Continuous infusion of cefepime and neurotoxicity: a retrospective cohort study

Abstract Objectives Neurotoxicity related to cefepime is increasingly reported in the literature but specific data concerning continuous infusion (CI) of the drug are still lacking. Our primary objective was to evaluate the incidence of neurotoxicity related to CI of cefepime and the associated risk factors. Our secondary objectives were to analyze the plasma cefepime concentrations and to define the threshold above which neurotoxicity occurs. Methods In this single-center retrospective cohort study, all adult patients who underwent at least one cefepime therapeutic drug monitoring (TDM) and were treated with CI of 4 g/day between January 2017 and June 2019 were included. Neurotoxicity was evaluated according to a strict definition and was correlated with steady-state concentration at the time of toxicity presentation. Results Ninety eight patients with 201 cefepime TDM were included, with an incidence of neurotoxicity of 14.3% (14/98). Patients with neurotoxicity had more often underlying brain disease [35.7% (5/14) vs 11.9% (10/84), p=0.030)] and higher steady-state concentrations (mean ± SD 71.8 ± 32.9 mg/L vs 49.6 ± 30.6, p=0.036) than the others. A receiver operating characteristic (ROC) curve analysis yielded a cefepime steady-state concentration of 63.2 mg/L as the best cutoff point between patients with or without neurotoxicity. A mean steady-state concentration of 46.4 mg/L was achieved if the dosages of cefepime were adapted to renal function (RF) which was under our threshold concentration but above our highest PK/PD target of 32-40 mg/L. Conclusions Our results suggest that 4 g/day of cefepime adapted to RF and infused over 24 hours is a trade-off for the risk/benefit ratio, when used empirically.