Overreporting of vitamin D levels by the roche elecsys vitamin D3 (25OH) method

Abstract

Vitamin D deficiency is common, and is associated with increased risk of osteoporosis and, increasingly, is shown to have association with many other illnesses. There are many ways to measure vitamin D, but all can suffer from precision and bias. Recently, Roche removed its Elecsys Vitamin D3 (25OH) electrochemilu-minescence immunoassay (ECLIA) from the market, citing deteriorating traceability to the reference method. We investigated the performance of two formulations of the Roche immunoassay against the reference method (liquid chromatography tandem mass spectrometry). A set of samples was assayed for vitamin D using the polyclonal formulation of the Elecsys electrochemiluminescence immunoas-say, and the monoclonal immunoassay. A subset of samples was assayed for vitamin D3 using LCMSMS. The correlation between the monoclonal formulation of ECLIA and LCMSMS was: ECLIA (monoclonal)=14; 0.31 LCMSMS + 23, r²=14; 0.27. The correlation between the monoclonal and polyclonal formulations of the ECLIA was: ECLIA (monoclonal)=14; 0.93 ECLIA (polyclonal) – 3, r²=14; 0.60. At the time of its removal from market, the Roche Elecsys vitamin D3 (25OH) electrochemiluminescence immunoassay had unacceptable performance, markedly under-reporting vitamin D levels. The bias likely preceded the introduction of the monoclonal formulation of the immunoassay.