冠心病患者经皮冠状动脉介入后的医院感染状况及其风险预测模型构建

Healthcare-associated infection status and construction of a risk prediction model for coronary heart disease patients after percutaneous coronary intervention

摘要:

目的探讨冠心病(CHD)患者经皮冠状动脉介入(PCI)治疗后医院感染状况及其影响因素,并建立感染风险预测模型。方法回顾性选取2019年5月—2023年10月在某院进行PCI治疗的CHD患者作为研究对象,分析CHD患者的感染状况。随后按7 ∶3比例随机分为建模集和测试集。对建模集资料进行单因素和多因素logistc回归分析确定患者医院感染的影响因素,采用R软件构建列线图模型并进行验证。结果共纳入858例CHD患者,分为建模集601例和测试集257例,建模集中感染组41例、非感染组560例。CHD患者PCI治疗后医院感染发病率为6.88%(59/858),感染部位以上呼吸道和泌尿道为主。共分离出病原菌74株,其中革兰阳性菌39株,革兰阴性菌31株,真菌4株。多因素分析结果显示,年龄大、合并糖尿病、心功能(NYHA)分级高、侵入性操作均是CHD患者PCI治疗后医院感染的危险因素(均P＜0.05),微型营养评定简表(MNA-SF)评分高是保护因素(P＜0.05)。基于上述5个指标构建列线图预测模型的受试者工作特征(ROC)曲线下面积(AUC)为0.894(95%CI:0.815~0.931)、灵敏度为89.0%、特异度为82.5%。引入测试集数据验证得出AUC值为0.879(95%CI:0.801~0.923)、灵敏度为87.5%、特异度为81.3%,与建模集效果相当,说明模型稳定。H-L拟合优度检验无统计学意义(P＞0.05),表明该模型无过拟合现象;校准曲线分析表明模型具有较好的一致性;决策曲线分析证实模型在临床具有实用价值。结论列线图模型对CHD患者PCI治疗后医院感染具有良好的预测能力,可为医务人员识别存在医院感染风险的个体提供简便、有效的评估工具。

Abstract:

Objective To evaluate healthcare-associated infection (HAI) status and influencing factors in coronary heart disease (CHD) patients after percutaneous coronary intervention (PCI) treatment, and construct a risk prediction model. Methods CHD patients who underwent PCI in a hospital from May 2019 to October 2023 were retrospectively selected as the research subjects. Infection status of the CHD patients was analyzed. Patients were randomly divided into a modeling set and a testing set in a 7 ∶3 ratio. Univariate and multivariate logistic regression analyses were performed to analyze the data in the modeling set and determine the influencing factors for HAI in patients. R software was used to construct and validate a nomogram model. Results A total of 858 CHD patients were included in the analysis, 601 in the modeling set and 257 in the testing set. In the modeling set, 41 cases were in the infected group and 560 cases in the non-infected group. The incidence of HAI in CHD patients after PCI treatment was 6.88% (59/858). Infection site were mainly upper respiratory tract and urinary tract. A total of 74 pathogens were isolated, including Gram-positive bacteria, Gram-negative bacteria, and fungi being 39, 31, and 4 strains, respectively. Multivariate analysis showed that old age, combined diabetes, high grade of New York Heart Association (NYHA) classification, and invasive procedures were all risk factors for HAI in CHD patients after PCI treatment (all P<0.05), while high mini-nutritional assessment short-form (MNA-SF) score was a protective factor (P<0.05). The area under the receiver operating characteristic (ROC) curve (AUC) of the nomogram prediction model constructed based on the above five indicators was 0.894 (95%CI: 0.815-0.931), with a sensitivity of 89.0% and a specificity of 82.5%. The testing set data validation showed an AUC value of 0.879 (95%CI: 0.801-0.923), with a sensitivity of 87.5% and a specificity of 81.3%, which were comparable to the modeling set and presented the stability of the model. The H-L goodness of fit test showed no statistical significance (P>0.05), indicating that the model didn’t exhibit overfitting. Calibration curve analysis showed that the model had good consistency. Decision curve analysis confirmed that the model had practical value in clinical practice. Conclusion The nomogram model has a good predictive ability for HAI in CHD patients after PCI treatment, and can provide a simple and effective evaluation tool for medical staff to identify HAI high-risk individuals.