Analytical Concordance of Total Vitamin D on a Fully Automated Random-Access LC-MS/MS Platform.

Abstract

The adoption of LC-MS/MS laboratory developed tests in the clinical laboratory is limited by many factors including the lack of automation. Recently, the Cascadion™ clinical analyzer was introduced as a fully automated random-access LC-MS/MS platform. Here, the analytical concordance between the platform and a Roche immunoassay analyzer was investigated for vitamin D analysis in human serum, including samples selected for high triglyceride levels. Analytical precision was evaluated on 3 levels of QC samples (10, 30, and 90 ng/mL) within days (n = 4, 5 days) and between days (20 days). Assay comparison to the Roche was performed using reference samples from the CDC and CAP programs for accuracy. Concordance was also monitored using routine patient samples, as well as samples selected for elevated triglyceride levels (>250 mg/dL). Precision met manufacturer specifications (<10% CV and <15% bias), whereas the accuracy evaluations showed a linear fit (y = 0.97x - 1.1, r = 0.995) with 1:1 correlation to reference samples, independent of C-3-epi-vitamin D levels. A mean positive bias (11%) was observed for the Roche measurements in normal patient samples, whereas a mean negative bias (-8%) was observed in samples selected for elevated triglyceride levels. Cascadion measurements of total vitamin D compared favorably with Roche results in our laboratory, although discordance was observed in the analysis of patient serum, which could be explained in terms of known differences between the 2 assays. However, operational issues need to be addressed to effect clinical adoption.