Development and evaluation of a time-resolved immunofluorimetric assay for thyrotropin

The assay described in this article is based on microtitre plate technology; it employs an europium label. The streptavidin-biotin system has been used and all components are commercially available. The lower detection limit of the assay is below 0.003 mU/l; the standards are made up in newborn calf serum. Correlation with a commercially available immunoluminometric assay (Berilux--Behringwerke) was excellent (r = 0.92, n = 201 data pairs, range covered 0-10 mU/l). The regression line using a double logarithmic transformation was: (log y) = 0.91 (log x)-0.08. The assay precision at the clinical limits of decision--i.e. between hyper- and euthyroidism (0.2 mU/l) and eu- and hypothyroidism (4 mU/l)--was acceptable. The median coefficient of variation was 1.93% in the range 0.02-1 mU/l and 2.11% in the range 1.0-3.5 mU/l, with both values being determined from precision profiles using 214 and 188 data pairs respectively. Inter-assay coefficients of variation determined in over 30 consecutive assays were under 6% in the range 1.3-20 mU/l. From 201 sera measured in both assays, 194 sera gave clinically identical values; 8 sera gave clinically discrepant values. The assay has a large dynamic range covering a concentration range of above 5 decades, with the count ratio between the 100 mU/l standard and the zero standard being in excess of 4150:1. A high dose hook effect was first seen in excess of 500 mU/l. The maximum signal was achieved around 150 mU/l, which registered around 8 x 10(6) counts per second, a figure more than 8000 times higher than that in the zero standard.