Association between antipseudomonal beta-lactam empiric therapy and oral switching: an observational study of community-acquired uncomplicated gram-negative bloodstream infections in a university hospital

Abstract Background Oral switching is a strategy involving the transition from intravenous to oral antibiotics in the treatment of infection. It reduces unnecessary prolongation of intravenous antibiotic therapy and is widely accepted for patients with uncomplicated gram-negative bloodstream infection (GN-BSI). Although previous studies have suggested that antipseudomonal beta-lactam (APBL) empiric therapy is associated with a lower likelihood of oral switching, this association has not been examined in patients with GN-BSI. The present study aimed to evaluate this association in patients with community-acquired uncomplicated GN-BSI. Methods This retrospective observational study was conducted at a tertiary-care university hospital in Japan. Patients hospitalized with community-acquired uncomplicated GN-BSI between January 1, 2021, and December 31, 2023, were included. Patients who received carbapenems, piperacillin/tazobactam, or fourth-generation cephalosporins as empiric therapy were categorized into the APBL group, whereas those who received other antibiotics (e.g. ceftriaxone, cefmetazole, or levofloxacin) were categorized into the non-APBL group. The primary outcome was the oral switching rate. Secondary outcomes included the duration of intravenous antibiotic therapy, total duration of antibiotic therapy, and length of hospital stay (LOS). Results Of 106 episodes in 96 patients, 32 and 74 were in the APBL and non-APBL groups, respectively. The rate of oral switching was significantly lower in the APBL group than in the non-APBL group (46.9% [15/32] vs. 75.7% [56/74], relative risk = 0.62 [95% confidence interval: 0.42–0.92], p = 0.006). The total duration of antibiotic therapy did not significantly differ between the groups (median [IQR]: 13 [9.5–15.5] days vs. 14 [11–15] days, p = 0.16). However, the duration of intravenous antibiotic therapy and LOS were significantly longer in the APBL group (median [IQR]: 8 [6–12.5] days vs. 7 [5–10] days, p = 0.040; and 12.5 [8.5–18.5] days vs. 8 [7–13] days, p = 0.029, respectively). Conclusions Compared to non-APBL empiric therapy, APBL empiric therapy was associated with a lower likelihood of oral switching and a longer duration of intravenous antibiotic therapy and LOS. These findings underscore the need for enhanced interventions to promote oral switching in patients with GN-BSI who receive APBLs as empiric therapy.