摘要
目的探讨早期帕金森病(ePD)轻度认知障碍(MCI)患者伴失眠的影响因素及预测模型。方法选取2023年2月至2024年12月ePD-MCI患者290例为研究对象。使用阿森斯失眠量表(AIS)结合多导睡眠图(PSG)评估失眠,依据失眠情况分为失眠组、无失眠组。统计ePD-MCI患者蒙特利尔认知评估量表(MoCA)、医院焦虑抑郁量表(HADS)、帕金森病(PD)自主神经症状量表(SCOPA-AUT)、左旋多巴(LD)等效剂量等资料。Logistic回归分析ePD-MCI患者失眠的影响因素,构建失眠列线图预测模型,受试者工作特征(ROC)曲线、校准曲线、决策与临床影响曲线,评价ePD-MCI患者失眠预测模型的区分度、一致性、临床实用性。结果本研究有效问卷回收率96.55%(280/290),280例ePD-MCI患者失眠发生率38.21%(107/280)。Logistic回归分析显示,年龄(OR:2.873,95%CI:1.253~6.588)、LD等效剂量(OR:1.010,95%CI:1.004~1.016)、首发症状(OR:5.296,95%CI:1.381~20.318)、病程(OR:3.031,95%CI:1.856~4.951)、MoCA评分(OR:0.542,95%CI:0.394~0.747)、SCOPA-AUT评分(OR:1.504,95%CI:1.296~1.745)、HADS-焦虑(HADS-A)评分(OR:1.795,95%CI:1.425~2.260)、HADS-抑郁(HADS-D)评分(OR:1.491,95%CI:1.170~1.901)是ePD-MCI患者伴失眠的影响因素(P<0.05)。依据影响因素构建失眠列线图预测模型,ROC曲线评价该模型对失眠预测的曲线下面积(AUC)为0.951。H-L检验显示失眠列线图模型预测值与实际观测值比较差异无统计学意义(χ^(2)=7.803,P=0.453)。决策与临床影响曲线评价该模型具有良好的临床应用价值。结论年龄、LD等效剂量、首发症状、病程、MoCA评分、SCOPA-AUT评分、HADS-A评分、HADS-D评分水平均可影响ePD-MCI患者失眠,据此构建的失眠列线图模型具有良好的预测效能、一致性及临床价值,可为临床早期风险管理提供依据。
Objective To investigate the influencing factors and predictive models of insomnia in early Parkinson's disease(ePD)patients with mild cognitive impairment(MCI).Methods 290 ePD-MCI patients from February 2023 to December 2024 were selected as research subjects.Athens insomnia scale(AIS)combined with polysomnography(PSG)was used to assess insomnia,and patients were divided into insomnia group and non-insomnia group.Data collected included Montreal cognitive assessment(MoCA),Hospital anxiety and depression scale(HADS),scales for outcomes in Parkinson's disease-autonomic(SCOPA-AUT),levodopa(LD)equivalent dose and other data.Logistic regression was used to analyze factors influencing insomnia in ePD-MCI patients.A nomogram prediction model for insomnia was constructed,and its discrimination,consistency,and clinical utility were evaluated using receiver operating characteristic(ROC)curve,calibration curve,decision and clinical impact curve.Results The effective questionnaire recovery rate was 96.55%(280/290),and the insomnia incidence in 280 ePD-MCI patients was 38.21%(107/280).Logistic regression analysis showed that age(OR:2.873,95%CI:1.253-6.588),LD equivalent dose(OR:1.010,95%CI:1.004-1.016),first symptoms(OR:5.296,95%CI:1.381-20.318),course of disease(OR:3.031,95%CI:1.856-4.951),MoCA score(OR:0.542,95%CI:0.394-0.747),SCOPA-AUT score(OR:1.504,95%CI:1.296-1.745),HADS-Anxiety(HADS-A)score(OR:1.795,95%CI:1.425-2.260),HADS-Depression(HADS-D)score(OR:1.491,95%CI:1.170-1.901)were influencing factors for insomnia in ePD-MCI patients(P<0.05).A nomogram prediction model for insomnia was constructed based on these factors.ROC curve evaluation showed that the area under curve(AUC)of this model for insomnia prediction was 0.951.Hosmer-Lemeshow test showed no statistical significance between predicted and observed values of the nomogram model(χ^(2)=7.803,P=0.453).Decision and clinical impact curves indicated that the model had good clinical application value.Conclusion Age,LD equivalent dose,first symptom,course of disease,MoCA score,SCOPA-AUT score,HADS-A score,HADS-D score can all influence insomnia in ePD-MCI patients.The insomnia nomogram model constructed based on these factors has good predictive performance,consistency,and clinical value,providing a basis for early clinical risk management.
作者
徐冰泉
刘晓卫
季灵
王婷
季静
刘卫国
潘全慧
Xu Bingquan;Liu Xiaowei;Ji Ling;Wang Ting;Ji Jing;Liu Weiguo;Pan Quanhui(Department of Traditional Chinese Medicine,Nanjing Medical University Affiliated Brain Hospital,Jiangsu 210029,China)
出处
《脑与神经疾病杂志》
2025年第12期734-739,共6页
Journal of Brain and Nervous Diseases
基金
2024江苏省社会发展项目(BE2021711)。
关键词
早期帕金森病
轻度认知障碍
失眠
影响因素
预测模型
Early Parkinson's disease
Mild cognitive impairment
Insomnia
Influencing factors
Prediction model
