Pharmacokinetic and Pharmacodynamic Modelling to Characterize the Tolerability of Alternative Up-Titration Regimens of Roflumilast in Patients with Chronic Obstructive Pulmonary Disease

Abstract

Background: In the OPTIMIZE study, 4 weeks of roflumilast 250 µg once daily before escalation to the approved 500 µg once daily maintenance dose reduced treatment discontinuations and improved tolerability to roflumilast among patients with chronic obstructive pulmonary disease (COPD). In this study, we present the pharmacokinetic (PK) results and PK/pharmacodynamic (PD) modelling data from OPTIMIZE.

Methods: OPTIMIZE was a multicentre, double-blind, phase III study in which patients with severe COPD were randomized 1:1:1 to receive oral roflumilast 250 μg once daily, 500 μg every other day, or 500 μg once daily for 4 weeks, followed by 500 μg once daily for 8 weeks. A population PK (popPK) model characterized roflumilast exposure levels (total phosphodiesterase-4 inhibition [tPDE4i]). Furthermore, models characterized the percentage of patients with adverse events (AEs) of interest (PK/AE model), and time to discontinuation due to such AEs (PK/time-to-event model).

Results: The popPK model adequately described average plasma concentrations and variability from 1238 patients. The percentage of patients with AEs of interest increased with predicted tPDE4i exposure (logit scale slope 0.484; confidence interval 0.262-0.706; p = 2 × 10-5). PK/time-to-event model analysis predicted that patients receiving the 250 μg up-titration regimen had significantly lower discontinuation rates and longer time to discontinuation compared with roflumilast 500 μg every other day or 500 μg once daily (p = 0.0014).

Conclusions: In this PK/PD model, a 4-week up-titration regimen with roflumilast 250 µg once daily was found to reduce discontinuations and improve tolerability, confirming the main clinical findings of the OPTIMIZE study. However, use of this lower dose as long-term maintenance therapy may not induce sufficient phosphodiesterase-4 inhibition to exert clinical efficacy, supporting the approval of 500 µg as maintenance dose.

Trial registration: OPTIMIZE: NCT02165826; REACT: NCT01329029.

Conflict of interest statement

Funding

This study was funded by Takeda Pharmaceuticals International GmbH, Zurich, Switzerland; AstraZeneca, Cambridge, UK, are the current study sponsors.

Conflict of interest

Henrik Watz has received consulting fees and payment for lectures from Takeda and AstraZeneca. He has received support for travel to meetings for the study and manuscript preparation from AstraZeneca, and provided writing assistance to AstraZeneca. Philip Gardiner is an employee of AstraZeneca. Axel Facius was employed as pharmacometrician at Takeda Pharmaceuticals, who executed the studies reported in this manuscript. He further received travel support from AstraZeneca for the ATS2017 conference, where parts of the current manuscript were presented. Gezim Lahu was employed by Takeda, who executed the studies reported in this manuscript. Eleonora Marostica was a paid consultant for Takeda.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study. Additional informed consent was obtained from all individual participants for whom identifying information is included in this article.

Figures

Fig. 1

PopPK model for roflumilast and roflumilast N-oxide. Subscripts indicate parameters belonging to parent (p) or metabolite (m). CL clearance, KA absorption rate constant, Q intercompartmental clearance, F relative oral bioavailability, popPK population pharmacokinetic

Fig. 2

Visual predictive checks showing variability in roflumilast and roflumilast N-oxide exposures. Visual predictive checks of 500 µg OD exposures for each treatment arm for roflumilast (top panels) and roflumilast N-oxide (bottom panels) for patients receiving a 500 μg OD from all treatment arms, b 500 μg EOD (up-titration arm 2), or c 250 μg OD (up-titration arm 3). Purple line and grey area represent the median prediction and 90% prediction interval, respectively; green and red dotted lines represent median observation and 5th and 95th percentiles of observations, respectively; grey dots represent observations from OPTIMIZE. EOD every other day, OD once daily, Conc concentration

Fig. 2

Visual predictive checks showing variability in roflumilast and roflumilast N-oxide exposures. Visual predictive checks of 500 µg OD exposures for each treatment arm for roflumilast (top panels) and roflumilast N-oxide (bottom panels) for patients receiving a 500 μg OD from all treatment arms, b 500 μg EOD (up-titration arm 2), or c 250 μg OD (up-titration arm 3). Purple line and grey area represent the median prediction and 90% prediction interval, respectively; green and red dotted lines represent median observation and 5th and 95th percentiles of observations, respectively; grey dots represent observations from OPTIMIZE. EOD every other day, OD once daily, Conc concentration

Fig. 3

Estimated covariate effects on total PDE4 inhibitory activity relative to the reference population. *Race was not tested in OPTIMIZE. †Not significant in the phase I–III popPK model. Derived covariate effects on steady state tPDE4i (red) compared with previously identified covariate effects (blue). The common reference is a male, formerly smoking, White COPD patient, aged 65 years with a body weight of 70 kg. Data are expressed as geometric means and 68% ranges. popPK population pharmacokinetic, tPDE4i total phosphodiesterase-4 inhibition, i.v. intravenous, COPD chronic obstructive pulmonary disease

Fig. 4

PK/AE model response from the logistic regression for the percentage of patients with AEs of interest as a function of tPDE4i (500 µg OD). Univariate model response for each significant covariate. Mean model response (straight line), 95% CIs (shaded area), locally averaged percentage of patients with AEs (curved line). Note, variability is not directly related to the 95% CIs (approximately 95% of the local fits are not expected to fit within the shaded area, as with standard visual predictive check plots). tPDE4i total phosphodiesterase-4 inhibition, PK pharmacokinetic, AE adverse event, OD once daily, CIs confidence intervals

Fig. 5

PK/time-to-event model predicted typical response during the 4-week up-titration phase. Predicted percentage of patients not having discontinued due to adverse events of interest when taking roflumilast 250 µg OD, 500 µg OD, or 500 µg EOD (straight lines). Kaplan–Meier estimates are also shown (dotted lines), indicating good agreement between the model and the observed events. PK pharmacokinetic, AEs adverse events, OD once daily, EOD every other day