Accelerometer-Derived Physical Activity, Sedentary Behavior, and Continuous Glucose Monitoring-Derived Glycemic Variability in Adults with Type 2 Diabetes: A Longitudinal Repeated-Measures Study.

Background: Day-to-day movement behavior may influence short-term glycemic variability in adults with type 2 diabetes, but free-living evidence integrating accelerometry with continuous glucose monitoring (CGM) remains limited. Objectives: This study examined whether daily moderate-to-vigorous physical activity (MVPA) and accelerometer-defined sedentary time were associated with 24-h within-day glycemic variability. Methods: In this longitudinal repeated-measures study, 140 adults with type 2 diabetes underwent 14 days of concurrent hip-worn ActiGraph GT3X+ accelerometry and FreeStyle Libre 2 CGM under free-living conditions in Lima, Peru. The primary outcome was a daily CGM-derived coefficient of variation (CV). Linear mixed-effects models separated within-person and between-person components, with Holm adjustment for the two primary within-person associations. Results: The fully adjusted primary analysis included 132 participants and 1728 participant-days. Higher within-person MVPA was associated with a lower daily CV (β = -0.27 percentage points per 10 min/day, 95% CI -0.45 to -0.09; Holm-adjusted p = 0.006), whereas greater accelerometer-defined sedentary time was associated with a higher CV (β = 0.05, 95% CI 0.01 to 0.09; Holm-adjusted p = 0.014). Secondary concurrent analyses showed favorable MVPA and opposite sedentary time associations for time in range, time above range, mean glucose, and glucose standard deviation (all false discovery rate q ≤ 0.018), with no statistically detectable association with time below range. Lagged associations were directionally consistent, but none remained statistically significant after multiplicity correction. Conclusions: Daily MVPA and accelerometer-defined sedentary times were associated with modest differences in within-day glycemic variability and concurrent CGM profiles. These observational findings do not establish causality or a specific exercise dose.