Ceftazidime-avibactam (CAZAVI) has been the therapy of choice for class A carbapenemase-producing Klebsiella pneumoniae (KPCP), and, combined with aztreonam (ATM), for class B KPCP or co-producers (A+B). Our objective was to compare 30-day mortality in patients treated with regimens containing CAZAVI ± ATM versus other regimens, in a Brazilian hospital that recommends CAZAVI±ATM as first-line therapy for KPCP.

Retrospective cohort study. Eligible were adult patients who had ≥1 blood culture with KPCP, collected > 2 days after hospital admission between Jan/14 and Dec/24 and who received ≥ 2 days of adequate therapy (≥ 1 in vitro active antimicrobial). Two univariate analyses were performed comparing 1) CAZAVI±ATM group vs other treatments, and 2) deaths vs survival at 30 days. A Cox regression model for the primary outcome was performed including, one by one, all independent variables with P ≤ 0.20 in the univariate analyses.

Excluding 79 duplicates, 172 patients were eligible, of whom 43 were excluded for absence of treatment/death ≤ 2 days, resulting in 129 analyzed: mean age = 68 ± 19 years, 73 (57%) male, and 61 (47%) were in ICU. The most common primary sites were CVC (31%), respiratory (23%), and urinary (22%). Class A, B, and A+B KPCP corresponded to 82%, 9%, and 9%, respectively. A total of 73 (57%) patients received CAZAVI±ATM (58 CAZAVI only) and 56 (43%) other treatments (79% with regimens containing polymyxins). Covariates were quite similar between the two groups. Selected for the model were: time to therapy initiation (P < 0.001), BMI (P = 0.17), baseline eGFR (P = 0.05), and high-risk primary site (P = 0.03). Mortality was 31% (n = 40): 23.3% (17/73) in the CAZAVI±ATM group and 41.1% (23/56) in the other treatments group (P = 0.05). Variables associated with death included in the model were age (P = 0.09), Pitt score (P = 0.01), mechanical ventilation (P = 0.09), high-risk site (P = 0.03), and baseline eGFR (P = 0.05). In the final model adjusted for high-risk site (P = 0.02) and age (P = 0.09), treatment with CAZAVI±ATM was independently associated with a lower risk of 30-day death (hazard ratio 0.47; 95% CI 0.25–0.89, P = 0.02).

CAZAVI±ATM was associated with lower 30-day mortality in patients with bloodstream infection due to KPCP who had quite similar baseline characteristics.