Analytical and clinical performance of six sensitive thyrotrophin kits

We have evaluated the analytical and clinical performance of six sensitive thyrotrophin (TSH) assay kits. The detection limits of the assay kits ranged from 0.02 mIU/L to 0.18 mIU/L. Except for one assay kit, the precision of the assays (coefficients of variation) at different concentrations was below or around 10%. Although results obtained from different kits showed good correlations (r = 0.82 to 0.95), large discrepancies were found when aliquots of the same control sera were assayed by different kits. Results from 33 hyperthyroid patients were all below the lower limit of reference ranges. However, 2 healthy subjects had low TSH results clearly distinguishable from the other 69 healthy subjects. Out of 31 hyperthyroid patients who had recovered after treatment, suppressed TSH values were found in 20% of patients. These results indicate that although TSH is a good test for the detection of hyperthyroidism, low TSH may be found in some euthyroid subjects and diagnosis of hyperthyroidism cannot be based on a suppressed TSH result alone.