Angela Abelo (1), Andreas Lindauer (1), Massimo Cella (2), Mirco Govoni (2), Anna Nandeuil (3), Koen Jolling (1)
1) SGS Exprimo, Mechelen, Belgium, (2) Chiesi Farmaceutici, Parma, Italy, (3) Chiesi Farmaceutici, Paris, France.
Objectives:
CHF6001 is a potent and selective phosphodiesterase-4 (PDE-4) inhibitor currently under development for treatment of chronic obstructive pulmonary disease (COPD). CHF6001 is being developed for inhalation to help overcome the well-known gastrointestinal side effects associated with this therapeutic class when given orally [1]. Plasma concentrations and Fridericia corrected QT interval (QTcF) data from 2 phase I studies in healthy volunteers at different dose levels were used to assess potential QT prolongation liability. The exposure-response relationship was investigated using observed QtcF as well as placebo- and baseline corrected QTcF (ΔQTcF, ΔΔQTcF), respectively [2-4].
Methods:
CHF6001 plasma concentrations and QTcF data (12-lead ECGs extracted from 24-hour Holter recordings just before the scheduled blood sampling in triplicate) were obtained from 2 phase I, dose escalation studies: study FIH and study Extension. Both studies were double-blind, randomized, placebo-controlled and included a single ascending dose (SAD) part and a multiple ascending dose part (MAD). In the SAD part, across the two studies, single doses ranging from 20 to 4800 µg (leading to concentrations up to 2700 pg/mL) were administered via Aerolizer® (FIH) or NEXThaler® (Extension). In the MAD part of FIH study, doses ranging from 100 to 1600 µg (leading to concentrations up to 2000 pg/mL) were administered once daily via the Aerolizer® inhaler, while in the Extension study administration was twice daily in doses of 1200-2400 µg (up to 7000 pg/mL) via the NEXThaler® inhaler. The number of QTcF observations were 1310 (613 placebo) from 100 subjects. Three approaches were used in the modelling: 1) using the observed QTcF measurements (mean of triplicate measurements) as dependent variable (DV), and estimation of baseline parameters including modeling of the circadian rhythm of QTcF over time, 2) using the placebo adjusted change from baseline QTcF (ΔΔQTcF) as DV and 3) using the change from baseline QTc (ΔQTcF) as DV. NONMEM V7.3.0 was used for all modelling.
Results:
1) Using only data from the subjects receiving placebo, a model describing the diurnal changes of QTcF was first developed. A function with three cosine terms described the data best. Next, using all data, a linear slope model was found to best describe the concentration-QTcF relationship. The slope was estimated at 0.19 (90% CI: -0.18 – 0.57) ms per fg/ml. 2) A slope model performed the best also when ΔΔQTcF was used as DV. A slope of -0.23 (90% CI: -0.71 – 0.25) ms per fg/ml was estimated. 3) A slope-intercept model was applied to the ΔQTcF data [3]. Time-after-dose, observed baseline QTcF and treatment (active/placebo) effects were added as covariates on the intercept. In addition, because of pooling 2 studies, a study effect was added on the residual error. A intercept of -1.09 (90% CI: -2.14 to -0.0485) ms was estimated. The corresponding slope was estimated at -0.58 (90% CI: -1.4 to -0.058) ms per fg/ml.
Simulations [5] with the model developed in 1), using uncertainty in the drug effect parameter, showed that the upper limit of the 90% CI of the mean QT-prolongation the mean QT-prolongation never exceeded 10 ms for concentrations up to 17500 pg/mL.
Conclusions:
The developed models adequately described the data of FIH and Extension studies, using QTcF, ΔΔQTcF and ΔQTcF as dependent variables. Similar slope estimates were obtained using the three approaches and in all cases the slope estimate was not significantly different from 0. Method 1) gave the most precise estimate of slope (narrowest confidence interval). Given the available data, CHF6001 is unlikely to show an increase of QTcF of >10ms in the dose range studied.
References:
[1] Pinner, NA et al. Roflumilast: a phosphodiesterase-4 inhibitor for the treatment of severe chronic obstructive pulmonary disease. Clin Ther. 2012 Jan;34(1):56-66.
[2] ICH E14 Guideline. The clinical evaluation of QT/QTc interval prolongation and proarrythmic potential for non-antiarrhytmic drugs. Questions & answers (R3). 2015
[3] Darpo B et al. Results from the IQ-CSRC prospective study support replacement of the thorough QT study by QT assessment in the early clinical phase. Clin Pharmacol Ther. 2015 Apr;97(4):326-35.
[4] Garnett et al. Scientific white paper on concentration_QTc modelling. J Pharmacokin Pharmacodyn. 2017
[5] Simulo software (www.exprimo.com/simulo)
Reference: PAGE 27 (2018) Abstr 8671 [www.page-meeting.org/?abstract=8671]
Poster: Drug/Disease Modelling - Safety