Category: Clinical Pharmacology
Purpose: Development of effective treatment for spinal cord injury (SCI) is challenging, partly due to the substantial progression and complications that arise after the initial mechanical damage that can potentially alter pharmacokinetics (PK) of therapeutics. The North American Clinical Trials Network (NACTN) phase 1 and the ongoing Riluzole in Acute Spinal Cord Injury Study (RISCIS) phase 2 clinical trials documented that riluzole PK changes significantly over the two-week treatment period. We previously established a disease progressive model using 198 observations from 46 subjects (34 from phase I and 12 from phase II) to characterize the dynamic PK changes that take place between Day 3 and Day 14 post SCI. A one compartment with first order elimination population PK model was developed with days-post-injury and concomitant administration of oxycodone as covariates. Recognizing the potential of guidance for rational dosing adjustments in patients who suffer from acute SCI, we present a novel method for translating a population PK model into an interactive online tool. Healthcare professionals can explore the effects of time after injury and oxycodone co-administration on the PK of riluzole and simulate an appropriate dosing regimen to maintain the desired therapeutic exposure.
Methods: Our tool is developed in R using Shiny package, a user-friendly framework that allows customization of the user-interface to provide model visualization with changing outputs in response to user input controls. The R Shiny application contains two components: a user‐interface component, which controls layout and appearance, and a server component, incorporating instructions for user‐input, processing data, and output by utilizing the R language and functions. The application features a plot of concentration-time profile for multiple dosing with prebuilt widgets.
Results: The display defaults dosing regimen for riluzole orally at 100 mg loading dose BID on the first day with the first dose given within 12 hours post SCI, followed by 50 mg BID for the remaining 13 days as defined by RISCIS trial. When slider values for dose change and/or radial button for oxycodone co-treatment is selected, the concentration-time simulation is updated. Nonreactive functions used were based on population PK model developed. Population estimates of absorption rate constant (ka), clearance (CL/F), and volume of distribution (V/F) were 5.9 h-1, 72 L·h-1, and 904 L, respectively. The progressive model determines that CL/F and V/F increases by the factors of N0.27 and N0.37, respectively, where N is the days post-injury. Reactive function defines an increase in CL/F by e0.54 if patient is receiving oxycodone for pain management.
Conclusion: Understanding dose-concentration and therapeutic outcome relationships is crucial to drug development and clinical use of the therapeutics, which is especially challenging given the progressive nature of SCI and the difficulty in communicating complex pharmacometric to health practitioners. With easy access and user-friendly interface, the R Shiny application allows users to virtually explore the effects of population variability on drug exposure. This interactive population pharmacokinetic application is intended to offer the guide for future rational dosing adjustments for SCI patients if riluzole is confirmed to be an effective therapy.
Lei Wu– Research Associate professor, University of Houston, Houston, Texas
Angela Teng– Houston, Texas
Mahua Sarkar– Post Doctoral Fellow, University of Houston, Houston, Texas
Elizabeth Toups– Houston, Texas
Robert Grossman– Professor, Department of Neurosurgery, Methodist Hospital Research Institute, Houston, Texas
Diana Chow– Professor of Pharmaceutics and Director, Institute of Drug Education and Research, University of Houston, Houston, Texas