Abstract
Background and Aims
The glucose-lowering effect of exercise has been well characterized in controlled settings. However, the extent to which real-world physical activities modulate glucose levels is less understood. This study evaluated sensor glucose (SG) change during logged activity events in individuals wearing continuous glucose monitor (CGM) and fitness tracker devices.
Methods
Data were uploaded between 6/2017-10/2017 by 5 individuals aged 27-63 years (mean 44.9±12.9 years) who used the MiniMedTM 530G system with the Sugar.IQTM diabetes assistant. Corresponding activity logs, steps, and metabolic equivalents (METs) were acquired with Fitbit® Charge 2. Change in SG during activities ranging 15 to 154 minutes in duration was calculated as the difference between the end and start SG level of within each logged activity event. Only events with a negative SG change, and no active insulin from bolus, were retained for analyses. The association of SG change with sum of METs, sum of steps, duration, and heart rate (HR), were assessed with Pearson’s correlation.
Results
There were 16 events out of a total 174 that met both criteria of no active insulin from bolus (21%) and negative SG change (63%), which included walking, running, cleaning, and gardening. Strong correlations were observed for SG change against METs (r= -0.986, p<0.001), duration (r= -0.899, p<0.001), steps (r= -0.829, p<0.001), and HR (p= -0.566, p=0.028).
Conclusions
Glucose reduction during exercise can be observed and characterized in real-world scenarios with CGM and fitness tracking. Future models of exercise impact on SG may optimize algorithm prediction accuracy.
