RESPIRE Score: Derivation and Validation of a New Risk Score for Prediction of Community-acquired Pneumonia Caused by Resistant Pathogens.

Background: Community-acquired pneumonia (CAP) remains a leading cause of infectious mortality worldwide. The increasing prevalence of multidrug-resistant (MDR) pathogens challenges the established empirical antibiotic therapy schemes. This study aimed to derive and validate a clinical score (RESPIRE) to identify CAP-MDR patients. Methods: We conducted an observational study in a setting of medium-high multidrug-resistant (MDR) endemicity. The primary outcome was the derivation of a clinical score predicting MDR-CAP. A retrospective derivation cohort (January 2022-December 2023) including adult patients hospitalized with microbiologically confirmed CAP by culture on respiratory samples was used for score derivation. A prospective cohort (January 2024-December 2025) including adult patients with the same clinical features was used for score validation. Results: The derivation cohort included 275 patients (mean age 68.3 years), with an MDR prevalence of 30.1% (n = 83). The RESPIRE score assigns 1.5 points for recent hospitalization and prior antibiotic use, and 1 point each for enteral feeding, poor functional status, and residence in long-term care facilities (maximum 6 points). A cutoff ≥ 2 showed optimal performance in predicting CAP-MDR (AUROC 0.85; 95% CI 0.80-0.91), outperforming other scores except DRIP. Validation in 141 patients (39% MDR) confirmed an AUROC of 0.88 (95% CI 0.74-0.89) showing better performance compared to other diagnostic models. Conclusions: The RESPIRE score outperformed HCAP criteria and other clinical scores in predicting CAP-MDR. Its implementation in protocols could support antimicrobial stewardship for empiric treatment of cases of CAP presenting at the Emergency Department. A multicenter validation is warranted to confirm the performance of the RESPIRE score. Trial registration number: NCT07425561.