便携式睡眠监测中复合低氧评分预测重度OSA的效能：BMI分层中正常体重组辨别力最高

Predictive Performance of a Composite Hypoxic Score for Severe OSA in Portable Sleep Monitoring: Highest Discriminative Ability in the Normal-Weight Group within BMI Stratification

摘要:

目的 基于最低血氧饱和度（SpO?）和最长呼吸暂停时间构建复合低氧评分，系统评估其对重度阻塞性睡眠呼吸暂停（OSA）的预测效能，并探索其在性别、BMI及年龄亚组中的差异。 方法 基于回顾性数据的横断面研究，纳入1943例成年（≥18岁）行便携式睡眠监测的患者。采用Spearman秩相关分析各低氧指标与呼吸暂停低通气指数（AHI）的相关性；采用ROC曲线评估其对重度OSA（AHI≥30）的预测效能；按性别、BMI和年龄进行亚组分析。 结果 AHI与最低SpO?呈强负相关（r=?0.786，P<0.001），与最长呼吸暂停时间呈强正相关（r=0.755，P<0.001）。复合低氧评分的AUC为0.904（95%CI：0.891~0.918），数值上高于最低SpO?（AUC=0.897，95%CI：0.883~0.911）和最长暂停时间（AUC=0.870，95%CI：0.854~0.885）。亚组分析显示，性别分层中复合低氧评分预测重度OSA的AUC在女性中为0.942、男性中为0.893；最低SpO?在女性中为0.917、男性中为0.889；最长暂停时间在女性中为0.916、男性中为0.855。BMI分层中复合低氧评分在正常体重组预测效能最高（AUC=0.918，95%CI：0.888~0.949），超重组为0.887（95%CI：0.864~0.909），肥胖组为0.914（95%CI：0.892~0.935）。年龄分层中成人组复合评分AUC为0.909（95%CI：0.896~0.922），老年组为0.819（95%CI：0.746~0.892）。 结论 复合低氧评分整合低氧深度与持续时间，对重度OSA的预测效能略优于单一指标（AUC=0.904）。在BMI分层中，其于正常体重（非肥胖）人群（AUC=0.918）辨别力最高，且该人群常规筛查最易漏诊，提示复合低氧评分可作为瘦型OSA床旁筛查的简便有效工具。

Abstract:

Objective To develop a composite hypoxic score based on lowest oxygen saturation (SpO?) and longest apnea duration, and to systematically evaluate its predictive performance for severe obstructive sleep apnea (OSA), as well as to explore its variations across sex, BMI, and age subgroups. Methods This was a retrospective cross-sectional study of 1,943 adult patients (≥18 years) who underwent portable sleep monitoring. Spearman''s rank correlation analysis was used to assess the correlations between each hypoxic indicator and the apnea-hypopnea index (AHI). Receiver operating characteristic (ROC) curves were employed to evaluate the predictive performance for severe OSA (AHI ≥ 30 events/h). Subgroup analyses were performed stratified by sex, BMI, and age. Results AHI showed a strong negative correlation with lowest SpO? (r = ?0.786, P < 0.001) and a strong positive correlation with longest apnea duration (r = 0.755, P < 0.001). The composite hypoxic score achieved an AUC of 0.904 (95% CI: 0.891–0.918), numerically higher than lowest SpO? (AUC = 0.897, 95% CI: 0.883–0.911) and the longest apnea duration (AUC = 0.870, 95% CI: 0.854–0.885). Subgroup analyses revealed that, in sex stratification, the AUC of the composite hypoxic score for predicting severe OSA was 0.942 in females and 0.893 in males; that of lowest SpO? was 0.917 in females and 0.889 in males, while that of longest apnea duration was 0.916 in females and 0.855 in males. In BMI stratification, the composite hypoxic score showed the highest predictive performance in the normal-weight group (AUC = 0.918, 95% CI: 0.888–0.949), followed by the obese group (AUC = 0.914, 95% CI: 0.892–0.935) and the overweight group (AUC = 0.887, 95% CI: 0.864–0.909). In age stratification, the AUC of the composite hypoxic score was 0.909 (95% CI: 0.896–0.922) in the adult group and decreased to 0.819 (95% CI: 0.746–0.892) in the elderly group. Conclusion The composite hypoxic score, integrating hypoxic depth and duration, demonstrates predictive performance for severe OSA slightly superior to single indicators (AUC = 0.904). Within BMI stratification, its discriminative value was highest in the normal-weight (non-obese) group (AUC = 0.918), a subgroup most prone to under-detection by conventional screening, suggesting the score serves as a simple, effective bedside tool for lean OSA screening.