Objectives: Toxic serum cefepime trough concentrations are not well defined in the current literature. We aimed to define a more precise plasma trough concentration threshold for this antibiotic's neurological toxicity and to identify individuals at risk for developing neurotoxic side effects. Methods: Retrospective study including all individuals who underwent cefepime therapeutic drug monitoring (TDM) between 2013 and 2017. Individuals with cefepime concentrations other than trough were excluded. The primary outcome was to assess the incidence of neurotoxicity and its relationship with cefepime plasma trough concentrations. Secondary outcomes were the relationship of renal function, cefepime daily dose, age, and cerebral and general co-morbidities with the occurrence of neurotoxicity. We also compared the mortality rate during hospitalization in individuals with and without neurotoxicity, and the possible impact of neuroprotective co-medications on outcomes. Results: Cefepime concentrations were determined in 584 individuals. Among 319 individuals with available trough concentrations included, the overall incidence of neurotoxicity was 23.2% (74 of 319 individuals). Higher cefepime plasma trough concentrations were significantly associated with risk of neurotoxicity (no neurotoxicity 6.3 mg/L (interquartile range (IQR) 4.1e8.6) versus with neurotoxicity 21.6 mg/L (IQR 17.0e28.6), p <0.001). Individuals with presumed cefepime neurotoxicity had a significantly lower renal function (estimated glomerular filtration rate 82.0 mL/min/1.73 m 2 (IQR 45.0e105.0) versus 35.0 mL/min/1.73 m 2 (IQR 23.3e53.3], p <0.001), and significantly higher in-hospital mortality (19 (7.8%) versus 26 (35.1%) individuals, p <0.001). No neurotoxic side effects were seen below a trough concentration of 7.7 mg/L. Levels 38.1 mg/L always led to neurological side effects. Conclusion: In individuals with risk factors for cefepime neurotoxicity, such as renal insufficiency, TDM should be systematically performed, aiming at trough concentrations <7.5 mg/L.

SummaryPiwi proteins and piRNAs protect eukaryotic germlines against the spread of transposons. During development in the ciliate Paramecium, two Piwi-dependent sRNA classes are involved in the elimination of transposons and transposon-derived DNA: scan RNAs (scnRNAs), associated with Ptiwi01 and Ptiwi09, and iesRNAs, whose binding partners we now identify as Ptiwi10 and Ptiwi11. scnRNAs derive from the maternal genome and initiate DNA elimination during development, whereas iesRNAs continue DNA targeting until the removal process is complete. Here, we show that scnRNAs and iesRNAs are processed by distinct Dicer-like proteins and bind Piwi proteins in a mutually exclusive manner, suggesting separate biogenesis pathways. We also demonstrate that the PTIWI10 gene is transcribed from the developing nucleus and that its transcription depends on prior DNA excision, suggesting a mechanism of gene expression control triggered by the removal of short DNA segments interrupting the gene.

During development in the ciliate Paramecium, excess DNA interspersed throughout the germline genome is deleted to generate a new somatic genome. In this process, most of the intervening DNA is excised by a Piggybac-derived transposase, assisted by small RNAs (scnRNAs and iesRNAs) and chromatin remodelling. As the list of genes involved in DNA elimination has been growing, a need for a general approach to discover functional relationships among these genes now exists. We show that deep sequencing-based comparisons of experimentally-induced DNA retention provide a sensitive, quantitative approach to identify and analyze functional associations among genes involved in native genome editing. This reveals two functional molecular groups: (i) iesRNAs/scnRNAs, the putative Piwi- and RNA-binding Nowa1/2 proteins, and the transcription elongation factor TFIIS4; and (ii) PtCAF1 and Ezl1, two proteins involved in chromatin remodelling. Comparative analyses of silencing effects upon the largely unstudied regions comprising most developmentally eliminated DNA in Paramecium suggests a continuum between precise and imprecise DNA elimination. These findings show there is now a way forward to systematically elucidate the main components of natural eukaryotic genome editing systems.