Epidemiologic factors associated with errors in blood pressure measurements obtained with an automated device compared with a mercury sphygmomanometer

Background: Automated Devices (AD) for measuring Blood Pressure (BP) are gradually replacing arcury Sphygmomanometers (MM), even in clinical settings. We compared measurements obtained using these two techniques, and investigated the factors associated with measurement errors.Methods: Two trained observers took three BP measurements in 439 subjects as part of an epidemiologic survey, alternately using a MM and an AD. BP measurement error was defined as BP (MM) minus BP (AD). Alarm Reactions (AR) were calculated by subtracting the third systolic BP (SBP) measurements from the first SBP.Results: The mean age of subjects was 50.92±15.66 years (n = 439). The proportion of female was 55.4%. The mean SBP (MM) was 116.84±15.9 mmHg, and the mean SBP (AD) was 114.85±15.34 mmHg. BP measurement errors were 1.98±6.92 mmHg for SBP (error-SBP), and 3.23±5.61 mmHg for diastolic BP (error-DBP) (P = 0.001). In a multiple regression model, factors independently associated with error-SBP were female gender (β = 0.237, P = 0.0005), arm circumference (β = -0.262, P < 0.0001), SBP (β = 0.337, P < 0.0001) and alarm reaction (β = -0.199, P < 0.0001) (adjusted R2 = 0.267). Factors associated with error-DBP were heart rate (β = -0.050, P = 0.0185), DBP (β = 0.193, P < 0.0001), and pulse pressure (β = -0.435, P < 0.0001) (adjusted R2 = 0.285).Conclusions: The observed errors in BP measurements taken with the MM and AD in a population survey are associated with blood pressure, gender, arm circumference, pulse pressure, and the alarm reaction