Concerns about study design and suggestion of additional analysis of therapeutic drug monitoring-guided piperacillin/tazobactam therapy for patients with sepsis. Author's reply.

We would like to thank for the opportunity to respond to the issues raised in the commentary by Shodai Yoshihiro and Shunsuke Taito [1]. We also greatly appreciate their interest in our paper [2]. Following are the concerns raised by them. First, they considered the therapeutic efficacy index used for the dose adjustment of piperacillin/tazobactam in our study was inappropriate. We agree with the colleagues that ß-Lactam antibiotics have a time-dependent antibacterial activity and increasing the time above the pathogens minimal inhibitory concentration is important for promoting antibacterial activity. In our trial, the predefined target plasma concentration of free piperacillin was defined as four times (range ± 20%) the minimal inhibitory concentration (MIC) of the pathogen causative of sepsis [2]. All patients in the control and experimental therapy group received a 24 h continuous infusion of piperacillin/tazobactam. Therefore, the once-daily measured piperacillin concentration represented a steady-state concentration. Hence, pharmacokinetic/pharmacodynamic (PK/PD) target was 100% f T>4MIC (range ± 20%) [percentage of time during a dosing interval that the free (f) drug concentration exceeded 4 times the MIC, (range ± 20%)]. This is in line with the proposed target of Shodai Yoshihiro and Shunsuke Taito. To illustrate, 41.3% of patients with piperacillin therapeutic drug monitoring (TDM) achieved the PK/PD target on day 2 after randomization. This means that in these patients, the piperacillin concentration was 100% of time (24 h) four times (range ± 20%) the MIC of the pathogen causative of sepsis. Second, the administration period was short since the study protocol involved de-escalation. This observation is correct and indeed complicates the interpretation of the overall study results. However, it would have been unethical not to permit de-escalation after identifying the pathogen causative of sepsis. For example, therapy with piperacillin/tazobactam compared to oxacillin/cefazolin is associated with a higher mortality in patients with S. aureus bloodstream infection [3]. Finally, sensitivity analyses were recommended for patients with, e.g., Pseudomonas aeruginosa infection. Analyses of subgroups have been performed; however, due to low patient numbers, they are not meaningful.