Exacerbations and intent-to-treat analyses in randomised trials

To the Editors: The recent paper by Keene et al. 1 that reviews different approaches for the analysis of exacerbation rates is valuable because of the central role of exacerbations in evaluating the benefits of drugs in chronic obstructive pulmonary disease (COPD) trials 1. The authors present reanalyses of the TRISTAN (Trial of Inhaled Steroids and Long-acting β2-agonists) and ISOLDE (Inhaled Steroids in Obstructive Lung Disease in Europe) trial data, which have been previously identified as incorrectly analysed because of the absence of consideration for between-subject variability in exacerbation rates 2. With the proper reanalyses, their conclusions from these trials remain the same; namely that inhaled corticosteroids (ICSs) with or without long-acting β-agonists reduce the frequency of …