Analysis of pathogen distribution, clinical features and factors associated with lower respiratory tract multidrug-resistant organism isolation in patients with infectious pneumonia

Objective To investigate pathogen distribution, clinical characteristics, and factors associated with lower respiratory tract multidrug-resistant organism (MDRO) isolation in patients with infectious pneumonia, and to develop an exploratory prediction-oriented nomogram. Methods We retrospectively analyzed 117 patients with infectious pneumonia at Tianjin Chest Hospital from September 2024 to September 2025. Bacterial culture was positive in 102 patients. Patients were divided into an MDRO group ( n = 38) and a non-MDRO group ( n = 79; 64 with non-MDRO isolates and 15 culture-negative). Pathogen culture, identification, and susceptibility testing were performed. Clinical features were compared, and univariate and binary logistic regression identified factors associated with MDRO isolation. A prediction-oriented nomogram excluding length of RICU stay was developed and internally validated using ROC and calibration curves and 1000 bootstrap resamples. Results MDRO isolation occurred in 38 patients: 37.25% (38/102) of culture-positive and 32.48% (38/117) of all enrolled patients. After deduplication, 81 MDRO isolates were identified; 75.31% were Gram-negative and 24.69% Gram-positive. The most common MDROs were Klebsiella pneumoniae (57.89%), Pseudomonas aeruginosa (42.11%), and Acinetobacter baumannii (34.21%); Gram-positive MDROs occurred in 9 patients (23.68%), mostly Staphylococcus aureus (21.05%). The MDRO group had longer RICU stays, higher rates of diabetes, hypoproteinemia, invasive mechanical ventilation, and combined antimicrobial therapy, and lower Glasgow Coma Scale (GCS) scores (all P < 0.05). Binary logistic regression identified diabetes (adjusted odds ratio [aOR] = 2.119), invasive mechanical ventilation (aOR=2.849), combined antimicrobial therapy (aOR=2.492), and prolonged RICU stay (aOR=1.137) as associated with MDRO isolation; higher GCS score (aOR=0.674) was associated with lower odds (all P < 0.05). The nomogram showed an apparent AUC of 0.862 (95% CI: 0.793–0.931), sensitivity 82.4%, specificity 87.8%, and an optimism-corrected AUC of 0.831 after 1000 bootstrap resamples. Calibration was acceptable (Brier score=0.142, intercept=−0.02, slope=0.88), with some underestimation at intermediate and overestimation at higher predicted probabilities. Conclusion MDRO isolation is common in infectious pneumonia, predominantly Gram-negative bacteria. Diabetes, invasive mechanical ventilation, combined antimicrobial therapy, and prolonged RICU stay were associated with MDRO isolation, while higher GCS score was associated with lower odds. An exploratory prediction-oriented nomogram incorporating these factors was constructed but requires prospective validation.