Wearable Systems for Continuous Blood Pressure Monitoring

Abstract

Cardiovascular diseases (CVDs) are the leading cause of death worldwide, posing significant risks to global health. Regular blood pressure (BP) monitoring plays a crucial role in CVDs prevention by providing comprehensive and accurate assessments of individual BP patterns. While clinically utilized cuff-based sphygmomanometers offer snapshot readings of systolic and diastolic BP, they struggle to capture dynamic BP variations. Invasive arterial catheters provide continuous, accurate BP readings but increase patients suffering and infection risk. In comparison, the rapid development of cutting-edge wearable electronics provides a noninvasive solution for continuous BP monitoring by integrating techniques from materials sciences, micro/nano fabrication, electrical engineering, and computer science. However, current continuous BP monitoring devices remain in early stages, primarily focusing on thin, soft, skin-interfaced pulse wave sensors. Poor BP estimation accuracy and low system integration levels persist as key challenges in translating these systems to daily life applications. This thesis aims to develop a comprehensive continuous BP monitoring solution compromising from flexible sensors, wireless data sampling/transmission to BP estimation model by addressing some key challenges in sensor stability, device-skin interface robustness, BP estimation accuracy and model long-term stability.

To address the challenges associated with sensor stability, we first introduce a transparent, crosstalk free triboelectric nanogenerator (CF-TENG) sensor array for precise pressure mapping and position tracking. Given that the crosstalk effect between adjacent sensing units in TENGs significantly limits the pixel density of sensor arrays and sensing performance. The CF-TENG employed a 3D printed substrate to create electrical isolation structure between sensing units to suppress the mutual interference caused by electrostatic induction. 3D printing of soft, transparent photocurable polymer substrate, spray-coated silver nanowire electrodes, and widely utilized dielectric silicone elastomer (polydimethylsiloxane, PDMS) endow the CF-TENG with great optical transparency and remarkable mechanical and electrical robustness and can be processed in a simple, low-cost, and scalable way. The CF-TENG sensor array with 100 sensing units in an overall size of 7.5 cm × 7.5 cm. All the sensing units show good sensitivity of 0.11 V/kPa with a wide range of pressure detection from 10 kPa to 65 kPa, which allows to accurately distinguish various tactile formats from gentle touching (as low as 2 kPa) to hard pressuring. The 3D printed electronic isolation structure allows to cast triboelectric layers of polydimethylsiloxane in an independent sensing manner for each unit, which greatly suppresses the cross talk arising from adjacent sensing units with the maximum crosstalk output of only 10.8%. The excellent uniformity and reproducibility of the sensor array offer precise pressure mapping for complicated pattern loadings, which demonstrates its potential in biomedical sensing and human-machine interfaces.

While the thin, soft, skin-interfaced sensors offer a promising solution for continuous BP monitoring, integrating with external signal processing and transmission modules presents significant challenges. The inherent bulkiness of rigid chips increases system complexity, thus making it difficult to miniaturize all components to micrometer scale in thickness. To address these challenges and improve interface robustness, we developed two innovative solutions for enhancing connections and electrical/mechanical stabilities between integrated systems and user skin. We first introduce a passive interfacial adhesion strategy that incorporates biocompatible hydrogel to form robust covalent bonds with both the encapsulation materials of integrated systems and biological tissues, such as skin. The adhesive hydrogel comprises a poly (acrylic acid) (PAA) network crosslinked with biodegradable gelatin methacrylate, along with a biodegradable gelatin network, showing strong adhesion to diverse materials and superior biocompatibility, without skin irritation caused on 2 volunteers after 12 h continuous covering. Moreover, the bio-adhesive demonstrates remarkable interface robustness, with over 35 kPa shear strength and over 1000 interface toughness achieved, respectively. Additionally, to address the need for interface pressure regulation between sensing systems and user skin, we developed a novel interface pressure adapter. This adapter significantly enhances compliance and stability between the users skin and the wearable system. It incorporates three key components: a micro-airbag, a pressure-regulating one-way valve, and a micro-pump. These elements work synergistically to provide sufficient pressure support to the sensing module, thereby improving the sensing stability and robustness of wearable systems. As a proof of concept, we demonstrated that the adapter could generate and regulate a maximum pressure of 16.2 kPa through five pumping phases, providing essential back pressure to a pressure sensor for pulse wave measurement. The proposed adapter demonstrated remarkable improvements in interfacial stability and measurement robustness against various joint deformation and motion artifacts, which presents a promising solution for enhancing the performance and interface stability of contact-based sensing systems.

Integrating these innovations, we present a thin, soft, miniaturized system (TSMS) for real-time, continuous wireless monitoring of ambulatory arterial BP. The TSMS combines a conformal pressure sensor array, an active pressure adaptation unit, a wireless signal processing module, and an advanced machine learning algorithm. The highly chemical/physical stabled piezoelectric sensor array, together with the active pressure adaptation unit, enables accurate and continuous monitoring of pulse signals. This system also facilitates in-situ transformation of piezoelectric responses through integration with a signal processing module that incorporates a built-in theoretical mathematical model. Advanced sampling and data transmission strategies facilitate high precise sampling and wireless transmission of pulse wave form and local pulse wave velocity (PWV) to the machine learning data model. The system achieves mean errors of 0.11 ± 3.68 mmHg for diastolic BP (DBP) and -0.05 ± 4.61 mmHg for systolic BP (SBP), meeting the criteria for Grade A classification according to standards for continuous BP monitoring devices. Initial trials conducted on diverse cohorts of 87 volunteers, including two individuals with hypertension, and 60 clinical patients demonstrate the efficacy of the TSMS. The results show comparable accuracy to both commercial noninvasive continuous monitoring systems and clinical invasive catheters, in terms of both mean values and dynamic range.

In conclusion, this thesis presents significant advancements in three key areas: skin-interfaced pressure sensors, device-skin interface enhancement strategies, and system-level integration for blood pressure monitoring. The development of the TSMS demonstrates superior overall BP estimation performance. We believe that the TSMS and the underlying technologies developed in this thesis will contribute significantly to improving patient outcomes. By providing accurate, continuous, and convenient blood pressure monitoring, our work presents a promising avenue for revolutionizing cardiovascular disease management. This comprehensive innovation has the potential to significantly enhance prevention strategies, facilitate earlier diagnosis, and improve overall disease management, ultimately contributing to a reduction in morbidity and mortality rates associated with cardiovascular conditions.

Date of Award	30 Dec 2024
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Xinge YU (Supervisor)