麻醉恢复期基于气管音的呼吸监测方法研究

Respiratory system complication is one of the most common complications during recovery from anesthesia. If effective remedy is not provided promptly, it will lead to severe sequelae and even death. Technologies of respiratory monitoring adopted in clinical practice are defective, which make the patients’ safety are potentially threatened during recovery from anesthesia. Our research team proved that it’s feasible to detect apnea using tracheal sounds during sedation in healthy volunteers, which opens up the opportunity of introducing tracheal sounds to respiratory monitoring during recovery from anesthesia. In this research, firstly, target controlled infusions technology and anesthetics pharmacodynamics response surface model will be used to simulate the physiological conditions of patients during recovery from anesthesia. Secondly, the researching on the relationship between the tracheal sounds and respiratory flow of volunteers in different sedation depth will be performed, and then apnea detection will be conducted using the proper detection algorithm. Thirdly, the central apnea and the obstructive apnea will be separated by identifying the special obstructive snored tracheal sounds which usually happens ahead of the obstructive apnea. Finally, the validity of the proposed respiratory monitoring technology will be evaluated through clinical trials. This research will not only decrease the patients’ compromises from respiratory system complication during recovery from anesthesia, but relieve the workload of medical staff, reduce the clinical mistakes and optimize the allocation of medical resources.

呼吸系统并发症是麻醉恢复期最常见的并发症之一，如得不到及时有效的处理，病人可产生严重的后遗症甚至死亡。目前临床中常用的呼吸监测方法仍有缺陷，使得病人在麻醉恢复期的生命安全受到潜在威胁。课题组前期研究证明，利用颈部气管音对镇静状态下志愿者出现的呼吸暂停进行早期检测是可行的，这一研究成果开辟了我们将气管音应用于麻醉恢复期进行呼吸监测的思路。本研究拟借助靶控输注技术及麻醉药物药效学模型对麻醉恢复期病人的生理状态进行全面模拟；在各种麻醉深度下研究人体气管音与通气状况的关系，并利用相应的检测算法实现呼吸暂停的检测；通过对阻塞型呼吸暂停发生前狭窄阻塞型鼾声的识别，实现中枢型与阻塞型呼吸暂停的分类；最后通过临床对照试验，对本研究提出的呼吸监测方法的临床有效性进行评估。本研究不但可有效降低麻醉恢复期呼吸系统并发症给病人带来的危害，在减轻医护人员工作压力、减少工作失误及优化医疗资源配置方面也有着重要意义。

呼吸系统并发症是麻醉恢复期较为常见的并发症，如不能被及时发现并予以纠正，病人在此阶段的致病及致死机率将大大增加。目前麻醉后恢复室（PACU）中常用的呼吸监测方法主要包括呼气末二氧化碳、脉搏血氧饱和度以及心电监测等，但这些方法均有一定的缺陷，如存在时延、敏感度不足或耐受性差等。本课题首先进行了气管音采集方法的研究，并在环境噪声较大时引入了自适应滤波技术提升信号质量，设计出一种拾音效果好、噪声抑制能力强、重量轻、无毒害且成本低的气管音采集系统，该系统不但可用于PACU中病人气管音的采集，还能够为其他相关研究提供新的思路和方法；之后我们使用已有的实验数据，从模拟呼吸气流速度信号效果的角度比较了不同气管音特征值模拟呼吸气流速度的性能，进行了最优气管音特征值的选择，并对PACU中采集的气管音数据进行此种特征提取，进而完成呼吸暂停检测算法的设计与优化，达到了较为理想的敏感度（92%）和特异度（98%）；此外，我们还通过噪声环境下使用气管音进行呼吸暂停检测的实验进一步证实了我们设计的具有自适应滤波功能的气管音采集系统在有效降低气管音中环境噪声的基础上，可有效地提升呼吸暂停检测算法的敏感度（81.1%→91.5%）及特异度（95.8%→96.4%）；最后，我们还证明了在呼吸率相对稳定时使用气管音监测呼吸率的可行性，进一步丰富了基于气管音的呼吸监测方法。本课题的研究成果证明了在PACU中使用气管音的方法对病人的呼吸状况进行监测（尤其是呼吸暂停的检测）是可行且安全的，其性能可满足临床要求，可有效地保障病患安全，提升医疗质量，在优化医疗资源的同时，还可减轻医护人员工作负荷、减少工作失误，从而进一步提升PACU中病人安全。