扁桃体腺样体肥大合并分泌性中耳炎患者睡眠异常风险预测模型的构建与验证

Construction and validation of a model for predicting sleep abnormalities in patients with adenotonsillar hypertrophy and secretory otitis media

摘要:

目的 建立并验证预测模型，评估扁桃体腺样体肥大合并分泌性中耳炎患者发生睡眠异常的风险。方法 回顾性选取2020年3月—2022年6月国药东风总医院耳鼻咽喉头颈外科收治的279例扁桃体腺样体肥大伴分泌性中耳炎患者作为研究对象。通过多导睡眠监测检测睡眠情况，根据患者是否出现睡眠异常分为异常组（86例）和正常组（193例）。通过分析两组患者的年龄、性别、病程、病史、临床症状等资料，利用多因素一般Logistic回归模型分析影响睡眠异常的关键因素。运用R4.3.0软件绘制列线图模型，并采用受试者工作特征曲线评估模型的区分度。使用Bootstrap方法进行1 000次重复采样，以验证列线图模型的预测效能。结果 异常组男性占比、年龄<10岁占比、病程≥12月占比、扁桃体炎症率、鼻塞率、哮喘病史率、腺样体切除手术史率、入睡困难率和睡眠焦虑率均高于正常组。多因素一般Logistic回归分析结果显示：年龄大［O^R=4.943（95%CI：1.424，17.154）］、男性［O^R=4.928（95%CI：1.357，17.895）］、病程长［O^R=81.286（95% CI：13.621，485.082）］、扁桃体炎症［O^R=26.767（95% CI：6.664，107.519）］、鼻塞［O^R=72.782（95% CI：17.008，311.466）］、哮喘病史［O^R=8.980（95% CI：2.117，38.087）］、腺样体切除手术史［O^R=5.606（95% CI：1.600，19.644）］、入睡困难［O^R=6.143（95% CI：1.829，20.631）］、睡眠焦虑［O^R=7.505（95% CI：1.829，20.631）］均是患者睡眠异常的危险因素（P <0.05）。校准曲线显示，预测模型预测扁桃体腺样体肥大伴分泌性中耳炎患者睡眠异常情况的校准曲线趋近于理想曲线。结论 通过年龄、性别、病程和临床症状等因素来构建扁桃体腺样体肥大伴分泌性中耳炎患者睡眠异常风险的预测模型是可行的，对于临床识别高风险患者并及时进行干预具有重要意义。

Abstract:

Objective To construct and validate a model for predicting the risk of sleep abnormalities in patients with adenotonsillar hypertrophy and secretory otitis media.Methods A total of 279 patients with adenotonsillar hypertrophy and secretory otitis media treated in the Department of Otolaryngology and Head and Neck Surgery at Guoyao Dongfeng General Hospital between March 2020 and June 2022 were retrospectively enrolled. Sleep was assessed by polysomnography (PSG), and the patients were divided into an abnormal-sleep group (n = 86) and a normal-sleep group (n = 193). Age, sex, disease duration, medical history, clinical symptoms, and other data were compared between the groups. Multivariable logistic regression was used to identify key factors associated with sleep abnormalities. R version 4.3.0 was used to construct a nomogram, and receiver operating characteristic (ROC) curves were used to evaluate its discrimination. Bootstrap resampling was performed 1,000 times to validate the predictive performance of the nomogram.Results The proportions of male patients, patients aged < 10 years, patients with a disease duration of ≥ 12 months, and patients with tonsillitis, nasal congestion, a history of asthma, a history of adenoidectomy, difficulty falling asleep, or sleep anxiety were higher in the abnormal-sleep group than in the normal-sleep group. Multivariable logistic regression showed that older age [O^R = 4.943 (95% CI: 1.424, 17.154) ], male sex [O^R = 4.928 (95% CI: 1.357, 17.895) ], longer disease duration [O^R = 81.286 (95% CI: 13.621, 485.082) ], tonsillitis [O^R = 26.767 (95% CI: 6.664, 107.519) ], nasal congestion [O^R = 72.782 (95% CI: 17.008, 311.466) ], a history of asthma [O^R = 8.980 (95% CI: 2.117, 38.087) ], a history of adenoidectomy [O^R = 5.606 (95% CI: 1.600, 19.644) ], difficulty falling asleep [O^R = 6.143 (95% CI: 1.829, 20.631) ], and sleep anxiety [O^R = 7.505 (95% CI: 1.829, 20.631) ] were risk factors for sleep abnormalities (P < 0.05). The calibration curve of the prediction model closely approximated the ideal curve.Conclusion A model based on age, sex, disease duration, and clinical symptoms can predict the risk of sleep abnormalities in patients with adenotonsillar hypertrophy and secretory otitis media and may facilitate the timely identification and intervention of high-risk patients.