Evaluation and correction of charcoal-stripped serum matrix bias in stable-isotope-dilution LC-MS/MS quantification of endogenous serum folates

Isotope-dilution LC-MS/MS is widely used for targeted quantification of endogenous metabolites, but surrogate-matrix calibration can remain biased when the surrogate matrix is noncommutable with native specimens. In this study, three endogenous serum folate vitamers were quantified by stable-isotope-dilution LC-MS/MS using charcoal-stripped serum (CSS) as a surrogate blank matrix. Post-extraction spiking and pooled-serum standard addition showed analyte-dependent and batch-dependent matrix bias between CSS and native serum. Standard addition comparison revealed matrix factors of 1.005 to 1.611 for 5-methyltetrahydrofolate (5-mTHF), 0.607 to 1.000 for folic acid (FA), and 0.902 to 1.008 for folinic acid, indicating a higher analyte-to-internal-standard normalized response in native serum relative to CSS for 5-mTHF, a lower normalized response for FA, and minimal matrix-dependent response difference for folinic acid. To address this bias, we developed a pooled-serum, batch-specific matrix-factor correction workflow based on standard addition. This approach estimates matrix factors in pooled native serum and applies them to CSS-calibrated concentrations generated under the same batch conditions, avoiding the sample-volume burden of conventional standard addition. Matrix-factor correction reduced validation bias for 5-mTHF and FA to within the predefined ± 15% criterion. In 64 clinical serum samples, correction changed classification relative to selected total folate reference decision limits (5-21 ng/mL) in ten samples (15.6%) for 5-mTHF and five samples (7.8%) for FA. These findings show that CSS-based stable-isotope-dilution LC-MS/MS can produce residual matrix-dependent bias despite normalization using stable-isotope-labeled internal standards (SIL-IS), and that pooled-serum standard addition provides a practical correction strategy.