The concentration-specific smallest detectable difference (D) between two concentrations (i.e., the sensitivity) was determined from the results of within-run and run-to-run reproducibility studies of a third-generation thyrotropin assay. Only low concentrations were considered (range 0-0.3 m U/l). Assays were performed in singlet or in duplicate. Findings confirmed the value of the assay: the analytic detection limit approximated 0.005 mU/l and the functional detection limit was around 0.01 mU/l. Expressed as a percentage of the concentration, D% was under 20% for levels of less than 0.02 mU/l in within-run studies or of less than 0.03 mU/l in run-to-run studies. With concentrations in excess of 0.05 mU/l, D% was under 10% in within-run studies and slightly above 10% in run-to-run studies. These new detection thresholds for this assay should be taken into account when interpreting clinical results from a single run (TRH test) or from several runs (hormone replacement therapy or suppression test).