Rodriguez-Villegas Esther, Chen Gwangwei, Radcliffe Jeremy, Duncan John
Department of Electrical and Electronic Engineering, Imperial College London, London, UK.
Department of Neuroanaesthesia, UCL Institute of Neurology, National Hospital for Neurology and Neurosurgery, London, UK.
BMJ Open. 2014 Oct 3;4(10):e005299. doi: 10.1136/bmjopen-2014-005299.
Current techniques for monitoring patients for apnoea suffer from significant limitations. These include insufficient availability to meet diagnostic needs, cost, accuracy of results in the presence of artefacts and difficulty of use in unsupervised conditions.
We created and clinically tested a novel miniature medical device that targets overcoming these limitations.
We studied 20 healthy control participants and 10 patients who had been referred for sleep apnoea diagnosis. The performances of the new system and also of the Food and Drug Administration (FDA) approved SOMNO clinical system, conventionally used for sleep apnoea diagnosis were evaluated under the same conditions. Both systems were tested during a normal night of sleep in controls and patients. Their performances were quantified in terms of detection of apnoea and hypopnoea in individual 10 s epochs, which were compared with scoring of signals by a blinded clinician.
For spontaneous apnoeas during natural sleep and considering the clinician scorer as the gold standard, the new wearable apnoea detection device had 88.6% sensitivity and 99.6% specificity. In comparison the SOMNO system had 14.3% sensitivity and 99.3% specificity. The novel device had been specifically designed to detect apnoea, but if apnoea and hypopnoea during sleep were both considered in the assessment, the sensitivity and specificity were 77.1% and 99.7%, respectively, versus 54% and 98.5%, respectively, for the SOMNO.
The performance of the novel device compares very well to the scoring by an experienced clinician even in the presence of breathing artefacts, in this small pilot study. This can potentially make it a real solution for apnoea home monitoring.
当前用于监测患者呼吸暂停的技术存在显著局限性。这些局限性包括无法充分满足诊断需求、成本高、存在伪影时结果的准确性以及在无人监督条件下使用困难。
我们研发并对一种新型微型医疗设备进行了临床测试，旨在克服这些局限性。
我们研究了20名健康对照参与者和10名被转诊进行睡眠呼吸暂停诊断的患者。在相同条件下评估了新系统以及美国食品药品监督管理局（FDA）批准的常用于睡眠呼吸暂停诊断的SOMNO临床系统的性能。在对照者和患者正常睡眠的夜晚对两个系统进行了测试。根据在10秒的单个时段内对呼吸暂停和呼吸不足的检测情况对它们的性能进行量化，并与一位不知情的临床医生对信号的评分进行比较。
对于自然睡眠期间的自发性呼吸暂停，并将临床医生评分作为金标准，新型可穿戴呼吸暂停检测设备的灵敏度为88.6%，特异性为99.6%。相比之下，SOMNO系统的灵敏度为14.3%，特异性为99.3%。该新型设备专门设计用于检测呼吸暂停，但如果在评估中同时考虑睡眠期间的呼吸暂停和呼吸不足，其灵敏度和特异性分别为77.1%和99.7%，而SOMNO系统的灵敏度和特异性分别为54%和98.5%。
在这项小型试点研究中，即使存在呼吸伪影，新型设备的性能与经验丰富的临床医生的评分相比也非常出色。这有可能使其成为呼吸暂停家庭监测的真正解决方案
