医学
布里氏评分
队列
逻辑回归
心内膜炎
内科学
感染性心内膜炎
队列研究
外科
回顾性队列研究
统计
数学
作者
Junjie Wang,Jian Hou,Kangni Feng,Huawei Wu,Quan Liu,Zhuoming Zhou,Huayang Li,Luo Lu,Guangguo Fu,Liqun Shang,Guangxian Chen,Sheng Huang,Zhongkai Wu
标识
DOI:10.1016/j.ijcard.2023.131432
摘要
Objectives Bleeding complications are one of the most serious postoperative complications after cardiac surgery and are associated with high mortality, especially in patients with infective endocarditis (IE). Our objectives were to identify the risk factors and develop a prediction model for postoperative bleeding complications in IE patients. Methods The clinical data of IE patients treated from October 2013 to January 2022 were reviewed. Multivariate logistic regression analysis was used to evaluate independent risk factors for postoperative bleeding complications and develop a prediction model accordingly. The prediction model was verified in a temporal validation cohort. The performance of the model was evaluated in terms of its discrimination power, calibration, precision, and clinical utility. Results A total of 423 consecutive patients with IE who underwent surgery were included in the final analysis, including 315 and 108 patients in the training cohort and validation cohort, respectively. Four variables were selected for developing a prediction model, including platelet counts, systolic blood pressure, heart failure and vegetations on the mitral and aortic valves. In the training cohort, the model exhibited excellent discrimination power (AUC = 0.883), calibration (Hosmer–Lemeshow test, P = 0.803), and precision (Brier score = 0.037). In addition, the model also demonstrated good discrimination power (AUC = 0.805), calibration (Hosmer–Lemeshow test, P = 0.413), and precision (Brier score = 0.067) in the validation cohort. Conclusions We developed and validated a promising risk model with good discrimination power, calibration, and precision for predicting postoperative bleeding complications in IE patients.
科研通智能强力驱动
Strongly Powered by AbleSci AI