authors’ reply We would like to thank Dr Hart for his useful comment. We agree that it is important that the absolute numbers of cells are presented when measuring differential cell count in a biological sample. In fact, fig 1 of our paper illustrated the reduction of sputum neutrophils as absolute cell numbers. We thank Dr Hart for the table where the data have been presented as mean absolute cell counts, showing no difference in the absolute number of macrophages after treatment and confirming that the increase in the proportion of sputum macrophages following treatment is attributable to the reduced number of sputum neutrophils. However, the presentation of data as absolute cell numbers did not change the major conclusion of our article that a two month course of treatment with high dose inhaled beclomethasone dipropionate significantly reduces the sputum neutrophil cell count in patients with clinically stable, smoking related COPD.

I was very interested to read the article by Confalonieri et al published recently inThorax.2-1 It is interesting that the sputum neutrophil count was reduced after two months of treatment with inhaled beclomethasone with no parallel improvement in spirometric parameters and blood gas data. My group has recently completed a study on the effects of inhaled fluticasone (500 mg twice daily) via the Accuhaler device on 24 patients with steady state bronchiectasis in a double blind, placebo controlled manner.2-2 After eight weeks of treatment we also found a significant reduction (p<0.05) in the sputum neutrophil density and the levels of interleukin (IL)-1, IL-8, and leukotriene B4, but no parallel changes in Sao ₂ or lung function indices. There is little doubt that tracheobronchial inflammation occurs in bronchiectasis, COPD and asthma, and plays an important role in the pathogenesis of these diseases.2-3 2-4 Although inhaled steroid therapy is undoubtedly efficacious in asthma, its use in COPD has not shown any clinical benefits from the trials reported to date.2-5-2-7Similarly, little is known of the efficacy of inhaled steroid therapy in bronchiectasis despite its anti-inflammatory effects.2-2 2-8 It is possible that the clinical benefits of inhaled steroid therapy in COPD and bronchiectasis will only be shown by long term studies in large numbers of subjects in view of the more “fixed” damage in these two conditions. The similarity of the findings of Confalonieri et al and my group is exciting and should lead to further research in the use of anti-inflammatory treatment in COPD and bronchiectasis.

authors’ reply We would like to thank Dr Tsang for his interesting comment. We appreciate his finding of a similar effect of inhaled corticosteroids both on cells and inflammatory mediators in a group of patients with bronchiectasis without any parallel changes in Sao ₂ or lung function indices. We agree with Dr Tsang on the necessity of long term trials with a sufficient number of subjects to show any beneficial effect of inhaled corticosteroids on inflammatory airway diseases other than asthma. In fact, as mentioned in our paper, Stanescuet al 3-1 showed that airway obstruction as well as accelerated decline in lung function are associated with increased numbers of neutrophils in the sputum. This suggests that a reduction in airway inflammation (neutrophils) might influence the decline in lung function only over a long period of time. Further research on the effect of corticosteroids on airway inflammation could also clarify the similarities and differences in distinct airway diseases with fixed obstruction.

We read with interest the effect of inhaled corticosteroids in reducing the neutrophil count in patients with chronic obstructive pulmonary disease (COPD).4-1 This highlights the value of sputum induction as a tool in the study of airway inflammation in a diverse range of airway diseases. The authors have concentrated on the effect of beclomethasone dipropionate on neutrophilic inflammation, but we note that in both the control and treatment groups the mean sputum eosinophil count was outside the normal range of our laboratory and others (sputum eosinophils 0–2%). The authors did not comment on whether this eosinophilia was significantly different from the normal subjects they studied. Do they have any explanation for this apparently high sputum eosinophil count? Did any of the subjects have a previous history of asthma?

We have recently described a population of patients with fixed airway obstruction and a marked sputum eosinophilia,4-2 and there is some evidence that such patients respond particularly well to corticosteroids.4-3 Although there was no overall change in the sputum eosinophil count, we wonder whether some of the patients reported by Confalonieri and co-workers fit into this category and whether the effect of beclomethasone dipropionate was different in these patients.

Until we clearly establish whether sputum evidence of an eosinophilic bronchitis predicts a response to corticosteroids and determine how common it is in patients with COPD, interpretation of trials of corticosteroid therapy in COPD will remain difficult.

authors’ reply We would like to thank Drs Brightling and Pavord for their interesting comments. As stated in our article, we enrolled only patients with stable COPD, diagnosed according to a recent European Consensus Conference, and none of them had a previous history of asthma.5-1 The percentage of sputum eosinophils in the global COPD study population (34 subjects; mean (SE) 2.7 (0.7)) was not significantly different from that of the healthy subjects (16 subjects; mean (SE) 0.98 (0.2)) by the Mann-Whitney U test (p = 0.08). Indeed, if we consider the treated and control groups separately, a significant increase in the proportion of sputum eosinophils is seen in both COPD groups compared with the healthy subjects (p = 0.02).

We suggest that the sputum eosinophilia in our patients with smoking related COPD could be explained by their current smoking habit. In fact, recent experimental and clinical data seem to support the hypothesis that exposure to cigarette smoke can induce eosinophilic airway inflammation both in animals and humans.5-2 5-3

Although there was no overall change in the sputum eosinophil count after two months of treatment with beclomethasone dipropionate, we have analysed separately the seven subjects with eosinophils of >2% in the treated group. In these subjects, not only neutrophils but also sputum eosinophils decreased (from a mean (SE) of 4.5 (1.2)% to 2.0 (0.4)%) after two months of treatment, although the difference did not reach statistical significance (p = 0.06). Moreover, these subjects did not show a significant increase in FEV₁ after two months of treatment with inhaled corticosteroids (from 60.1 (5.6)% to 64.9 (4.1)% predicted).

We also analysed separately the subgroup of treated patients with COPD with sputum eosinophils <2% in order to verify the changes in sputum neutrophils after two months of treatment with inhaled beclomethasone dipropionate. These patients showed a significant reduction in both total cell and neutrophil counts after treatment. In fact, the mean difference from baseline of the total cell count (cells/ml × 10⁴) was 191 (51.8) (95% CI 68.5 to 314), and the mean difference from baseline of the neutrophils was 27 (1.7) (95% CI 22.9 to 31.1).

We are grateful to the authors of this letter for their careful consideration that provides a good insight into our paper. Nevertheless, the results of our study do not change since a reduction in sputum neutrophils also occurred after treatment with high dose inhaled beclomethasone dipropionate in the subgroup of patients with COPD without sputum eosinophilia.