Six Novel Biomarkers Aid in Type 2 Diabetes Predictions in At-Risk Patients

risk for type 2 diabetes
risk for type 2 diabetes, t2d
Researchers sought to identify the most predictive set of biomarkers for T2D and validate them in an independent cohort study.

The evaluation of 6 novel biomarkers in addition to standard, noninvasive clinical modeling can significantly improve prediction of type 2 diabetes development in at-risk patients, according to study results published in The Journal of Clinical Endocrinology and Metabolism.

The newly identified novel biomarkers include interleukin-1 receptor antagonist (IL-1RA), insulin growth factor binding protein-2 (IGFBP-2), soluble E-selectin (sE selectin), decorin, adiponectin, and high-density lipoprotein cholesterol (HDL-C).

In order to assess whether any single novel biomarker, or a combination of several novel biomarkers, could improve prediction of type 2 diabetes, researchers conducted a large population-based case-cohort study. Specifically, investigators sought to identify the most predictive set of biomarkers and validate these predictors in an independent cohort study.

A total of 689 incident cases and 1850 noncases were used for the discovery portion of the study; the independent cohort study comprised 262 incident cases and 2549 noncases for validation. Ultimately, the efficacy of 47 novel biomarkers was evaluated.

Within the discovery study, the application of the basic adapted German Diabetes Risk Score (GDRS) model did “reasonably well” in predicting the 14-year risk for type 2 diabetes (C-index, 0.775; 95% CI, 0.755-0.790). In combination, the 6 selected biomarkers strongly improved C-index delta by 0.053 (95% CI, 0.039-0.066), with an additional significant improvement in both overall and case vs noncase category-free net reclassification index.

In the validation study, the C-index was “significantly improved” when all 6 biomarkers were simultaneously added to the adapted GDRS model (delta of C-index, 0.034; 95% CI, 0.019-0.053) and the adapted GDRS plus HbA1c model (delta C-index, 0.023; 95% CI, 0.009-0.039).

Study limitations include the initial selection of biomarkers without taking HbA1c into account, an inability to assess predictive value in addition to fasting glucose concentrations, and the use of oral glucose tolerance tests in the KORA validation cohort.

The investigators also noted that the continual emergence of new technologies, such as proteomics profiling and aptamer-based techniques, that allow for “simultaneous measurement of even larger biomarker panels” may further improve risk prediction.

“Risk prediction models including these markers may help to improve the identification of persons at high risk of developing type 2 diabetes in order to effectively target preventive efforts to those who are most in need,” the researchers concluded.

Disclosure: This clinical trial was partly supported by Tethys Bioscience and Singulex. Please see the original reference for a full list of authors’ disclosures.


Thorand B, Zierer A, Büyüközkan M, et al. A panel of six biomarkers significantly improves the prediction of type 2 diabetes in the MONICA/KORA study population. Published online December 31, 2020. J Clin Endocrinol Metab. doi:10.1210/clinem/dgaa953