Skip to main content
Premium Trial:

Request an Annual Quote

Blood-Based Omics Signatures Reflect 'Biological' BMI, Metabolic Health

BMI graphic alison_art_0.png

NEW YORK – Using multiomic, blood-based biomarkers, a team from the Institute for Systems Biology and the University of Washington has developed signatures of obesity and metabolic disease that can find affected individuals more accurately than body mass index based on height and weight.

"Biological BMI calculated from metabolomics data is more responsive to lifestyle changes, offering valuable feedback to monitor the effects of health interventions," senior and corresponding author Noa Rappaport, a researcher at the Institute for Systems Biology, said in an email. "This can be crucial in keeping people engaged in lifestyle changes by showing health benefits, even before or without weight loss."

As they reported in Nature Medicine on Monday, Rappaport and her colleagues studied some 1,111 blood markers, ranging from proteins or small molecule metabolites to traditional clinical lab measures, in more than 1,200 participants in the Arivale wellness program. They analyzed the data alongside the individuals' polygenic risk scores and gut microbiome profiles from 16S ribosomal RNA.

By analyzing the multiomic data with machine learning approaches, the team put together a metabolomics-based BMI score and other biological BMI prediction scores that tracked more closely with metabolic features and gut microbial community composition than traditional BMI scores — results that were subsequently validated with data for another 1,834 individuals from the TwinsUK study.

"[W]e trained machine learning models to predict baseline BMI for each of the omics platforms (metabolomics, proteomics, and clinical labs) or in combination: metabolomics-based BMI (MetBMI), proteomics-based BMI (ProtBMI), clinical labs (chemistries)-based BMI (ChemBMI), and combined omics-based BMI (CombiBMI) models," the study's authors explained, noting that the "resulting models retained 62 metabolites, 30 proteins, 20 clinical laboratory tests, and 132 analytes" across all of the models.

The omics-based bBMI models appeared to explain a large proportion of the variability found for classical BMI, the researchers explained, while maintaining associations with physiological, genetic, and lifestyle factors previously linked to classical BMI. 

Even so, the investigators saw intriguing patterns in the subset of individuals whose bBMI patterns differed from their classical BMI estimates. In particular, their results pointed to poorer metabolic health in individuals who had a "normal" classical BMI classification, but a higher-than-expected bBMI. On the other hand, metabolic health was better than expected in individuals with a low bBMI based on the omics-based signatures, despite an elevated classical BMI.

Likewise, in participants followed over time, lifestyle coaching and interventions around improved diet and exercise appeared to prompt pronounced metabolic responses, reflected in the MetBMI, even in the absence of weight loss. That appeared particularly true for individuals with higher bBMI scores who had normal BMI classifications by traditional measures.

"Although weight loss success rates are often low following lifestyle changes, there is evidence that these interventions can help prevent diabetes up to 20 years later, even if weight is regained," Rappaport added, noting that the long-term protective effect associated with improved metabolomic BMI measurements "underscores the importance of comprehensive molecular profiling in precision medicine to better understand the complex relationship between obesity, metabolic health, and chronic diseases."

With the available data, the team also confirmed blood biomarkers linked in the past to satiety and obesity, including specific metabolites such as uric acid, along with metabolism-, energy balance-, and inflammation-related proteins, including leptin, adiponectin, or FABP4.

"Our study not only confirmed these known associations, but also highlighted new candidate factors not previously measured, such as different forms of lipids," Rappaport said, explaining that the team's approach "further created a unique molecular signature of obesity, showing the diverse and complex connections between these molecules and how they influence our overall metabolic health."