Analysis of risk factors for multidrug-resistant bacterial infections in intensive care units and application of a nomogram prediction model.

内容摘要

Multidrug-resistant (MDR) bacterial infections remain a major challenge in intensive care units, leading to prolonged hospitalization and increased mortality. Although several clinical factors have been associated with MDR infections, there is still a lack of practical and individualized predictive tools to facilitate early risk stratification in intensive care unit (ICU) patients. This retrospective single-center study included adult patients admitted to the ICU of Leshan People's Hospital between January 2020 and December 2025 who stayed for at least 48 hours and had complete microbiological records. The primary outcome was the occurrence of MDR bacterial infection, defined as resistance to at least 3 classes of antibiotics confirmed by microbiological culture and susceptibility testing. Patients were randomly divided into a training cohort (70%) and a validation cohort (30%). Univariable and multivariable logistic regression analyses were performed to identify independent risk factors. A nomogram prediction model was constructed based on significant predictors. Model performance was evaluated using receiver operating characteristic curves, area under the curve (AUC), calibration curves, the Hosmer-Lemeshow goodness-of-fit test, decision curve analysis, and internal validation using Bootstrap resampling combined with 10-fold cross-validation. A total of 876 ICU patients were included (mean age: 68.4 ± 14.7 years; 58.2% male), among whom 161 (18.38%) developed MDR bacterial infections. Multivariable logistic regression identified age > 75 years (OR: 3.350, 95% CI: 2.916-3.792), endotracheal intubation (OR: 2.079, 95% CI: 1.740-3.130), ICU stay > 1 week (OR: 3.428, 95% CI: 2.553-4.417), coma (OR: 2.735, 95% CI: 2.000-4.469), and central venous catheterization (OR: 2.438, 95% CI: 1.464-4.295) as independent risk factors. The nomogram demonstrated good discriminative ability, with an AUC of 0.827 in the training cohort and 0.812 in the validation cohort. Internal validation yielded an AUC of 0.820, with a sensitivity of 83.2% and a specificity of 82.8%. Advanced age, invasive procedures, prolonged ICU stay, and impaired consciousness are associated with an increased risk of MDR bacterial infections in ICU patients. The developed nomogram provides a practical tool for individualized risk prediction and may assist clinicians in early identification and targeted preventive strategies. However, due to the retrospective nature of the study, these findings suggest associations but do not establish causality.