Objectively measured physical activity and all cause mortality: A systematic review and meta-analysis

Current physical activity recommendations have been based on evidence from systematic reviews of questionnaire-based data. Questionnaire-based physical activity data are subject to both random and non-random error. If the estimated association between physical activity and health outcomes was different when a more accurate, objective measure was used, this would have important health policy implications for physical activity. We conducted a systematic review and meta-analysis of published cohort studies that investigated the association between an objective measure of physical activity and all cause mortality. We searched PubMed, Scopus, Embase, Cochrane library, and SPORTDiscus for prospective cohort studies that examined the association between objectively measured (accelerometer, pedometer, or doubly labeled water method) physical activity and mortality in adults aged≥18 years, of either sex. Summary hazard ratios and 95% confidence interval [CI]s were computed using random-effects models. Thirty-three articles from 15 cohort studies were identified that together ascertained 3903 deaths. The mean years of follow-up ranged from 2.3-14.2 years. Individuals in the highest category of light, moderate-to-vigorous, and total physical activity had 40% (95%CI 20% to 55%), 56% (95%CI 41% to 67%), and 67% (95%CI 57% to 75%), respectively, lower risk for mortality compared to individuals in the lowest category of light, moderate-to-vigorous, and total physical activity. The summary hazard ratio for objectively measured physical activity and all cause mortality is lower than previously estimated from questionnaire based studies. Current recommendations for physical activity that are based on subjective measurement may underestimate the true reduction in mortality risk associated with physical activity.