Theresa Cain (1), Adrian Barnett (1), Masoud Jamei (1)
(1) SIMCYP Limited (a Certara Company), Sheffield, UK
Objectives: To develop a Simcyp R library package facilitating the simulation of virtual clinical trials using the Simulator via R. To use the R package to predict the compliance of Metoprolol from a patient’s single PK sampling point, and to determine the optimal time point for correctly identifying the compliance scenario in two populations of CYP2D6 Extensive Metabolisers (EM) and CYP2D6 Poor Metabolisers (PM).
Methods: We developed an R library package where Simcyp is run via R and facilitates simulating various compliance scenarios. The Barriere et al. [1] method, incorporating a Bayesian framework, was used to predict patient’s compliance after a scheduled 100 mg Metoprolol BID for 6 days. Prior in vitro and physicochemical parameters for Metoprolol and the Healthy Volunteer population were used to generate plasma concentration profiles of 500 CYP2D6 EM and PM patients. The first 10 doses were taken and the final two doses varied over five scenarios: full compliance, missing the first dose, missing the second dose, taking both doses together and missing both doses. Compliance scenarios were predicted for each patient using plasma concentrations taken at 0, 12, 24, 36, 48, 60 and 72 hours after the scheduled final dose. A ROC curve was used to determine the optimal sampling time for predicting compliance scenarios.
Results: The probability of a true positive in the EM population was greatest at the time of the final dose for all scenarios and decreased over time. In the PM population the probability of a true positive is lowest at the time of the final dose, increases rapidly by 12 hours and then remains fairly constant over the remainder of the sampling times. In both populations, the probability of a true negative is high for all sampling times. The ROC curve shows that for the EM population, the concentration taken at 12 hours is the best at predicting all compliance scenarios apart from where two doses are taken together; in this case the 0, 12 and 24 hour samples are all equally good predictors. For the PM population the optimal times varied by compliance scenario.
Conclusions: An R library package for Simcyp is developed that enables running virtual clinical trials from within R. This was used to show that the optimal time for correctly predicting compliance in the EM population is 12 hours after the final dose. However, the optimal sampling time in the PM population depends on the compliance scenario.
References:
[1] Barriere O et al., J Pharmacokinet Pharmacodyn. 2011 Jun 1;38:333-351
Reference: PAGE 24 (2015) Abstr 3538 [www.page-meeting.org/?abstract=3538]
Poster: Methodology - Other topics