We are grateful for Dr. D’Urzo’s comments about our article, and agree with his conclusions about the gap between clinical trial data and how this relates to normal clinical practice. One of the purposes of our article was to highlight the conservative nature of any estimate of differences between treatment limbs with regard to rate of change of lung function or indeed symptomatic outcomes. If the study size had been much larger, it is possible that a difference in treatment might have emerged. As Dr. D’Urzo points out, this would still be a very conservative estimate of a “true” effect. The changes in lung function and indeed in symptomatic outcomes that we have seen in the ISOLDE study and other investigations we have undertaken suggest that patients cannot readily be classified into “responders” and “non-responders,” nor can these states be identified with commonly recommended tests such as acute treatment with bronchodilators or oral corticosteroids.1–2 As such, we will be more cautious than Dr. D’Urzo in suggesting that there was no benefit of treatment in those patients who withdrew from the 3 year trial while receiving fluticisone. What did appear to be the case, and which the trial was better constructed to address, was that taking the inhaled corticosteroid reduced the chances of withdrawing due to ill effects.