Abstract WE435: Metabolomic Signatures Linking Continuous Glucose Monitoring–Derived Fasting Glucose Variability to Increased Type 2 Diabetes Risk

Introduction: Fasting plasma glucose is a cornerstone for diagnosing prediabetes and diabetes. However, a single measurement overlooks substantial intraindividual variability, as revealed by continuous glucose monitoring (CGM). A recent study showed that two-week CGM-based fasting glucose assessments reclassified ~40% of individuals previously deemed normoglycemic as having prediabetes. Yet, the systemic metabolic features underlying short-term fasting glucose variability, and its prospective association with type 2 diabetes risk, remain unknown. Methods: We analyzed 1059 Israeli adults from the Human Phenotype Project (HPP) who wore a CGM device for two weeks between 2018-2020 and had available Nightingale metabolomics and real-time dietary logging. Fasting glucose was defined as CGM readings between 6-9 AM with prior fasting duration 8 hours. Short-term fasting glucose variability was quantified as the standard deviation (SD) of fasting glucose. Associations between individual metabolites and fasting glucose SD were assessed using linear regression, adjusting for age, sex, education, body mass index (BMI), smoking, alcohol intake, physical activity, diet, CGM-derived mean glucose, and mean fasting glucose. Elastic-net regression with ten-fold cross-validation (100 repeats) identified metabolites jointly associated with fasting glucose SD, with those selected 70 times were included in the metabolomic signature. The signature was then evaluated for incident type 2 diabetes among 6246 diabetes-free adults in the UK Biobank (2006–2010 baseline; follow-up through 2022) using Cox models. Results: In the HPP, 81 metabolites were associated with fasting glucose SD (FDR<0.05), with 67 having inverse associations (e.g., average diameter for HDL particles and phospholipids in Large LDL) and 14 having positive associations (e.g., concentration of medium and small LDL particles). A 24-metabolite panel was jointly associated with fasting glucose SD. In the UK Biobank, each 1-SD increase in the metabolomic signature corresponded to an 18% higher risk of type 2 diabetes (HR, 1.18, 95% CI, 1.07–1.31; P <0.001) independent of BMI, HbA1c, and fasting plasma glucose. Conclusions: Short-term fasting glucose variability signals systemic metabolic dysregulation and predicts type 2 diabetes risk beyond HbA1c and fasting glucose, suggesting that it may capture unique aspects of glycemic dysregulation beyond traditional metrics.