Prediabetes is not a uniform condition, and individuals differ widely in their likelihood of progressing to type 2 diabetes or developing complications. Previous research had classified prediabetes into six distinct clusters, including three linked to a high risk of adverse outcomes. However, identifying these clusters required detailed clinical and metabolic testing, limiting their practicality in routine care.
In the present study, Amandeep Singh from the Department of Experimental Diabetology at the German Institute of Human Nutrition Potsdam-Rehbrücke (DIfE) and colleagues aimed to develop a simpler, blood-based method for identifying high-risk prediabetes clusters. Using a machine learning workflow, the researchers analysed DNA methylation patterns in blood cells—epigenetic changes known to be relatively stable over time and influenced by metabolic health.
The analysis revealed the following findings:
- The analysis included individuals from four prediabetes clusters: cluster 2 (low risk) and clusters 3, 5, and 6 (high risk).
- In the discovery cohort of 187 participants, researchers identified 1,557 CpG sites that could predict prediabetes cluster membership.
- These epigenetic markers were validated in an independent replication cohort of 146 individuals.
- The markers distinguished high-risk clusters with an accuracy of approximately 92%, demonstrating the robustness of the approach.
- Each high-risk cluster displayed a distinct epigenetic signature.
- Between 300 and 339 CpG sites were specific to individual high-risk clusters.
- Genes linked to cluster 3 were associated with TGF-β receptor and calcium signalling pathways.
- Cluster 5–specific CpG sites were linked to the MAPK cascade and extracellular matrix organisation.
- Cluster 6 showed enrichment in genes involved in Wnt and SMAD signalling pathways.
- These pathway-specific differences mirrored the varying patterns and severity of metabolic deterioration across clusters, supporting the biological relevance of the identified epigenetic markers.
Clinically, the use of blood-based epigenetic markers could reduce reliance on time- and resource-intensive tests, such as the oral glucose tolerance test. This may allow risk stratification to be applied more broadly and enable earlier, targeted preventive interventions for those most likely to progress to diabetes or complications.
The authors note several limitations. The study populations were of German ancestry, which may limit generalisability to other ethnic groups. The relatively small cohort size required careful cross-validation, and additional functional studies are needed to confirm causal mechanisms. Moreover, while blood DNA methylation provides stable markers, linking these changes directly to specific metabolic tissues remains challenging.
Overall, the study highlights the potential of blood-based epigenetic profiling as a practical tool for identifying high-risk prediabetes subtypes and supporting more personalised diabetes prevention strategies.
Reference:
Singh, A., Schwartzenberg, R.Jv., Wagner, R. et al. Stratifying high-risk prediabetes clusters using blood-based epigenetic markers. Biomark Res 14, 19 (2026). https://doi.org/10.1186/s40364-025-00887-8
Disclaimer: This website is primarily for healthcare professionals. The content here does not replace medical advice and should not be used as medical, diagnostic, endorsement, treatment, or prescription advice. Medical science evolves rapidly, and we strive to keep our information current. If you find any discrepancies, please contact us at corrections@medicaldialogues.in. Read our Correction Policy here. Nothing here should be used as a substitute for medical advice, diagnosis, or treatment. We do not endorse any healthcare advice that contradicts a physician's guidance. Use of this site is subject to our Terms of Use, Privacy Policy, and Advertisement Policy. For more details, read our Full Disclaimer here.
NOTE: Join us in combating medical misinformation. If you encounter a questionable health, medical, or medical education claim, email us at factcheck@medicaldialogues.in for evaluation.