Broad-spectrum antibiotics (BSAs) are routinely used to rapidly stabilize patients hospitalized with sepsis. While early administration is essential, extended use of BSAs has been linked to adverse effects and secondary infections.1 Current sepsis guidelines advise narrowing antibiotic therapy when multidrug-resistant organisms are not detected, but uncertainty persists about the ideal timing for de-escalation in clinical practice.2 Recently, a target trial emulation study assessed the safety of de-escalating BSAs in adults hospitalized with community-onset sepsis who did not have evidence of multidrug-resistant organisms on culture.
The trial, Antibiotic De-Escalation in Adults Hospitalized for Community-Onset Sepsis, was conducted across 67 Michigan hospitals in partnership with the Michigan Hospital Medicine Safety Consortium.
“We noticed in our population of hospitals that even though guideline recommendations advise the de-escalation of antibiotics based on culture results, there was wide variation in who was de-escalated and who was not,” said lead author Ashwin B. Gupta, MD, Section Chief of Hospital Medicine at the Veterans Affairs Ann Arbor Healthcare System and Associate Professor of Internal Medicine at the University of Michigan Medical School. “That observation led us to ask what outcomes are associated with the de-escalation of therapy.”
Dr. Gupta and colleagues studied 36,924 adults hospitalized with community-onset sepsis who received empiric BSAs and had no evidence of multidrug-resistant organisms.One emulation examined the de-escalation of antimethicillin-resistant Staphylococcus aureus (MRSA), and the other focused on anti-Pseudomonas aeruginosa (PSA).2 Of the patients, 2,993 (43.2%) and 2,493 (22.4%), respectively, were de-escalated from anti-MRSA and anti-PSA coverage on day 4.2
The primary outcome measured was 90-day all-cause mortality, with secondary outcomes including in-hospital mortality, length of stay, and total antibiotic days.2 Researchers found that the de-escalation of BSAs on hospital day 4 was not associated with increased mortality compared with continued BSA therapy. Ninety-day all-cause mortality was similar among patients who underwent the de-escalation of anti-MRSA coverage (OR, 1.00) and anti-PSA coverage (OR, 0.98) compared with those who continued therapy.2
De-escalation was also associated with stewardship-related benefits. Patients who underwent anti-MRSA or anti-PSA de-escalation had fewer antibiotic days through day 14 (anti-MRSA: RR 0.91, 95% CI, 0.89-0.93; anti-PSA: RR 0.91, 95% CI, 0.88-0.93) and shorter lengths of hospitalization (anti-MRSA: RR 0.88, 95% CI, 0.85-0.92; anti-PSA: RR 0.88, 95% CI, 0.80-0.96). All other secondary outcomes were similar between groups.2
Among exploratory outcomes, only anti-PSA de-escalation was additionally associated with lower 90-day readmission (RR 0.87, 95% CI, 0.76-0.96).
To control for confounding, the investigators “used inverse probability of treatment weighting to balance the de-escalated and continued BSA therapy groups on likelihood of being de-escalated.”2
“We effectively ‘randomized’ patients on day 4 to either de-escalation or continued broad-spectrum therapy. We used inverse probability of treatment weighting to balance the groups as closely as possible,” Dr. Gupta said.
He also acknowledged the risk of residual confounding. “One might argue that patients who are sicker are more likely to remain on antibiotics longer and are more likely to have negative effects,” he said.
To address this concern, the researchers conducted sensitivity analyses limited to clinically stable patients, which yielded similar results. In that subgroup, the de-escalation of anti-MRSA therapy was associated with a mortality difference favoring de-escalation.
Taken together, the findings suggest that the de-escalation of empiric BSAs in clinically stable or improving patients with community-onset sepsis may be safely pursued without compromising mortality outcomes.
While the study was not randomized or controlled, the consistency of results across multiple analyses provides reassurance regarding the safety of earlier antibiotic narrowing. Dr. Gupta emphasized that these data may help address lingering clinician hesitancy surrounding the de-escalation of antibiotics in this patient population.
“Antibiotics are not without consequences, whether at the population level through increasing antimicrobial resistance or at the individual level through antibiotic-associated adverse effects,” he said. “I am hopeful that these data will help physicians become more comfortable in reducing antibiotic exposure.”
References
1. Tamma PD, Avdic E, Li DX, et al. Association of adverse events with antibiotic use in hospitalized patients. JAMA Intern Med. 2017;177(9):1308-1315. doi:10.1001/jamainternmed.2017.1938
2. Gupta AB, Heath M, Walzl E, et al. Antibiotic de-escalation in adults hospitalized for community-onset sepsis. JAMA Intern Med. doi:10.1001/jamainternmed.2025.6919
