Evaluation of amoxicillin and benzylpenicillin therapy in early-onset neonatal sepsis: a pharmacometric external...

Early-onset sepsis (EOS) poses a significant morbidity and mortality risk in neonates, for which early diagnosis and adequate antibiotic therapy is crucial. Amoxicillin and benzylpenicillin combined with aminoglycosides are often prescribed empirically for neonatal EOS but optimal dosing regimens are lacking. To evaluate the pharmacokinetics (PK), PTA and toxicity of amoxicillin and benzylpenicillin in (pre)term neonates with EOS, and define optimal dosing regimens.

One hundred forty-five neonates [gestational age (GA): 24-42 weeks] with EOS treated with intravenous amoxicillin or benzylpenicillin, dosed as per the Dutch Pediatric Formulary (DPF), were included. Amoxicillin and benzylpenicillin were quantified in left-over samples during the first 48 h of life. First, the performance of nine paediatric amoxicillin and benzylpenicillin population PK models was evaluated. Second, the most appropriate models were used for simulation-based PTA and toxicity analyses, evaluating eight international neonatal dosing regimens. Third, simulation-based dose optimization was conducted.

The Bijleveld (amoxicillin) and Padari (benzylpenicillin) models adequately described the obtained PK data (N = 252). For amoxicillin, all regimens showed >90% PTA up to 100%fT > MIC but displayed GA-dependent toxicity potential (concentrations >110 mg/L), the DPF regimen excepted. By contrast, all benzylpenicillin regimens showed suboptimal PTA, often accompanied with GA-dependent toxicity potential (concentrations >50 mg/L). Simulations indicated GA-based intermittent dosing or continuous infusion as options to further optimize benzylpenicillin therapy.

(Pre)term neonates with EOS can be adequately treated with amoxicillin dosed as per the DPF regimen. By contrast, further optimization is warranted for benzylpenicillin, for which GA-based intermittent dosing or continuous infusion pose potential alternatives.