OBJECTIVES: To examine post-pandemic population, treatment, and outcomes of severe community-acquired bacterial pneumonia (sCABP) as identified via two common definitions, and to compare these two groups. DESIGN: A retrospective quasi-cohort study within Premier Healthcare Database, 2021-2024. SETTING: 849 U.S. acute care hospitals (253 reporting microbiology data). PATIENTS: Adults hospitalized with sCABP, as defined by the need for ICU (ICU-sCABP) or by the American Thoracic Society (ATS) major criteria. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Among 67,439 patients who met either definition of sCABP (24.5% of all CAPB admissions), 51.2% met ICU-sCABP and 78.0% ATS-sCABP definitions. The groups were similar with respect to age, gender, and race, and the burden of chronic illness. While similar proportions in both groups required vasopressors (17.0% ICU-sCABP vs. 15.8% ATS-sCABP), fewer in the ICU-sCABP than the ATS-sCABP received mechanical ventilation (MV, 51.9% vs. 91.0%). Noninvasive positive pressure ventilation predominated in both (27.9% ICU-sCABP vs. 65.6% ATS-sCABP), while invasive MV (IMV) was used in under ¼ of all patients in either group. Among patients with a known pathogen (n = 851 ICU-sCABP and n = 959 ATS-sCABP), the distribution of organisms was similar. Staphylococcus aureus (30.2% ICU-sCABP and 31.7% ATS-sCABP) and Pseudomonas aeruginosa (24.5% ICU-sCABP vs. 27.0% ATS-sCABP) were most common. Streptococcus pneumoniae caused 11.0% of ICU-sCABP and 9.8% of ATS-sCABP. Antibiotics classes used were similar as well. Hospital mortality (15.6% ICU-sCABP vs. 15.0% ATS-sCABP), 30-day readmission (14.8% ICU-sCABP vs. 15.7% ATS-sCABP), median hospital length of stay (LOS, 7 d for each), median ICU LOS (3 d for each), and median MV duration (3 d for each) were also similar. CONCLUSIONS: The two commonly used sCABP definitions identify a similar population of patients, which is large and resource-intensive. MV practices have evolved since the pandemic, with much less IMV used. Despite these changes, mortality risk from sCABP remains unacceptably high.
Zilberberg et al. (2026) studied this question.