L. Pelligand1, D. Renard2, A. Soubret2, J. N. King3, J. Elliott1, J. Mochel2
1 Royal Veterinary College, Hawkshead Lane, Hatfield, UK; 2 Integrated Quantitative Sciences, Novartis Pharmaceuticals, Werk Saint Johann 4056 Basel Switzerland.; 3 Clinical Development, Elanco Animal Health, c/o Novartis, 4002 Basel Switzerland
Objectives: In veterinary medicine, characterization of pharmacokinetic (PK) information is usually performed using a so-called 2 stages approach, thereby limiting most of the analyses to rich data. The objective of the study was to model the PK of robenacoxib in cats by pooling data from diverse sources in order to leverage the richness of the intensively sampled individuals to inform parameter estimates of the more sparsely sampled patients.
Methods: Data from 83 cats were pooled from 7 preclinical (laboratory cat) and 1 field (client-owned cat, perioperative sampling) robenacoxib PK studies. Cats received robenacoxib subcutaneously (SC) and/or intravenously (IV). Sampling was rich for 47 laboratory cats (24 SC, 9 IV, 14 both) and sparse for 36 clinical cats (SC, 1 to 2 samples per cat). The exact dose ranged from 1.6 to 2.3 mg/kg. Data from both routes were modelled simultaneously with NLMEs in Monolix 4.3.2, using a combined additive and proportional error model. Standard goodness-of-fit diagnostics, NPDEs, as well as VPCs were performed to evaluate the adequacy of the selected model. The influence of individual characteristics on population parameter estimates was assessed from the visual inspection of the full posterior distribution of the random effects.
Results: A two-compartment mammillary model with first-order absorption and elimination best described the PK of robenacoxib in blood. The precision of the final models parameters was considered satisfactory (RSE<20% for most parameters). Total body clearance was estimated to be moderate (0.518 L/kg/h) and the global extraction ratio E was 0.06. The SC bioavailability was high (83.6%) and the steady state volume of distribution was 0.280 L/kg. The absorption constant (Ka=0.85 h-1) was lower than the elimination constant of the combined model (K10=2.19 h-1), thus confirming flip-flop SC PK. None of the population characteristics, in the investigated ranges herein, was found to explain the between-subject variability observed in the present studies.
Conclusions: Joint modelling of the IV and SC routes enabled to unveil the flip-flop disposition kinetics of robenacoxib in cats. Estimates of exposures from perioperative (sparse) and conscious (rich) sampled cats did differ substantially. The use of population modelling to leverage information from densely sampled cats to estimate PK parameters of anaesthetized (sparse) patients is innovative in veterinary pharmacology.
Reference: PAGE 24 (2015) Abstr 3539 [www.page-meeting.org/?abstract=3539]
Poster: Methodology - Other topics