Abstract
Objective To construct and validate an individualized nomogram early warning model for stroke-associated pneumonia (SAP) in patients with acute cerebral hemorrhage (ACH). Methods A total of 390 patients with ACH admitted to our hospital from January to December 2023 were retrospectively enrolled and divided into an observation group (n=200) and a control group (n=190 ) based on whether they developed SAP. Potential risk factors were collected, including basic patient information, clinical indicators, and treatment data. Independent predictors were screened using univariate and multivariate logistic regression analyses. A nomogram model was constructed using R 3. 6. 3 software and the rms package. The discrimination, calibration, and clinical utility of the model were evaluated using the receiver operating characteristic (ROC) curve, calibration curve, decision curve analysis (DCA), and integrated discrimination improvement (IDI) index. Results Multivariate analysis identified five independent predictors: dysphagia (OR=3. 45), duration of mechanical ventilation (OR=2. 78), hypoproteinemia (OR=2. 15), NIHSS score (OR=1. 23), and age (OR=1. 05). The nomogram model constructed based on these factors achieved an area under the curve (AUC) of 0. 85 in the training set and 0. 83 in the validation set, with good calibration (Hosmer-Lemeshow test, P=0. 125). DCA confirmed that the model provided a clinical net benefit within the threshold probability range of 10%–80%, with an IDI of 0. 05 (P=0. 003). Conclusion The nomogram model constructed in this study can effectively achieve individualized and visualized prediction of SAP risk in ACH patients during hospitalization, providing a reliable quantitative tool for early clinical identification of high-risk populations and implementation of targeted interventions. (目的 构建并验证急性脑出血(ACH)患者并发卒中相关性肺炎的个体化列线图预警模型。方法 回顾性纳入2023年1月—12月本院收治的390例ACH患者, 根据是否并发卒中相关性肺炎分为观察组(n=200)与对照组(n=190)。收集患者基本信息、临床指标、治疗数据等中的潜在危险因素, 通过单因素及多因素Logistic回归筛选独立预测因子, 利用R 3. 6. 3软件及rms程序包构建列线图模型。采用受试者工作特征(ROC)曲线、校准曲线、决策曲线分析(DCA)及综合判别改善指数(IDI)评估模型的区分度、校准度和临床实用性。结果 多因素分析确定5项独立预测因子: 吞咽障碍(OR=3. 45)、机械通气时间(OR=2. 78)、低蛋白血症(OR=2. 15)、NIHSS评分(OR=1. 23)、年龄(OR=1. 05)。基于此构建的列线图模型显示训练集曲线下面积(AUC)为0. 85, 验证集AUC为0. 83, 校准曲线拟合良好(Hosmer-Lemeshow检验, P=0. 125)。DCA证实模型在阈值概率10%～80%范围内具有临床净获益, IDI为0. 05(P=0. 003)。结论 本研究所构建的列线图模型能有效实现ACH患者住院期间肺炎风险的个体化、可视化预测, 为临床早期识别高危人群并实施针对性干预提供了可靠的量化工具。)