Abstract
Serum neurofilament light chain (sNfL) reflects neuro‑axonal injury, and is an emerging biomarker in multiple sclerosis (MS). This prospective cross-sectional study compared analytical agreement and clinical applicability among three analytical platforms. Serum samples from adult MS patients were analysed by single molecule array (Simoa) (frozen samples), high-sensitivity ELISA (hsELISA) (frozen samples), and fully automated chemiluminescent immunoassay (CLIA) (fresh and frozen samples). Simoa and hsELISA were strongly correlated (rₛ = 0.796) without systematic bias. CLIA with frozen samples showed higher sNfL levels (relative bias: 39.53% vs. Simoa; 29.56% vs. hsELISA). CLIA with fresh samples correlated very strongly with Simoa (rₛ = 0.820) and strongly with hsELISA (r(s) = 0.764) (relative bias: -7.7% vs. Simoa; -14.3% vs. hsELISA), though broad limits of agreement indicated notable individual variability. Comparison of fresh vs. frozen CLIA values indicated an influence of pre-analytical conditions. Age was positively associated with sNfL determined by Simoa and hsELISA, whereas EDSS correlated only weakly with Simoa-derived sNfL. Simoa and hsELISA yield interchangeable sNfL results with consistent biological correlations, supporting their clinical and research application. CLIA can be automated but is affected by pre-analytical factors. Assay harmonization is essential before routine clinical implementation.