A rigorous UHPLC-HRMS untargeted workflow for bipolar disorder metabolic fingerprinting

Bipolar Disorder (BD) is a complex psychiatric condition characterized by alternating episodes of mania and depression. Diagnosis relies exclusively on clinical evaluation, frequently resulting in misdiagnosis with other psychiatric disorders due to overlapping symptomatology. This diagnostic uncertainty necessitates the development of objective laboratory-based tools to complement clinical assessment. The absence of reliable biomarkers represents a critical gap that hinders accurate diagnosis and timely therapeutic intervention. This study addresses the need for objective metabolic signatures capable of distinguishing BD patients from healthy individuals through urinary metabolomics profiling. Ultra High-Performance Liquid Chromatography coupled to High Resolution Mass Spectrometry (UHPLC-HRMS) was employed to analyze urine samples from 20 BD patients and 20 healthy controls in both positive and negative ionization modes. A rigorous untargeted metabolomics workflow incorporating blanks, internal standards, and pooled quality control samples ensured high data quality and reproducibility throughout the analytical sequence. Features were extracted and putatively annotated using Compound Discoverer® v3.3 software, followed by implementation of a structured filtering pipeline designed to address high-dimensional data and potential confounding effects from pharmacological treatments. Exploratory principal component analysis revealed clear group separation, while partial least squares discriminant analysis achieved excellent classification performance with accuracy exceeding 90%. The retained metabolic features demonstrated robust discriminative power, successfully distinguishing BD patients from healthy controls based on characteristic urinary metabolic profiles. This research demonstrates that a characteristic urinary metabolic fingerprint distinguishes BD patients from healthy individuals with high accuracy. The validated analytical and chemometric workflow provides a foundation for developing objective laboratory-based diagnostic tools that could complement clinical evaluation, potentially reducing misdiagnosis rates. These findings represent a significant step toward translating metabolomics research into clinically actionable biomarker strategies for psychiatric disorders. • UHPLC-HRMS untargeted metabolomics workflow to discover disease-associated markers. • A structured filtering pipeline ensures high-quality, robust feature selection. • Pharmacological-treatment effects were minimized through dedicated filtering steps. • PCA showed a clear BD vs HC separation, PLS-DA yielded good classification. • Identification of a urinary metabolic fingerprint for BD.