Tolerance and Microbiological Efficacy of Cefepime or piperacillin/tazobactam in Combination With Vancomycin as Empirical Antimicrobial Therapy of Prosthetic Joint Infection: A Propensity-Matched Cohort Study
The use of piperacillin/tazobactam with vancomycin as empirical antimicrobial therapy (EAT) for prosthetic joint infection (PJI) has been associated with an increased risk of acute kidney injury (AKI), leading us to propose cefepime as an alternative since 2017 in our reference centre.
To compare microbiological efficacy and tolerance of these two EAT strategies.
All adult patients with PJI empirically treated with vancomycin+cefepime (n = 89) were enrolled in a prospective observational study and matched with vancomycin+piperacillin/tazobactam-treated historical controls (n = 89) according to a propensity score including age, baseline renal function and concomitant use of other nephrotoxic agents. The two groups were compared using Kaplan-Meier curve analysis, and non-parametric tests regarding the proportion of efficacious empirical regimen and the incidence of empirical therapy-related adverse events (AE).
Among 146 (82.0%) documented infections, the EAT was considered efficacious in 77 (98.7%) and 65 (98.5%) of the piperacillin/tazobactam- and cefepime-treated patients, respectively (P = 1.000). The rate of AE, particularly AKI, was significantly higher in the vancomycin+piperacillin/tazobactam group [n = 27 (30.3%) for all AE and 23 (25.8%) for AKI] compared with the vancomycin+cefepime [n = 13 (14.6%) and 6 (6.7%)] group (P = 0.019 and <0.001, respectively), leading to premature EAT discontinuation in 20 (22.5%) and 5 (5.6%) patients (P = 0.002). The two groups were not significantly different regarding their comorbidities, and AKI incidence was not related to vancomycin plasma overexposure.
Based on the susceptibility profile of bacterial isolates from included patients, microbiological efficacy of both strategies was expected to be similar, but vancomycin + cefepime was associated with a significantly lower incidence of AKI.