RESUMO
BACKGROUND: Inappropriate empiric antimicrobial treatment (IET) contributes to worsened outcomes. While IET's differential impact across types of nosocomial pneumonia (NP: non-ventilated [nvHABP], ventilated [vHABP] hospital-acquired and ventilator-associated [VABP] bacterial pneumonia) is established, its potential interaction with the bacterial etiology is less clear. METHODS: We conducted a multicenter retrospective cohort study in the Premier Healthcare Database using an administrative algorithm to identify NP. We paired respective pathogens with empiric treatments. Antimicrobial coverage was appropriate if a drug administered within 2 days of infection onset covered the recovered organism(s). All other treatment was IET. RESULTS: Among 17,819 patients with NP, 26.5% had nvHABP, 25.6% vHABP, and 47.9% VABP. Gram-negative (GN) organisms accounted for > 50% of all infections. GN pathogens were ~ 2 × as likely (7.4% vHABP to 10.7% nvHABP) to engender IET than Gram-positive (GP, 2.9% vHABP to 4.9% nvHABP) pathogens. Although rare (5.6% nvHABP to 8.3% VABP), GN + GP infections had the highest rates of IET (6.7% vHABP to 12.9% nvHABP). Carbapenem-resistant GNs were highly likely to receive IET (33.8% nvHABP to 40.2% VABP). Hospital mortality trended higher in the IET group, reaching statistical significance in GN + GP vHABP (47.8% IET vs. 29.3% non-IET, p = 0.016). 30-day readmission was more common with IET (16.0%) than non-IET (12.6%, p = 0.024) in GN VABP. Generally post-infection onset hospital length of stay and costs were higher with IET than non-IET. CONCLUSIONS: IET is ~ 2 × more common in GN than GP infections. Although the magnitude of its impact varies by NP type, IET contributes to worsened clinical and economic outcomes.