We evaluated the technical performance of two new commercial automated immunoassay systems of third-generation assays for thyrotropin. The interassay CV was 2.8% for AutoDELFIA and 3.25% for Access at thyrotropin concentrations of approximately 1.3 mIU/1. The lower detection limits of the assays were 0.023 mIU/1 for AutoDELFIA and 0.0096 mIU/1 for Access, and the functional sensitivity for a CV of 20% was 0.027 mIU/1 and 0.028 mIU/1, respectively. Sample to sample carry over was negligible (0.0016% for AutoDELFIA and 0.005% for Access). The range of linearity was acceptable for Access (102-115%) but not for low thyrotropin concentrations in AutoDELFIA (143% for a thyrotropin value of 0.48 mIU/1). Correlation between AutoDELFIA and Access was adequate (r = 0.999). We conclude that both automated immunoassays offer good reliability, practicability and performance characteristics.The technology used is an automated adaptation of the DELFIA manual equipment. The improved lower detection limit obtained with the thyrotropin test is due to the use of three monoclonal antibodies directed against three separate antigenic determinants of the thyrotropin molecule.Summary: The automated immunoassay analyser, IMMULITE, developed by DPC, was evaluated. IMMULITE is an automated system that allows random access in combination with immediate and continuous access. In this study we evaluated the IMMULITE on four panels of analytes: thyroid, fertility, tumour and "non-routine" markers.We observed good within-run reproducibility (ranging from 2.3-15.9%CV, for low controls, to 2.7-8.7%CV for high controls) as well as between-day imprecision (ranging from 3.7-24.6%CV for low controls, to 2.5-11.8%CV for high controls).The analytical sensitivity of the assays ranged generally from very sensitive to acceptable.The dilution curves for all assays were nearly linear i. e. the maximum deviation of the observed from the expected 1 recovery was 8%. .
I
\Correlation between IMMULITE and other assays (AxSYM, IMx, TDx, AIA-1200, DPC-c. a. c., Medgenix) varied i, from r = 0.931 to r = 0.994, except for lutropin and parathyrin with coefficients of correlation of r = 0.594 and r = 0.591. The slopes of the regression lines ranged from 0.745 to 1.327, except for parathyrin where a slope of 2.389 was found. Inter-laboratory correlation was very good between the two locations (Sittard and Apeldoorn) and varied from r = 0.984 to r = 0.999; slopes of the regression lines varied from 0.924 to 1.086. We conclude that the DPC JMMULITE system is suitable for testing routine analytes in random access mode as well as for testing "non-routine" analytes on an automated immunoassay analyser.