Résumé : This report describes the results of the second part of the collaborative study organized by a working group sponsored by the Community Bureau of Reference of the European Community Commission. The whole study was designed to understand the causes of discrepancy among LH immunoassay methods. In the parent report we described the characteristics of LH monoclonal antibodies. In the present work we focused on the comparison of 11 commercially available monoclonal antibody based kits and one polyclonal antibody based RIA. Recovery experiments of the second IS 80/552 and of a related LH preparation showed that curve calibration differed among kits. The ratio between the LH recovery of the two most different kits was 2.30. LH concentrations were determined, using the 12 assay methods in the sera of prepubertal children (n = 46), normal women (n = 26) and men (n = 39) before and after LHRH stimulation, post-menopausal women (n = 29) and patients with renal failure (n = 71) or polycystic ovaries (n = 28). In children LH was detected in 0 to 80% of the sera depending on the kit. In adults, the mean LH concentration provided by the 12 kits varied from 6.0 to 13.55 IU/L showing a ratio of 2.26 between the two most different kits. A thorough statistical analysis allowed to distinguish two groups of kits. The first groups consisting of 6 kits provided results close to those obtained by the RIA. The 5 other kits misrecognized circulating LH in subjects with various clinical status. These 5 kits were the only ones to use, as labelled probes, monoclonal antibodies specific for the holohormone (anti-alpha beta) whereas the 6 other kits used, as labelled probes, monoclonal antibodies directed to the alpha (1 kit) or the beta-subunit (5 kits). In both groups, the coated monoclonal antibodies are directed to either the beta-subunit (1/6 and 2/5 kits) or the holohormone (5/6 and 3/5 kits). Taken together these data suggest that discrepancy among LH assay kits is related to variation in standard curve calibration and in epitope specificity of monoclonal antibodies used in the kits. These findings may prove to be instrumental to alleviate differences among LH assay methods.