From the authors:
We would like to thank B.J. Lipworth for his letter and interest in the SHINE study. B.J. Lipworth incorrectly describes the conclusion of the SHINE study article. Nowhere in the article is it stated that QVA149 “demonstrated clinically relevant superiority versus its respective monotherapy components in COPD”. Rather, we concluded that “dual bronchodilation with once-daily QVA149 demonstrated superior and clinically meaningful outcomes versus placebo and superiority versus treatment with a single bronchodilator” [1]. The conclusion on QVA149 versus placebo is justified by improvements in trough forced expiratory volume in 1 s (FEV1) being twice the accepted minimal clinically important difference (MCID) of 100 mL [2] and that for the transition dyspnoea index (TDI) exceeding the MCID of >1 unit improvement [3]. The improvement in the St George's Respiratory Questionnaire (SGRQ) total score (least squares mean difference versus placebo -3.99 units at week 12) fell just short of the MCID of 4 units reduction [4].
Further, we acknowledged in the discussion section of the article, that a limitation of the study is the “difficulty in evaluating the clinical significance of spirometric and other clinical end-points (e.g. TDI and SGRQ total scores) versus active (i.e. monocomponent) treatments” [1]. Current standards for MCID thresholds were proposed for comparisons versus placebo or for improvements versus baseline, but have not been confirmed to be applicable or attainable in comparisons either between active comparators, or between combination products and their monotherapy components [2].
Based on this, and as recommended in the new European Medicines Agency's Committee for Medicinal Products for Human Use guidance [5], a predefined clinically relevant mean treatment difference of 0.06 L for trough FEV1 was set in the SHINE study protocol for QVA149 monotherapy comparisons. This is consistent with published data comparing free long-acting β2-agonist (LABA)+long-acting muscarinic antagonist (LAMA) combinations with individual LABA or LAMA therapies [6, 7]. Along with the statistical significance seen for QVA149 versus all monotherapies, the observed treatment differences for trough FEV1 also met, and exceeded, the predefined clinically relevant threshold of 0.06 L (QVA149 versus indacaterol, 0.07 L, glycopyrronium, 0.09 L, and tiotropium, 0.08 L). It is also important to note that the MCID is defined as the smallest difference that patients can perceive (at the individual level) as beneficial. The focus should not only be on mean differences between treatments but, therefore, also on the number and percentage of patients who experience relevant improvement (i.e. responder analyses). Responder analyses carried out for this study also corroborate the clinical relevance of these data. A significantly higher proportion of patients on QVA149 demonstrated an improvement of >100 mL in trough FEV1 from baseline (64.3%) than all the monotherapies (indacaterol 46.2%, glycopyrronium 43.2% and tiotropium 46.6%) and placebo 18.9% (p<0.001 for all treatment comparisons).
Because no published consensus on MCID thresholds between active comparators exists, we did not predefine clinically relevant differences in TDI or SGRQ total scores when comparing QVA149 with its monotherapy components or tiotropium. The reduction in SGRQ total score in the placebo group was unusually high after 26 weeks of treatment (>7 units, two to three times the placebo effect seen in most clinical trials), which may have diminished the opportunity to observe MCID differences between QVA149 and active comparators. Despite this, QVA149 demonstrated the greatest mean reduction in SGRQ total score out of all the active treatments tested; 64% of patients achieved an improvement of ≥4 units and >50% an improvement of ≥8 units, compared with baseline values. Furthermore, a greater proportion of patients achieved the MCID in TDI focal score of >1 unit with QVA149 (68.1%), compared with indacaterol (64.6%), glycopyrronium (63.7%), tiotropium (59.2%) and placebo (57.5%). Therefore, we believe that our original summary is both accurate and justified.
Footnotes
Conflict of interest: Disclosures can be found alongside the online version of this article at www.erj.ersjournals.com
- Received September 6, 2013.
- Accepted September 25, 2013.
- ©ERS 2014