HUANG Wenbin 1 , LIU Wei 1 , XIA Mengmei 1 , WANG Yuzhe 1 , XU Ping 1,2
  • 1. Department of Emergency, Zigong Fourth People’s Hospital, Zigong, Sichuan 643000, P. R. China;
  • 2. Institute of Medical Big Data, Zigong Medical Big Data and Artificial Intelligence Research Institute, Zigong, Sichuan 643000, P. R. China;
XU Ping, Email: 58124453@qq.com
Carbon monoxide poisoning; delayed encephalopathy; nomogram; Glasgow Coma Scale
10.7507/1002-0179.202310019
Export PDF Favorites Scan Get Citation

Abstract Full text Figures/Tables References Cited by

Objective  To construct a nomogram model for predicting delayed encephalopathy after acute carbon monoxide poisoning (DEACMP) in emergency departments. Methods  All patients with acute carbon monoxide poisoning who visited the Department of Emergency of Zigong Fourth People’s Hospital between June 1st, 2011 and May 31st, 2023 were retrospectively enrolled and randomly divided into a training set and a testing set in a 6∶4 ratio. LASSO regression was used to screen variables in the training set to establish a nomogram model for predicting DEACMP. The discrimination, calibration, and clinical practicality were compared between the nomogram and Glasgow Coma Scale (GCS) in the training and testing sets. Results  A total of 475 patients with acute carbon monoxide poisoning were included, of whom 41 patients had DEACMP. Age, GCS and aspartate aminotransferase were selected as risk factors through LASSO regression, and a nomogram model was constructed based on these factors. The areas under the receiver operating characteristic curves for nomogram and GCS to predict DEACMP in the training set were 0.897 [95% confidence interval (CI) (0.829, 0.966)] and 0.877 [95%CI (0.797, 0.957)], respectively; and those for nomogram and GCS to predict DEACMP in the testing set were 0.925 [95%CI (0.865, 0.985)] and 0.858 [95%CI (0.752, 0.965)], respectively. Compared with GCS, the performance of nomogram in the training set (net reclassification index=0.495, P=0.014; integrated discrimination improvement=0.070, P=0.011) and testing set (net reclassification index=0.721, P=0.004; integrated discrimination improvement=0.138, P=0.009) were both positively improved. The calibration of nomogram in the training set and testing set was higher than that of GCS. The decision curves in the training set and testing set showed that the nomogram had better clinical net benefits than GCS. Conclusion  The age, GCS and aspartate aminotransferase are risk factors for DEACMP, and the nomogram model established based on these factors has better discrimination, calibration, and clinical practicality compared to GCS.

Citation： HUANG Wenbin, LIU Wei, XIA Mengmei, WANG Yuzhe, XU Ping. Construction of a nomogram prediction model for delayed encephalopathy after acute carbon monoxide poisoning. West China Medical Journal, 2023, 38(11): 1648-1654. doi: 10.7507/1002-0179.202310019 Copy

  • Previous Article

    Interpretation of the updates in the pediatric and neonatal life support in 2022 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science with Treatment Recommendations

  • Next Article

    Construction of nosocomial pulmonary infection prediction model for acute pesticide poisoning