NEW YORK – Using nuclear magnetic resonance (NMR) spectroscopy, a team from Germany and other international centers has uncovered circulating metabolite clusters coinciding with nearly two dozen common diseases. They then used these results to develop disease-specific metabolomic profiles for predicting disease risk.
"Taken together, our work demonstrates the potential and limitations of NMR-derived metabolomic profiles as a multi-disease assay to inform the risk of many common diseases simultaneously," corresponding and co-senior author Roland Eils, a researcher affiliated with Charité-University Medicine Berlin and Heidelberg University Hospital, and his colleagues wrote in Nature Medicine on Thursday.
With NMR metabolomics measurements taken at baseline for almost 118,000 UK Biobank participants, the investigators established training, validation, and testing sets spanning 22 centers where the individuals were enrolled. From there, they turned to machine learning to gauge potential ties between 168 metabolomic markers and 24 common conditions — ranging from respiratory, vascular, or musculoskeletal diseases to cancer, neurological conditions, or metabolic maladies.
With roughly 1.4 million person years of clinical and other data collected from UK participants since they enrolled in the effort, the team explained, it was possible to tease out "integrative metabolomic states" coinciding with 23 of the 24 conditions considered. Just one disease, breast cancer, did not follow this pattern.
The researchers explored these disease-specific metabolomic features further using proton NMR (1H-NMR) assay profiles for almost 11,700 more individuals from four other population cohorts, validating their UK Biobank findings.
"In our perspective, 1H-NMR metabolomics profiling is an attractive candidate for a single-domain, multi-disease assay," the authors noted. "Because many countries already recommend regular check-ups entailing blood tests in the prevention of selected common diseases, our results indicate the potential of NMR metabolomic profiling in combination with simple demographic, but also with comprehensive laboratory predictors, to estimate disease risk."
By digging into clinical data spanning 10 years for 15 of the conditions, they demonstrated that metabolomic states can help to flag individuals with enhanced disease risk. While type 2 diabetes occurred in nearly 22 percent of individuals classified as being at high risk based on metabolomic state profiling, for example, T2D was documented in just 0.36 percent of individuals in the lowest metabolomic risk group.
The team saw dramatic disease risk differences between high- and low-risk metabolomic states for several other conditions, including heart failure and abdominal aortic aneurysm, while other conditions such as asthma or glaucoma had more modest metabolomic state-related risk ratios.
Along with analyses into the specific metabolites contributing to risk profiles for different diseases, the researchers looked at whether metabolomic states improved risk prediction over conventional demographic or clinical predictors. There, they saw eight conditions where metabolomics significantly improved risk prediction over established risk factors.
"These findings largely translate into potential clinical utility for NMR-based metabolomic profiling, both as a replacement for comprehensive laboratory examinations and as an additional source of discriminatory information to refine comprehensive risk assessments for multiple diseases simultaneously," they noted.