Charité-Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Institute for Translational Physiology, Berlin, Germany.
Blood Press. 2022 Dec;31(1):254-269. doi: 10.1080/08037051.2022.2128716.
Hypertension diagnosis is one of the most common and important procedures in everyday clinical practice. Its applicability depends on correct and comparable measurements. Cuff-based measurement paradigms have dominated ambulatory blood pressure (BP) measurements for multiple decades. Cuffless and non-invasive methods may offer various advantages, such as a continuous and undisturbing measurement character. This review presents a conceptual overview of recent advances in the field of cuffless measurement paradigms and possible future developments which would enable cuffless beat-to-beat BP estimation paradigms to become clinically viable. It was refrained from a direct comparison between most studies and focussed on a conceptual merger of the ideas and conclusions presented in landmark scientific literature. There are two main approaches to cuffless beat-to-beat BP estimation represented in the scientific literature: First, models based on the physiological understanding of the cardiovascular system, mostly reliant on the pulse wave velocity combined with additional parameters. Second, models based on Deep Learning techniques, which have already shown great performance in various other medical fields. This review wants to present the advantages and limitations of each approach. Following this, the conceptional idea of unifying the benefits of physiological understanding and Deep Learning techniques for beat-to-beat BP estimation is presented. This could lead to a generalised and uniform solution for cuffless beat-to-beat BP estimations. This would not only make them an attractive clinical complement or even alternative to conventional cuff-based measurement paradigms but would substantially change how we think about BP as a fundamental marker of cardiovascular medicine.
高血压诊断是日常临床实践中最常见和最重要的程序之一。其适用性取决于正确和可比的测量。基于袖带的测量范式已经主导了多年的动态血压（BP）测量。无袖带和非侵入性方法可能具有各种优势，例如连续和无干扰的测量特性。这篇综述介绍了无袖带测量范式领域的最新进展和可能的未来发展的概念概述，这将使无袖带的逐拍血压估计范式在临床上可行。它避免了对大多数研究的直接比较，而是专注于在具有里程碑意义的科学文献中提出的思想和结论的概念融合。无袖带逐拍 BP 估计在科学文献中有两种主要方法：首先，基于对心血管系统生理理解的模型，主要依赖于脉搏波速度结合其他参数。其次，基于深度学习技术的模型，在其他各种医学领域已经显示出很好的性能。这篇综述旨在介绍每种方法的优点和局限性。在此之后，提出了将生理理解和深度学习技术的优势结合起来用于逐拍 BP 估计的概念性想法。这可能会为无袖带逐拍 BP 估计提供一个通用和统一的解决方案。这不仅会使它们成为传统袖带测量范式的有吸引力的临床补充甚至替代方案，而且还会实质性地改变我们对血压作为心血管医学基本标志物的看法。
