Thresholds for Diagnosing Hypertension Based on Automated Office Blood Pressure Measurements and Cardiovascular Risk

Abstract

The risk of cardiovascular events in relation to blood pressure is largely based on readings taken with a mercury sphygmomanometer in populations which differ from those of today in terms of hypertension severity and drug therapy. Given replacement of the mercury sphygmomanometer with electronic devices, we sought to determine the blood pressure threshold for a significant increase in cardiovascular risk using a fully automated device, which takes multiple readings with the subject resting quietly alone. Participants were 3627 community-dwelling residents aged >65 years untreated for hypertension. Automated office blood pressure readings were obtained in a community pharmacy with subjects seated and undisturbed. This method for recording blood pressure produces similar readings in different settings, including a pharmacy and family doctor's office providing the above procedures are followed. Subjects were followed for a mean (SD) of 4.9 (1.0) years for fatal and nonfatal cardiovascular events. Adjusted hazard ratios (95% confidence intervals) were computed for 10 mm Hg increments in blood pressure (mm Hg) using Cox proportional hazards regression and the blood pressure category with the lowest event rate as the reference category. A total of 271 subjects experienced a cardiovascular event. There was a significant (P=0.02) increase in the hazard ratio of 1.66 (1.09, 2.54) at a systolic blood pressure of 135 to 144 and 1.72 (1.21, 2.45; P=0.003) at a diastolic blood pressure of 80 to 89. A significant (P=0.03) increase in hazard ratio of 1.73 (1.04, 2.86) occurred with a pulse pressure of 80 to 89. These findings are consistent with a threshold of 135/85 for diagnosing hypertension in older subjects using automated office blood pressure.