1472. Antibiotic De-escalation Compared with Continued Empirical Treatment in Non-Ventilated Hospital-Acquired Pneumonia

BACKGROUND: Antibiotic de-escalation is an important component of antimicrobial stewardship programs. Nosocomial pneumonia is the most common healthcare-associated infection with nonventilated hospital-acquired pneumonia (HAP) comprising the majority of cases. We aimed to compare antibiotic de-escalation with continued empirical treatment in terms of clinical outcomes in nonventilated HAP. METHODS: A retrospective cohort study was conducted including patients meeting the American Thoracic Society criteria for HAP. This compared de-escalated HAP patients to those continued on empirical treatment across three hospitals in West London over 3 months. The primary outcome was the length of stay (LOS), and secondary outcomes were duration of treatment and cost of hospital stay. Effects were adjusted for confounders using multivariate linear regression models. RESULTS: Eighty patients with HAP were identified. Overall, 22/80 (27.5%) had therapy de-escalated and 47/80 (58.8%) continued empirical treatment. A total of 58 patients survived and were included in the analysis, 20 in de-escalation and 38 in continued empirical treatment. Length of stay was shorter in de-escalation by −7.2 (95% CI −12.2, −3.0) days, P < 0.01, with an adjusted difference of −3.2 (95% CI −8.3, 1.9) days, P = 0.21. The duration of treatment was shorter in de-escalation by −3.4 (95% CI −5.8, −0.9) days, P < 0.01, with an adjusted difference of −2.6 (95% CI −5.2, 0.1) days, P = 0.06. The cost of hospital stay was lower in de-escalation by £-2, 907.37 (95% CI −4,865.31, −949.43), P < 0.01, with an adjusted difference of £-1,290.00 (95% CI −3,320.75, 740.74), P = 0.21. CONCLUSION: In HAP, 27.5% of patients were de-escalated. There was no difference in LOS, duration of treatment, and cost of hospital stay between de-escalation and continued empirical treatment on adjustment for confounders. Future work should explore the relationship between de-escalation and antimicrobial resistance in HAP. DISCLOSURES: All authors: No reported disclosures.