Construction of a Nomogram Prediction Model for the Occurrence of Pulmonary Infection in Patients with Chronic Obstructive Pulmonary Disease Combined with Respiratory Failure

doi:10.12104/j.issn.1674-4748.2025.09.018

Abstract：Objective: To construct a nomogram prediction model for the occurrence of pulmonary infection in patients with chronic obstructive pulmonary disease (COPD) combined with respiratory failure (RF). Methods: A convenience sampling method was used to select 214 patients with COPD combined with RF who were treated in the hospital from March 2021 to March 2023 as the study subjects, and they were divided into a modeling set (150 cases) and a validation set (64 cases). According to the diagnostic criteria for pulmonary infection, whether the patients developed pulmonary infection was determined. The modeling group was divided into an infected group (83 cases) and a non-infected group (67 cases). General patient data were collected and analyzed, and multivariate logistic regression analysis was performed to identify risk factors for pulmonary infection in these patients, and a risk prediction model for pulmonary infection was constructed. Results: C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-α) were independent risk factors for the occurrence of pulmonary infection in patients with COPD combined with RF, while serum albumin levels were protective factors (P < 0.05). In predicting the risk of pulmonary infection in patients with COPD combined with RF, the nomogram model demonstrated excellent performance, with a C-index of 0.918 [95% CI (0.892, 0.944)]. The calibration curve showed a high level of consistency between observed and predicted data; the decision curve indicated that when the risk threshold was set at >0.03, the model could effectively provide standardized net benefit. ROC curve analysis showed that the modeling set had an AUC of 0.923 [95% CI (0.869, 0.960)], indicating a certain predictive value for pulmonary infection in patients with COPD combined with RF (P < 0.001); the validation set had an AUC of 0.920 [95% CI (0.864, 0.958)]. Conclusion: This study constructed a nomogram prediction model for the occurrence of pulmonary infection in patients with COPD combined with RF, and the model demonstrated excellent predictive performance.