Improved Recognition of 25-Hydroxyvitamin D2 by 2 Automated Immunoassays

Background: Despite recommendations to limit vitamin D testing to specific clinical scenarios, test volume remains high in many clinical laboratories. Automated total vitamin D immunoassays frequently under- or over-recover 25-hydroxyvitamin D-2 [25(OH)D-2], making accurate assessment of vitamin D status difficult in patients taking high-dose 25(OH)D-2 supplements. Mass spectrometry-based methods offer excellent recovery of 25(OH)D-2 but are not practical for use in all laboratories. In this study, we evaluated 2 automated immunoassays against an LC-MS/MS method performed at a national reference laboratory. Methods: A method comparison against LC-MS/MS was performed for the Roche Elecsys Vitamin D total II assay and the IDS-iSYS 25 VitD(S) immunoassays using 49 patient specimens submitted for clinical 25(OH)D measurement. Mean bias was calculated, and vitamin D status was determined for each specimen according to the 2011 Endocrine Society clinical practice guidelines. Results: Theil-Sen regression lines relative to LC-MS/MS were y = 0.88x+2.94 for Roche and y = 1.03x+2.48 for IDS. Mean bias (+/- SD) in samples with 25(OH)D-2 concentrations less than 5 ng/mL was -0.25 ng/mL (+/- 6.30) for Roche and -1.45 ng/mL (+/- 6.82) for the IDS. Mean bias (+/- SD) in samples with 25(OH)D-2 concentrations greater than 5 ng/mL was -3.19 ng/mL (+/- 6.61) for Roche and 5.52 ng/mL (+/- 6.36) for IDS. Median percentage recovery of 25(OH)D-2 was 87.1% (interquartile range 76.0-111.3) for Roche and 120.6% (interquartile range: 105.3-133.4) for IDS. Vitamin D status was misclassified in 7 samples by the Roche assay and 3 by the IDS assay. For all but one of the discordant pairs, the immunoassay result was within 1.7 ng/mL of the diagnostic cutoff. Conclusions: The automated immunoassays evaluated here demonstrate improved recovery of 25(OH)D-2 relative to previous generations. Both are acceptable for use in the determination of vitamin D status.