Eva Germovsek (1, 2), Irja Lutsar (3), Karin Kipper (3), Mats O Karlsson (2), Joseph F Standing (1) on behalf of the NeoMero consortium
(1) Department of Infection, Inflammation and Rheumatology, Institute of Child Health, University College London, UK; (2) Department of Pharmaceutical Biosciences, Uppsala University, Sweden; (3) Department of Microbiology, University of Tartu, Estonia
Objectives: Meropenem is used off-label in the treatment of late-onset sepsis (LOS) and meningitis in infants. We aimed to develop a population pharmacokinetic (PK) model for meropenem plasma and CSF concentrations, and use it for dosing recommendations in infants <3 months. We also aimed to investigate if the PKPD target of 40% of time above the minimal inhibitory concentration (MIC) (t>MIC) [1] is sufficient for infants.
Methods: Meropenem was given at 20 mg/kg (LOS) or 40 mg/kg (meningitis) as a 30-min infusion every 8, or every 12h in infants <32 weeks gestational age (GA) and <2 weeks postnatal age (PNA). NONMEM 7.3 was used. Firstly, different structural and residual error models were fitted to the plasma data. Allometric weight scaling and a function describing renal function maturation [2] were used a priori, and serum creatinine (SCr) and PNA were tested on clearance (CL). Then, a CSF compartment was added, and plasma-CSF clearance (Q_CSF) and fraction of meropenem penetration from plasma to CSF (f_CSF) were estimated. The effect of markers of central nervous system inflammation on f_CSF was investigated. Monte-Carlo simulations (n=1000) were used to generate the probability-of-target-attainment (PTA) curves. Finally, the PK model was combined with an in vitro PKPD model [3], describing the effects of meropenem on Pseudomonas aeruginosa.
Results: 167 infants with median (range) GA 33.3 (22.6-41.9) weeks, and PNA 13 (1-90) days at enrolment provided 401 plasma and 78 CSF samples. The final PK model for plasma was a 1-compartment model, SCr proved significant on CL, and CSF protein concentration on f_CSF. A Box-Cox power transformation [4] was used to address the non-normally distributed residuals. The parameter values (mean (%SE)) for a typical infant (weight=2.12 kg, postmenstrual age=37.4 weeks, SCr=32 μmol/L) were: CL=0.48 (5.2%) L/h, volume of distribution 1.18 (5.6%) L, Q_CSF=0.0012 (24.3%) L/h, f_CSF=0.085 (18.8%). Diagnostic plots showed adequate descriptive and simulation properties of the model. Dosing regimens used in the study were appropriate if MIC <4mg/L (LOS) or <1mg/L (meningitis) for a target of 40% t>MIC; however, this target overpredicted meropenem effect, when compared to the in vitro target of 2log kill at 24h.
Conclusion: A PK model for plasma and CSF meropenem data in young infants with LOS and/or meningitis was developed. The dosing regimen used in the study was shown appropriate for susceptible pathogens, but 40% t>MIC appeared insufficient for severely immunocompromised infants.
Acknowledgements: The NeoMero study was funded as a part of the European FP7 Programme.
References:
[1] Drusano, Nat Rev Microbiol. 2004; 2(4):289-300.
[2] Rhodin et al., Pediatr Nephrol, 2009. 24(1):67-76.
[3] Mohamed et al., J Antimicrob Chemother. 2016; 1-12.
[4] Dosne et al., J Pharmacokinet Pharmacodyn. 2015; 1-15.
Reference: PAGE 25 () Abstr 6014 [www.page-meeting.org/?abstract=6014]
Poster: Drug/Disease modeling - Paediatrics