Journal of Capital Medical University ›› 2026, Vol. 47 ›› Issue (4): 783-791.doi: 10.3969/j.issn.1006-7795.2026.04.019

Previous Articles     Next Articles

Predictors of pulmonary infection during hospitalization in patients with chronic heart failure and development of a predictive model

Sun Yao, Hao Jingjing, Liu Yufeng, Guo Hebing, Zhong Xiaoxi, Liu Jingyuan*   

  1. Department of Critical Care Medicine, Beijing Ditan Hospital , Capital Medical University, Beijing 100015,China
  • Received:2026-02-06 Revised:2026-04-14 Online:2026-08-21 Published:2026-06-30
  • Supported by:
    This study was supported by the Capital Health Development Research Special Project (Shoufa 2024-1-1202).

Abstract: ObjectiveTo investigate the risk factors for pulmonary infection (PI) during hospitalization in patients with chronic heart failure (CHF) and to develop a predictive model based on these findings, thereby providing a reference for early identification of high-risk individuals and formulation of personalized intervention strategies. MethodsA retrospective study was conducted enrolling 181 CHF patients admitted from March 2022 to March 2025. Patients were grouped based on whether they developed PI during hospitalization, forming a PI group and a non-PI group. Baseline characteristics and relevant examination data were compared between groups. Variables showing significant differences were subjected to least absolute shrinkage and selection operator (LASSO) regression for variable selection. A multivariate Logistic regression model analyzed factors influencing PI occurrence during hospitalization. The SHAP-interpreted CatBoost model ranked variable importance, and a receiver operating characteristic (ROC) curve risk prediction model was constructed based on the finalized key variables. ResultsAmong 181 CHF patients hospitalized, 51 (28.18%) developed PI. The PI group exhibited statistically significant differences compared to the non-PI group in smoking history, comorbid pulmonary diseases, invasive procedures, heart function classification, and mechanical ventilation (P < 0.05). The PI group exhibited higher white blood cell  (WBC), N-terminal pro-B-type natriuretic peptide (NT-proBNP), C-reactive protein (CRP), and procalcitonin (PCT) levels compared to the non-PI group, while the left ventricular ejection fraction (LVEF) was lower in the PI group (P < 0.05). Seven important influencing factors were obtained by least absolute shrinkage and selection operator(LASSO) regression : pulmonary disease, mechanical ventilation, WBC, NT-proBNP, CRP, PCT, LVEF, and there was no collinearity problem, which were all important variables affecting the occurrence of PI in CHF patients during hospitalization. NT-proBNP (OR=1.044,95 % CI : 1.017-1.071), CRP (OR=1.386,95 % CI : 1.139-1.686), PCT (OR=29.223,95 % CI : 1.512-564.776) and LVEF (OR=0.814,95 % CI : 0.719-0.921) were independent influencing factors of PI in CHF patients during hospitalization (NT-proBNP>CRP > LVEF > PCT). The area under the ROC curve of combined prediction reached 0.937 (95 % CI : 0.894-0.979), which was significantly higher than that of single index (P<0.05). ConclusionThere is a risk of PI in CHF patients during hospitalization. The study suggests that it is related to NT-proBNP, CRP, PCT and LVEF. Based on this, it is feasible to construct a prediction model, which can help identify high-risk patients in the early stage of clinical practice and provide scientific guidance for individualized intervention. 

Key words: chronic heart failure, pulmonary infection, influencing factors, predictive model, left ventricular ejection fraction, mechanical ventilation

CLC Number: