Continuous, accurate blood pressure (BP) monitoring is crucial in various medical settings, yet current wearable technologies lack the clinical accuracy of bulky, invasive methods. Existing wearable approaches often rely on estimations from pulse transit time or oscillometric readings, suffering from limitations in accuracy and requiring recalibration. This paper addresses this gap by presenting CNAP2GO, a novel method for continuous noninvasive BP measurement aiming for high accuracy and miniaturization for wearable applications. The potential benefits of such a technology are significant, including revolutionizing patient monitoring in both ambulatory and in-hospital settings, improving patient care, and potentially mitigating intensive care unit overload during crises such as the COVID-19 pandemic. The success of CNAP2GO hinges on its ability to overcome the challenges of existing methods by providing clinically acceptable accuracy, precision, reliability, and ease of use in a miniaturized form. Ideally, the system should operate unobtrusively without user interaction, a key requirement for widespread adoption in diverse clinical settings.

The paper reviews existing methods for continuous noninvasive BP monitoring, highlighting their limitations. Pulse transit time (PTT) methods, while simple, suffer from inaccuracies due to confounding factors. Other time-based methods using PPG or piezoelectric pulse detection also struggle to achieve clinically acceptable accuracy and require frequent recalibration. Oscillometric approaches, like the Omron HeartGuide, provide intermittent measurements and require user interaction. Existing volume clamp or P0 devices based on vascular unloading techniques (VUT), like Finapres and CNAP devices, offer continuous measurement but are too bulky for wearable applications. The authors challenge the common assumption that constant blood volume in the finger is essential for VUT-based BP measurement.

CNAP2GO utilizes a novel volume control technique (VCT) implemented in software on existing hardware (CNAP Monitor HD). The system uses a PPG probe with an actuator that varies the contact pressure of the light elements (LED and photodiode) on the finger. The VCT algorithm uses the PPG signal and its components (filtered to represent different physiological aspects such as rhythm and pulse) to control the actuator pressure. An initial open-loop phase determines the initial mean blood pressure (mBP). The closed-loop phase involves a controller that adjusts the actuator pressure to maintain a balanced blood volume in the finger, compensating for vasomotor changes and mBP fluctuations. A proportional-integral (PI) control approach enables this adaptive pressure adjustment. The system uses a Gaussian-style envelope curve to derive mBP from the PPG signal and then uses this along with the pulsatile components to derive the complete pulsatile BP waveform. The accuracy of the CNAP2GO software prototype was validated against invasive BP measurements in two studies: a lab study with healthy subjects and a clinical study with 46 patients undergoing neurosurgery under general anesthesia. The clinical validation uses the CNAP Monitor HD's sphygmomanometer for calibration and compares CNAP2GO's mean BP to that obtained from a radial arterial catheter. Miniaturization was explored using a fluid-filled bladder actuator in a finger-ring prototype, tested on 7 subjects. Power consumption estimates for a wearable version were also made, considering different operating modes and components.

The CNAP2GO software prototype demonstrated excellent accuracy in comparison to invasive BP measurements in both lab and clinical settings. In the clinical study, the mean difference in mBP between CNAP2GO and the invasive reference method was -1.0 ± 7.0 mmHg, well within the acceptable limits of ±8 mmHg (ISO 81060-2). The system proved robust to motion artifacts. The miniaturized prototype using a fluid-filled bladder actuator in a finger-ring produced oscillometric signals comparable to the standard CNAP system. Power consumption estimates suggested that a 24-hour operation on a single battery should be feasible. The study also demonstrated that CNAP2GO can accurately track physiological BP changes and that the adaptive anti-oscillatory measures successfully prevent system resonance, ensuring accurate and stable measurement.

CNAP2GO successfully addresses the limitations of existing continuous noninvasive BP measurement techniques by combining direct BP measurement with a novel VCT and self-calibration. The high accuracy demonstrated in both lab and clinical studies, combined with the feasibility of miniaturization, strongly suggests that CNAP2GO represents a significant advancement in this field. The robustness to motion artifacts and the capability to function in diverse physiological conditions further strengthen its potential for real-world applications. While the study focused on neurosurgical patients under general anesthesia, the results imply the potential for broader applications across various patient populations and healthcare settings. The relatively low power consumption also supports the development of a truly portable and user-friendly device.

CNAP2GO offers a promising solution for accurate and continuous noninvasive BP monitoring. Its high accuracy, self-calibration capabilities, and potential for miniaturization make it a strong candidate for revolutionizing patient monitoring. Future work should focus on completing the miniaturization and conducting further clinical validation across various patient populations and clinical settings. Additional research should investigate the integration of additional sensors and the exploration of advanced analytical capabilities using the obtained continuous BP waveforms.

The study's clinical validation was conducted in a specific population (neurosurgical patients under general anesthesia). Further validation in broader populations is needed to assess the generalizability of the findings. The miniaturized prototype was tested on a limited number of subjects and used a commercially available infusion pump as a fluid-compressing driver. A dedicated miniature actuator needs to be developed and tested. While motion artifacts were observed, the impact on long-term continuous monitoring requires further investigation in conditions involving significant movement.