国际医药卫生导报 ›› 2025, Vol. 31 ›› Issue (6): 1027-1032.DOI: 10.3760/cma.j.cn441417-20241012-06029

• 临床研究 • 上一篇    下一篇

肝硬化患者发生急性肾损伤的危险因素分析及Nomogram预测模型构建

张海梅1  于国英2  胡汉军3   

  1. 1青海省第五人民医院消化内科,西宁 810000;2青海省第四人民医院肝病二科,西宁 810007;3青海省康复医院消化内科,西宁 810000

  • 收稿日期:2024-10-12 出版日期:2025-03-15 发布日期:2025-03-17
  • 通讯作者: 胡汉军,Email: 839008132@qq.com
  • 基金资助:

    青海省卫生健康委指导性计划(2022-wjzdx-73)

Risk factors of acute kidney injury in patients with cirrhosis and establishment of a nomogram prediction model

Zhang Haimei1, Yu Guoying2, Hu Hanjun3   

  1. 1 Department of Gastroenterology, The Fifth People's Hospital of Qinghai Province, Xining 810000, China; 2 Second Liver Disease Department, The Fourth People's Hospital of Qinghai Province, Xining 810007, China; 3 Gastroenterology Department, Qinghai Provincial Rehabilitation Hospital, Xining 810000, China

  • Received:2024-10-12 Online:2025-03-15 Published:2025-03-17
  • Contact: Hu Hanjun, Email: 839008132@qq.com
  • Supported by:

    Guiding Plan of Qinghai Health Commission (2022-wjzdx-73)

摘要:

目的 分析肝硬化患者发生急性肾损伤的危险因素,并构建其Nomogram预测模型。方法 采用回顾性研究,选取2024年1月至5月青海省第五人民医院和青海省第四人民医院收治的135例肝硬化患者为建模集,根据是否发生急性肾损伤分为急性肾损伤组(50例)和非急性肾损伤组(85例)。另取2024年6月至7月45例肝硬化患者为验证集。收集患者临床资料,以单因素分析差异有统计学意义的因素,行多因素logistic回归法分析,得到肝硬化患者发生急性肾损伤的危险因素,构建肝硬化患者发生急性肾损伤的Nomogram风险预警模型。统计学方法采用t检验、χ2检验。结果 单因素结果显示,年龄、高血压、腹水、高尿酸血症(HUA)、感染、糖尿病、Child-Pugh分级均是肝硬化患者发生急性肾损伤的影响因素,差异均有统计学意义(均P<0.05)。多因素结果显示,年龄、HUA、腹水、感染、Child-Pugh分级均是影响肝硬化患者发生急性肾损伤的独立危险因素,差异均有统计学意义(均P<0.05)。受试者操作特征曲线(ROC)显示曲线下面积为0.877[95%置信区间(CI) 0.818~0.936],区分度良好,最大约登指数为0.587,灵敏度为0.740,特异度为0.847。外部验证结果显示,曲线下面积为0.921(95%CI 0.844~0.998),最大约登指数为0.647,灵敏度为0.647,特异度为1.000,校准曲线的理论值和实际值有较好一致性。结论 年龄、HUA、腹水、感染、Child-Pugh分级是影响肝硬化患者发生急性肾损伤的独立危险因素,以此构建预测模型具较高预测价值。

关键词:

肝硬化, 急性肾损伤, Nomogram模型

Abstract:

Objective To analyze the risk factors of acute kidney injury in patients with cirrhosis and establish a nomogram prediction model. Methods A total of 135 patients with cirrhosis treated at the Fifth People's Hospital of Qinghai Province and the Fourth People's Hospital of Qinghai Province from January to May 2024 were selected as the modeling set, and were divided into an acute kidney injury group (50 cases) and a non-acute kidney injury group (85 cases) according to whether they had acute kidney injury. Another 45 patients with cirrhosis from June to July 2024 were selected as the validation set. The patients' clinical data were collected. The multivariate logistic regression analysis was performed on the variates with statistical differences in the univariate analysis to obtain the risk factors of acute kidney injury in the patients, and a nomogram risk warning model for acute kidney injury in the patients was constructed. t and χ2 tests were used for the statistical analysis. Results The univariate analysis results showed that age, hypertension, ascites, hyperuricemia (HUA), infection, diabetes, and Child-Pugh grade were the influencing factors for acute kidney injury in the patients with cirrhosis, with statistical differences (all P<0.05). The results of multiple factors showed that age, HUA, ascites, infection, and Child-Pugh grade were independent risk factors for acute kidney injury in the patients with cirrhosis, with statistical differences (all P<0.05). The area under the curve was 0.877 (95%CI 0.818-0.936); the differentiation was good; the maximum approximate entry index was 0.587; the sensitivity was 0.740; the specificity was 0.847. The external verification results showed that the area under the curve was 0.921 (95%CI 0.844-0.998); the maximum approximate entry index was 0.647; the sensitivity was 0.647; the specificity was 1.000. The theoretical value and the actual value of the calibration curve were in good agreement. Conclusion Age, HUA, ascites, infection, and Child-Pugh grade are independent risk factors for acute kidney injury in cirrhotic patients, and this model has high predictive value.

Key words:

Cirrhosis, Acute kidney injury, Nomogram model