Article Text
PDF
Statistics from Altmetric.com
We read with interest the paper by Dr Wedzicha and colleagues on cell and cytokine measurements in exacerbations of COPD.1 We feel there are two factors which are worthy of comment. The authors described their patients as being predominantly those with chronic bronchitis. They presented with an exacerbation that consisted of the cardinal features described by Anthonisen and colleagues—namely, combinations of increased breathlessness, increased sputum volume, and increased sputum purulence.2 In view of the fact that increased sputum purulence is a feature of exacerbations, we are surprised at the lack of increase in neutrophils seen in the exacerbations described in the paper. Even if half the patients did not have purulent sputum, we would have expected to have seen an overall increase in the number of neutrophils in the samples obtained.
In addition, the authors used sputum induction when most of the patients must have presented with spontaneous sputum production in order to have increased sputum volume or sputum purulence. Surely the use of sputum induction in such patients would lead only to dilution of the bronchial secretion obtained, and this may explain some of the negative neutrophil results. Indeed, in our own studies of approximately 140 exacerbations in a similar setting, two thirds of them were purulent in nature and were associated with increases in cytokines in the spontaneously expectorated sample. The purulent samples were associated with an increase in myeloperoxidase concentration and in neutrophil numbers seen on Gram staining. We would have expected the same findings in the paper by Wedzicha and colleagues if the exacerbations were similar.
We consider that it may be appropriate to stratify exacerbations, particularly when trying to assess the role of intervention treatments and the nature of the cytokines present.
authors' reply We appreciate the comments of Professor Stockley and Dr Hill that neutrophils should increase during COPD exacerbations and their hypothesis that this was not found in our study because of a possible dilutional effect of the induced sputum technique on lower airway secretions.
Eleven of the 37 sampled exacerbations (29.7%) in our study1-1 were associated with purulent sputum, whereas 24 (64.7%) were associated with increased sputum volume and 10 patients had no sputum production at exacerbation. This indicates that about 33% of COPD exacerbations are not associated with sputum production.
We have shown previously that the number of viable cells was greater in induced sputum than in spontaneous sputum (65% versus 41.2%, p = 0.001).1-2 In the latter study we found no difference in total or differential cell counts between spontaneous and induced sputum. There is therefore no evidence for a diluting effect of induced sputum relative to spontaneous sputum in patients with COPD. Rather, the use of the induced sputum technique allowed us to obtain standardised samples from all our patients at exacerbation, whether or not they were sputum producers.
In our study there was a tendency for patients with purulent sputum to have a greater increase in neutrophils at exacerbation (rho = 0.416, p = 0.068) but there was clearly no significant overall change in neutrophils (p = 0.771). Furthermore, there was great variability in the neutrophil counts at COPD exacerbation (IQR 1.18–4.67 × 106 cells/g sputum). We sampled our patients early in the course of the exacerbation (median of three days after onset), so a later rise in neutrophil count may not have been detected in our study.
Examination of induced sputum has been used for some years as a diagnostic technique to investigate lower airway inflammation.1-3 1-4 In asthma this technique is now well established as a relatively safe, non-invasive, repeatable, and valid method1-5 and we have shown that it is useful and safe in patients with COPD.1-2 It is unlikely that a dilutional effect of induced sputum is important in patients with COPD. The variability in neutrophil counts may reflect heterogeneity in the exacerbations sampled as well as the timing of sampling in the course of an exacerbation.