目的：分析初诊未服药帕金森病伴抑郁（Depression in Parkinson's disease，DPD）患者的临床特征及相关影响因素，构建列线图风险预测模型。方法：选取2021年10月至2025年7月南京医科大学附属江宁医院神经内科连续收治的268 例初诊、未服药PD患者，收集其一般临床资料及相关量表评估结果。依据汉密尔顿抑郁量表24项（HAMD-24）进行分组：HAMD-24评分＜8分纳入非DPD组，≥8分纳入DPD组。通过单因素分析比较2组在一般临床资料及临床量表评分上的差异，单因素分析筛选出的统计学差异指标，均纳入多因素Logistic回归模型，以此明确初诊未服药PD患者并发抑郁的独立危险因素。并基于上述因素绘制列线图，构建DPD风险预测模型。结果：单因素分析显示，2组 UPDRS-Ⅱ、UPDRS-Ⅲ、H-Y 分期、HAMA、MMSE、MoCA、PDSS、NMSQ、RBD、嗅觉评分、KPPS 评分比较差异有统计学意义（P＜0.05），据此将上述存在组间差异的指标纳入多因素Logistic回归模型行进一步分析，结果表明 HAMA、PDSS、RBD、嗅觉评分及 KPPS 评分可独立影响初诊 PD 患者抑郁的发生（均P＜0.05）。基于上述5项独立危险因素建立的预测模型，ROC 曲线分析显示，曲线下面积为 0.796（95% CI 0.755～0.837），提示模型具有良好的临床区分度；校准曲线显示模型预测值与实际观测值吻合度较高；决策曲线分析表明其临床净获益显著，适用性良好。结论：HAMA、PDSS、RBD、嗅觉评分及 KPPS 评分是初诊未服药PD患者发生抑郁的独立影响因素，基于上述指标构建的预测模型具备良好预测效能。

Objective: To analyze the clinical characteristics and related influencing factors in drug‑naive patients with newly‑diagnosed Parkinson's disease accompanied by depression (Depression in Parkinson's disease, DPD), and to construct a nomogram risk‑prediction model. Methods: A total of 268 consecutive drug‑naive patients with newly‑diagnosed Parkinson's disease (PD) admitted to the Department of Neurology, Jiangning Hospital Affiliated to Nanjing Medical University from October 2021 to July 2025 were enrolled. General clinical data and results of relevant scale assessments were collected. Patients were divided into two groups according to the 24‑item Hamilton Depression Rating Scale (HAMD‑24): patients with HAMD‑24 score <8 points were assigned to the non‑DPD group, and those with HAMD‑24 score ≥8 points to the DPD group. Univariate analysis was performed to compare inter‑group differences in general clinical data and clinical scale scores. Variables with statistically significant differences screened by univariate analysis were included in the multivariate Logistic regression model to identify independent risk factors for depression in drug‑naive newly‑diagnosed PD patients. A nomogram was plotted based on these factors to establish the DPD risk‑prediction model. Results: Univariate analysis revealed statistically significant inter‑group differences in UPDRS-Ⅱ, UPDRS-Ⅲ, Hoehn‑Yahr (H‑Y) stage, HAMA, MMSE, MoCA, PDSS, NMSQ, RBD, olfactory score and KPPS score (all P<0.05). These differentially distributed variables were further entered into the multivariate Logistic regression model. The results demonstrated that HAMA, PDSS, RBD, olfactory score and KPPS score were independent influencing factors for depression in newly‑diagnosed PD patients (all P<0.05). The prediction model built on the five independent risk factors yielded an area under the receiver‑operating characteristic (ROC) curve of 0.796 (95%CI 0.755～0.837), indicating favorable discriminatory power of the model. The calibration curve showed high consistency between model‑predicted values and actual observed values. Decision‑curve analysis suggested prominent clinical net benefit and satisfactory applicability of the model. Conclusion: HAMA, PDSS, RBD, olfactory score and KPPS score are independent influencing factors for depression in drug‑naive newly‑diagnosed PD patients. The prediction model constructed from the above indicators exhibits good predictive performance.