A retrospective study of US hospital patients with gram-negative bloodstream infections (GN-BSIs) suggests that extended-infusion beta-lactam (EI-BL) antibiotic therapy may benefit those who are severely ill or infected with non-susceptible organisms, researchers reported today in JAMA Network Open.
The study, led by researchers with the Johns Hopkins University School of Medicine, examined outcomes in 4,861 GN-BSI patients treated at 24 US hospitals in 2019. All patients were treated with beta-lactam antibiotics for at least 72 hours, but one group of patients received intravenous beta-lactam agents for 3 or more hours at a time (EI-BL), while the other received beta-lactam antibiotics for 1 hour or less (intermittent-infusion, or II-BL).
As the study authors explain, extending the infusion of a beta-lactam antibiotic is thought to benefit patients infected with gram-negative organisms that have elevated minimum inhibitory concentration (MIC) levels. But the results of clinical trials investigating the effect of EI-BL therapy on patient outcomes have been mixed, with some showing improved outcomes and others finding no benefit.
The primary outcome of the study was mortality within 90 days of blood culture collection. Secondary outcomes included recurrent infection with the same bacterial species, emergence of antibiotic resistance, and antibiotic-related adverse events. To ensure patients in the exposed group (EI-BL therapy) and unexposed group (II-BL therapy) were as similar as possible, the EI-BL and II-BL groups underwent 1:3 nearest-neighbor propensity score matching (PSM) without replacement.
Findings support a more targeted approach
Of the 4,861 GN-BSI patients in the study (median age, 67 years; 52.4% male), 352 (7.2%) received EI-BL therapy, and 4,509 (92.7%) received II-BL therapy. The most common bacterial species among patients were Escherichia coli (50.8%), Klebsiella pneumoniae (17.3%), and Pseudomonas aeruginosa (8.7%). Patients who received EI-BL were more likely to be severely immunocompromised (38% vs 29%), receive care in the intensive care unit (46% vs 31%), and have a Pitt bacteremia score of 4 or higher (26% vs 19%).
The matched cohort consisted of 352 in the EI-BL 1:3 PSM group and 1,056 in the II-BL 1:3 PSM group. Among 1,408 matched patients, 79 (22%) in the EI-BL group died by day 90, compared with 294 (28%) in the II-BL group (adjusted odds ratio [aOR], 0.91; 95% confidence interval [CI], 0.52 to 0.97). But in a stratified analysis, the mortality benefit associated with EI-BL therapy was seen only in patients with critical illness (aOR; 0.47; 95% CI, 0.28 to 0.81) and those with an elevated beta-lactam MIC (aOR, 0.06; 95% CI, 0.01 to 0.66).
There was no difference in the odds of recurrent infection between the groups (aOR, 0.96; 95% CI, 0.64 to 1.45), and among patients who had a recurrent infection with the same bacterial species, emergence of resistance was similar in the EI-BL group (2.9%) and the II-BL group (7.2%). But the study authors note that with only 10% of patients having a recurrent infection with the same bacterial species, the study was underpowered to investigate the effects of EI-BL on subsequent emergence of resistance.
Adverse events were low in the entire PSM cohort (5%), but EI-BL therapy was associated with increased odds of catheter complications (aOR, 3.14; 95% CI, 1.66 to 5.96) and antibiotic discontinuation because of adverse events (aOR, 3.66; 95% CI, 1.68 to 7.95).
"Taken together, these results suggest that while EI-BL therapy may be associated with positive outcomes, the benefits may not surpass the risks if applied to all patients," the study authors wrote. "Rather, a more targeted approach may be necessary for patients who are severely ill or those known to have or be at reasonable risk for infection with an elevated β-lactam MIC."
In an accompanying commentary, Miranda So, PharmD, MPH, of the University of Toronto, says the findings are important because they support prior knowledge about EI-BL therapy while also addressing important knowledge gaps for antimicrobial stewardship and healthcare safety practitioners.
"Although EI is not novel, Karaba et al characterized the circumstances in which EI will be advantageous, while exploring important balancing measures," she wrote. "In turn, their data will help patients and clinicians to get more out of the so-called antibiotic buck."
