S.M. Illamola (1,2), J.G Coen van Hasselt (3), L.Pou (4), H.Colom (2)
(1) Biochemistry Laboratory. Corporació Sanità ria Parc TaulÃ. Sabadell, Barcelona. Spain. (2) Pharmacy and Pharmaceutical Technology Department. School of Pharmacy, University of Barcelona. Spain. (3) Department of Clinical Pharmacology, Netherlands. Cancer Institute, Amsterdan, Netherlands. (4) Biochemistry Service. Hospital Universitario Vall Hebron, Barcelona. Spain.
Objectives: The aim of this study was to establish the population pharmacokinetics (PK) of amikacin in newborns from serum concentration data obtained during the routine therapeutic drug monitoring and to explore the influence of patient covariates on drug disposition.
Methods: Data were retrospectively collected from a study in newborns with postnatal age less than 90 days admitted in the neonatal unit of Vall Hebron (July 2000 to July 2006) who were treated with amikacin, and with at least two serum concentration samples. Amikacin was administered by intravenous (IV) infusion over 30 to 60 min. Blood samples were collected prior to administration, and after 1 hour start of the infusion. Amikacin serum concentrations were quantified using a fluorescence polarization immunoassay (TDx; Abbott Laboratories). Subsequently, the data was analyzed using a nonlinear mixed-effect modeling approach in NONMEM 7. Between subjectpatient variability (BSV) was modeled exponentially, and was evaluated for tested in all the PK parameters. The First order conditional estimation method (FOCE) with interaction (FOCE-I) was used through all the model building process. Parameter precision was evaluated using the bootstrap (n=200).
Results: A total of 451 amikacin serum levels from 148 newborns were included in the analysis. The PK of amikacin after IV administration was best described by a two-compartment linear disposition model. All parameters were estimated with adequate precision (RSE<41%). BSV was estimated for clearance (CL) (34.50 CV%), central compartmental distribution volume (V1) (21.07 CV%) and distributional clearance (Q) (70.21 CV%). Residual variability was modeled using a combined error model. The final model included creatinine clearance (CLCR) and body weight (WGT) on CL, and WGT on V1. All of them reduced the BSV about 39% of the variability on CL and 37% on V1 from the base model. The final covariate relationships identified were: TVCL=(0.094·(CLCR/24.75)^0.649)*((WGT/1470)^0.752) and TVV1=0.640*(WGT/1470)^1.090. The bootstrap indicated that estimates of the fixed and random effects in the final model were estimated with good precision.
Conclusions: The developed population PK model for amikacin in newborns adequately described the observed data. CLCR and WGT were identified as the best predictor for BSV in CL, and WGT was for V1.
References:
[1] Botha J.H, Preez M.J, Miller R, Adhikari M. “Determination of population pharmacokinetic parameters for amikacin in neonates using mixed-effect models”. Eur J Clin Pharmacol. 53(5). 337-341 (1998).
[2] Bleyzac N, Varnier V, Labaune JM, et al. “Population pharmacokinetics of amikacin at birth and interindividual variability in renal maturation”. Eur J Clin Pharmacol 2001; 57:499-504.
Reference: PAGE 21 (2012) Abstr 2439 [www.page-meeting.org/?abstract=2439]
Poster: Paediatrics