This study aimed to develop and validate a predictive model to determine the risk of in-hospital mortality in patients with acute paraquat poisoning. This retrospective observational cohort study included 724 patients with acute paraquat poisoning whose clinical data were collected within 24 h of admission. The primary outcome was in-hospital mortality. Patients were randomly divided into training and validation cohorts (7/3 ratio). In the training cohort, the least absolute shrinkage and selection operator regression models were used for data dimension reduction and feature selection. Multivariate logistic regression was used to generate a predictive nomogram for in-hospital mortality. The prediction model was assessed for both the training and validation cohorts. In the training cohort, decreased level of consciousness (Glasgow Coma Scale score < 15), neutrophil-to-lymphocyte ratio, alanine aminotransferase, creatinine, carbon dioxide combining power, and paraquat plasma concentrations at admission were identified as independent predictors of in-hospital mortality in patients with acute paraquat poisoning. The calibration curves, decision curve analysis, and clinical impact curves indicated that the model had a good predictive performance. It can be used on admission to the emergency department to predict mortality and facilitate early risk stratification and actionable measures in clinical practice after further external validation.
Read full abstract