Mortality in Culture-Negative Sepsis Not Linked to Early Discontinuation of Antibiotics

Early discontinuation of antibiotics among patients with culture-negative severe infections and sepsis is not associated with increased risk of mortality, according to study results published in Open Forum Infectious Disease. 

Researchers conducted a retrospective cohort study to assess the effects of early antibiotic discontinuation on survival among patients hospitalized with culture-negative severe infections and sepsis between January 2009 and December 2014. Data were sourced from a multi-hospital electronic health record database. Early discontinuation of antibiotic therapy was defined as discontinuing therapy after 3 to 4 consecutive days. Late discontinuation was defined as discontinuing therapy after at least 5 days. The primary outcome was the adjusted odds ratio (aOR) of in-hospital mortality or discharge to hospice care.

A total of 20,714 patients were included in the final analysis, of whom the median age was 64 (IQR, 49-77) years, 72.2% were White, 4211 (20.3%) had early sepsis indicators, and pulmonary sites of infection (n=4826) were the most common. 

Early discontinuation of antibiotics occurred among 1862 (9.0%) patients overall, as well as 269 (6.4%) patients with early sepsis indicators. Of note, patients who discontinued antibiotics early vs late had similar demographics and comorbidity distributions.

In-hospital mortality or discharge to hospice care occurred among 740 (4%) patients overall and 299 (7%) patients with early sepsis indicators. However, no statistically significant association between early discontinuation of antibiotics and increased mortality risk was observed (aOR, 1.27; 95% CI, 0.98-1.65).

Logistic regression analysis showed no significant association between early discontinuation of antibiotics and increased mortality risk among patients with (aOR, 1.39; 95% CI, 0.88-2.20) and without (aOR, 1.17;95% CI, 0.81-1.69) early sepsis indicators. Similar findings were observed among patients with (aOR, 1.23; 95% CI, 0.65-2.34) and without (aOR, 1.30; 95% CI, 0.99-1.72) pulmonary sites of infection.

In regard to secondary outcomes, patients who discontinued antibiotics early vs late had similar odds for both Clostridioides difficile infection (aOR, 0.63; 95% CI, 0.34-1.14) and late sepsis indicators (aOR, 1.15; 95% CI, 0.58-2.27).

Limitations of this study include the inability to determine whether patients continued antibiotic therapy after hospital discharge, as well as the inclusion of only inpatients in the early antibiotic discontinuation group.

According to the researchers, “[E]arly antibiotic discontinuation in culture-negative serious infection is relatively infrequent and variable, and the safety of this practice remains unclear.”