By Richard R. Watkins, MD, MS, FACP, FIDSA, FISAC
Professor of Medicine, Division of Infectious Diseases, Northeast Ohio Medical University, Rootstown, OH
SYNOPSIS: Patients with Enterobacterales bacteremia were assigned randomly to receive a course of an anti-pseudomonas beta-lactam or de-escalation to a narrower agent based on sensitivity results. De-escalation was non-inferior to continuing empiric therapy.
SOURCE: López-Cortés LE, Delgado-Valverde M, Moreno-Mellado E, et al. Efficacy and safety of a structured de-escalation from antipseudomonal β-lactams in bloodstream infections due to Enterobacterales (SIMPLIFY): An open-label, multicentre, randomized trial. Lancet Infect Dis 2024;24:375-385.
Antibiotics with antipseudomonal activity often are prescribed empirically to patients with serious infections. However, antipseudomonal agents can exert selection pressure by inducing the expression of resistance mechanisms in Pseudomonas aeruginosa. They also increase the risk for Clostridioides difficile infection and carbapenem resistance in Enterobacterales. De-escalation of a broad-spectrum antibiotic to a narrow-spectrum one once the pathogen has been identified and susceptibility results are known is widely believed to be beneficial, despite a paucity of evidence. López-Cortés and colleagues sought to determine whether de-escalation to an active non-antipseudomonal drug in patients with Enterobacterales bacteremia was non-inferior to continuing empiric antipseudomonal therapy.
The study was a multicenter, randomized controlled trial conducted at 21 Spanish hospitals between October 2016 and January 2020. Inclusion criteria were age 18 years or older, having monomicrobial Enterobacterales bacteremia, receipt of empiric monotherapy with an antipseudomonal beta-lactam with activity against the causative microorganism (i.e., meropenem, imipenem, piperacillin-tazobactam, cefepime, ceftazidime, or aztreonam), susceptibility of the microorganism to one of the narrow-spectrum antibiotics prespecified for de-escalation, and treatment planned for at least five days from receipt of the first active drug. Exclusion criteria were women who were pregnant, patients with a life expectancy less than 30 days, recruitment more than 48 hours after the antimicrobial susceptibility report was available, patients with neutropenia, or a planned antibiotic duration of more than 28 days.
Patients were randomized 1:1 to continue empiric antipseudomonal therapy (control group) or to be switched to a drug showing in vitro activity from a predefined de-escalation list (de-escalation group). The primary endpoint was clinical cure, which was assessed three to five days after completion of therapy. Patients who died during the study were considered not to have reached clinical cure. Recurrence was considered to be a new bacteremia episode caused by the same bacterial species as the initial episode within 60 days of randomization.
There were 331 patients in the modified intention to treat (mITT) population: 164 in the de-escalation group and 167 in the control group. The median age of the patients was 72 years (range, 64-80 years). The most common comorbid illnesses were diabetes (36% of patients) and solid organ cancer (33%). Approximately half (168/331; 51%) had a community-acquired infection. The biliary tract and urinary tract (39% and 38%, respectively) were the most common sources of bacteremia. Escherichia coli (65%) and Klebsiella pneumoniae (16%) were the most frequently identified pathogens. The median time taking antipseudomonal drugs was two days (range, two to three days) in the de-escalation group and seven days (range, six to nine days) in the control group. Furthermore, the median total days of therapy was 11 in both groups (range, nine to 14 days). Piperacillin-tazobactam was the most frequent antipseudomonal drug used in both groups. Cefotaxime and ceftriaxone were the most commonly used drugs for de-escalation.
Clinical cure was achieved in 148/164 patients (90%) in the de-escalation group and 148/167 patients (89%) in the control group (difference in percentage points, 1.6; 95% confidence interval [CI], -5.0 to 8.2). Therefore, de-escalation met the prespecified non-inferiority criterion. The 60-day mortality rate was 5% (7/153) in the de-escalation group and 6% (9/160) in the control group. Furthermore, at day 60, relapse occurred in 6% (9/153) in the de-escalation group and 11% (18/160) in the control group. There was only one case of C. difficile infection in each group.
COMMENTARY
This pragmatic clinical trial is especially relevant for clinicians who treat hospitalized patients with Enterobacterales bacteremia. De-escalation from an antipseudomonal beta-lactam to a narrower-spectrum antibiotic was non-inferior in effectiveness compared to maintaining empiric therapy. This is an important finding because it provides evidence for the safety of de-escalation. The investigators were clever for making bacteremia their central focus rather than using a more nebulous diagnosis, such as sepsis. That decision strengthened the study design and increased its clinical usefulness.
The study had a few limitations. First, the median duration of antibiotic therapy was 11 days. Current guidelines recommend seven days of therapy for uncomplicated gram-negative bloodstream infections. Using longer treatment courses than necessary may have overestimated the benefits of de-escalation. Second, the study was underpowered to assess the ecological impact of de-escalation, such as the emergence of multidrug resistance. Third, approximately half of the patients had a community-acquired infection. Details were not provided in the study as to why these patients would have needed an antipseudomonal drug. Finally, a large number of the patients screened (2,030) were excluded based on eligibility criteria (1,265; 62%) or because of the physician’s discretion (317; 16%). Had more patients been included, it is unclear whether this would have led to similar results.
This study should reassure clinicians that de-escalation is a safe and effective strategy for many patients with Enterobacterales bacteremia. Further large, randomized clinical trials are needed to address antibiotic de-escalation for other patient populations (e.g., neutropenic patients, pregnant women) and pathogens (e.g., polymicrobial bacteremia, multidrug-resistant organisms).
