Benchtop NMR as a clinical tool for acute respiratory illness: metabolic signatures associated with disease severity

BACKGROUND: Acute respiratory failure remains a major challenge in critical care, and early identification of patients at risk of clinical deterioration is essential. Metabolomics provides a systems-level characterization of disease severity but its clinical implementation is limited by the need for high-resolution Nuclear Magnetic Resonance (NMR) infrastructure. Benchtop NMR offers a compact and potentially scalable alternative. This study aimed to identify serum metabolic signatures associated with respiratory severity and evaluate agreement between benchtop and high-resolution NMR platforms. METHODS: Serum samples from COVID-19 patients with acute respiratory failure were classified by respiratory severity (mild, moderate, severe, and very severe). Metabolomic profiling was performed using 500 MHz high-resolution and 80 MHz benchtop NMR systems. Multivariate analyses were used to identify severity-associated metabolic signatures, assess cross-platform agreement, and evaluate associations with clinical variables. RESULTS: Both platforms identified consistent metabolic alterations across severity groups. Supervised models achieved accurate discrimination between hospitalized and non-hospitalized patients and between moderate and very severe cases. Twenty-one metabolites, including glutamine, citrate, lactate, phenylalanine, myo-inositol, glycerol, and trimethylamine N-oxide, contributed to group separation. Altered pathways included carbohydrate metabolism, amino acid turnover, lipid metabolism, and tricarboxylic acid cycle intermediates, consistent with hypoxia and immune activation. Several metabolites correlated with inflammatory markers, immune cell counts, and oxygenation parameters. Importantly, benchtop NMR reproduced the key metabolic patterns observed with high-resolution NMR. CONCLUSIONS: Serum NMR metabolomics identifies metabolic signatures associated with respiratory severity. The agreement between high-resolution and benchtop NMR supports the feasibility of benchtop systems for clinically accessible patient stratification in acute respiratory disease, with potential applicability beyond COVID-19.