Abstract Background Although mass spectrometry (MS) is considered the reference measurement procedure for measuring 25-OH vitamin D (25-OH-D), there are limitations in expanding the application of MS due to high costs, technical difficulties, and long testing times. Various manufacturers produce upgraded reagents that can be compatible with MS. The present study aimed to compare the results of 25-OH-D concentrations determined by Maglumi X8 (X8, Snibe Diagnostics, Shenzen, China) and MS. Methods Forty samples, namely 30 samples with 25-OH-D concentration of around 1.00-29.99 ng/mg (Group 1), 10 samples with concentration above 30.00 ng/mg (Group 2), were prepared by pooling surplus samples at GC labs, Korea, during January 2025. For each pooled sample, four aliquots were made, and two aliquots were designated for X8 and MS, separately. For two days, each aliquot was tested in duplicate, resulting in being tested four times. Laboratory developed test MS test was performed using protein precipitation, solvent extraction, and derivatization steps with reagents from the PerkinElmer (Turku, Finland) MSMS Vitamin D kit, on Waters ACQUITY UPLC BEH C18 column (2.1×50 mm, 1.7 µm), and Xevo TQD (Waters Corporation, Milford, MA, USA) MS/MS system. Results The average coefficient of variation (CV) for all samples with X8 was 2.92% (Min-Max: 0.94%-5.41%), and that of Group 1 and Group 2 were 2.63% (Min-Max: 1.28%-4.59%) and 3.68% (Min-Max: 0.94%-5.41%). The linear regression analysis of all samples revealed the following equation: X8=0.9208*MS-0.1523 (r²=0.9678). The mean % bias and its 95% CI were -2.28% (-6.70-2.12%). The linear regression and the mean % bias with its 95% CI of Group 1 and Group 2 were X8=1.0168*MS+1.3347 (r²=0.959) and -1.66% (-4.39-1.08%), and X8=0.7552*MS+13.532 (r²=0.825) and -3.58% (-9.82-2.66%), respectively. Conclusion The assessment of X8 indicated that its performance is sufficient for the qualification of the Vitamin D Standardization Certification Program (VDSCP) run by the Centers for Disease Control and Prevention (CDC), i.e., CV = 10% and mean bias = |±5%|. However, the difference between methods increased as the 25-OH-D level increased, as implied by the mean bias of group 2 being higher than that of group 1. The higher variability in Group 2 points to the possible need for further evaluation or adjustments to optimize accuracy and consistency of X8.
Sunhyun Ahn (Wed,) studied this question.