Continuous glucose monitor–derived glucotypes and cardiovascular risk scores in individuals without diabetes

Abstract Introduction Dysglycemia is a well-established risk factor for cardiovascular disease (CVD); yet traditional glycemic traits, including fasting plasma glucose (FPG) and HbA1c, do not capture dynamic glucose fluctuations that may inform CVD risk. We cross-sectionally investigated the association of continuous glucose monitor (CGM)-derived metrics and 2-h post-prandial glucose (2-h PPG) with estimated 10-year CVD risk among individuals without diabetes. Methods We included 1,360 Framingham Heart Study participants (Third Generation, New Offspring Spouse, and Omni 2 cohorts at exam 4) without prevalent diabetes or CVD who had ≥3 days of CGM data and completed a mixed meal tolerance test (MMTT) with corresponding 2-PPG. We included 7 CGM summary metrics and defined data-driven glucotypes according to CGM measures of glycemic burden and variability. The 10-year CVD risk was estimated using the Predicting Risk of CVD EVENTs (PREVENT) base equations. We performed linear regression on standardized glycemic traits and glucotypes with log-transformed PREVENT risk scores and multinomial regression to relate standardized CGM metrics and 2-h PPG with PREVENT categories (low <5%[reference], borderline 5-<7.5%, intermediate/high ≥7.5%). All models were adjusted for FPG and body mass index (BMI). Results Among participants (55.9% women, 43.4% with prediabetes), mean age was 59.3 years, and mean BMI was 27.9 kg/m 2 . All CGM-derived metrics and 2-h PPG were positively associated with higher overall 10-year CVD risk (per 1 SD increase of each exposure variable, β range: 0.06-0.16, all p<0.001). A glucotype representing high glycemic burden and high glycemic variability was associated with higher overall 10-year CVD risk, compared with the glucotype representing low glycemic burden and low glycemic variability. Higher CGM-derived metrics and 2-h PPG were also associated with higher odds being in the intermediate/high CVD risk (OR range: 1.20-1.65, all p<0.001), adjusting for FPG and BMI. Conclusion Dynamic glycemic traits, including novel glucotypes that capture glycemic burden and variability, may provide novel insights into CVD risk prevention among individuals without T2D.