Interpretations of five monoclonal immunoassays of lutropin and follitropin: effects of normalization with WHO standard.

Five mono(oligo)clonal immunometric assays for lutropin (LH) and follitropin (FSH)--bioMérieux, IRE-Medgenix, Serono Diagnostics, Diagnostics Products Corp. (DPC), and LKB--were evaluated in comparison with two polyclonal RIAs (DPC and Amersham). Detection limits varied from 0.04 to 0.32 int. unit/L and 0.06 to 0.86 int. unit/L for LH and FSH, respectively. Intra- and interassay precision (CV) at three concentrations varied from 2.0% to 29.8%, showing that not all kits tested gave acceptable results, especially for LH. Linearity and parallelism were acceptable, except for the DPC FSH kit and the bioMérieux LH kit. High-dose "hook" effects were seen in some kits at LH concentrations of 250 int. units/L, but not in the FSH kits up to concentrations of 350 int. units/L. Reagents in some kits cross-reacted with choriogonadotropin. The clinical validity of the assays was tested in 25 pre- and 25 postmenopausal healthy women and in 66 patients with polycystic ovary disease. In contrast to FSH, LH values varied significantly not only between polyclonal and monoclonal assays but also between the various monoclonal assays, despite the fact that all manufacturers state that their kits are calibrated on the same standards: WHO International Reference Preparation (IRP) 68/40 for LH and 78/549 for FSH. We normalized the results by using new WHO standards: IRP 80/552 for LH and IRP 83/575 for FSH. This decreased significantly the between-kit differences in LH results for individuals. The much-used LH/FSH ratio greater than 3 for diagnosing patients with polycystic ovary disease is not valid when monoclonal assays are used, and is kit-dependent. However, using the normalized results yields a "new" LH/FSH ratio, which is kit-independent and differs significantly between patients and healthy subjects.