24-HOUR AMBULATORY BLOOD PRESSURE MEASUREMENT USING A PULSE ARRIVAL TIME BASED BLOOD PRESSURE MODEL BY A PROTOTYPE WEARABLE SENSOR

Abstract

Objective: Methods to enable cuffless blood pressure (BP) have gained widespread interest. If proven to be comparable to standard cuff measurements and more accepted by patients, these methods can improve hypertension care. We investigated accuracy and patient acceptability of a pulse arrival time (PAT) based model using a prototype cuffless device on the chest compared to standard 24-hour ambulatory blood pressure monitoring (24hABPM). Design and method: PAT was calculated using the electrocardiogram (ECG) and photoplethysmography (PPG) sensors incorporated in the chest belt. The model to predict BP from PAT was developed from a general population cohort during isometric exercise induced BP changes. Ninety-four normotensive and hypertensive adults underwent simultaneous 24hABPM with the prototype chest belt sensor and a validated cuff-based oscillometric device on the non-dominant arm. PAT-based BP predictions were compared to reference BPs using 2-minute time windows corresponding to reference measurements. Statistical analyses were performed using Stata® v. 17.0 (Statacorp., Texas, USA) with continuous data presented as means (SD) and compared using paired t-test. Correlation analysis and agreement were evaluated by Bland-Altman plots. Results: Mean age (SD) was 48.1 (14.3) years and 48 % had a prior hypertension diagnosis. Comparisons between reference and PAT-based mean 24-hour, daytime and nighttime BP showed that the PAT-based model generally overestimated BP (Table). The largest difference was seen during nighttime; the mean difference (SD) was 19.6 (22.2) mmHg and 17.0 (12.2) mmHg for systolic BP (SBP) and diastolic BP (DBP). Ninety-five % limits of agreements were [-29.8, 36.0 mmHg] and [-23.9, 63.0 mmHg] for daytime and nighttime SBP. The prototype sensor was reported to be more comfortable and less disturbing during daily activities and sleep compared to the reference device during 24hABPM (Figure). Conclusions: The present study demonstrated that a general PAT-based BP model significantly overestimated 24hABPM in normotensive and hypertensive subjects. The largest discrepancy was observed during nighttime. Although the PAT-based model did not provide accurate measurements of BP, the potential of cuffless BP technology is promising due to patient acceptability.