Comparison of oxcarbazepine metabolite concentrations measured by EMIT-based Siemens Viva-ProE® System and LC-MS/MS in Chinese patients

Ming Chen^1,2Rong-Qi Lin³Yun-Yi Mao^1,2Ying-Bin Huang^1,2Jun-Nan Wu^1,2Xue-Yong Li¹Xue-Mei Wu^1,2Yu Cheng^1,2*Hong-Qiang Qiu^1,2*

• ¹Department of Pharmacy, Fujian Medical University Union Hospital, Fuzhou, China
• ²Department of Pharmacy, Fujian Medical University, Fuzhou, China
• ³Shanghang County Hospital, Shanghang, China

Background: Therapeutic drug monitoring (TDM) of oxcarbazepine's active metabolite, the monohydroxy derivative (MHD), is essential for effective seizure management. Although liquid chromatography-tandem mass spectrometry (LC-MS/MS) is considered the gold standard for MHD quantification, its technical complexity restricts widespread clinical utility. The Siemens Viva-ProE® System (SVPS), an automated immunoassay platform, presents a promising alternative. However, its comparability with LC-MS/MS warrants thorough and systematic evaluation. Objective: This study established and validated an LC-MS/MS method for quantifying MHD in plasma and assessed the correlation and concordance of SVPS measurements using concentration-specific Deming regression. The objective was to evaluate the feasibility of replacing LC-MS/MS with SVPS for TDM LC-MS/MS in clinical practice. Methods: A validated LC-MS/MS method (linear range: 0.18–39.30 µg/mL; intra/inter-day RSD < 15%) and SVPS (measurable range: 0.00–50.00 µg/mL) were applied to analyze 158 plasma samples. Correlation and concordance between the methods were assessed using Spearman's correlation, intraclass correlation coefficient (ICC), linear regression and Deming regression, Bland–Altman analysis, and Wilcoxon signed-rank tests. Stratified subgroup analyses, classified as low(< 12 µg/mL), medium(12–22 µg/mL), and high(> 22 µg/mL) concentration ranges, were conducted to evaluate the clinical acceptability of corrected SVPS values. Results: SVPS demonstrated a concentration-dependent positive bias (+13.04%) relative to LC-MS/MS. Despite this bias, strong overall correlation and concordance were observed (r = 0.9547, ICC = 0.952; p < 0.001). The overall Deming regression was defined by the equation: [LC-MS/MS] = 0.9763 × [SVPS] – 1.336. After correction, SVPS exhibited clinically acceptable concordance with LC-MS/MS within the low and medium concentration ranges, but not at higher concentrations. Conclusion: While uncorrected SVPS results exhibit a systematic bias that produces direct interchangeability with LC-MS/MS, applying a concentration-specific Deming correction enables clinically reliable TDM of MHD at concentrations below 22 µg/mL. However, method optimization is still required for accurate quantification in the high-concentration range.