International Medicine and Health Guidance News ›› 2024, Vol. 30 ›› Issue (22): 3779-3785.DOI: 10.3760/cma.j.issn.1007-1245.2024.22.018

• Treatises • Previous Articles     Next Articles

Establishment and validation of a prediction model for contrast-induced nephropathy in elderly patients with coronary heart disease undergoing PCI

Yang Tianfen1, Li Lei1, Zhang Nana2   

  1. 1 Department of Cardiovascular Medicine, The First People's Hospital of Xianyang, Xianyang 712000, China; 2 Department of Cardiovascular Medicine, People's Hospital of Tongchuan, Tongchuan 727031, China

  • Received:2024-06-27 Online:2024-11-15 Published:2024-11-20
  • Contact: Zhang Nana, Email: 341698049@qq.com
  • Supported by:

    Key Research and Development Plan of Shaanxi Province (2021SF-329)

老年冠心病患者PCI治疗相关造影剂肾病的预测模型建立及验证

杨天芬1  李磊1  张娜娜2   

  1. 1咸阳市第一人民医院心血管内科,咸阳 712000;2铜川市人民医院心血管内科,铜川 727031

  • 通讯作者: 张娜娜,Email:341698049@qq.com
  • 基金资助:

    陕西省重点研发计划(2021SF-329)

Abstract:

Objective To explore the influencing factors of contrast-induced nephropathy (CIN) in elderly patients with coronary heart disease undergoing percutaneous coronary intervention (PCI), and construct and validate a prediction model. Methods The medical records of 192 elderly patients with coronary heart disease who underwent PCI in the First People's Hospital of Xianyang from January 2020 to January 2023 were retrospectively analyzed, including 128 males and 64 females, aged 60-80 (71.73±3.59) years. According to the 80/20 law, they were randomly divided into a training set (153 cases) and a validation set (39 cases). According to the changes of serum creatinine (Scr) within 72 h after PCI, the patients were divided into a CIN group and a non-CIN group. The risk factors for CIN related to PCI in elderly patients with coronary heart disease were screened, and a risk prediction model was constructed and validated. χ2 test and t test were used; logistic regression model was used to analyze the influencing factors. Results In the training set, 22 cases (14.38%) of 153 elderly patients with coronary heart disease treated by PCI developed PCI-related CIN. In the validation set, 5 cases (12.82%) of 39 elderly patients with coronary heart disease treated by PCI developed PCI-related CIN. In the CIN group, the proportion of use of diuretics 24 h before PCI, high-sensitivity C-reactive protein (hs-CRP) 24 h after surgery, N-terminal pro B-type natriuretic peptide (NT-proBNP) 24 h after surgery, serum creatinine (Scr) 24 h after surgery, and urinary neutrophil gelatinase-associated lipocalin (NGAL) 24 h after surgery were higher than those in the non-CIN group (all P<0.05), and the estimated glomerular filtration rate (eGFR) was lower than that in the non-CIN group (P<0.05). Use of diuretics 24 h before surgery (OR=3.216, 95%CI 1.472-7.025), hs-CRP 24 h after surgery (OR=2.971, 95%CI 1.360-6.491), NT-proBNP 24 h after surgery (OR=3.341, 95%CI 1.528-7.297), serum Scr 24 h after surgery (OR=3.271, 95%CI 1.497-7.145), and urinary NGAL 24 h after surgery (OR=3.582, 95%CI 1.639-7.826) were risk factors for CIN in elderly patients with coronary heart disease undergoing PCI (all P<0.05), and eGFR (OR=0.253, 95%CI 0.115-0.553) was a protective factor for CIN in elderly patients with coronary heart disease undergoing PCI (P<0.05). The sensitivity, specificity, and area under the curve of the nomogram model for predicting CIN after PCI in elderly patients with coronary heart disease in the training set were 0.815 (95%CI 0.702-0.891), 0.743 (95%CI 0.674-0.856), and 0.834 (95%CI 0.718-0.915). The sensitivity, specificity, and area under the curve of the nomogram model for predicting CIN after PCI in elderly patients with coronary heart disease in the validation set were 0.754 (95%CI 0.662-0.875), 0.772 (95%CI 0.681-0.889), and 0.807 (95%CI 0.743-0.915). Conclusion The use of diuretics 24 h before PCI, hs-CRP, NT-proBNP, Scr, and urinary NGAL 24 h after PCI, as well as eGFR, are associated with the occurrence of CIN in elderly patients undergoing PCI for coronary heart disease. Developing a risk prediction model helps to forecast the likelihood of CIN in these patients.

Key words:

Coronary heart disease, Old age, Percutaneous coronary intervention, Contrast-induced nephropathy, Prediction model

摘要:

目的 探讨老年冠心病患者经皮冠状动脉介入术(PCI)治疗相关造影剂肾病(CIN)的影响因素,构建并验证预测模型。方法 回顾性分析2020年1月至2023年1月在咸阳市第一人民医院接受PCI治疗的192例老年冠心病患者病历资料,其中男128例,女64例,年龄60~80(71.73±3.59)岁。依据80/20定律随机分为训练集(153例)和验证集(39例)。根据患者PCI术后72 h内血肌酐(Scr)变化分为CIN组和非CIN组,筛查老年冠心病患者PCI治疗相关CIN发病的危险因素,构建并验证风险预测模型。采用χ2检验、t检验;影响因素分析采用logistic回归模型。结果 训练集153例老年冠心病患者PCI治疗后发生CIN 22例(14.38%)。验证集39例老年冠心病患者PCI治疗后发生CIN 5例(12.82%)。CIN组PCI术前24 h使用利尿剂占比、术后24 h超敏C反应蛋白(hs-CRP)、术后24 h N末端B型利钠肽前体(NT-proBNP)、术后24 h Scr、术后24 h尿中性粒细胞明胶酶相关脂质转载蛋白(NGAL)均高于非CIN组(均P<0.05),估算肾小球滤过率(eGFR)低于非CIN组(P<0.05)。术前24 h使用利尿剂(OR=3.216,95%CI 1.472~7.025)及术后24 h hs-CRP(OR=2.971,95%CI 1.360~6.491)、NT-proBNP(OR=3.341,95%CI 1.528~7.297)、Scr(OR=3.271,95%CI 1.497~7.145)、尿NGAL(OR=3.582,95%CI 1.639~7.826)均是老年冠心病患者PCI治疗发生CIN的危险因素(均P<0.05),eGFR(OR=0.253,95%CI 0.115~0.553)是老年冠心病患者PCI治疗发生CIN的保护因素(P<0.05)。列线图模型预测训练集老年冠心病患者PCI治疗发生CIN的灵敏度为0.815(95%CI 0.702~0.891),特异度为0.743(95%CI 0.674~0.856),曲线下面积为0.834(95%CI 0.718~0.915)。列线图模型预测验证集老年冠心病患者PCI治疗发生CIN的灵敏度为0.754(95%CI 0.662~0.875),特异度为0.772(95%CI 0.681~0.889),曲线下面积为0.807(95%CI 0.743~0.915)。结论 术前24 h使用利尿剂、eGFR以及术后24 h hs-CRP、NT-proBNP、Scr、尿NGAL均与老年冠心病患者PCI治疗发生CIN有关。构建风险预测模型有助于预测老年冠心病患者PCI治疗相关CIN。

关键词:

冠心病, 老年, 经皮冠状动脉介入治疗, 造影剂肾病, 预测模型