Presented at the Neonatal Society 2017 Spring Meeting.
Osborne L1, Lounis SA2, Germovsek E3, Gunaratnam F2, Busquets FB3, Standing JF3, Sinha A1,2
1 Neonatal Unit, Royal London Hospital, London, UK
2 Barts and the London School of Medicine and Dentistry, London, UK
3 Inflammation, Infection and Rheumatology section, Institute of Child Health, University College London, London, UK
Background: Vancomycin is commonly used for nosocomial bacterial pathogens causing late-onset septicaemia in preterm infants. There is limited data on optimum dosing regimens for vancomycin based on population pharmacokinetic analyses. We therefore aimed to develop and externally validate a pharmacokinetic model for continuous and intermittent vancomycin dosing regimens in neonates.
Methods: New-borns receiving vancomycin for suspected or confirmed late onset sepsis were included. Peak and trough levels for intermittent vancomycin dosing and a random level for continuous vancomycin dosing were measured. The vancomycin assay was performed using COBAS 702 platform (linear range was 1.7-80ug/mL). NONMEM 7.3 was used to perform the population pharmacokinetic analysis. Initial data collected over 12 months from September 2014 was used for model development and subsequent data collected between January and May 2016 for evaluation. The study was approved by Barts Health Clinical Effectiveness.
Results: Patient demographic and number of vancomycin samples for both model development stage and evaluation stage are given in Table. The final model was a 1-compartment model. Weight and PMA were included a priori; and after that no additional covariate significantly improved the model fit. Vancomycin volume of distribution (V) was fixed to the value from the intermittent administration; and clearance (CL) was estimated using both intermittent and continuous data. Internal evaluation showed the model is able to adequately describe the data. Final model parameter estimates (mean (relative standard error)): CL 5.4 L/h/70kg (7.0%), V 31.6 L/70kg. Values for a typical neonate from the studied population (weight 1.7kg, PMA 35.7 weeks): CL 0.10 L/h, V 0.77 L.
Visual predictive check on evaluation dataset using 1,000 simulated datasets confirmed that the model is able to predict external data.
Conclusion: A population PK model for continuous and intermittent vancomycin administration in neonates was developed. External evaluation showed that the model can predict external data that were not used in the model development. This confirms that the model could be used for simulations. This could be used to develop a new dosing scheme which could be prospectively evaluated.
Corresponding author: email@example.com
Janssen et al., 2016. Towards rational dosing algorithms. Antimicrob Agents and Chemo, 60(2);1013–21.