Comparison of oscillometric blood pressure measurements at the wrist with an upper-arm auscultatory mercury sphygmomanometer

1. Oscillometric devices for blood pressure (BP) measurement at the wrist a re becoming more widely used in clinical practice. However, systematic comp arisons with standard auscultatory BP measurement at the brachial artery ar e scarce. Therefore, we compared two such devices, the Boso-Mediwatch (Bosc h & Sohn GmbH U, Co., Jungingen, Germany) and the Omron R3 (Omron Corp., To kyo, Japan), with upper-arm auscultatory mercury sphygmomanometry, 2, In 20 normotensive subjects and 20 treated hypertensive,; subjects, the Boso-Mediwatch was applied to the left wrist by observer I and was compared with mercury sphygmomanometry of the right upper arm by observer 2, Each o bserver swapped sides and the procedure was repeated. The Boso-Mediwatch wa s then applied to the right wrist by observer 1 and was compared with mercu ry sphygmomanometry of the left upper arm by observer 2, Each observer once again swapped sides and the procedure was repeated. An identical protocol was followed for the Omron R3 in a further 20 treated hypertensive subjects and 20 normotensive subjects, 3. There Here no significant differences between observers or left,versus r ight arm for either oscillometric device or when measurements were performe d by mercury sphygmomanometry, In normotensive subjects, the Boso-Mediwatch readings were higher than mercury sphygmomanometer readings, with mean dif ferences (95% confidence intervals) of 3.9 (0.1, 7.6; P = 0.045) and 7.0 mm Hg (4.7, 9.2; P < 0.001) for systolic and diastolic BP, respectively, In hy pertensive subjects, the Boso-Mediwatch readings were lower for systolic BP (mean difference - 6.0 mmHg (- 11.6, - 0.3; P = 0.04) but were higher for diastolic BP (mean difference 3.8 mmHg (1.4, 6.3; P < 0.01), 4. In normotensive subjects, the Omron R3 readings were higher, with mean d ifferences of 3.2 (0.6, 5.8; P = 0.018) and 4.2 mmHg (1.6, 6.7; P = 0.003) for systolic and diastolic BP, respectively. In hypertensive subjects, the Omron R3 readings were lower for both systolic and diastolic BP (mean diffe rences: - 5.8 (- 8.8, - 2.8; P = 0.001) and - 5.5 (-9.3, - 1.6; P = 0.008), respectively). 5. For the Boso-Mediwatch, weighted least products analysis confirmed the p resence of both fixed and proportional error for systolic BP but not for di astolic BP. For the Omron R3, fixed or proportional error was not detected for either systolic or diastolic BP. 6. These wrist oscillometric devices, although offering portability and con venience, give BP measurements that frequently differ substantially (by at least 5 mmHg) from readings simultaneously measured at the upper arm by a m ercury sphygmomanometer, The magnitude and direction of differences detecte d are dependent on both the device used and the underlying level of BP.