Abstract
Serum thyroid stimulating hormone (TSH) measurements are useful for detecting clinical and subclinical primary hypo- and hyperthyroidism in ambulatory patients. For diagnosis of hyperthyroidism, the functional sensitivity (FS) is an important performance criterion, and current guidelines recommend an FS of ≤0.02m IU/l for “third” generation performance.
We evaluated TSH FS for the Access 2, Advia Centaur, Architect i2000, Dimension ExL Modular Analytics E170, Immulite 2000 and Dimension Vista 1500 automated immunoassays using serum pools tested over a 6week period using 2 reagent lots and 2 calibrations. FS was determined by fitting a power function to the imprecision data using KaleidaGraph software.
The FS (m IU/l) for Access 2, ADVIA Centaur, ARCHITECT i2000, Dimension ExL, Modular Analytics E170, Immulite 2000, and Dimension Vista 1500 systems were determined to be 0.039, 0.006, 0.007, 0.003, 0.008, 0.003, and 0.003, respectively. The lowest and next to lowest pools had overall mean TSH concentrations of 0.012m IU/l and 0.020m IU/l, respectively and a range of concentrations of 0.005 to 0.022m IU/l and 0.007 to 0.077m IU/l, respectively.
All assays showed excellent performance in FS consistent with a “third” generation claim except for the Access 2 system. Further harmonization of TSH immunoassays is required, especially at lower concentrations.
► The functional sensitivity of 7 TSH methods ranged from 0.003 to 0.0039 mIU/L. ► Different software was used to fit power curves to the imprecision data generated. ► The software used can impact the estimate of TSH assay functional sensitivity. ► The lowest serum pool showed substantial inter-method variation in TSH results. ► Opportunities exist for improvement in TSH assay harmonization.