| 程金海,许吉哲.高血压危象急诊救治期间死亡的风险预测列线图模型构建研究[J].安徽医药,2025,29(10):2057-2061. |
| 高血压危象急诊救治期间死亡的风险预测列线图模型构建研究 |
| Study on the establishment of a nomogram model to predict the risk of death during emergency treatment of hypertensive crisis |
| |
| DOI：10.3969/j.issn.1009-6469.2025.10.029 |
| 中文关键词: 高血压危象 颅内出血，高血压性 死亡 危险因素 预测 列线图模型 |
| 英文关键词: Hypertensive crisis Intracranial hemorrhage, hypertensive Death Risk factors Forecasting Nomogram model |
| 基金项目: |
|
| 摘要点击次数: 44 |
| 全文下载次数: 41 |
| 中文摘要: |
| 目的构建高血压危象急诊救治期间死亡的风险预测列线图模型。方法回顾性分析 2019年 1月至 2022年 12月南阳张仲景医院收治的 378例高血压危象病人的临床资料，按 2∶1的比例将病人按随机数字表法分为建模组( 252例)和验证组(126例)，并根据病人救治期间是否死亡将建模组分为死亡组( 34例)和存活组( 218例)。采用多因素 logistic回归分析高血压危象急诊救治期间死亡的危险因素，建立风险预测列线图模型，并绘制受试者操作特征曲线( ROC曲线)、校准曲线和决策性曲线( DCA)评估列线图模型的预测效能、校准度和临床效益，并使用验证组评估列线图的可行性。结果 378例病人中死亡 52例，病死率为 13.76%；死亡组转运时间［(1.54±0.34)h］、合并糖尿病［47.06%(16/34)］、入院 6h内血压控制不佳［61.76%(21/ 34)］及并发颅内出血［35.29%(12/34)］、蛛网膜下腔出血［11.76%(4/34)］、急性脑梗死［8.82%(3/34)］、急性主动脉夹层［14.71%(5/34)］、急性心力衰竭［20.59%(7/34)］、急性肾功能衰竭［11.76%(4/34)］占比均高于存活组［( 1.42±0.32)h、29.36%(64/218)、 25.69%(56/218)、 12.39%(27/218)、 0.92%(2/218)、 1.38%(3/218)、 0%(0/218)、 6.88%(15/218)、 1.38%(3/218)］(P<0.05)；多因素 logistic回归分析结果显示死亡组入院 6h内血压控制不佳、颅内出血、蛛网膜下腔出血、急性主动脉夹层、急性心力衰竭、急性肾功能衰竭均是高血压危象急诊救治期间死亡的危险因素(P<0.05)； ROC曲线分析结果显示，建模组列线图预测模型的曲线下面积( AUC)为 0.86［95%CI:(0.81，0.90)］，验证组列线图预测模型 AUC为 0.85［95%CI:(0.78，0.91)］；建模组和验证组模型校准曲线的一致性指数分别为 0.80、0.79，两组校准曲线均贴近标准曲线； DCA曲线显示，建模组阈概率为 0~0.90时、验证组阈概率为 0~0.85时模型均具有良好的临床获益。结论基于入院 6h内血压控制不佳、颅内出血、蛛网膜下腔出血、急性主动脉夹层、急性心力衰竭、急性肾功能衰竭构建高血压危象急诊救治期间死亡的列线图模型有助于临床早期识别死亡高风险病人，指导干预。 |
| 英文摘要: |
| Objective To construct a Nomogram model for predicting the risk of death during emergency treatment of hypertensive crises. Methods The clinical data of 378 hypertensive crisis patients admitted to Nanyang Zhang Zhongjing Hospital from January2019 to December 2022 were retrospectively analyzed, and the patients were randomly divided into the modeling group (252 cases) andthe validation group (126 cases) in a 2∶1 ratio by a random digital table. The modeling group was divided into the death group (34 cas-es) and the survival group (218 cases) based on whether the patients died during treatment. Multivariate Logistic regression was used toanalyze the risk factors of death during emergency treatment of hypertensive crisis, and the risk prediction Nomogram model was estab-lished. Receiver operating characteristic curves (ROC curves), calibration curves, and decision making curves (DCA) were drawn toevaluate the predictive efficacy, accuracy, and clinical benefits of the Nomogram model, and the validation group was used to evaluatethe feasibility of the Nomogram.Results 52 of the 378 patients died, with a mortality rate of 13.76%. The transport time [(1.54±0.34)h] and the proportion of patients with diabetes [47.06% (16/34)], poor blood pressure control within 6 hours of admission [61.76% (21/34)], intracranial hemorrhage [35.29% (12/34)], subarachnoid hemorrhage [11.76% (4/34)], acute cerebral infarction [8.82% (3/34)],acute aortic dissection [14.71% (5/34)], and acute heart failure [20.59% (7/34)] and acute renal failure [11.76% (4/34)] in the deathgroup were higher than those in the survival group [(1.42±0.32) h, 29.36% (64/218), 25.69% (56/218), 12.39% (27/218), 0.92% (2/218), 1.38% (3/218), 0% (0/218), 6.88% (15/218), 1.38% (3/218)] (P < 0.05). Multivariate Logistic regression analysis showed that poorblood pressure control within 6 hours of admission, intracranial hemorrhage, subarachnoid hemorrhage, acute aortic dissection, acute heart failure, and acute renal failure were all risk factors for death during emergency treatment of hypertensive crises in the death group(P < 0.05). The ROC curve analysis results showed that the area under the curve (AUC) of Nomogram prediction model in the modelinggroup was 0.86 [95%CI: (0.81, 0.90)], while that in the validation group was 0.85 [95%CI: (0.78, 0.91)]. The consistency indices of thecalibration curves in the modeling and validation groups were 0.80 and 0.79, respectively, and both calibration curves were close to thestandard curve. The DCA curve showed that the model had good clinical benefits when the threshold probability of the modeling groupwas 0-0.90 and the threshold probability of the validation group was 0-0.85.Conclusion Based on poor blood pressure control within6 hours of admission, intracranial hemorrhage, subarachnoid hemorrhage, acute aortic dissection, and acute heart failure and acute re-nal failure, constructing a Nomogram model for death during emergency treatment of hypertensive crisis can help identify high-risk pa-tients for death in clinical early stages and guide interventions. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
