Systematic review and meta‐analysis on population attributable fraction for physical inactivity to dementia

Abstract INTRODUCTION The number of cases of dementia attributable to physical inactivity remains unclear due to heterogeneity in physical inactivity definitions and statistical approaches used. METHODS Studies that used population‐based samples to estimate the population attributable fraction (PAF) of physical inactivity for dementia were included in this review. Weighted PAFs were adjusted for communality among the risk factors (i.e., inactive persons may also share other risk factors) analyzed. Values were reported as percentage (%) of cases of dementia attributable to physical inactivity. RESULTS We included 22 studies. The overall impact of physical inactivity, defined by any criteria, on dementia ranged from 6.6% (95% CI: 3.6%, 9.6%; weighted) to 16.6% (95% CI: 14.4%, 18.9%; unweighted). Studies using the WHO criterion for physical inactivity estimated a higher unweighted impact (β = 7.3%; 95% CI: 2.0%, 12.6%) than studies using other criteria. DISCUSSION Conservatively, one in 15 cases of dementia may be attributable to physical inactivity, defined by any criteria.