View Abstract

A7428 - COPDCompEx: A Novel Composite Endpoint to Accelerate Early Clinical Development of New Agents for Treatment of COPD
Author Block: C. A. Da Silva1, T. Bengtsson2, S. Peterson2, N. Karlsson3, M. Fageras4, A. Jauhiainen5; 1Respiratory, Inflammation and Autoimmunity Translational Medicine Unit - IMED Biotech Unit, AstraZeneca, Gothenburg, Sweden, 2StatMind - Innovation and Design AB, Lund, Sweden, 3Patient Centricity - Global Medical Affairs, AstraZeneca, Gothenburg, Sweden, 4Respiratory Global Medical Affairs - Global Product and Portfolio Strategy, AstraZeneca, Gothenburg, Sweden, 5Early Clinical Biometrics - IMED Biotech Unit, AstraZeneca, Gothenburg, Sweden.
Rationale: Reducing the frequency of moderate and severe COPD exacerbations is a key goal of treatment. Clinical trials using exacerbation as primary outcome are lengthy and require large sample sizes to have sufficient power to show differences between interventions, and are therefore traditionally not studied until phase III. Our objective was to establish an endpoint of mixed diary variables capturing clinically relevant COPD deteriorations that, when combined with exacerbations and drop-outs, creates a composite outcome (COPDCompEx) allowing the design of shorter clinical phase II trials. Methods: Data from eight 3-12-month trials (>9,000 subjects) evaluating the effect of budesonide/formoterol using standardized collection of COPD exacerbations and diary card variables were investigated. All available diary variables [i.e. peak expiratory flow (PEF), reliever medication use, breathlessness-cough-sputum score (BCSS), chest tightness and night-time awakening] were tested against predefined threshold values of deterioration from baseline, and slopes to assess rate and direction of change, to define diary events. A COPDCompEx event was defined as the first occurrence of either a diary event, a moderate or severe exacerbation or drop-out, and was evaluated using a Cox proportional hazards model. The performance of CompEx was assessed by comparing treatment effects as hazard ratios relative to exacerbations. Results: A COPDCompEx algorithm, based on morning PEF, evening PEF, total (daytime + nighttime) reliever medication use, COPD symptoms (for those available) and subject drop-out, reflected the treatment effect on exacerbation with the best statistical power. When censored at 3 months, 43.7% (range 24.1%-68.0%) of subjects receiving budesonide/formoterol experienced COPDCompEx events versus 12.7% (range 6.5%-30.3%) for exacerbations alone. The corresponding numbers across the comparator arms were 60.2% (range 42%-78.1%) of subjects experiencing a COPDCompEx event versus 19.9% (range 10%-42%) for exacerbations. The treatment effect for COPDCompEx at 3 months reflected the one on exacerbations with an average HR at 3 months of 0.59 (range 0.39-0.75) for exacerbations and 0.62 (range 0.49-0.72) for COPDCompEx. The increased number of events in combination with the sustained effect, resulted in a decreased sample size by at least 50% on average (range 35%-88%, except for one trial). COPDCompEx showed a similar effect profile as exacerbations over time. Conclusions: This is the first report of a composite endpoint, COPDCompEx, combining COPD moderate and severe exacerbations, events from diary card variables, and drop-outs. Using COPDCompEx will enable evaluation of exacerbation risk reduction in shorter trials involving fewer patients, therefore facilitating the accelerated development of new COPD therapies.