Methotrexate (MTX) remains the first-choice treatment for rheumatoid arthritis (RA), but individual variability in response and adherence underscores the need for reliable biomarkers of long-term drug exposure. Intracellular methotrexate polyglutamates (MTXPGs), typically measured in red blood cells (RBCs), fulfill this role but require invasive venous sampling. This study aimed to develop and validate a multi-matrix LC–MS/MS method for measuring MTXPGs in capillary blood samples obtained via volumetric absorptive microsampling (VAMS) and dried blood spots (DBS), and to compare these methods with traditional matrices. The method was validated in accordance with ICH M10 guidelines across RBC, whole blood (WB), VAMS, and DBS samples. MTX and MTXPG2–5 and total MTXPG were measured in 40 matched clinical samples. MTXPG6–7 were not detected across the tested clinical samples. Validation using Passing–Bablok regression, Bland–Altman analysis, and Spearman correlation showed strong agreement between VAMS and DBS (slopes 0.95–1.07; bias −4.21% to 0.36%; SRCC ≥ 0.969), with up to 100% of samples within ±20% of the agreement limits for total MTXPG. Significant differences were observed between capillary matrices and RBCs, with higher MTXPG levels in erythrocytes (bias up to −28%). Whole blood showed closer agreement with microsampling methods. ISR pass rates ranged from 84% to 95%, and stability tests indicated matrix- and chain length-dependent degradation, particularly for long-chain MTXPGs. These findings show that VAMS and DBS yield comparable results and can be considered interchangeable within a capillary-sampling framework. However, interpretation must account for matrix-specific differences when relating measurements to RBC-based reference values. This validated method could support the analytical feasibility of decentralized MTXPG monitoring in RA. However, prospective studies linking matrix-specific thresholds with disease activity, adherence, and toxicity are required before implementation for therapeutic decision-making.
Kocur et al. (Fri,) studied this question.