This study explores the potential of 1H-NMR spectroscopy-based metabolic profiling in various biofluids as a diagnostic and predictive modality to assess disease severity in individuals with 5q spinal muscular atrophy. A total of 213 biosamples (urine, plasma, and CSF) from 153 treatment-naïve patients with SMA across five German centers were analyzed using 1H-NMR spectroscopy. Prediction models were developed using machine learning algorithms which enabled the patients with SMA to be grouped according to disease severity. A quantitative enrichment analysis was employed to identify metabolic pathways associated with disease progression. The results demonstrate high sensitivity (84-91%) and specificity (91-94%) in distinguishing treatment-naïve patients with SMA from controls across all biofluids. The urinary and plasma profiles differentiated between early-onset (type I) and later-onset (type II/III) SMA with over 80% accuracy. Key metabolic differences involved alterations in energy and amino acid metabolism. This study suggests that 1H-NMR spectroscopy based metabolic profiling may be a promising, non-invasive tool to identify patients with SMA and for severity stratification, potentially complementing current diagnostic and prognostic strategies in SMA management.
Keywords: 1H-NMR spectroscopy; metabolic profiling; metabolomics; spinal muscular atrophy.