Metabolomic biomarkers improve diabetes risk prediction accuracy

Utilizing cutting-edge metabolomics, this study reveals a cost-effective and accurate diabetes prediction model that has the potential to revolutionize early detection and prevention strategies.

study: A new type 2 diabetes prediction score based on traditional risk factors and circulating metabolites: model derivation and validation in two large cohort studies.. Image credit: Shutterstock AI

In a recent study published in the journal e-clinical medicineResearchers evaluated the incremental value of adding metabolomic-derived biomarkers to the traditional Cambridge Diabetes Risk Score (CDRS) in predicting 10-year diabetes risk. This study utilized data from two large cohorts, the UK Biobank and the German ESTHER cohort, to ensure robust model development and validation. 11 biomarkers significantly improve CDRS accuracy (0.815 to 0.834) from metabolite data from over 86,000 UK Biobank (training and internal validation) and approximately 4,400 German ESTHER cohort (external validation) participants has become clear.

Notably, a concise predictive model using only four low-cost and easily available metabolite biomarkers achieved comparable accuracy, highlighting its usefulness in routine diabetes risk assessment.

background

Type 2 diabetes (T2D) is a chronic medical condition characterized by unhealthy blood sugar levels, leading to potentially life-threatening complications such as cardiovascular disease (CVD), kidney disease, and vision loss. The condition is caused by the body's inability to secrete or utilize sufficient amounts of insulin and is thought to be caused by genetics, health behaviors (sleep and physical activity levels), and high body weight.

Alarmingly, the prevalence of T2D is increasing at unprecedented levels, resulting in significant economic, quality of life, and mortality burdens for patients and their families. Unfortunately, no cure exists for diabetes, and traditional clinical interventions aim to reduce or delay the onset of T2D. Early detection or prediction of T2D risk is essential to prepare clinicians and potential patients for chronic T2D. Unfortunately, current predictive approaches, although able to distinguish between low and high risk, lack specificity and can be confounded by combinations of risk factors.

About research

Advances in nuclear magnetic resonance (NMR) spectroscopy and its application in metabolomics studies provide a more comprehensive and nuanced view of the various metabolomic changes that precede T2D onset, highlighting its potential in T2D predictive modeling. . The present study leverages this concept in deriving a new T2D risk score using NMR metabolomics data from two longitudinal population-based datasets.

Study data for model derivation and internal validation were from the United Kingdom Biobank (UKB; 70% training, 30% validation) comprising 502,493 participants aged 37-73 years across 22 sites in Scotland, Wales and England. I got it. The resulting model was externally validated using the ESTHER cohort, a German (Saarland)-centered population dataset of 9,940 participants aged 50 to 75 years. Participants with a history of diabetes and missing data were excluded from modeling and analysis.

NMR metabolomics data were acquired using a high-throughput Nightingale Health platform containing 250 plasma-derived metabolites. During the analysis, one of the metabolites (glycerol) was missing from the dataset for most participants and was therefore excluded from the final model derivation. Model selection identified the least absolute shrinkage and selection operator (LASSO) as the most reliable. To further improve model performance, a logarithmic transformation of the input data (which ensures normality and accounts for outliers) was performed.

The study also performed subgroup analyzes to examine the robustness of the new model across different groups, including age, gender, and obesity status. To compare and improve the current gold standard for T2D prediction, variables from the Cambridge Diabetes Risk Score (CDRS) were included in the model derivation, and the 249 metabolites evaluated herein were added to the CDRS variables in a stepwise manner. Now. This allows us to calculate the relative improvement in predictive power for each additional metabolomic variable and identify the variable with the highest predictive power (receiver operating characteristic curve). [ROC] analysis).

Research results

Of the over 512,000 participants comprising the UKB and ESTHER cohorts, 86,232 (UKB) and 4,383 (ESTHER) participants met study inclusion criteria and were included in the analysis (model derivation and validation) . These participants had a similar age (59.9 and 60.2 years) and gender distribution (44.3% and 42.7% male) in the UKB and ESTHER datasets, respectively. Notably, baseline body mass index (BMI) and hemoglobin A1c (HbA1c) levels were nearly identical between both cohorts, highlighting their comparability.

LASSO analysis revealed 11 metabolites with the highest predictive power among the 249 metabolites analyzed. These metabolites included glycolysis-related metabolites (n = 4), ketone bodies (n = 2), amino acids (n = 2), lipoprotein-related metabolites (n = 2), and fatty acid-related metabolites ( n = 1). Key examples include glucose, pyruvate, lactic acid, and citric acid. Of note, the 10-year prediction accuracy of these metabolites independent of CDRS variables was high (C-index = 0.733 and 0.735 in internal and external validation datasets, respectively). Stepwise combination of UKB-derived metabolites with CRDS significantly improved prediction accuracy (baseline C-index = 0.815) (combined C-index = 0.834).

Similar improvements were observed in the external ESTHER validation cohort (C-index increased from 0.770 to 0.798). Impressively, the new 'UK Biobank Diabetes Risk Score (UKB-DRS)' model utilized only 4 of the 11 identified metabolites to achieve comparable improvements.

The robustness of the simplified model was evident in the calibration curves and showed similar predictive performance as the full model in both cohorts.

conclusion

This study represents the most extensive dataset used to derive T2D prediction models and highlights the value of NMR-derived metabolite data in assessing an individual's 10-year risk of developing T2D. Advances in NMR imaging have significantly reduced the economic burden of these evaluations, and the ease of data collection (small amounts of blood-derived plasma) highlights the methodological and clinical utility of this approach.

This study presents a new UKB-DRS model that significantly improves the predictive accuracy of the current gold standard (CDRS), thereby providing clinicians and potential patients with an opportunity to reduce, delay, or avoid T2D onset. Provide additional time. Future studies are recommended to test this model across ethnically diverse populations and younger age groups.