contrast to the NACB guidelines, our model for TPOAb and TgAb is more robust, as it is independent of the characteristics of the reference population.