Fentanyl pharmacokinetic and pharmacodynamic (PK/PD) estimation in neonates and infants using allometric and ontogeny methods
E. Encinas (1), M. Rodríguez (2), V. Vozmediano (2), J. Lukas (2), E. Suárez (1), R. Calvo (1)
(1) Department of Pharmacology, University of the Basque Country, Spain; (2) Dynakin S.L., Parque Tecnológico de Bizkaia, 801-B, 1st floor, 48160 Derio, Vizcaya. Spain.
Background: Use of fentanyl for prolonged sedation in neonates and infants has become increasingly widespread. Usually, dose schedules are extrapolated from adults, but this is questionable because dramatic age-related changes on PK/PD take place after birth.
Objectives: The aim of this study was to develop and validate a predictive PK/PD model of fentanyl for sedation in neonates and infants, based on the integration of knowledge on the drug behaviour in adults and physiological changes during development.
Methods: Two approaches were used for estimation of tricompartmental PK parameters  . A) Allometric equations based on body weight and postmenstrual age were used for all volumes of distribution (V) and clearances (Cl). B) Physiologically based ontogeny: V1 and V2 were estimated from the extracellular and total body water content, respectively. V3 was obtained from the adult ratio V1/Vss and V3/Vss (assumed to be constant at all ages). Intercompartmental clearances (CLd1 and Cld2) were related to cardiac output. Factors accounting for age-related changes on a1-glicoprotein (AAG), CYP3A4 and liver blood flow were estimated  and applied for obtaining the intrinsic and systemic Cl in each age range. Simulations on different ages and dosing protocols were performed in NONMEM using the estimated parameters. A semiparametric model  was used for PD, after scaling of the effect compartment equilibrium rate constant (Keo) according to brain perfusion.
Results: Parameters obtained by approaches A and B were compared with those described in the literature , although published data were scarce and heterogeneous. In contrast to the allometric method, physiologically based model seemed a good predictor as bias was statistically different from cero. For a representative newborn, the following parameters were obtained: 1.26, 2.63, 22.02 and 25.90 L for V1, V2, V3 and Vss, respectively; 0.03, 0.44 and 0.21 L/min for Cls, Cld1 and Cl2, respectively. PK/PD integration suggests that neonates and infants could not reach the target concentration in the effect compartment early after general schedules for sedation.
Conclusions: The developed physiologically based model satisfactorily predicts PK/PD in the studied ages, which allows proposals of new dosing regimens. However, given the important variability observed in this subpopulation, monitoring of plasma levels is highly recommended. In this sense, the present model could aid in optimizing the sampling protocol.
 Ebling W., Lee E., Stanski D. Understanding pharmacokinetics and pharmacodynamics through computer simulation: I. The comparative clinical profiles if fentanyl and alfentanil. Anesthesiology 1990;72:650-658
 Hughes M., Glass P, Jacobs J. Context-sensitive half-time in multicompartment pharmacokinetic models for intravenous anesthetic drugs. Anesthesiology 1992;76:334-341
 Johnson T., Rostami-Hodjegan A., Tucker G. Prediction of the clearance of eleven drugs and associated variability in neonates, infants and children. Clinical Pharmacokinetics 2006;45(9):931-956
 Verotta D, Beal SL, Sheiner LB. Semiparametric approach to pharmacokinetic-pharmacodynamic data. American Journal of Physiology 1989 Apr;256(4 Pt 2):R1005-10.
 Gauntlett I., Fisher D., Hertzka R., Kuhls E., Spellman M., Rudolph C. Pharmacokinetics of fentanyl in neonatal humans and lambs: effects of age. Anesthesiology 1988;69:683-687