摘要
目的建立复杂先天性心脏病(complex congenital heart disease,CCHD)患儿术后死亡风险的列线图预测模型,并验证其预测效能。方法回顾性分析2018年1月至2019年12月重庆医科大学附属儿童医院胸心外科收治并行体外循环下开胸手术的354例CCHD患儿临床资料。运用单因素分析、多因素Logistic回归分析筛选CCHD患儿术后死亡的独立危险因素。根据筛选结果建立列线图预测模型,并评价模型效能。结果经单因素分析和多因素Logistic回归分析发现,正性肌力药物评分升高(OR=1.113,95%CI:1.055~1.175)、白细胞计数增多(OR=1.140,95%CI:1.044~1.245)、血红蛋白降低(OR=0.973,95%CI:0.949~0.998)、白蛋白减少(OR=0.907,95%CI:0.838~0.983)和术前有心肺复苏史(OR=36.656,95%CI:5.187~259.037)是CCHD患儿术后死亡的独立危险因素。以上述5项独立危险因素为输入变量建立列线图预测模型,输出变量为总分,总分值越大,死亡概率越大,反之越小;经计算该模型C-index为0.898,受试者工作特征曲线下面积(area under curve,AUC)为0.899(95%CI:0.821~0.976),校正曲线贴近于理想曲线,说明该列线图模型具有良好的预测效能,其最佳截取值为36分。结论正性肌力药物评分高、白细胞计数增高、血红蛋白降低、白蛋白减少和有术前心肺复苏史是CCHD患儿术后死亡的独立危险因素,根据危险因素建立的列线图模型具有良好的预测效能。当患儿的列线图总分≥36分时应引起高度重视,采取更为积极的诊疗干预措施以改善患儿预后。
Objective To construct a nomogram model for predicting the mortality risk of postoperative in children with complex congenital heart disease(CCHD)and validate its predictive potence.Methods From January 2018 to December 2019,a retrospective review was conducted for 354 CCHD children undergoing cardiopulmonary bypass(CPB)thoracotomy.And independent risk factors for postoperative mortality were screened by univariate and multivariate Logistic regression analyses and constructing a nomogram prediction model.Results The independent postoperative mortality risk factors for CCHD children were as follows:inotropic score increased(IS,OR=1.113,95%CI:1.055-1.175),white blood cell increased(WBC,OR=1.140,95%CI:1.044-1.245),hemoglobin decreased(HB,OR=0.973,95%CI:0.949-0.998),albumin decreased(ALB,OR=0.907,95%CI:0.838-0.983)and cardiopulmonary resuscitation-preoperative(CPR-pre,OR=36.656,95%CI:5.187-259.037).A nomogram prediction model was constructed with the above risk factors as input variables and output variable was total score.The higher total scores,the greater of mortality and vice versa.This model verification showed that C-index and area under curve(AUC)were 0.898 and 0.899 respectively.And calibration curve was close to ideal,indicating decent prediction and discriminant validity.Also the optimal cut-off value was 36.Conclusion IS,WBC,HB,ALB and CPR-pre are independent postoperative mortality risk factors for CCHD children.The nomogram model constructed on the basis of five risk factors offers excellent predictive accuracy.Individuals with a total score≥36 should raise a high alert and more aggressive intervention is required for a better prognosis.
作者
周雅婷
易敏
邱宏翔
郭芸
石洁
吴春
潘征夏
Zhou Yating;Yi Min;Qiu Hongxiang;Guo Yun;Shi Jie;Wu Chun;Pan Zhengxia(Department of Cardiothoracic Surgery,Ministry of Education's Key Laboratory of Children's Development&Disorders,National Clinical Research Center for Children's Health&Disorders,China International Science&Technology Cooperation Base of Children's Development&Critical Disorders,Chongqing Municipal Key Laboratory of Pediatrics,Affiliated Children's Hospital,Chongqing Medical University,Chongqing 400014,China)
出处
《临床小儿外科杂志》
CAS
CSCD
2023年第1期48-55,共8页
Journal of Clinical Pediatric Surgery
基金
重庆市科卫联合医学科研项目(2019MSXM006)
关键词
婴儿
新生
心脏病
死亡率
列线图
统计学
Infant,Newborn
Heart Diseases
Mortality
Nomograms
Statistics
作者简介
通信作者:易敏,Email:744173239@qq.com.
