Serum Metabolomic Profiling Across Five Oligoclonal Band (OCB) Patterns: A Targeted 1H-NMR Study in Serum

Abstract

Cerebrospinal fluid (CSF) oligoclonal band (OCB) analysis remains central to the diagnostic evaluation of neuroinflammatory diseases of the central nervous system (CNS), as it reflects intrathecal immunoglobulin synthesis. However, its reliance on lumbar puncture limits its applicability for screening and repeated longitudinal assessment. Serum metabolomics offers a minimally invasive strategy to explore peripheral biochemical correlates of central immune activity. Building on previous binary OCB comparisons, the present study extends serum metabolomic analysis to encompass all five classical OCB patterns, thereby capturing a broader immunological spectrum. A total of 92 adults undergoing diagnostic evaluation for suspected CNS inflammatory disorders were retrospectively stratified according to OCB type (Types 1-5). Serum samples were analysed using targeted H-1-NMR spectroscopy on a Bruker Avance Neo 600 MHz platform and processed using Bruker's IVDr pipeline. Group-wise differences were assessed using non-parametric statistical testing with false discovery rate (FDR) correction, complemented by effect size estimation, exploratory multivariate analyses, and Receiver Operating Characteristic (ROC) modelling. Distributional characteristics were further examined using boxplots and violin plots. Across analytical approaches, several metabolites-most prominently leucine, 2-oxoglutaric acid, histidine, threonine, and glycerol-exhibited nominal variation and moderate effect sizes across OCB patterns. Rather than discrete metabolic separation, these metabolites demonstrated graded shifts in central tendency accompanied by substantial overlap between groups. Unsupervised principal component analysis did not reveal robust clustering, while supervised multivariate models highlighted amino acid- and tricarboxylic acid cycle-related metabolites as contributors to partial differentiation. Post hoc power analysis indicated limited sensitivity to detect small-to-moderate effects under multiple-testing correction, supporting an exploratory interpretation of the findings. Taken together, this first targeted serum H-1-NMR metabolomic evaluation spanning all classical OCB patterns suggests that peripheral metabolic profiles may reflect graded immunometabolic variation associated with intrathecal immune activity. While not intended for diagnostic classification, these findings provide a spectrum-based framework for integrating serum metabolomics with OCB phenotyping and identify candidate metabolites for future prospectively powered and clinically characterised studies.