Combined effects of diet, physical activity, obesity on continuous glucose monitoring-derived measures
OA Version
Citation
Abstract
BACKGROUND: Diabetes is a prevalent disease that affects millions of people of all ages globally. The prevention of this disease is a crucial step in improving population health and longevity. While factors such as physical activity (PA), diet, and body mass index (BMI) are individually linked to traditional measures of diabetes status (including hemoglobin A1c and fasting blood glucose), their collective impact on more complex glycemic features that can be measured using continuous glucose monitoring (CGM) has not been fully elucidated. The primary objective of this study was to assess the joint association of diet, PA, and obesity with CGM-derived traits in individuals with prediabetes and normoglycemia.
METHODS: The data utilized in this study were derived from participants in the Third Generation of the Framingham Heart Study, Omni 2 and New Offspring Spouse cohorts who wore a CGM and Fitbit monitor for a minimum of three complete days. Individuals with diabetes mellitus, those taking glucose-lowering medications, and those with less than two days of diet record completion were excluded from the study. The final sample size consisted 655 individuals, including 377 with normoglycemia and 278 with prediabetes. We used multivariable linear regression to examine the association of diet quality (AHEI – Alternative Healthy Eating Index), PA (average steps per day) and BMI with CGM-derived measures (mean glucose, coefficient of variation (CV), mean difference between glucose values obtained at the same time of day (MODD), mean amplitude of glycemic excursions (MAGE), continuous overall net glycemic action (CONGA-1) and time spent >140 mg/dL). The models were adjusted for age, sex, CGM device lot number, CGM device location (arm vs. abdomen), alcohol and smoking status. We assessed models additionally including the other predictors (BMI, PA, and diet quality) in a stepwise fashion.
RESULTS: Our cohort of 655 individuals demonstrated that those with prediabetes (N=278) had a higher BMI (29.2 vs. 26.5), and higher levels of CGM metrics compared to participants with normoglycemia (N=377). The study revealed significant associations between PA, diet, and BMI with glycemic control. The results showed that each additional 2,000 steps taken per day by the prediabetic cohort was associated with a reduction in mean glucose levels of approximately 1 mg/dL (Parameter Est. = -0.60, p=0.004). A similar reduction in mean glucose of 1 mg/dL was associated with a 5-6 point improvement in diet quality (Parameter Est. = -0.17, p=0.006), and a decrease in BMI of 3 kg/m² (Parameter Est. = 0.37, p=0.03). Notably, higher PA and better diet quality were linked to lower mean glucose levels and improved glucose variability, with these associations being more pronounced in normoglycemic individuals. In contrast, the prediabetic cohort exhibited fewer significant associations, with diet and PA showing moderate effects on mean glucose levels and J-index. Importantly, the interaction between PA and diet approached significance in the prediabetic group (p=0.05), indicating a potential combined effect on glycemic outcomes.
CONCLUSION: Overall, the findings of this study underscore the importance of integrating both PA and a high-quality diet to achieve optimal glycemic control. While PA alone was associated with improved glucose metrics, the combination of regular exercise and a healthy diet amplifies these associations, suggesting a synergistic effect. These results highlight the need for tailored lifestyle interventions that consider both PA and diet quality to effectively manage and prevent glucose dysregulation.
Description
2025