Steltner Holger, Staats Richard, Timmer Jens, Vogel Michael, Guttmann Josef, Matthys Heinrich, Christian Virchow J
Center for Data Analysis and Modeling, University of Freiburg, Freiburg, Germany.
Am J Respir Crit Care Med. 2002 Apr 1;165(7):940-4. doi: 10.1164/ajrccm.165.7.2106018.
Detecting and differentiating central and obstructive respiratory events is an important aspect of the diagnosis of sleep-related breathing disorders with respect to the choice of an appropriate treatment. The purpose of this study was to evaluate the performance of a new algorithm for automated detection and classification of apneas and hypopneas, compared with visual analysis of standard polysomnographic signals. The algorithm is based on time series analysis of nasal mask pressure and a forced oscillation signal related to mechanical respiratory input impedance, measured at a frequency of 20 Hz throughout the night. The method was applied to all-night measurements on 19 subjects. Two experts in sleep medicine independently scored the corresponding simultaneously recorded polysomnographic signals. Evaluating the agreement between two scorers by a weighted kappa statistic on a second-by-second basis, we found that inter-expert variability and the discrepancy between automatic analysis and visual analysis performed by an expert were not significantly different. Implementation of this algorithm in a device for home monitoring of breathing during sleep might aid in the differential diagnosis of sleep-related breathing disorders and/or as a means for follow-up and treatment control.
对于睡眠相关呼吸障碍的诊断而言，检测并区分中枢性和阻塞性呼吸事件是选择合适治疗方法的一个重要方面。本研究的目的是评估一种用于自动检测和分类呼吸暂停及低通气的新算法的性能，并与标准多导睡眠图信号的视觉分析进行比较。该算法基于鼻罩压力的时间序列分析以及与机械呼吸输入阻抗相关的强迫振荡信号，整夜以20赫兹的频率进行测量。该方法应用于19名受试者的整夜测量。两名睡眠医学专家独立对相应同时记录的多导睡眠图信号进行评分。通过逐秒加权kappa统计量评估两位评分者之间的一致性，我们发现专家间的变异性以及自动分析与专家视觉分析之间的差异并无显著不同。将该算法应用于睡眠期间呼吸的家庭监测设备中，可能有助于睡眠相关呼吸障碍的鉴别诊断和/或作为随访及治疗控制的一种手段。
