Detecting rapid elevations in blood pressure using a novel pulse arrival time and machine learning technique.

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Wearable Technology

Background:

• Malignant hypertensive crises require rapid detection for effective intervention.
• Current methods for continuous blood pressure (BP) monitoring are limited, especially for wearable applications.
• Wearable devices offer potential for non-invasive, real-time physiological monitoring.

Purpose of the Study:

• To assess the initial feasibility of an acute hypertension detection platform (AHDP) for wearable devices.
• To determine if the AHDP can accurately detect laboratory-simulated elevations in blood pressure (BP).
• To explore the utility of pulse arrival time (PAT) for BP monitoring in a wearable context.

Main Methods:

• 42 healthy young adults underwent isometric handgrip exercise with blood flow restriction to induce BP elevation.
• Continuous electrocardiogram (ECG) and photoplethysmography (PPG) data were collected to calculate pulse arrival time (PAT).
• A logistic regression model using PAT z-scores classified normotensive and elevated BP states, with performance evaluated by PR-AUC.

Main Results:

• Isometric exercise significantly increased BP (Δ18.8±15.1 mmHg) and decreased PAT (Δ-5.6±5.7 ms) (p<0.001).
• The AHDP model achieved a precision-recall area under the curve (PR-AUC) of 0.68 (p<0.001), indicating successful detection beyond random chance.
• An optimal PAT z-score threshold of -3.87 was identified for classifying elevated BP.

Conclusions:

• The developed AHDP demonstrated feasibility in detecting laboratory-simulated acute hypertension.
• PAT, calculated from ECG and PPG, shows promise as a biomarker for elevated BP in wearable systems.
• Further development of the AHDP could lead to wearable devices for early detection of hypertensive crises.