Delayed neuropsychiatric sequelae (DNS), which seriously affect the daily lives of patients, are the most common complications of carbon monoxide (CO) poisoning. No uniform screening tool is available for identifying high-risk groups. Therefore, in this study, we aimed to explore whether conventional laboratory indicators and imaging data from primary hospitals could predict the occurrence of DNS.
This retrospective observational study was conducted in a single-center primary hospital from January 1, 2021 to May 31, 2023. Participants included patients aged >18 years with acute CO poisoning. Patients with complete recovery in the acute phase were followed up by telephone and outpatient visits, and the presence of DNS was determined according to the occurrence of new neurological symptoms within 6 weeks after discharge. We obtained demographic, laboratory, and imaging data from the medical records and performed a univariate analysis. A multivariate logistic regression model was used to identify independent clinical predictors of DNS.
A total of 73 patients were included in the study, of whom 25 (34.2%) developed DNS. Multivariate logistic regression analysis revealed that a longer duration of CO exposure (adjusted odds ratio (AOR): 1.262, 95% confidence interval (CI): 1.069–1.490) and the presence of acute brain lesions on diffusion-weighted imaging (DWI) (AOR: 5.117, 95% CI: 1.430–18.315) were independent risk factors for DNS. Receiver operating characteristic analyses of the duration of CO exposure were performed (area under the curve (AUC): 0.825; 95% CI: 0.731–0.918) with a cut-off value of 5.5 h, and DNS was predicted with a sensitivity of 96% and a specificity of 66.7%.
High cranial DWI signal within 24 h and duration of poisoning longer than 5.5 h are independent predictors of DNS. The predictive effects of conventional laboratory indicators require further standardized and large-sample studies.