author’s reply We thank Dr Sherwood Burge for his comments on our study. Indeed, the results of their study have been presented not only in abstract form but also in more detail.1-1 We apologise for this omission. As we stated in the discussion section of our study,1-2 Weiret al investigated the long term effect of inhaled beclomethasone 1500 μg daily on the decline in FEV₁ in a therapeutic trial. This (uncontrolled) treatment followed an original randomised controlled trial of 107 patients including a group of patients with moderately severe COPD (FEV₁ <70% predicted), comparable to our study. In their original study patients were treated with prednisone 40 mg daily, inhaled beclomethasone 1500 μg daily, and placebo in a crossover design to assess the short term steroid response (defined as an increase in FEV₁, FVC or PEF of ⩾20% from baseline). Seventy four patients were reassessed approximately two and five years after this original trial. Thirty two patients received beclomethasone only during the second follow up period, the remaining patients were treated with beclomethasone during the whole follow up period. In the former group of 32 patients the authors showed that the decline in FEV₁ decreased significantly when the first follow up period (without the use of inhaled beclomethasone) was compared with the second follow up period (during treatment with inhaled beclomethasone). This result, although not achieved in a randomised controlled design and therefore excluded from our meta-analysis, supports our finding of a significant two year treatment effect of inhaled beclomethasone or budesonide in a daily dose of 1500/1600 μg compared with placebo of +0.034 l/year.1-2

Dr Sherwood Burge also pointed to the absence of a relationship between the response to the initial steroid trial and the subsequent decline in FEV₁ in their study. Indeed, an interesting finding of their study was that, although the majority (75%) of the original 23 steroid responders had been treated with beclomethasone during the whole follow up period, the mean (SE) decline in FEV₁ of 51.5 (10.3) ml/year did not differ from the 51 non-responders of which only 50% were treated with beclomethasone during the whole follow up period (53.1 (8.7) ml/year). The authors therefore concluded that the acute steroid response did not seem to be an adequate predictor of the long term response to inhaled steroids in patients with moderate to severe COPD. In our meta-analysis1-2 only patients in the study by Renkema et al had been treated with a course of steroids in a dose of 40 mg daily during the eight days preceding the trial in order to assess steroid responsiveness (response as defined by an increase in FEV₁ of ⩾20% from baseline). In their study only three steroid responders (of 58 patients with severe COPD) were identified.1-3 In two studies in patients with mild to moderate COPD the presence of “asthmatic features” such as allergy and reversibility of obstruction resulted in a better long term treatment effect with inhaled corticosteroids.1-4 1-5 In our meta-analysis we did not find this relationship.1-2 This result could be explained by our selection of subjects with strictly diagnosed non-asthmatic and irreversible COPD, comparable to the subjects in the study of Weiret al. Forthcoming data of the ISOLDE trial which was initiated by Dr Sherwood Burge (ISOLDE is a three year randomised placebo controlled trial in 990 patients with severe COPD treated with 1000 μg fluticasone daily) may give a more definite answer to the question of the predictive value of a short course of oral prednisone (or other asthmatic features such as reversibility and allergy) on the clinical effects of long term treatment of patients with moderately severe COPD with inhaled corticosteroids.