Abstract
OBJECTIVE: We assessed whether timing of physical activity, independent from the total activity amount, – which we refer to as chronoactivity – is associated with type 2 diabetes (T2D) risk.
METHODS: We included UK Biobank participants with valid accelerometry data (UK, exposure measurement: 2013-2015, follow-up till November 2023) and without diabetes mellitus at baseline (N = 89,439; mean age: 61.7 [SD:7.9] years). Relative hourly physical activity was calculated by dividing the average hourly clock time physical activity by the average hourly physical activity in a week. Participants were categorized into different chronoactivity clusters using k-means cluster analysis on relative hourly physical activity. We used multivariable-adjusted cox-proportional hazard regressions to examine associations between relative hourly physical activity, chronoactivity clusters and T2D, adjusted for potential confounders, including BMI as a potential mediator.
RESULTS: Over 7.8 (interquartile range: 7.2 to 8.3) years of follow-up, 2240 participants developed T2D. Higher relative hourly activity amounts during late morning (8:00-10:59) and late afternoon (15:00-15:59, 17:00-17:59) were associated with approximately 5 %-10 % lower T2D risk. Four clusters of chronoactivity patterns were identified, notably: midday (reference), early morning peak, late morning peak, and evening peak. Compared with participants exhibiting a midday pattern, those with a late morning peak had a lower T2D risk (Hazards Ratio: 0.88, 95 %CI: 0.79, 0.98). Overall, all observations attenuated after additional BMI adjustment.
CONCLUSIONS: Independent of the total amount of physical activity, specific timing of physical activity represents an additional dimension in T2D risk.