Myriam Chartoire2, Leon Aarons1, Jacklyn Smith2 3, Kayode Ogungbenro1
1Centre for Applied Pharmacokinetics Research, Manchester Pharmacy School, The University of Manchester, Oxford Road, Manchester M13 9PT, United Kingdom; 2Division of Infection Immunity and Respiratory Medicine, School of Biological Sciences, University of Manchester, Manchester, UK; 3 University Hospital of South Manchester
Introduction: The capsaicin cough challenge test has been used for many years to elucidate the mechanisms of cough and evaluate the efficacy of antitussive therapies. This test involves four consecutive inhalations of capsaicin at increasing doses to effectively induce and measure the cough response. The data obtained from the capsaicin cough challenge test have been previously analysed using a population PD model [1]. Recently, a joint model of the cough challenge test and the dropout occurring during the capsaicin cough challenge test was developed.
Objective: The aim of this study was to evaluate the design of the capsaicin cough challenge test using simulations. The simulated data allowed to test the effect of one, two, three or four inhalations of capsaicin at each dose level on the estimation of the parameters of the population PD model.
Methods: The dropout and the number of coughs at each inhalation of each dose of capsaicin in each patient were simulated in MATLAB using the joint model of the population PD and dropout model. The dropout from the capsaicin cough challenge test was simulated using a binomial distribution and the number of cough at each inhalation of capsaicin was predicted using a Poisson distribution based using an Emax model.
In total, 12 000 datasets were simulated. They consisted of sets of 4000 simulations for 150 patients, 100 patients and 50 patients. For each set of simulations, the 4000 datasets were created with a design of four inhalations of capsaicin at each dose. For 1000 of these datasets, the fourth inhalation at each dose of capsaicin was deleted in order to create the datasets of three inhalations at each dose of capsaicin. The same methodology was used to create the one and two inhalations’ datasets. The simulation of the 4000 datasets at the same time allowed to have the same number of coughs at each dose of capsaicin for each design of the capsaicin cough challenge test.
The simulated data were analysed using the population PD model. All population PD models were implemented in Monolix 2020R1 using R and the package lixoftConnectors to automate the process.
Result: For each set of simulations, the mean of the parameters for the different number of inhalations (one – four) were compared.
Overall, the models provided good estimates and good precision of the parameters. The parameters estimates were close to the original estimates obtained from the joint model used for the simulations. The interindividual variability and parameter estimates of Emax (maximum cough response evoked by capsaicin) was approximately similar for the four designs. Finally, the one inhalation design showed that the interindividual variability of ED50 (the capsaicin dose inducing half-maximal response) was lower than for the other designs.
Conclusion: Based on this work, the best design for the capsaicin cough challenge test is to give only one inhalation of capsaicin at each dose levels. This design allows to accurately describe the cough response to capsaicin and require simpler modelling compared to the other designs. Such design is also more ethical for the volunteers and easier to perform during the capsaicin cough challenge test. The next step will be to find the most appropriate doses of capsaicin to perform the capsaicin cough challenge test.
References:
- Satia, et al., Capsaicin-evoked cough responses in asthmatic patients: Evidence for airway neuronal dysfunction. J Allergy Clin Immunol, 2017. 139(3): p. 771-779 e10.
Reference: PAGE 30 (2022) Abstr 10114 [www.page-meeting.org/?abstract=10114]
Poster: Drug/Disease Modelling - Other Topics