A 44-blood metabolite signature may help predict type 2 diabetes (T2D) beyond traditional risk factors, according to a new study based on data from more than 20,000 individuals without baseline diabetes.

“Integrating these data in large samples allowed us to identify early metabolic pathway changes that precede type 2 diabetes, providing new insights into disease mechanisms,” said Jun Li, MD, epidemiologist and assistant professor of medicine at Harvard Medical School in Boston.

“By further examining how genetics and modifiable lifestyle factors, such as diet, physical activity, and body weight, are related to these metabolites, our findings help shift the field toward more precise and potentially more effective strategies for preventing type 2 diabetes,” Li told Medscape Medical News.

Metabolites and T2D Pathophysiology

The human metabolome — the composite of all small molecules in a biological sample — provides insight into biochemical processes, disease states, and drug responses. However, “despite decades of research, the precise biological mechanisms underlying type 2 diabetes are still not fully understood, and effective prevention strategies remain limited,” said Li.

“At the same time, recent advances in metabolomic technologies, together with the growing availability of large, long-term population studies, have made it possible to move beyond single biomarkers and systematically map entire metabolic pathways related to diabetes risk,” she said.

Li and her colleagues sought to examine how circulating metabolites, along with genetics and lifestyle, might predict T2D risk.

In their study, published in Nature Medicine, the researchers used data from 23,634 adults from 10 prospective cohorts with up to 26 years of follow-up and no baseline T2D. The researchers analyzed 469 metabolites in blood samples, along with genetic, diet, and lifestyle data. A total of 235 metabolites were associated with a higher or lower risk of developing T2D.

Additional genetic analyses connected the metabolites to features of T2D pathophysiology including insulin resistance, glucose-insulin response, ectopic fat deposition, energy-lipid regulation, and liver function. When the researchers added lifestyle factors to the mix, they found that lifestyle factors, notably physical activity, obesity, and diet, had a stronger effect on variations in T2D-associated metabolites than on variations in metabolites not associated with T2D.

Based on their findings, the researchers created a signature consisting of 44 metabolites that was an effective predictor of T2D risk in independent testing cohorts, with an area under the receiver operating characteristic curve ranging from 0.62 to 0.86. In a multivariable analysis of all the cohorts in the study, individuals in the highest decile of the metabolomic signature had approximately five times the risk of developing T2D compared with those in the lowest decile (risk ratio, 5.07; 95% CI, 4.02-6.39).

Surprise Findings and Applications

The overall results of the research were not unexpected, Li said. “For example, our integrative analyses of metabolites and genetics showed that metabolites associated with type 2 diabetes are genetically linked to organ systems central to the disease, such as the pancreas, liver, and adipose tissue.”

However, the researchers were surprised by how strongly and independently many of the metabolites associated with T2D were linked to modifiable lifestyle factors, especially diet, physical activity, and body weight. “This finding underscores the substantial influence that lifestyle factors may have on human metabolism and highlights their potential role in more effective strategies for preventing type 2 diabetes,” said Li.

She acknowledged that additional studies are needed before the results can be translated into clinical practice but noted that several potential applications can be explored. “One promising direction is the use of our 44-metabolite signature for improved type 2 diabetes risk assessment,” said Li. “In our study, this signature enhanced risk prediction beyond traditional clinical risk factors across multiple cohorts, suggesting it could eventually help identify individuals at higher risk earlier and more precisely,” she explained.

“Another potential application is in dietary and lifestyle intervention trials,” said Li. Given that many of the metabolites identified were related to modifiable risk factors, “the signature could be used to monitor individual metabolic responses to interventions, helping to evaluate effectiveness and potentially tailor strategies to maximize benefit.”

Study Limitations

Several limitations should be considered when interpreting the study findings, Li told Medscape Medical News. The study’s observational design prevents conclusions about causality or the biological mechanisms linking specific metabolites to T2D risk, she said. Experimental and mechanistic studies will be needed to clarify whether and how these metabolites play a causal role in disease development.

In addition, although the study included participants from multiple racial and ethnic backgrounds, approximately 77% of participants were non-Hispanic White, which limited the researchers’ ability to fully examine potential differences in associations across ethnic groups. More clinical trials are needed to validate the signature in diverse populations, Li said.

Studies are also needed to assess the performance of signature as a predictor in real-world clinical settings and to determine how the changes in a person’s metabolite profile over time align with response to interventions and risk reduction, she added.

Clinical Implications and Next Steps

“Diabetes prevalence is rising globally, and current risk models incompletely capture the chronic disease’s complexity,” said Charles Leonard, PharmD, MSCE, MPH, associate professor at the Perelman School of Medicine, University of Pennsylvania in Philadelphia, who was not involved in the current study.

Metabolomics makes possible the integration of metabolic, genetic, and lifestyle data to improve risk stratification and prevention, he told Medscape Medical News.

Although the new findings were not unexpected, “the breadth of associations was striking,” said Leonard. “The findings are an important reminder that type 2 diabetes is a systemic metabolic disorder, not just an issue with glucose.”

Ultimately, metabolomic signatures could complement traditional diabetes risk factors in identifying high-risk patients and supporting tailored prevention strategies, he said.

Looking ahead, “Replication in diverse populations is key, followed by investigations of whether metabolomic-guided interventions reduce diabetes incidence,” said Leonard. “Implementation research would likely be critical for integrating new tools into clinical practice.”

The study was funded by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), the National Heart, Lung, and Blood Institute; the New York Regional Center for Diabetes Translation Research via a grant from NIDDK; the National Institute on Aging; and a cooperative agreement from the National Human Genome Research Institute.

Li and Leonard reported having no financial conflicts to disclose.