An analysis of patients with suspected sepsis at US hospitals found that antibiotic de-escalation was infrequent and varied widely across hospitals but was associated with reduced risk of adverse outcomes, researchers reported today in Clinical Infectious Diseases.
Using data from the PINC AI Healthcare Database, a team led by researchers from the Harvard Pilgrim Healthcare Institute retrospectively analyzed antibiotic therapy in all adults admitted to 236 US hospitals from 2017 through 2021 with suspected community-onset sepsis.
They focused their analysis on patients who were initially treated with 2 or more days of anti–methicillin-resistant Staphylococcus aureus (MRSA) or anti-pseudomonal antibiotics but had no resistant organisms requiring these agents through day 4 of hospitalization. They defined de-escalation as stopping anti-MRSA and anti-pseudomonal antibiotics or switching to narrower antibiotics by day 4.
Sepsis occurs when the body overreacts to an infection, setting off a chain of events that can lead to tissue damage, organ failure, and death. Sepsis guidelines recommend starting treatment quickly with broad-spectrum antibiotics to cover a wide range of potential organisms and then reassessing after 48 to 72 hours based on clinical trajectories and microbiologic results. But the study authors note that little is known about real-world practice in patients with suspected sepsis.
"De-escalation rates have ranged widely in studies, but many were single-center, limited to intensive-care unit (ICU) settings, or focused on specific infections," they wrote. "There are little published data on variability in de-escalation practices across hospitals and factors associated with de-escalation."
In addition to examining rates of antibiotic de-escalation, the researchers also looked at clinical predictors of de-escalation, as well as outcomes associated with de-escalation, including hospital-onset acute kidney injury (AKI), Clostridioides difficile infection (CDI), ICU admission after day 4, and in-hospital mortality.
De-escalation occurred in less than 30% of patients
Of the 124,577 patients (mean age, 64; 43.7% female; 71.6% White) in the study, antibiotics were de-escalated by day 4 in 36,806 (29.6%), including 27,177 patients (21.8%) in whom antibiotics were narrowed and 9,629 (7.7%) in whom they were stopped. The de-escalation rate varied widely across hospitals, with a median rate of 29.4%. The highest de-escalation rates were found in large hospitals and teaching hospitals in urban areas and the northeast.
Clinical predictors of de-escalation, after adjusting for covariates, included less severe disease on days 3 and 4, positive cultures for non-resistant organisms, present-on-admission infection diagnoses, and negative/absent MRSA nasal swabs.
Propensity score matching of 36,803 de-escalated patients with 32,964 non–de-escalated patients revealed that de-escalation was associated with lower adjusted risks of AKI (odds ratio [OR], 0.80; 95% confidence interval [CI], 0.74 to 0.84), ICU admission after day 4 (OR, 0.59; 95% CI, 0.52 to 0.66), in-hospital mortality (OR, 0.92; 95% CI, 0.86 to 0.99), and a trend toward less CDI (OR, 0.84; 95% CI, 0.71 to 1.01).
The study authors say that while the large number of hospitals in the study amplifies previous findings on antibiotic de-escalation rates in sepsis patients, the observed lower risk of adverse outcomes "substantially augments the evidence in favor of de-escalation." The potential mechanisms for reducing mortality, they suggest, could include lowering the risk of antibiotic toxicities and antibiotic-resistant infections.
"Early antibiotic cessation or narrowing may also induce less disruption of the gut microbiome, which is increasingly recognized as an important modulator of the immune response to sepsis and patient outcomes," they wrote.
Because of the difficulties in interpreting observational data, the authors suggest that well-designed randomized controlled trials could more definitively evaluate the clinical outcomes of antibiotic de-escalation.
