Population pharmacokinetics of Phenobarbital in neonates and infants.
Amélie Marsot (1), Renaud Vialet (2), Audrey Boulamery (1), Bernard Bruguerolle (1), Nicolas Simon (1)
(1) Laboratoire de Pharmacologie Médicale et Clinique APHM, Université de la Méditerranée, Marseille, France; (2) Département Réanimation Pédiatrique, Hôpital Nord, Marseille, France.
Objectives: Phenobarbital is widely used for treatment of neonatal seizures and the prevention of neonatal hyperbilirubinaemia. A pharmacokinetic model of phenobarbital in 35 neonates and infants was described in 2005 by Yukawa et al., studying oral and rectal routes. A new model in a similar population is proposed for intravenous administration.
Methods: 27 neonates and infants (weight: 1.2-10.0 kg; PNA: 0-466 days) hospitalized in a pediatric intensive care unit, were studied. Total mean dose of 194 mg (40-450mg) was administered by 30-min infusion. Blood phenobarbital concentrations were determined by immunoassay method. Pharmacokinetic analysis was performed by using a non linear mixed-effect population model. Data analysing included calculation of performance error (PE), median performance error (MDPE) and median absolute performance error (MDAPE). A bootstrap was used as internal validation.
Results: Data were modelled with an allometric pharmacokinetic model using three-fourths scaling exponent for clearance and parameters. This one-compartment model gave the following results. The population typical mean (percent relative standard error (%RSE)) values for clearance (CL) and apparent volume of distribution (Vd) were 8.70 mL/H/kg (10.4%) and 1.44 L/kg (15.1%), respectively. The interindividual variability of Vd (%RSE) and intraindividual variability (%RSE) were 60% (25.0%) and 42% (27.9%), respectively. The indicators of predictive performance gave the following results: MDPE (range) was -9.5% (-74.4 to 131.7%) and MDAPE (range) was 30.1% (3.9 to 131.7%).
Conclusions: The pharmacokinetic parameters of intravenous phenobarbital in neonates and infants were estimated. The predictive performance was acceptable with a small bias. These intermediate results should be confirmed by the inclusion of new patients.
 Yukawa et al. Journal of Clinical Pharmacy and Therapeutics 2005; 30: 159-163.