Background: While previous studies have reported lower risk of coronary heart disease with physical activity, limited data exist on the relation of physical activity with heart failure risk in a community setting.
Objective: To test the primary hypothesis that physical activity is inversely associated with HF risk in a cohort of apparently healthy women and a secondary hypothesis that walking time and usual walking pace are associated with lower risk of HF.
Methods: Prospective cohort study of 39,635 participants from the Women’s Health Study. Information on leisure-time physical activity, including walking time and usual pace, was self-reported at baseline and HF was ascertained through annual follow-up questionnaires with validation through review of medical records. We used Cox regression to estimate multivariable adjusted hazard ratios of HF.
Results: During a mean follow-up of 17.2 years, 635 cases of HF occurred. The mean age at baseline was 54.6 ± 7.0 years. Median physical activity was 8.4 MET-h/week (IQR: 2.7-20.2); about 18% of women did not walk regularly while 32% reported walking pace of 3+ miles per hour. In a multivariable adjusted model, hazard ratios (95% CI) for HF were 1.0 (ref), 0.73 (0.57-0.95), 0.86 (0.69-1.07), 0.70 (0.54-0.92), and 0.93 (0.66-1.05) for physical activity of <3.75, 3.75-7.4, 7.5-14.9, 15-22.4, and 22.5+ MET-h/week, respectively (p trend 0.043). However, the association was attenuated upon additional adjustment for usual walking pace (corresponding HR: 1.0 (ref), 0.77 (0.60-1.01), 0.93 (0.73-1.18), 0.77 (0.58-1.02), and 0.94 (0.73-1.22), p trend 0.47). Walking time showed a J-shaped relation with risk, that also was attenuated after adjustment for pace and MET-h/wk (HR: 1.0 (ref), 1.19 (0.94-1.50), 1.06 (0.80-1.39), 0.69 (0.50-0.97), 0.84 (0.63-1.12), and 0.87 (0.63-1.20) for walking time of 0-19 min/wk, 20-59 min/wk, 1h/wk, 1.5h/wk, 2-3h/wk, and 4+h/wk, respectively). In contrast, usual walking pace was inversely associated with HF risk even after adjustment for total physical activity and walking time with multivariable adjusted HR of 1.0 (ref), 0.97 (0.75-1.24), 0.88 (0.70-1.10), and 0.74 (0.55-0.98) for walking pace of 0, <2, 2-2.9, and 3+ miles per hour, respectively, p trend 0.038).
Conclusions: Our data show inverse associations of total leisure-time physical activity with HF risk that was attenuated after adjustment for usual walking pace. In contrast, usual walking pace was inversely and robustly associated with HF risk while walking time showed a J-shaped relation.