Introduction: Myalgic Encephalomyeltis/ Chronic Fatigue Syndrome (ME/CFS) is a poorly understood and disabling illness. The heterogenous symptom patterns, which can overlap with other diseases, has prompted our research into identifying shared and unique biological signatures for ME/CFS when compared to comorbidities. Traditional clinical measurements often lie within the normal range for ME/CFS patients and do not provide diagnostic insight, therefore our biological signatures have combined NMR metabolomic biomarkers and baseline characteristics to probe the disease from both biomolecular and clinical angles. Furthermore, metabolomics provides the ideal interface where internal biological and external environmental interactions can be observed and tracked.
Methods: This study consisted of 493 ME/CFS participants with NMR metabolomics data from the UK Biobank, and 7 other comorbid cohorts as positive control groups. The dataset consisted of 249 metabolic measures including lipoprotein subclasses, lipids, fatty acids, and low molecular weight metabolites generated by the Nightingale Health platform. A standardised data processing workflow and epidemiological statistical pipeline was followed.
Results: Significant biomarker associations that were common in ME/CFS and comorbidities were dominated by VLDL-related measurements and inverse associations with HDL and ApoA1, indicating a disruption in the transportation of phospholipids, triglycerides, and cholesterol. Low molecular weight metabolites provided the required discriminatory power, identifying unique associations of alanine and acetone (inverse) in ME/CFS. This finding supported an altered consumption of amino acids which is preferentially utilised over the more efficient carbohydrate metabolism. Furthermore, the variation in potential ME/CFS biomarkers were correlated and quantitated to physical and clinical measurements such as basal metabolic rate, reticulocyte count and testosterone levels to explore actionable antecedents.
Conclusion: Multifaceted illnesses such as ME/CFS pose a challenge for diagnosis, treatment, and research, however, presents the opportunity for precision medicine applications where NMR metabolomics can play a key role. The ability to detect various macromolecules in a single run, and the reproducible and quantitative nature of NMR are essential for the translation of metabolomics into clinical practise. This work showcases the integration of two different data types with an emphasis on biological context, laying the foundation for the development of future multi-omics and disease prediction models.