[OP.LB03.04] FIRST IN MAN MEASUREMENT OF ARTERIAL STIFFNESS USING A CONNECTED BATHROOM SCALE: CALIBRATION AGAINST SPHYGMOCOR. (September 2016)

• Record Type: Journal Article
• Authors: Khettab, H.; Campo, D.; Yu, R.; Buard, N.; Boutouyrie, P.
• Is Part Of: Journal of hypertension. Volume 34:(2016) Supplement 2
• Publication Date: 2016-09
• Subjects: Hypertension -- Periodicals; Hypertension -- Periodicals; 616.132005
• DOI: 10.1097/01.hjh.0000491626.52108.3c
• Languages: English
• ISSNs: 1473-5598

Abstract

Abstract : Objective: Measurement of arterial stiffness (AS) is still hampered by the technical skill necessary to obtain reliable measurements. It is generally considered that AS is not prone to self-measurement. We developed a non-invasive technique to assess AS from a connected bathroom scale, based on ballistocardiography (BCG) and impedance plethysmography (IPG). Figure. No caption available. Design and method: We included 76 healthy subjects and 35 essential hypertensive, 50% women. The scale includes high sensitivity transducers measuring oscillations of weight due to cardiac contraction (BCG). BCG signal comprises G, H, I, J, K, L, M waves. The HIJ waves are hallmarks of systole and were subsequently tracked. The scale comprises electrodes to measure the blood flow in a foot (IPG), used as a hallmark of flow wave arrival in the foot. The scale pulse transit time (WS-PTT) was calculated by linear transformation of the time intervals between the BCG systolic waves and the foot of the IPG. Several estimates where combined to improve the robustness of the measurement in a home setting. Carotid to femoral transit time (CF-PTT) was measured using Sphygmocor (Sydney, Australia). Data are presented as mean ± SD (min-max). Spearman and robust multivariate regressions were used. Results: The age was 45 ± 17 years (20–75), SBP 125 ± 18 mmHg (92–174), CF-PTT 66 ± 15 (28.3–100.9). The WS-PTT correlated well with CF-PTT with R 2 = 0.62. The root mean square error was 8 ms, and the slope of the correlation significantly differed from 1 (0.73, 95% CI [0.62 to 0.84]), indicating a significant bias. Correlations of WS-PTT with age and blood pressure were similar or stronger (R 2 = 0.73 and 0.37, resp.) than CF-PWV (R 2 = 0.48 and 0.37, resp.). This good correlation is non-trivial given the differences in wave paths, the fact that measurements are made in orthostatic position and totally investigator-free. Conclusions: A simple user-oriented instrument such as a connected bathroom scale can estimate arterial stiffness with accuracy close to healthcare-oriented systems. The availability of arterial stiffness data on very large, non-medicalized populations will change our management of well-being and health.