Validation of a novel cuffless photoplethysmography-based wristband for measuring blood pressure according to the regulatory standard

Abstract

Abstract Background Blood pressure is a key paramater in acute and chronic cardiovascular diseases (CVD) [1]. However, obtaining reliable and reproducible cuff blood pressure (BP) measurements is challenging. We demonstrate that a wristband-based PPG measurement in combination with a custom developed BP algorithm represents a possible alternative to cuff BP measurements. Purpose This study aimed to validate a novel cuffless photopletysmography (PPG)-based non-invasive wristband for continuous BP monitoring in accordance with ISO 81060-2:2019. Methods The study compared PPG-guided BP algorithm predictions with subclavian arterial reference measurements taken during cardiac catheterization. Eligible patients were consecutively included, and eligibility was screened following ISO 81060-2:2019 requirements. Reference measurements were performed using a validated invasive BP monitoring device with a sampling rate of 100Hz. PPG signals were collected using six light emission diodes and two photodiodes at a sampling rate of 128Hz. Three sequential initialization measurements were taken using a validated blood pressure cuff before the cardiac catheterization exam. These measurements, along with approximately 100 additional features (PPG-derived and based on patient demographics), were used as input for the machine learning-based BP algorithm. Correlation, mean error, and standard deviation (SD) were determined for systolic and diastolic BP measurements between the BP algorithm predictions and invasive reference measurements Results The study included 97 patients from whom 420 individual 30-second samples were obtained. The mean age, weight, and height of the analysed subjects were 67.1 (SD 11.1), 83.4 (SD 16.1), and 174.1 (SD 10.0), respectively. In 48 samples (11%), systolic BP was ≤100mmHg, while in 106 samples (25%), systolic BP was ≥160mmHg. Diastolic BP was ≤70 mmHg in 222 samples (53%) and ≥85 mmHg in 99 samples (24%). The BP algorithm predictions showed a high correlation with invasive reference measurements for systolic (r = 0.985) and diastolic (r = 0.961) BP measurements. The mean error of the BP algorithm predictions compared to the invasive reference measurements was ±3.7 mmHg (SD 4.4 mmHg) and ±2.5 mmHg (SD 3.7 mmHg) for systolic and diastolic BP, respectively. Results were similar within each gender and skin colour category (Fitzpatrick I-VI). Conclusion This study demonstrates that a wristband-based PPG measurement in combination with the developed BP algorithm can provide accurate continuous BP measurements across a wide range of BP distributions. Therefore, wristband BP monitoring may serve as a valid and less burdensome alternative to cuff BP measurements for both in-hospital and at-home BP monitoring. However, further research is necessary to evaluate the precision of the BP algorithm during movement and the stability of the predictions over time.figure of BP