Association between device-based measured physical activity timing and depression risk: a prospective cohort study of UK Biobank participants.

Introduction: Physical activity (PA) decreases the risk of mental diseases. Although PA timing has been hypothesised to affect mental health outcomes, no large-scale epidemiological cohort studies have been performed. We investigated the association between PA timing and incident depression in the general population. Methods: A total of 89 934 European participants (62.4±7.9 years, age range: 43-79 years, 44.8% men) without a history of depression were analysed. Hourly PA levels were derived from accelerometry data and standardised for the total daily amount of PA. Participants were followed for a maximum of 8.5 years. Risks of incident depression dependent on individually standardised hourly PA level were assessed using Cox proportional-hazards models adjusted for potential confounders. Results: During follow-up, 1748 participants developed depression. Using the hourly standardised PA levels, we observed higher risks for incident depression with higher night-time PA (eg, at 4:00 HR 1.35; 95% CI 1.28, 1.41), and lower depression risks with higher morning-time PA (eg, at 9:00 HR 0.85; 95% CI 0.81, 0.89). Also in the subgroup analysis, compared with participants who were most active during midday, participants with most PA in the early morning had a lower risk for incident depression (HR 0.80; 95% CI 0.68, 0.95). No differences were observed when analyses were stratified for sex, chronotype or for the overall objective PA level. Conclusions: Morning PA was associated with lower risk and night-time PA was associated with increased risk of depression. These findings may suggest that future intervention studies should include timing as an additional dimension of PA as a possible treatment and prevention strategy for depression.