摘要
目的分析帕金森病(PD)患者并发异动症的危险因素,并构建列线图预测模型。方法收集114例PD患者的临床资料,根据是否并发异动症分为异动症组(n=18)和非异动症组(n=96)。采用logistic回归模型分析PD患者并发异动症的独立危险因素并构建列线图预测模型,列线图模型的内部验证及预测效能分别用校正曲线、决策曲线评估。结果异动症组发病年龄、维生素B12水平均低于非异动症组(t分别=-3.55、-2.57,P均<0.05),女性比例、病程、左旋多巴等效剂量(LEDD)、H-Y分级、帕金森病统一评分量表(UPDRS)评分、Beck抑郁量表(BDI)评分、Back焦虑量表(BAI)评分均高于非异动症组,差异均有统计学意义(χ^(2)=6.15,t分别=3.80、2.44、4.74、2.41、3.91、2.72,P均<0.05)。ROC结果显示:发病年龄、病程、LEDD、H-Y分级、UPDRS评分、BDI评分、BAI评分、维生素B12的最佳截断值分别为65岁、8年、449 mg/d、3级、22分、17分、20分、323.48 pmol/L。多元logistic回归模型结果显示:发病年龄、性别、病程、LEDD、H-Y分级、UPDRS评分是PD患者并发异动症的独立危险因素(OR分别=1.57、1.36、1.73、1.88、1.70、1.76,P均<0.05)。内部验证结果显示,列线图模型对PD患者并发异动症预测的C-index为0.90。列线图模型的阈值>0.19,列线图模型提供的临床净收益均高于发病年龄、性别、病程、LEDD、H-Y分级、UPDRS评分。结论本次研究基于发病年龄、性别、病程、LEDD、H-Y分级、UPDRS评分所构建的列线图模型,对PD患者并发异动症的预测价值较好,可为临床预防PD患者并发异动症提供依据。
Objective To analyze the risk factors for dyskinesia in patients with Parkinson's disease and construct a predictive model with the nomogram.Methods Clinical data of 114 patients with Parkinson's disease(PD)were collect-ed.They were divided into LID group(n=18)and non-LID group(n=96)according to whether they had dyskinesia or not.Logistic regression was used to analyze the independent risk factors of Parkinson's disease patients with dyskinesia and to construct the predictive model of the nomogram.The internal validation of the nomogram and the predictive effica-cy were assessed by the calibration curve and the nomogram decision curve,respectively.Results The age of onset and VitB12 level of the LID group were lower than those of the non-LID group(t=-3.55,-2.57,P<0.05).The proportion of females,duration of disease,LEDD,H-Y grading,UPDRS score,BDI score,and BAI score of the LID group were higher than those of the non-LID group,and the difference was significant(χ^(2)=6.15,t=3.80,2.44,4.74,2.41,3.91,2.72,P<0.05).The ROC results showed that the optimal cut-off values for age of onset,disease duration,LEDD,H-Y classifica-tion,UPDRS score,BDI score,BAI score,and VitB12 were 65 years,8 years,449 mg/d,grade 3,22 points,17 points,20 points,and 323.48 pmol/L,respectively.The results of logistic multiple regression model showed that age at onset,gender,duration of disease,LEDD,H-Y classification,and UPDRS score were independent risk factors for dyskinesia in patients with Parkinson's disease(OR=1.57,1.36,1.73,1.88,1.70,1.76,P<0.05).Internal validation showed that the C-index of the nomogram for the prediction of dyskinesia was 0.90.The threshold for the nomogram model was>0.19,and the column-line diagram model provided a net clinical benefit over age at onset,gender,disease duration,LEDD,H-Y classification,and UPDRS score.Conclusion In this study,the nomogram model constructed on the basis of age at onset,gender,disease duration,LEDD,H-Y grading,and UPDRS scores has a good predictive value of dyskinesia in patients with Parkinson's disease,which can provide a referrence for clinical prevention of dyskinesia in patients with Parkinson's disease.
作者
徐雯
冯耀耀
张长国
XU Wen;FENG Yaoyao;ZHANG Changguo(Department of Neurology,Huzhou Third People's Hospital,Huzhou 313000,China)
出处
《全科医学临床与教育》
2024年第3期219-223,F0002,共6页
Clinical Education of General Practice
基金
湖州市科学技术局项目(2022GYB52)。
关键词
帕金森病
异动症
列线图预测模型
Parkinson's disease
dyskinesia
nomogram predictive model
