WALKING AND INCIDENCE OF HYPERTENSION IN POSTMENOPAUSAL WOMEN: THE WOMEN’S HEALTH INITIATIVE STUDY
MetadataShow full item record
BACKGROUND: Though considerable evidence exists regarding the hypothesized anti-hypertensive influence of physical activity in adults, limited research has been conducted within postmenopausal women, particularly regarding walking-specific activity. Women aged 60 and older is a rapidly growing demographic among U.S. adults that is highly susceptible to both hypertension and low levels of physical activity. Therefore, if physical activity, walking specifically, were shown to be associated with lower risk of primary incident hypertension, such a finding could have important public health implications.OBJECTIVES: (1) To examine whether self-report measures of walking-specific physical activity (walking amount [MET-hr/wk] and walking speed [mph]) at baseline is associated with the incidence of hypertension in a defined cohort of postmenopausal women, aged 50-79 years, and to assess the independence of such associations beyond relevant confounding factors and non-walking physical activity levels. (2) To evaluate whether time-varying measures of walking-specific physical activity (walking amount [MET-hr/wk] and walking speed [mph]) are associated with the incidence of hypertension, and to contrast findings with those based solely on baseline walking exposure measures. (3) To explore potential effect modification of the association between walking and hypertension risk by age (50-59, 60-69, ≥70 years), race-ethnicity (White, Black, Hispanic, Asian/Pacific Islander), baseline measured blood pressure (≤120/≤80, >120/>80 mmHg), and baseline BMI (<25, 25 to <30, ≥30 kg/m2). METHODS: This prospective study was conducted in a cohort of postmenopausal women who are enrolled in the Women’s Health Initiative (WHI) Study. Included in this analysis were members of the WHI cohort who, at baseline: (a) completed the physical activity questionnaire, including walking questions; (b) were without a history of physician diagnosed-treated hypertension and whose measured systolic blood pressure was <140 mmHg and diastolic blood pressure <90 mmHg; (c) reported being able to walk at least 1 block unassisted; (d) reported no history of heart failure, coronary heart disease, and stroke; and (e) had at least 1 year of post-baseline follow-up for incident hypertension. Walking exposures were assessed by responses to self-administered questionnaires at baseline, and in much of the cohort again at least once during follow up. Walking-specific physical activity amount (0, >0-3.5, 3.6-7.5, >7.5 metabolic equivalent [MET] hours per week; MET-hr/wk) and speed (None/rarely, <2, 2-3, 3-4, >4 miles per hour [mph]) served as the primary exposure variables. History of diagnosed-treated hypertension at baseline was assessed by questionnaire; resting blood pressure by standard auscultatory methods. The study outcome variable was post-baseline physician diagnosed-treated incident hypertension, assessed by self-report on annual mailed health update questionnaires during the follow-up interval. The primary hypotheses were examined using time-to-event statistical analyses, with person-time follow-up defined as the time interval between date of baseline examination to date of either death, loss to follow-up, health update questionnaire documenting hypertension diagnosis, or August 29th, 2014. Follow-up was censored at this date to avoid overlap with a large WHI clinical trial involving walking intervention (WHISH Trial) that begun in late 2014, and changes in national blood pressure guideline in 2017. Primary hypothesis tests were conducted using Cox proportional hazard regression models to estimate hazard ratios, 95% confidence intervals, and associated p-values, in univariate and multivariate models, with walking amount (and separately, walking speed) as a baseline exposure. Several individual-level characteristics assessed at baseline were considered as potential confounding factors, using as a guide to their selection knowledge of previous studies, the biology of the exposure-disease of interest, and standard epidemiologic measures for their empirical evaluation (change-in-estimate approach). Secondary analyses pertaining to time-varying exposure measures also utilized Cox proportional hazard regression models to estimate hazard ratios, 95% confidence intervals, and associated p-values. Effect modification by pre-specified factors was evaluated using cross-product interaction terms in the multivariable Cox regression models, and stratified analysis to understand the nature of an interaction identified during statistical analyses.RESULTS: After exclusions, there were 83,435 participants comprising the analytic cohort, whose mean age at baseline was 61.9 years. The majority of participants were white (86.6%), educated beyond high school (81.4%), and from middle- and upper socioeconomic strata (66.7%). The cohort was of relatively good health, with the prevalence of obesity, physical inactivity, current smoking, diagnosed-treated diabetes and diagnosed-treated hyperlipidemia lower at WHI baseline compared with estimates for women of similar ages at that time. A total of 72,580 (87.0%) women reported some participation in recreational physical activity at baseline enrollment, of which walking activity accounted for a substantial proportion (37.9%) of the activity reported. There were 24,253 (29.1%) participants who reported rarely or never walking outside of their home for more than 10 minutes without stopping; while 20,269 (24.3%) reported expending > 7.5 MET-hr/wk (comparable to guideline recommended activity levels) through walking alone. Among walkers, usual walking speeds of 2-3 mph were most frequently reported (35.7%). There were 38,230 (45.8%) incident cases of diagnosed-treated hypertension documented over a mean follow-up of 10.8 years (42.4 cases per 1000 person-years). In an unadjusted model, the HR (95% CI) for incident hypertension were 0.96 (0.93-0.99), 0.91 (0.88-0.93), and 0.82 (0.80-0.84) for participants who reported walking 0-3.5, 3.6-7.5 and > 7.5 MET-hr/wk compared to non-walkers (trend p < 0.001). After adjusting for potential confounders (including non-walking physical activity), the association was attenuated yet remained statistically significant. Participants who reported a usual walking speed > 4 mph had a 35% lower risk of developing hypertension when compared with non-walkers (HR= 0.65; 95% CI: 0.59-0.71, trend p < 0.001) in an unadjusted model. The magnitude of association was lower but remained statistically significant after adjustment for confounding factors. Secondary analyses, which modeled time-varying walking exposures, produced similar patterns of association and point estimates to primary baseline models. Evidence of effect modification was observed between walking and age and race-ethnicity, with stronger associations observed among younger participants for both walking amount and speed exposures (interaction p values < 0.001), while stratification by race-ethnicity suggested that the association between hypertension and walking amount was more pronounced in white participants than other race-ethnicities, though statistical tests of interaction yielded non-significant findings. Similar associations between walking exposures and hypertension risk were observed across categories of baseline blood pressure, physical functioning status, and multiple measures of adiposity for both walking amount and speed exposures.CONCLUSIONS: These data from a large prospective cohort study suggest that an association exists between self-reported measures of walking-specific physical activity and hypertension incidence in older women, independent of non-walking physical activity. The association between walking speed and hypertension incidence was particularly strong. Walking speed’s correlation with cardiorespiratory fitness and physical functioning, both of which are associated with better blood pressure regulation and lower hypertension risk, may partly account for the stronger magnitude of association than seen for walking amount. An important observation in the present study was that no significant differences were observed in the associations between walking and hypertension incidence across categories of baseline blood pressure or measures of adiposity, suggesting that anti-hypertensive benefits attributable to walking-specific physical activity are potentially accessible even for persons with elevated hypertension risk due to elevated blood pressure or overweight/obesity. If these observational study results are confirmed in other diverse older cohorts, there may be reasonable evidence with which to contemplate design of a randomized clinical trial evaluating walking in the primary prevention of hypertension in later life.