Abstract
Respiratory tract infections in pneumonia patients can rapidly progress to severe pneumonia (SP). However, the clinical relationship between eosinophils (EOS) and SP prognosis remains underexplored. This study aimed to assess this relationship to enhance prevention and treatment strategies for SP. A retrospective analysis was conducted on 325 SP patients from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. The patients were divided into four groups (Q1, Q2, Q3, Q4) based on ascending EOS levels. Baseline characteristics were analyzed to compare clinical indicators. Kaplan-Meier (K-M) survival analysis and weighted multivariate logistic regression models were employed to examine the association between EOS and SP. Additionally, univariate Cox regression analysis identified prognostic factors, and a nomogram based on these factors was validated. Restricted cubic spline (RCS) models were used to explore the relationship between EOS levels and SP. The analysis of patients in the Q1-Q4 groups revealed significant differences in platelet-lymphocyte ratio (PLR) and SP (p < 0.05). K-M survival analysis showed significant differences in in-hospital mortality among the groups (Q1-Q4) (p < 0.05). Higher EOS levels were significantly associated with a reduced risk of SP (odds ratio (OR) = 0.3802, 95% confidence intervals (CI) = 0.1546-0.8928, p = 0.0295) in logistic regression analysis. Prognostic factors, including minimum mean blood pressure (MBP), atrial fibrillation (AF), and EOS group, were identified, with the nomogram demonstrating strong predictive accuracy. RCS models confirmed a linear relationship between EOS and SP. This study highlighted the critical role of EOS in SP prognosis. The nomogram proved to be a highly effective tool for predicting in-hospital mortality in SP patients.