Current assays overestimate 25-hydroxyvitamin D3 and underestimate 25-hydroxyvitamin D2 compared with HPLC:: need for assay-specific decision limits and metabolite-specific assays

Background: Clinical demand for quick, cheap, precise and accurate 25-hydroxyvitamin D (25(OH)D) results has led to the development of a variety of assay methods. Lack of standardization of these methods has resulted in inter-method disagreement and challenged whether current assays recognize 25(OH)D-2 and 25(OH)D-3 equally. Methods: We studied 172 patient samples from hip fracture cases using DiaSorin (DS) and IDS radioimmunoassays and the Nichols Advantage-automated protein binding assay (NA-CLPBA) in comparison to high-performance liquid chromatography (HPLC). 52 patient samples were analysed before and after three months treatment with 1000 IU of daily ergocalciferol (vitamin D-2). Results: Linear regression analysis in pre-treatment samples demonstrated a positive Y-intercept for each immunoassay compared with HPLC, and a slope that varied from 0.64 (IDS) to 0.97 (DS, NA-CLPBA). Bland Altman analysis demonstrated that all the three assays had a proportional positive bias relative to HPLC at values from 20 to 50 nmol/L. Regression analysis of post-treatment samples demonstrated a slope that was not significantly different from zero for the IDS and NA-CLPBA and 0.2 for the DS method, with a positive intercept for all assays of between 8 and 22, indicating less than 50% of 25(OH)D-2 measured by HPLC was detected. Conclusions: These results demonstrate the need for assay-specific decision limits for 25(OH)D-3 in order to define appropriate thresholds for treatment institution. Treatment with vitamin D-2 may not be accurately monitored with any of the three commercial assays studied. Clinicians and biochemists who continue to use 25(OH)D assays need to be urgently informed of these issues.