多道睡眠图参数自动分析与人工分析误差产生原因的探讨

Analysis of the differences between autoanalysis and manual analysis of parameters in polysomnogram

目的分析多道睡眠图(polysomnogram,PSG)呼吸事件自动分析与人工分析产生误差的原因,探讨提高阻塞性睡眠呼吸暂停低通气综合征诊断结果准确性的要素。方法选择376例打鼾伴有睡眠憋气或可能有睡眠憋气者进行7 h以上的PSG监测,对检测结果先用软件自动分析,再进行人工分析,对数据进行统计学分析。结果 PSG呼吸事件经人工分析得出的呼吸暂停指数、低通气指数、最长呼吸暂停时间、呼吸暂停低通气指数,与自动分析结果差别有统计学意义,人工分析结果更符合患者病情。结论自动分析结果对PSG呼吸事件病情程度的诊断有误差。目前,应该常规对PSG行人工分析。睡眠呼吸软件应及时更新内容,以提高临床诊断的准确性。

OBJECTIVE To investigate the differences between autoanalysis and manual analysis of respiration events in polysomnogram. To explore how to improve the accuracy of clinic diagnosis on obstructive sleep anea hypopnea syndrome. METHODS 376 snoreing human were monitored by sleep monitor for more than seven hour in the evening. The polysomnogrma was first analyzed by the software of sleep monitor and then by the artificial operator. Finally datas were performed statistical analysis. RESULTS There were differences between autoanalysis and manual analysis of polysomnogram on apnea hyponea index. The datas of manual analysis were more consistent with clinic symptoms of patients than autoanalysis. CONCLUSION The result of autoanalysis of polysomnogram had a deviation from the diagnosis of status of an illness. In order to improve the accuracy of clinic diagnosis, artificial analysis should be utilized in polysomnogrma universally. The software of sleep monitor should be updated in time.