作者：杨月，黄双，柏惠，朱德慧

单位：

江苏省淮安市第二人民医院神经内科

Units：

Department of Neurology, Huai'an Second People's Hospital, Huaian 223022, China

关键词：

帕金森病; 抑郁; 危险因素; 预测模型;

Keywords：

Parkinson disease; Depression; Risk factors; Predictive model

CLC：

DOI：

10.12114/j.issn.1008-5971.2023.00.128

Funds：

摘要：

目的 探讨帕金森病（PD）患者伴发抑郁的影响因素，构建并验证其风险预测列线图模型。方法选取2020年1月至2022年4月淮安市第二人民医院收治的PD患者126例，根据是否伴发抑郁将其分为帕金森病伴发抑郁（d-PD）组（n=60）与帕金森病未伴发抑郁（nd-PD）组（n=66）。收集两组患者临床资料。采用多因素Logistic回归分析探讨PD患者伴发抑郁的影响因素。采用R 3.6.3软件与rms程序包构建PD患者伴发抑郁的风险预测列线图模型；采用ROC曲线、校准曲线及Hosmer-Lomoshow拟合优度检验评估该列线图模型的区分度和拟合程度。结果 d-PD组男性占比、病程<5年者占比、早期PD者占比、基础血浆催乳素（PRL）低于nd-PD组，统一帕金森病评分量表（UPDRS）Ⅰ评分、UPDRSⅡ评分、UPDRSⅢ评分、UPDRSⅣ评分、血清胱抑素C、血清IL-6、血清C反应蛋白（CRP）、匹兹堡睡眠质量指数（PSQI）评分、日间过度嗜睡者占比高于nd-PD组（P<0.05）。多因素Logistic回归分析结果显示，女性、病程≥5年、中晚期PD、UPDRSⅢ评分升高、血清IL-6升高、血清CRP升高是PD患者伴发抑郁的独立危险因素（P<0.05）。基于多因素Logistic回归分析结果构建PD患者伴发抑郁风险预测列线图模型。ROC曲线分析结果显示，该列线图模型预测PD患者伴发抑郁的AUC为0.971[95%CI(0.948,0.993)]，最佳截断值为0.303，灵敏度为96.7%，特异度为84.8%。校准曲线分析结果显示，该列线图模型预测PD患者伴发抑郁的校准曲线接近于理想曲线。Hosmer-Lemeshow拟合优度检验结果显示，该列线图模型预测PD患者伴发抑郁的发生率与患者实际伴发抑郁的发生率比较，差异无统计学意义（χ2=3.209,P=0.921）。结论 基于性别、病程、Hoehn-Yahr分级、UPDRSⅢ评分、血清IL-6、血清CRP构建的列线图模型可有效预测PD患者伴发抑郁的风险。

Abstract：

Objective To explore the influencing factors of depression in patients with Parkinson disease (PD) , and construct and verify a nomogram model for predicting its risk. Methods A total of 126 PD patients admitted to Huai'an Second People's Hospital from January 2020 to April 2022 were selected and divided into depression in PD (d-PD) group (n=60) and no depression in PD (nd-PD) group (n=66) according to whether they had depression. Clinical data of the two groups were collected. Multivariate Logistic regression analysis was used to explore the influencing factors of depression in PD patients. R 3.6.3 software and rms package were used to construct a nomogram model for predicting depression in PD patients. ROC curve, calibration curve and Hosmer Lomoshow goodness of fit test were used to evaluate the discrimination and fitting degree of the nomogram model. Results The proportion of males, the proportion of patients with disease course < 5 years, the proportion of early PD, basic plasma prolactin (PRL) in d-PD group were lower than those in nd-PD group, Unified Parkinson's Disease Rating Scale (UPDRS) Ⅰ score, UPDRS Ⅱ score, UPDRS Ⅲ score, UPDRS Ⅳ score, serum cystatin C, serum IL-6, serum C-reactive protein (CRP) , Pittsburgh Sleep Quality Index (PSQI) score and the proportion of excessive daytime sleepiness were higher than those in nd?PD group (P < 0.05) . Multivariate Logistic regression analysis showed that female, disease duration ≥ 5 years, middle and late PD, increased UPDRS Ⅲ score, increased serum IL-6 and increased serum CRP were independent risk factors for depression in PD patients (P < 0.05) . Based on the results of multivariate Logistic regression analysis, the nomogram model for predicting depression in PD patients was constructed. ROC curve analysis results showed that the AUC of nomogram model for predicting depression in PD patients was 0.971 [95%CI (0.948, 0.993) ] , the optimal cut-off value was 0.303, the sensitivity was 96.7%, and the specificity was 84.8%. The calibration curve analysis results showed that the calibration curve of nomogram model for predicting depression in PD patients was close to the ideal curve. Hosmer-Lemeshow goodness of fit test results showed that there was no statistically significant difference between the incidence of depression in PD patients predicted by the nomogram model and the actual incidence of depression in PD patients (χ2 =3.209, P=0.921) . Conclusion The nomogram model constructed based on gender, course of disease, Hoehn-Yahr classification, UPDRS Ⅲ score, serum IL-6, and serum CRP can effectively predict the risk probability of depression in PD patients.

ReferenceList：