Sex Differences in Blood Pressure Associations With Cardiovascular Outcomes

Abstract

For decades, 120 mm Hg has been considered the normal upper limit for adult systolic blood pressure (SBP). Practice guidelines have long referred to this threshold for classifying ranges of blood pressure (BP) elevation and treatment targets, given the consistent epidemiologic finding that cardiovascular disease (CVD) risk is continuously increased from the SBP level of 120 mm Hg and upwards.1 Amid previous studies using outcomes associations to determine a normal SBP range, there remain limited data on potential sex differences. It is well known that BP levels in adulthood are on average lower in women than men in the healthy state2; however, whether or not a lower range of SBP might be considered normal for women versus men is unclear. We studied 27 542 participants (54% women) without baseline CVD who had standardized SBP measurements performed in 1 of 4 community-based cohort studies: the Framingham Heart Study, Multi-Ethnic Study of Atherosclerosis, Atherosclerosis Risk in Communities Study, and Coronary Artery Risk Development in Young Adults Study.3 Age and race distributions were similar between sexes (standard mean difference <0.1). Over 28±12 years, 7424 participants (44% women) developed nonfatal or fatal CVD: 3405 myocardial infarction (MI), 4081 heart failure (HF), and 1901 stroke events. We related SBP category (defined by 10 mm Hg increments from <100 mm Hg to ≥160 mm Hg) with incident CVD using cohort-stratified Cox proportional hazards models accounting for competing risks and adjusting for traditional risk factors (Figure); we observed no important violations of the proportional hazards assumptions. We tested for sex interactions and analyzed the MI, HF, and stroke outcomes separately. We also constructed models stratified by age, race, and cohort. We used R version 3.5.1 to perform analyses, with a 2-sided P<0.05 considered statistically significant. All participants provided written informed consent, institutional review boards approved all protocols, and National Heart, Lung, and Blood Institute approved data access. All data are available through public access policies of the National Heart, Lung, and Blood Institute BioLINCC repository, which does not contain information that could compromise research participant privacy.