Antibiotic De-escalation and 30-Day Mortality in Culture-Negative Sepsis

Introduction: Microbial cultures for workup of sepsis can often yield negative results up to 30-50%; with initial bacterial cultures even as low as 15%. As a result, broad spectrum antibiotics used initially for sepsis management are often challenging to determine for de-escalation. This study aims to examine the association of antibiotic de-escalation with 30-day mortality in sepsis patients with negative microbial cultures.


Methods: We conducted a retrospective cohort study from Duke university hospital and 2 Duke Health community hospitals from January 2016 to March 2023. We included adult patients with community onset sepsis within two days of admission with negative microbial cultures. Sepsis was defined using the U.S. Centers for Disease Control Adult Sepsis Event Surveillance criteria. Antibiotics were ranked into 4 categories: no antibiotic, narrow-spectrum, broad-spectrum, and extended-spectrum antibiotics including anti-MRSA and anti-pseudomonal antibiotics. De-escalation was defined as a reduction in the rank within 5 days from the onset. The study cohort was stratified by changes in sequential organ failure assessment (SOFA) score as improved (<-1), unchanged (+1), or deteriorated (>+1) from day 1 to day 3. Multilevel logistic regression models were used to estimate the association between de-escalation and 30-day mortality.


Results: We identified 3855 patients with culture negative sepsis. Of those, antibiotic de-escalation was performed in 1295 (34%) and escalation was performed in 273 (7%) patients. The median age was 61 (IQR, 49 – 71) years, 56% were male, and 34% were mechanically ventilated. The crude proportion of 30-day mortality was 16% for de-escalation, 16% for no change, and 19% for escalation. For the outcome analysis, antibiotic de-escalation (aOR, 0.98, 95% CI, 0.79 – 1.21) and escalation (aOR, 1.24, 95% CI, 0.87 – 1.78) were not associated with 30-day mortality in patients with culture-negative sepsis. In the subgroup analysis, de-escalation was not associated with 30-day mortality in the improved SOFA group, the unchanged SOFA group, and the deteriorated SOFA group.


Conclusions: Only one third of culture-negative sepsis patients had antibiotic de-escalation. De-escalation was not associated with 30-day mortality, irrespective of SOFA score changes.