作者：宋晶晶，文秦秦，朱雅文

单位：

223001江苏省淮安市第二人民医院神经内科

单位（英文）：

Department of Neurology, Huai'an Second People's Hospital, Huai'an 223001, China

关键词：

癫痫；睡眠障碍；影响因素分析；列线图

关键词（英文）：

Epilepsy; Sleep disorders; Root cause analysis; Nomograms

中图分类号：

R 742.1

DOI：

10.12114/j.issn.1008-5971.2023.00.151

基金项目：

江苏省卫生健康委2019年度医学科研立项项目（H2019030）

摘要：

目的　探讨癫痫患者发生睡眠障碍的影响因素，构建其风险预测列线图模型并进行内部和外部验证。方法　选取2019年1月至2020年8月淮安市第二人民医院收治的癫痫患者196例作为建模集，另选取2020年9月至2021年10月淮安市第二人民医院收治的癫痫患者84例作为验证集。收集患者的临床资料，根据是否发生睡眠障碍将建模集患者分为睡眠障碍组和无睡眠障碍组。癫痫患者发生睡眠障碍的影响因素分析采用多因素Logistic回归分析；采用R 4.1.0软件包及rms程序包建立癫痫患者发生睡眠障碍的风险预测列线图模型；采用Hosmer-Lemeshoe拟合优度检验评价该列线图模型的拟合程度；绘制校准曲线以评估该列线图模型预测建模集和验证集癫痫患者发生睡眠障碍的效能；采用ROC曲线分析该列线图模型对建模集和验证集癫痫患者发生睡眠障碍的预测价值。结果　建模集中，发生睡眠障碍84例（42.86%），未发生睡眠障碍112例（57.14%）。两组年龄、发作类型、发作频率、用药种类、用药数量、抑郁自评量表评分、焦虑自评量表评分、疲劳量表（FS-14）评分比较，差异有统计学意义（P＜0.05）。多因素Logistic回归分析结果显示，发作类型、发作频率、用药种类、用药数量、抑郁自评量表评分、焦虑自评量表评分是癫痫患者发生睡眠障碍的影响因素（P＜0.05）。基于多因素Logistic回归分析结果，构建癫痫患者发生睡眠障碍的风险预测列线图模型。Hosmer-Lemeshoe拟合优度检验结果显示，在建模集和验证集中该列线图模型的拟合程度较好（χ2=7.904，P=0.518；χ2=8.107，P=0.453）。校准曲线分析结果显示，该列线图模型预测建模集和验证集癫痫患者睡眠障碍发生率与实际发生率基本吻合。ROC曲线分析结果显示，该列线图模型预测建模集和验证集癫痫患者发生睡眠障碍的AUC分别为0.867〔95%CI（0.814，0.920）〕、0.880〔95%CI（0.811，0.949）〕。结论　全面发作、发作频率≥1次/月、传统AEDs、用药数量≥2种、抑郁自评量表评分升高、焦虑自评量表评分升高是癫痫患者发生睡眠障碍的危险因素，而基于上述因素构建的列线图模型对癫痫患者发生睡眠障碍具有一定预测价值。

英文摘要：

Objective To analyze the influencing factors of sleep disorders in patients with epilepsy, construct anomogram model for predicting its risk, and conduct internal and external validation. Methods A total of 196 patients withepilepsy in Huai'an Second People's Hospital from January 2019 to August 2020 were selected as modeling set; 84 patients withepilepsy in Huai'an Second People's Hospital from September 2020 to October 2021 were selected as validation set. Clinical dataof patients were collected, and the patients in the modeling set were divided into sleep disorders group and non-sleep disordersgroup according to the occurrence of sleep disorders. The multivariate Logistic regression analysis was used to analyze theinfluencing factors of sleep disorders in patients with epilepsy. The nomogram model for predicting the risk of sleep disorders inpatients with epilepsy was constructed by using the R 4.1.0 software package and rms package. Hosmer-Lemeshow goodness offit test was used to evaluate the fitting degree of the nomogram model. Calibration curve was used to evaluate the reliability of thenomogram model for predicting sleep disorders in patients with epilepsy in modeling set and validation set, and the ROC curvewas used to analyze the predictive value of the nomogram model for sleep disorders in patients with epilepsy in modeling set andvalidation set. Results In the modeling set, 84 (42.86%) patients had sleep disorders, and 112 (57.14%) patients had no sleepdisorders. There were significant differences in age, seizure type, seizure frequency, type of medication, number of medications, Self-rating Depression Scale score, Self-rating Anxiety Scale score, Fatigue Scale-14 (FS-14) score between the two groups(P < 0.05) . Multivariate Logistic regression analysis showed that seizure type, seizure frequency, type of medication, number ofmedications, Self-rating Depression Scale score, Self-rating Anxiety Scale score were the influencing factors of sleep disorders inpatients with epilepsy (P < 0.05) . The nomogram model for predicting sleep disorders in patients with epilepsy was constructedbased on the results of multivariate Logistic regression analysis. The results of Hosmer-Lemeshow goodness of fit test showedthat the nomogram model fit well in modeling set (χ2=7.904, P=0.518) and validation set (χ2=8.107, P=0.453) . The results ofcalibration curve analysis showed that the incidence of sleep disorders in patients with epilepsy predicted by the nomogram modelwas basically consistent with the actual incidence of sleep disorders in patients with epilepsy in modeling set and validation set.The results of ROC curve analysis showed that the AUC of the nomogram model for predicting sleep disorders in patients withepilepsy in modeling set and validation set was 0.867 [95%CI (0.814, 0.920) ] , 0.880 [95%CI (0.811, 0.949) ] , respectively.Conclusion Comprehensive seizures, seizure frequency ≥ 1 time/month, traditional AEDs, number of medications ≥ 2, increasedSelf-rating Depression Scale score, increased Self-rating Anxiety Scale score are the risk factors of sleep disorders in patients withepilepsy. The nomogram model constructed based on the above factors has a certain predictive value for sleep disorders in patientswith epilepsy.

参考文献：