Target Trial Emulation of Empiric Antibiotics on Clinical Outcomes in Moderately Immunocompromised Patients Hospitalized With Pneumonia

BACKGROUND: Immunocompromised patients are often excluded from pneumonia trials, guidelines, and stewardship interventions. The objective of this study was to evaluate whether empiric broad-spectrum antibiotic (BSA) treatment impacts mortality and other clinical outcomes in moderately immunocompromised patients without risk factors for multidrug-resistant organisms (MDROs) hospitalized with community-acquired pneumonia. METHODS: This was a target trial emulation including moderately immunocompromised (asplenia, hematologic malignancies, solid organ malignancy receiving chemotherapy, kidney transplant >1 year prior, congenital/acquired immunodeficiency, and receiving immunosuppressive medications) patients with pneumonia without risk factors for MDROs at 69 hospitals in the Michigan Hospital Medicine Safety Consortium. This study compared the receipt of empiric BSAs against antibiotics targeting typical respiratory pathogens on hospital day 1 or 2. The primary outcome was mortality. Secondary outcomes included length of stay, transfer to the intensive care unit (ICU) and 30-day readmission, emergency department visit, Clostridioides difficile infection, and antibiotic-associated adverse events. RESULTS: Of 2706 moderately immunocompromised patients with pneumonia, 59% (n = 1596) received empiric BSAs. Methicillin-resistant Staphylococcus aureus and resistant gram-negative bacteria were rare (94/2706 [3.5%]). After adjustment, empiric BSA treatment was not associated with mortality, but was associated with readmission (adjusted hazard ratio [aHR], 1.32 [95% confidence interval {CI], 1.05-1.66]), transfer to ICU (aHR, 2.65 [95% CI, 1.32-5.30]), and longer hospitalization (adjusted rate ratio, 1.14 [95% CI, 1.10-1.19]). CONCLUSIONS: Immunocompromised patients hospitalized with pneumonia often receive empiric BSAs despite low rates of MDROs. Empiric BSA use was not associated with mortality, but was associated with harm, including 30-day readmission, transfer to ICU, and longer duration of hospitalization.