Monica Simeoni, Chao Chen
GlaxoSmithKline R&D
Objectives: Remifentanil is a selective mu-opioid receptor agonist indicated as an analgesic. Plasma concentrations of its active metabolite, remifentanil acid, during short duration treatments were characterised by a single-compartment model in some patients and by a two-compartment model in others. The exposure to both remifentanil and this metabolite during prolonged treatment remained to be established. This work aimed to quantify steady-state metabolite-to-drug ratio (metabolic ratio) for this drug by a population model, using non-steady-state data collected during highly variable adaptive dosing.
Methods: Concentrations of remifentanil and remifentanil acid were available from a three-day study with frequent sampling and a 10-day study with sparse sampling. A mixture population pharmacokinetic model was developed using the data from the three-day study. The model was assessed in terms of observed concentrations and the derived endpoint of interest (the metabolic ratio) by an individual-based evaluation method, to accommodate the high variation in dosing regimen, both among patients and over time. The data from the 10-day study were then added to the dataset; and the most probable model for patients in this study was determined using individual objective function values.
Results: The mixture model adequately described the concentrations of both remifentanil and its metabolite from both trials. The individual-based evaluation allowed informative assessment of the model despite of the highly variable dosing regimen. Generally, the kinetics of the metabolite was better described by two compartments in patients with normal or mildly impaired renal function, and by one compartment in patients with more severely impaired renal function. Preliminary results from the dense dataset suggested that the geometric mean of the metabolic ratio was 16 in the former group and 78 in the latter group.
Conclusions: The joint parent-metabolite mixture population model enabled the integration of all relevant data, to maximise analytical power and to preserve the correlation among parameters. The model can be used to simulate the distribution of the concentration profiles of both compounds in any proposed dosing regimen.
Reference: PAGE 21 () Abstr 2521 [www.page-meeting.org/?abstract=2521]
Poster: Other Modelling Applications