Zamarrón C, Romero P V, Gude F, Amaro A, Rodriguez J R
Division of Respiratory Medicine, Hospital Clínico Universitario, Santiago de Compostela, Spain.
Respir Med. 2001 Sep;95(9):759-65. doi: 10.1053/rmed.2001.1128.
Using heart rate spectral analysis of nocturnal pulse oximetry, we prospectively evaluated the utility of this methodology in patients clinically suspected of having obstructive sleep apnoea (OSA). A hundred and ninety-seven outpatients referred with symptoms compatible with the diagnosis of OSA were studied. All participants had nocturnal pulse oximetry performed simultaneously with conventional polysomnography. Power density of heart rate obtained by nocturnal pulse oximetry was analysed using fast Fourier transformation of a Hamming-windowed signal. Recording test results were classified as abnormal (suspicion of OSA) in the presence of a peak in the periodogram between period boundaries 30-70 sec. A normal test result was defined as the absence of the 30-70 sec peak in the periodogram. The total area of the periodogram (S(TOT)), the area enclosed in the periodogram between the period boundaries 30-70 sec (S(30-70)), the area enclosed in the period boundaries 30-70 sec with respect to the total area of the periodogram (S) and the peak amplitude 30-70 sec (PA) were measured. The presence of a peak in the periodogram has a sensitivity of 81.3%, a specificity of 91.5% a positive predictive value of 89.1% and a negative predictive value of 85.1% for OSA diagnosis. The OSA patients were found to have higher values of S(TOT), S(30-70), S and PA than the non OSA patients. Receiver operating characteristics (ROC) curve was constructed at different thresholds of S(TOT), S(30-70) S and PA. For a PA threshold of 10(%)2, heart rate spectra analysis sensitivity for OSA was 58% and specificity was 92%. Furthermore, the positive and negative predictive values for diagnosis of OSA were 87 and 72% respectively. Apnoea hypopnea index (AHI) correlated significantly with S(TOT) (r=0.44; P<0.001), S(30-70) (r=0.59: P<0.001), S (r=0.58; P<0.001) and PA (r=0.58; P<0.001). According to our results, heart rate spectral analys s obtained by nocturnal pulse oximetry and identification of peak in the periodogram between period boundaries 30-70 sec could be useful as a diagnostic technique for OSA patients.
我们采用夜间脉搏血氧饱和度的心率谱分析方法，对临床疑似阻塞性睡眠呼吸暂停（OSA）患者进行了该方法实用性的前瞻性评估。研究对象为197例因症状符合OSA诊断而转诊的门诊患者。所有参与者均同时进行夜间脉搏血氧饱和度监测和传统多导睡眠图监测。通过对汉明窗信号进行快速傅里叶变换，分析夜间脉搏血氧饱和度获得的心率功率密度。在周期图中30 - 70秒的周期边界之间出现峰值时，记录测试结果被分类为异常（怀疑OSA）。正常测试结果定义为周期图中不存在30 - 70秒的峰值。测量周期图的总面积（S(TOT)）、周期图中30 - 70秒周期边界之间的面积（S(30 - 70)）、30 - 70秒周期边界之间的面积相对于周期图总面积的比例（S）以及30 - 70秒的峰值幅度（PA）。周期图中出现峰值对OSA诊断的敏感性为81.3%，特异性为91.5%，阳性预测值为89.1%，阴性预测值为85.1%。发现OSA患者的S(TOT)、S(30 - 70)、S和PA值高于非OSA患者。在S(TOT)、S(30 - 70)、S和PA的不同阈值下构建受试者工作特征（ROC）曲线。对于PA阈值为10(%)2，心率谱分析对OSA的敏感性为58%，特异性为92%。此外，OSA诊断的阳性和阴性预测值分别为87%和72%。呼吸暂停低通气指数（AHI）与S(TOT)（r = 0.44；P < 0.001）、S(30 - 70)（r = 0.59：P < 0.001）、S（r = 0.58；P < 0.001）和PA（r = 0.58；P < 0.001）显著相关。根据我们的结果，通过夜间脉搏血氧饱和度获得的心率谱分析以及识别周期图中30 - 70秒周期边界之间的峰值，可作为OSA患者的一种诊断技术。
