Tamara van Donge

Efficacy and safety of amoxicillin in neonates 69 4 Amoxicillin toxicity under different dosing recommendations Maximum amoxicillin exposures at steady state when considering different dosing regimens for the model population are shown in Figure 2. For only four regimens (100 mg/kg every 12 hours, 60/30 mg/kg every 12/8 hours, 50 mg/kg every 12/8/6 hours and 50 mg/kg every 12/8/4 hours) more than 10% of neonates experienced Cmax >140 mg/L. Median Cmax for these four regimens were 205.7 mg/L, 123.7 mg/L, 104.6 mg/L, 108.5 mg/L, respectively. Neonates exposed to three regimens (25 mg/kg every 12/8 hours (according to PMA and PNA), 25 mg/kg every 12/8 hours (according to GA and PNA) and 10/20 mg/kg every 12 hours) were not expected to experience Cmax >140 mg/L. Discussion Simulating amoxicillin exposure in neonates under various real-life dosing regimens yields three key findings: First, at low PKPD targets (<70% f T>MIC) and MICs (≤1 mg/L) all regimens perform well, and regimen selection could primarily be driven by avoidance of high and potentially neurotoxic concentrations and by simplicity of dosing regimen for readiness of implementation. Second, at high PKPD targets (70% f T>MIC or greater) or MICs (>2 mg/L) none of the regimens are expected to result in PTA ≥90%, indicating that amoxicillin may not be effective enough in such situations. Third, for enterococci (MIC 4 mg/L), amoxicillin may be acceptable under certain dosing regimens if the PKPD target is <70% f T>MIC, but appropriate regimens may result in a large proportion of neonates experiencing high Cmax concentrations >140 mg/L. Our analysis highlights the paradox of identifying optimal amoxicillin dosing regimens in neonates: the acceptability of regimens will differ according to selected PKPD targets (for efficacy and toxicity) and MICs, the selection of which in turn is challenging and highly influenced by microbiological epidemiology of neonatal sepsis (dominant pathogens and their resistance patterns). Rather than specifying dosing regimen based on individual patient characteristics such as weight, PNA or GA, it could be more appropriate to determine dosing decisions on the exact indication and its microbiological epidemiology: the assumptions (PKPD targets, MIC levels) informed by the target infection contribute to a large extent to target attainment compared to the individual patient characteristics. As amoxicillin efficacy is time-dependent, maximisation of amoxicillin concentrations above MIC throughout the dosing interval is theoretically desirable. However, the exact duration of the dosing interval that needs to be targeted remains a matter of debate. 19,20 In clinical practice, higher % f T>MIC are often targeted than indicated by experimental data. 20 This may be especially relevant for preterm neonates given their relative immunological immaturity resulting in altered pharmacodynamics compared with studied adult populations. 23 Data