Agreement of seven 25-hydroxy vitamin D-3 immunoassays and three high performance liquid chromatography methods with liquid chromatography tandem mass spectrometry

Background: Several recent studies have shown some discrepancies between 25-hydroxyvitamin D [25(OH)D] assay methods, despite some improvement in the past few years. The accuracy of 25(OH)D assay methods is still a real challenge for clinical laboratories. The aim of this study was to assess the agreement between a large panel of routine assays and a two-dimensional liquid chromatography/tandem mass spectrometry (2D LC-MS/MS) method, selected as the reference method. Methods: Forty-nine human plasma samples with only endogenous 25(OH)D-3 were analyzed with 11 different methods, especially with three LC-UV methods that differed in the extraction step. Seven routine immunoassays were also tested: two manual (RIA and EIA from IDS) and five fully-automated methods. The results of the 25(OH) D-3 assays were compared with those of the 2D LC-MS/MS method using weighted Deming regression analysis, Bland-Altman plots and concordance correlation coefficient (CCC). The ability of these methods to properly classify patients was evaluated by sorting results depending on vitamin D status. Results: The CCC was >0.90 for the three LC-UV methods and for most of the automated IA, meaning substantial agreement with 2D LC-MS/MS results. The ability to properly classify patients according to their vitamin D status was overall satisfactory for most of the methods tested (concordance >90%). Conclusions: The immunoassays available on Liaison, Isys, Architect and Elecsys, together with our in-house LC-UV method preceded by an SLE step met the minimum requirements for the assessment of vitamin D status in clinical laboratories.