Methods: In this retrospective, cohort study conducted at two tertiary care hospitals in Japan, we determined consecutive patients with EB identified from the laboratory databases between January 2010 and December 2017. We enrolled patients with 3GC-susceptible EB treated with 3GC or BSA (defined as fourth-generation cephalosporins, carbapenems, and piperacillin/tazobactam) as definitive therapy. The primary outcome was 28-day mortality. The secondary outcome was the emergence of antimicrobial-resistant strain during antimicrobial therapy. We compared outcomes using the propensity scores and inverse-probability-weighting (IPW) adjustment to decrease the confounding by indication.
Results: We identified 320 patients with EB; of these, 191 cases were eligible (86 treated with 3GC and 105 treated with BSA). All the measured covariates were well balanced after the IPW adjustment. We observed no significant differences in the unadjusted mortality [5.8% in the 3GC group vs. 13.3% in the BSA group; risk difference, −7.5%; 95% confidence interval (CI): -15.7–0.6; p = 0.09], and the IPW-adjusted mortality (5.1% vs. 9.4%; risk difference −4.3%; 95% CI: −12.2–3.5; p = 0.3) between the groups. The results of the propensity score-matched analysis and sensitivity analysis were similar. Furthermore, we did not observe the emergence of antimicrobial resistance during antimicrobial therapy in both groups.
Conclusion: Definitive therapy with 3GC for susceptible EB was not associated with an increased risk of the 28-day mortality after adjustment for potential confounders with the propensity score analysis or with the emergence of antimicrobial-resistant strain.
R. Hase, None
A. Toguchi, None
Y. Otsuka, None
N. Hosokawa, None