Abstract
OBJECTIVES: To investigate the associations between device-measured and self-reported physical activity (PA) and incident common mental disorders in the general population.</p>
DESIGN AND SETTING: Large-scale prospective cohort study.</p>
PARTICIPANTS: Using the UK Biobank data, a validated PA questionnaire was used to estimate self-reported weekly PA in 365 656 participants between 2006 and 2010 while 91 800 participants wore wrist-worn accelerometers for 7 days in 2013-2015 to derive objectively measured PA. All the participants were followed up until 2021.</p>
MAIN OUTCOME MEASURES: Incidences of depression and anxiety were ascertained from hospital inpatient records. Cox proportional hazards models and restricted cubic splines were used to assess the associations between subjectively and objectively measured PA and common mental disorders.</p>
RESULTS: During a median follow-up of 12.6 years, 16 589 cases of depression, 13 905 cases of anxiety and 5408 cases of comorbid depression and anxiety were documented in the questionnaire-based cohort. We found J-shaped associations of self-reported PA with incident risk of depression and anxiety, irrespective of PA intensities. The lowest risk for depression occurred at 550, 390, 180 and 560 min/week of light-intensity PA (LPA), moderate-intensity PA (MPA), vigorous-intensity PA (VPA) and moderate-to-vigorous PA (MVPA), respectively. During a median follow-up of 6.9 years, a total of 2258 cases of depression, 2166 cases of anxiety and 729 cases of comorbid depression and anxiety were documented in the accelerometer-based cohort. We found L-shaped associations of device-measured MPA and VPA with incident depression and anxiety. MPA was adversely associated with incident depression and anxiety until 660 min/week, after which the associations plateaued. The point of inflection for VPA occurred at 50 min/week, beyond which there was a diminished but continued reduction in the risks of depression and anxiety.</p>
CONCLUSION: Different patterns of associations between self-reported and device-measured PA and mental health were observed. Future PA guidelines should fully recognise this inconsistency and increasingly employ objectively measured PA standards.</p>