【摘要】 目的 分析急性感染性心内膜炎(IE)患者的临床特征，构建复杂IE的风险预测模型。方法 2020年1月—2023年6月郑州大学第一附属医院诊治急性IE患者400例，均给予经验性广谱抗菌药物治疗，后续根据病原菌及药物敏感试验结果调整用药，同时给予支持治疗，符合手术指征者行手术治疗，记录患者住院期间复杂IE发生情况。400例中2020年1月—2022年12月诊治300例为训练集，2023年1—6月诊治100例为验证集。记录患者心脏基础病史、合并症、入院48 h内实验室检查结果、发热持续时间等临床资料；采用单因素及多因素logistic回归分析训练集患者进展为复杂IE的影响因素，构建急性IE进展为复杂IE的风险预测模型；绘制ROC曲线，评估风险模型在2组预测进展为复杂IE的效能；采用Hosmer-Lemeshow检验构建的风险预测模型的拟合度。结果 (1)训练集年龄、性别比例、体质量指数、心脏基础疾病史占比、合并症占比、C反应蛋白(CRP)、CRP≥100 mg/L占比、红细胞沉降率、血红蛋白、血红蛋白<110 g/L占比、白细胞计数、血培养阳性占比、发热持续时间、发热持续时间≥7 d占比与验证集比较差异均无统计学意义(P>0.05)。(2)训练集患者进展为复杂IE 85例，验证集进展为复杂IE 28例。年龄≥60岁(OR=3.214,95%CI:1.872～5.501,P=0.030)、有心脏基础疾病史(OR=2.563,95%CI:1.458～4.512,P=0.030)、CRP≥100 mg/L(OR=3.506,95%CI:2.011～6.102,P=0.040)、血培养阳性(OR=2.895,95%CI:1.623～5.164,P=0.030)是急性IE患者进展为复杂IE的危险因素。(3)基于年龄、心脏基础疾病史、CRP、血培养结果构建的风险预测模型在训练集和验证集预测急性IE进展为复杂IE的AUC分别为0.944(95%CI:0.918～0.969,P<0.001)和0.899(95%CI:0.866～0.932,P<0.001),灵敏度分别为80.0%、82.1%,特异度分别为88.4%、83.3%。Hosmer-Lemeshow检验显示，模型拟合度良好(χ~2=4.12,P=0.846)。结论 根据急性IE患者年龄、血培养结果、CRP、心脏基础疾病史构建的风险预测模型对早期预测复杂IE具有较高价值。更多还原

【Abstract】 Objective To analyze the clinical characteristics of patients with acute infective endocarditis(IE) and to construct a risk prediction model for complex IE.Methods From January 2020 to June 2023,400 patients with acute IE were diagnosed and treated in the First Affiliated Hospital of Zhengzhou University.All of them were given empirical broad-spectrum antibacterial drug treatment.Subsequently,the medication was adjusted based on the results of pathogenic bacteria and drug sensitivity tests,and the supportive treatment was provided.Those who met the surgical indications underwent surgical treatment.The occurrence of complex IE was recorded during hospitalization.Among these 400 patients,300 from January 2020 to December 2022 were assigned to the training set,while 100 from January to June 2023 were designated as the validation set.The clinical data such as the history of underlying heart diseases,comorbidities,laboratory test results within 48 h of admission and duration of fever were recoreded.Univariate and multivariate logistic regression analyses were used to analyze the factors influencing the progression to complex IE in the training set,and a risk prediction model for the progression of acute IE to complex IE was constructed.The ROC curves were plotted to evaluate the performance of the constructed risk model to predict the progression to complex IE in two groups.The fitting degree of the risk prediction model was tested by Hosmer-Lemeshow.Results(1) The age,gender ratio,body mass index,proportions of history of underlying heart diseases and comorbidities,C-reactive protein(CRP),proportion of CRP ≥ 100 mg/L,erythrocyte sedimentation rate,hemoglobin,proportion of hemoglobin <110 g/L,white blood cell count,proportion of positive blood culture,duration of fever,and proportion of fever duration≥7d showed no statistically significant differences between two sets(P>0.05).(2) Complex IE developed in 85 patients in the training set,and in 28 in the validation set.The age of≥60 years old(OR=3.214,95%CI:1.872-5.501,P=0.030),history of underlying heart diseases(OR=2.563,95%CI:1.458-4.512,P=0.030),CRP≥100 mg/L(OR=3.506,95%CI:2.011-6.102,P=0.040) and positive blood culture(OR=2.895,95%CI:1.623-5.164,P=0.030)were the risk factors of the progression of acute IE to complex IE.(3) The AUCs of the risk prediction model constructed based on age,history of underlying heart diseases,CRP,and blood culture results for predicting progression to complex IE in the training set and validation set were 0.944(95%CI:0.918-0.969,P<0.001) and 0.899(95%CI:0.86 6-0.932,P<0.001),the sensitivities were 80.0 % and 82.1%,and the specificities were 88.4% and 83.3%,respectively.The Hosmer-Lemeshow test showed that the model had a good fit(χ~2=4.12,P=0.846).Conclusion The risk prediction model constructed based on the age,blood culture results,CRP,and history of underlying heart diseases has a high value for the early prediction of complex IE in patients with acute IE.更多还原