Dose-response relationship of total and leisure time physical activity to risk of heart failure: a prospective cohort study.

BACKGROUND The nature of the association between levels of physical activity and risk of heart failure is little known. We investigated nonlinear associations of total and leisure time physical activity with risk of heart failure. METHODS AND RESULTS In 1997, 39 805 persons without heart failure completed a questionnaire of lifestyle factors and medical history. We used Cox regression models to investigate total (adjusting for education and previous myocardial infarction) and direct (multivariable-adjusted) effects of self-reported total and leisure time physical activity on risk of heart failure of any cause and heart failure of nonischemic origin. Heart failure diagnoses were obtained until December 31, 2010. Higher leisure time physical activity was associated with lower risk of heart failure of any cause; hazard ratio of the total effect of leisure time physical activity was for fifth versus first quintile 0.54; 95% confidence interval was 0.44 to 0.66. The direct effect was similar. High total daily physical activity level was associated with lower risk of heart failure, although the effect was less pronounced than for leisure time physical activity (total effect hazard ratio, 0.81; 95% confidence interval, 0.69-0.95; fifth versus first quintile). A similar direct effect observed. CONCLUSIONS Leisure time physical activity was inversely related to risk of developing heart failure in a dose-response fashion. This was reflected in a similar but less pronounced association of total physical activity with risk of heart failure. Only part of the effects appeared to be mediated by traditional risk factors.