Prediction Score of Pseudomonas Aeruginosa Isolation in Acute Exacerbations of Chronic Obstructive Pulmonary Disease

Background: Pseudomonas aeruginosa ( PA ) is an important pathogen in acute exacerbations of chronic obstructive pulmonary disease (AECOPD), yet empirical anti- PA coverage is often prescribed inappropriately. Accurate prediction tools are needed to guide targeted therapy. Methods: We conducted a retrospective cohort study of hospitalized AECOPD patients at a regional referral hospital in Thailand (2015– 2018). Clinical, radiographic, and laboratory data were analyzed to identify independent predictors of PA isolation. Multivariable logistic regression was used to develop a prediction score, which was internally validated using bootstrap resampling. Model performance was assessed by the area under the receiver operating characteristic curve (AuROC). Results: Among 1,348 admissions (771 patients), P. aeruginosa was isolated in 167 cases (12.4%). A simplified prediction score was developed based on four independent predictors: home nebulizer use, prior anti- PA antibiotic exposure within 1 year, radiographic consolidation, and elevated serum bicarbonate. The prediction score achieved an AuROC of 0.66 (95% CI 0.62– 0.70) with good calibration after internal validation. Risk groups were defined as low (0– 1 points, < 12% risk), intermediate (2– 3 points, 16– 23% risk), and high (4– 6 points, ≥ 32% risk). Conclusion: This simplified prediction score, developed in hospitalized AECOPD patients, may aid clinicians in guiding empirical antibiotic selection by reducing unnecessary anti- PA coverage in low-risk cases while ensuring timely treatment for those at high risk. Keywords: AECOPD, acute exacerbations, chronic obstructive pulmonary disease, pseudomonas aeruginosa , predictive model, prediction score