Tadakatsu Nakamura, Kei Ogawa, Megumi Furukawa, Shinsuke Inoue, Atsuhiro Kawaguchi
Mitsubishi Tanabe Pharma Corporation
Objectives: Model-based meta-analysis (MBMA) allows to compare the clinical outcomes with different drugs that have not been compared directly using published aggregate data from many clinical studies. By modelling the response as a parametric function of time, MBMA allows the integration of information from trials of different durations and with different sampling time-points. The aim of this analysis was to develop a MBMA model for postmenopausal hot flush 1) to characterize the placebo and drug response, and 2) to provide a quantitative comparison of drugs which are currently developed as well as used for the treatment of vasomotor-symptoms (VMS).
Methods: Literature was searched in the public databases to extract data of clinical trials on drugs for postmenopausal VMS, including estradiol (oral estrogen, transdermal estrogen), conjugated estrogens/bazedoxifene (Duavee), paroxetine, gabapentin and neurokinin 3 receptor (NK3R) antagonists (fezolinetant, pavinetant, NT-814). Included studies were double-blind, randomized controlled trials that evaluated the efficacy in the reduction of moderate to severe hot flush frequency and set the criteria of a minimum of 7 hot flushes per day at baseline. For hot flush frequency, longitudinal MBMA was developed using a placebo and drug response (exploring Emax dose-response) models incorporating covariate effects. Simulation was performed to compare the treatment effects of the drugs included in this analysis. Model development and simulation were performed using NONMEM 7.3.
Results: Thirty three trials in 32 articles, involving 11,543 women, were included in the analysis. Our developed model indicated that the placebo and drug responses for hot flushes increased in a time-dependent manner. As a placebo run-in period is generally set to reduce the placebo response in clinical study, the effect of placebo run-in period was explored and identified as a covariate for the placebo response. The estimated placebo responses were smaller in the trials with placebo run-in period than without it (-3.7 vs -5.6 as change from baseline at 12 weeks). For the effect of estrogen (both oral and transdermal), which are current standard therapy for VMS, Emax dose-response model were implemented with a common Emax value and different ED50 values, but dose-response model was not identified for other drugs. Comparing the estimated drug effects, the change from placebo of daily hot flush frequency at 12 weeks for oral estrogen (1.0 mg), transdermal estrogen (0.050 mg), Duavee, paroxetine, gabapentin and NK3R antagonists were -3.9, -4.6, -3.3, -1.5, -3.0 and -6.3 per day, respectively, and time to achieve the decrease of 2 hot flushes per day from placebo (clinically relevant change) were within 4, 3, 4, 6, 1 and 1 weeks, respectively. Change from baseline of daily hot flush frequency were also presented in time-course plot to compare the drug effects and its response rate.
Conclusions: We performed a MBMA for the efficacy of different drugs which are under development as well as used for VMS. The developed model was found to describe adequately the time course of hot flushes reduction after administration of placebo and drugs. Our model also indicated that setting placebo run-in period in clinical trials decreased the placebo response. This is the first time to implement placebo run-in period into placebo response and characterize the Emax dose-response of estrogen for hot flush in meta analysis. This analysis provided a quantitative framework for comparison of treatment effects of VMS drugs.
Reference: PAGE 28 (2019) Abstr 8991 [www.page-meeting.org/?abstract=8991]
Poster: Drug/Disease Modelling - Endocrine