De-escalating empiric broad spectrum antibiotics (BSA) at day 4 for clinically stable patients with community-onset sepsis without positive culture for multidrug-resistance was associated with fewer days of antibiotic therapy and of hospitalization, and similar all-cause mortality compared to longer periods of BSA, in two trial emulation studies.1
"Despite guideline recommendations to de-escalate antibiotics, the incidence of antibiotic de-escalation and impact on clinical outcomes in sepsis remain unclear," observe study lead author Ashwin Gupta, MD, Medicine Service, VA Ann Arbor Healthcare System, Ann Arbor Michigan, and colleagues.
To compare outcomes of the shortened duration of empiric BSA to longer periods, the investigators conducted two trial emulation studies from data of 67 hospitals in the Michigan Hospital Medicine Safety Consortium Sepsis Initiative (HMS-Sepsis).One trial targeted empiric BSA for methicillin-resistant Staphyloccocus aureus (MRSA) and the other for Pseudomonas aeruginosa (PSA) or other resistant gram-negative bacteria.
A total of 2,993 patients in the trial of empiric BSA targeting MRSA and 2,493 targeting PSA were de-escalated on day 4, and their outcomes compared to 8,126 and 12,996 with longer periods of BSA treatment, respectively. After weighting variables, the de-escalated and longer BSA treatment groups were deemed comparable on propensity scores, patient characteristics and treating hospital resources.
What You Need to Know
De-escalating empiric broad-spectrum antibiotics at day 4 in clinically stable, culture-negative community-onset sepsis was not associated with increased 90-day all-cause mortality compared with longer courses, for both anti-MRSA and anti-Pseudomonas regimens.
Patients de-escalated at day 4 had fewer total days of antibiotic therapy and shorter lengths of hospitalization across both trial emulations, with anti-Pseudomonas de-escalation also linked to fewer readmissions.
Fewer than half of eligible patients were de-escalated by day 4, with wide inter-hospital variation; de-escalation was more common for anti-MRSA therapy, likely reflecting greater availability of rapid diagnostics compared with resistant gram-negative infections.
Across the 67 hospitals, the investigators noted, fewer than one-half of eligible patients were de-escalated on day 4, and there was more than 2-fold variation in de-escalation in both the anti-MRSA and anti-PSA groups.They also found a higher proportion de-escalated from BSA regimens targeting MRSA
"This may reflect both greater clinical certainty and greater ease of de-escalating anti-MRSA therapy," they posited."It has become easier to rule out MRSA pneumonia with a nasal swab and MRSA bacteremia with molecular testing, while development and dissemination of rapid tests for diagnosis of resistant gram-negative infections have been slower."
In the primary comparison of clinical outcomes associated with de-escalation and longer periods of empiric BSA in the two trials, Gupta and colleagues found similar rates of 90-day all-cause mortality (anti-MRSA: Odds Ratio 1.00 [95% CI 0.88-1.14]; anti-PSA: OR 0.98 [0.86-1.13]).In both trials, however, de-escalation at day 4 was associated with fewer days of any antibiotic (anti-MRSA: risk ratio 0.92 [0.89-0.93]; anti-PSA (RR 0.91 [0.88-0.92]), as well as shorter lengths of hospitalization (anti-MRSA: RR 0.88 [0.85-0.92]; anti-PSA: RR 0.91 [0.89-0.93]).
In a secondary measure, the anti-PSA de-escalation was associated with reduced hospital readmission, while this was not found with anti-MRSA de-escalation.The investigations offer the possibility that anti-PSA coverage often includes anti-anaerobic spectrum which is deleterious to the gut microbiome, and so may increase downstream risk of infection and mortality.Shortening the period of the anti-MRSA BSA, they suggest, might avoid those deleterious effects that contribute to readmission.
The investigators also comment on a hypothesis that improvement in short-term mortality with de-escalation in some observational studies is due to "reverse causation," where an improving clinical trajectory increases likelihood of de-escalation.This likely reflects the absence of adjusting for clinical stability, they argue, which was not the case in the design of these target trial emulations.
"In total, our data lend further evidence suggestive that antibiotic de-escalation is both safe and appropriate in patients hospitalized with community-onset sepsis, bolstering guideline recommendations to do so in clinical practice," Gupta and colleagues advise.
Reference
1.Gupta AB, Heath M, Walzl E, et al. Antibiotic de-escalation in adults hospitalized for community-onset sepsis.JAMA Intern Med. 2025. Online December 22. doi:10.1001/jamainternmed.2025.6919.
