Construction of a clinical mortality risk prediction model for extracorporeal membrane oxygenation based on nomogram
10.3760/cma.j.issn.1671-0282.2023.10.011
- VernacularTitle:基于列线图构建体外膜肺氧合临床死亡风险预测模型
- Author:
Jianchao LI
1
;
Xiaoliang QIAN
;
Jiaxin HUANG
;
Fanwei MENG
;
Leiyi YANG
;
Junjie SUN
;
Junlong HU
;
Zhaoyun CHENG
Author Information
1. 河南省人民医院 郑州大学人民医院 阜外华中心血管病医院体外循环科，郑州 451450
- Keywords:
Extracorporeal membrane oxygenation;
Risk of death;
Nomogram;
Predictive model
- From:
Chinese Journal of Emergency Medicine
2023;32(10):1353-1360
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the risk factors of death in patients receiving ECMO treatment and to construct a nomogram prediction model.Methods:The clinical data of 412 consecutive patients with acute heart and (or) pulmonary failure who received ECMO treatment between April 2018 and June 2022 were retrospectively included.According to the patients' in-hospital survival, univariate correlation analysis was used to select risk factor variables, and then Lasso regression was used to screen all variables, combined with common variables, combined with clinical practice, plotted a nomogram to predict the probability of early mortality, using the area under the ROC curve (AUC), Harrell C index and calibration curve were used to evaluate and internally validate the performance of the model.Decision curve analysis was applied to assess its clinical utility.Results:Cerebral infarction, diabetes, history of cardiopulmonary resuscitation, neurological complications, acute kidney injury, lactate, hemoglobin, albumin, and platelet count were risk factors for death in patients receiving ECMO ( P<0.05).At the same time, according to the actual situation and difference variables, we constructed a nomogram with high reliability to predict the probability of death. Conclusions:The study identified the risk factors of death in patients receiving ECMO, successfully constructed and validated a nomogram prediction model, and provided a simple and reliable tool for ECMO death prediction, which is of great significance for individualized treatment of patients.
