Project Grant F31DK146694

CHARACTERIZING GLYCEMIC HETEROGENEITY IN PREDIABETES AND ITS RELATIONSHIPS WITH LIFESTYLE PATTERNS AND CARDIOMETABOLIC RISK FACTORS - PROJECT SUMMARY PREDIABETES (PD) AFFECTS NEARLY 98 MILLION U.S. ADULTS AND POSES A SIGNIFICANT PUBLIC HEALTH CHALLENGE DUE TO ITS HIGH PREVALENCE, FREQUENT UNDER-RECOGNITION, AND INCREASED RISK OF PROGRESSION TO DIABETES AND RELATED COMPLICATIONS. DESPITE NOT HAVING ESTABLISHED TYPE 2 DIABETES (T2D), CARDIOMETABOLIC RISK FACTORS (E.G. DYSLIPIDEMIA, HYPERTENSION) ARE HIGHLY PREVALENT IN INDIVIDUALS WITH PD, CREATING A GROWING BURDEN OF MACROVASCULAR DISEASE. PD IS HIGHLY HETEROGENEOUS, WITH INDIVIDUALS EXHIBITING DISTINCT GLYCEMIC PROFILES AND VARYING SUSCEPTIBILITY TO PROGRESSION AND COMPLICATIONS. WHILE CONTINUOUS GLUCOSE MONITORING (CGM) ENABLES THE CHARACTERIZATION OF GLYCEMIC PROFILES BEYOND TRADITIONAL METRICS LIKE HBA1C, THE EXTENT TO WHICH LIFESTYLE FACTORS CONTRIBUTE TO THIS VARIABILITY AND THE IDENTIFICATION OF PROFILES ASSOCIATED WITH ELEVATED CARDIOMETABOLIC RISK REMAIN UNDEREXPLORED. UNDERSTANDING THE ASSOCIATIONS BETWEEN GLYCEMIC PROFILES, LIFESTYLE FACTORS, AND CARDIOMETABOLIC RISK IS ESSENTIAL FOR REVEALING THE DRIVERS OF GLYCEMIC HETEROGENEITY AND ASSOCIATED HEALTH OUTCOMES. IDENTIFYING CLINICALLY MEANINGFUL GLYCEMIC SUBTYPES OF PD USING CGM DATA CAN FURTHER ENABLE PERSONALIZED RISK STRATIFICATION AND TARGETED INTERVENTIONS. THIS INVESTIGATION IS NOW MADE POSSIBLE BY THE RECENT RELEASE OF THE ARTIFICIAL INTELLIGENCE READY AND EQUITABLE ATLAS FOR DIABETES INSIGHTS (AI-READI) DATASET, WHICH PROVIDES MULTIMODAL DATA FROM CGM, SMARTWATCHES, AND CLINICAL TESTS. THE OVERALL HYPOTHESIS IS THAT CGM-DERIVED MEASURES OF GLUCOSE DYNAMICS ARE ASSOCIATED WITH SMARTWATCH-DERIVED LIFESTYLE FACTORS AND CARDIOMETABOLIC RISK FACTORS, AND THAT THESE MEASURES CAN BE USED TO DEFINE GENERALIZABLE GLYCEMIC SUBTYPES OF PD. THE SPECIFIC AIMS OF THIS PROJECT ARE: AIM 1. CHARACTERIZE RELATIONSHIPS BETWEEN SMARTWATCH-DERIVED LIFESTYLE FACTORS (E.G., PHYSICAL ACTIVITY, SLEEP, AND PSYCHOLOGICAL STRESS) AND GLYCEMIC PROFILES IN PD THROUGH STATISTICAL MODELING. AIM 2. IDENTIFY GLYCEMIC PATTERNS ASSOCIATED WITH CARDIOMETABOLIC RISK FACTORS AND MACROVASCULAR COMPLICATIONS, SUCH AS STROKE AND MYOCARDIAL INFARCTION. AIM 3. DISCOVER GENERALIZABLE AND CLINICALLY MEANINGFUL PD GLYCEMIC SUBTYPES USING UNSUPERVISED LEARNING METHODS AND CHARACTERIZE THEIR UNIQUE GLYCEMIC PROFILES, LIFESTYLE PATTERNS, AND CARDIOMETABOLIC RISKS. THIS RESEARCH WILL CONTRIBUTE TO THE FIELD'S LONG-TERM GOAL OF INTRODUCING A MORE PROACTIVE, INDIVIDUALIZED APPROACH TO METABOLIC HEALTHCARE ENABLED BY WEARABLE DEVICES.