Cardiometabolic Risk-Related Blood Pressure Trajectories Differ by Sex
Hongwei Ji,Andy Kim,Susan Cheng,Teemu J. Niiranen,Brian Claggett,C. Noel Bairey Merz
DOI: https://doi.org/10.1161/hypertensionaha.119.14570
2020-01-01
Abstract:Sex differences in the manifestations of clinical cardiovascular disease become more evident with advancing age and are likely related to sexual dimorphism in risk factors such as high blood pressure (BP). Accordingly, we and others have observed marked differences between females and males in trajectories of BP elevation, a surrogate of vascular aging, beginning early in life. In effect, previously reported evidence has shown that females tend to exhibit a more accelerated rise in BP in early-to-mid life such that their BP levels catch up to those of men by later life. Such overt sex differences in patterns of BP elevation are likely related to a variety of underlying mechanisms, including variable associations with cardiometabolic risk factors, which are also known to differ between women and men. Therefore, we sought to examine sex differences in the extent to which cardiometabolic risk traits are related to BP elevation over the life course. We used serially examined BP measurements collected longitudinally from participants of 4 community cohorts with study designs described previously: FHS (Framingham Heart Study) offspring cohort (Exams 1–9), ARIC Study (Atherosclerosis Risk in Communities; Exams 1–4), CARDIA Study (Coronary Artery Risk Development in Young Adults; Exams 1–8), and MESA (Multi-Ethnic Study of Atherosclerosis; Exams 1–4). Each participant provided informed consent, and institutional review boards approved the protocol at each study site. We excluded observations if concurrent data were missing for BP measures, antihypertensive medication, or key covariates: body mass index, smoking status, diabetes mellitus, and total cholesterol. Thus, our final study sample included N=32 833 unique participants who contributed 136 869 observations over a 43-year period (1971–2014) with ages spanning 5 to 98 years. As part of harmonizing data across cohort studies, we corrected systolic BP measures to readings from mercury column sphygmomanometer after adjusting for previously described between-method differences. To account for the effects of antihypertensive therapy, we imputed BP values to be 10 mm Hg higher in estimated systolic BP and 5 mm Hg higher in estimated diastolic BP for individuals on antihypertensive therapy. In secondary analyses, we applied alternate approaches to imputing effects of medication therapy, and we also conducted analyses without imputation but with adjustment for medication use in multivariable modeling. Higher versus lower risk burden was defined as presence of ≥2 versus <2 of the following cardiometabolic risk factors assessed at each visit: smoking status (current smoker), diabetes mellitus, obesity (body mass index ≥30 kg/m), or hypercholesterolemia (total cholesterol ≥200 mg/dL). We categorized participant observations (N=136 869) into 4 categories: (1) women with higher risk burden, (2) women with lower risk burden, (3) men with higher risk burden, (4) men with lower risk burden. For each category, we fit multilevel linear regression models to display longitudinal BP trajectories, with age used as the timescale for all analyses. Each model included age as a fixed effect, participant IDs as random intercepts, and BP measure as the outcome. Given that the associations of the predictor variable (age) with the outcome variable (systolic BP) were nonlinear, we used restricted cubic splines with 4 knots to allow for nonlinearity of relationships. Under the premise of sex-specific physiology, we then calculated BP change from the baseline BP level (ie, mean BP level at 18 years of age) for each category. Differences between higher and lower risk burden in the relationships between BP measures and age were tested via likelihood ratio test between models with and without parameters representing the interaction between risk burden and the cubic spline variable representing age. Differences in systolic BP trajectories between higher and lower risk burden over time were shown as the area between the BP curves with higher and lower risk burden. Graphically quantitative difference was then plotted as area chart above x axis. In all fitted splines, data were truncated over an age range of 18 (0.5th percentile) to 85 (99.5th percentile) years. We then categorized participant observations according to age tertiles (ie, 18–49; 50–59; 60–85). To quantify the difference of BP trajectories between higher and lower risk burden, in each age tertile, we compared rates of BP increase with age (mm Hg per decade) in the presence of higher versus lower risk burden using multilevel linear regression models adjusted for diabetes mellitus, smoking status, obesity, hypercholesterolemia, race, and cohort. We then compared differences by sex in the risk-related differences in BP rate of increase From the Cardiovascular Division, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA (H.J., A.K., B.L.C., S.C.); Department of Cardiology, Shanghai Tenth People’s Hospital, Tongji University, Shanghai, China (H.J.); Framingham Heart Study, MA (H.J., A.K., S.C.); Barbra Streisand Women’s Heart Center (A.K., C.N.B.M., S.C.) and Hypertension Center of Excellence (J.E.B.), Smidt Heart Institute, Cedars-Sinai Medical Center, Los Angeles, CA; National Institute for Health and Welfare, Helsinki, Finland (T.J.N.); and Department of Medicine, University of Turku and Turku University Hospital, Finland (T.J.N.). Correspondence to Susan Cheng, Smidt Heart Institute, Cedars-Sinai Medical Center, Los Angeles, CA. Email susan.cheng@cshs.org Cardiometabolic Risk-Related Blood Pressure Trajectories Differ by Sex