Statistics from Altmetric.com
We read with interest the paper by Niimi et al1 published in February’s edition of Thorax in which the authors identified a small number of patients with unexplained recurrent pneumonias and reduced cough reflex sensitivity to capsaicin compared with an age and sex matched control group. It does seem intuitively reasonable that the risk of aspiration pneumonia might be greater in those with diminished cough responses, and this has already been demonstrated in post-stroke patients.2
Our concerns with this study are twofold. The first is the finding that unexplained recurrent pneumonia is uncommon (seven cases over a 5 year period) whereas insensitivity to tussive stimuli is relatively common in normal healthy volunteers. Indeed, it has been a sufficiently regular occurrence that the majority of cough challenges developed have had to incorporate an in-built method of assigning a theoretical cough threshold to non-coughers. This is usually taken as either the greatest concentration inhaled or the next incremental concentration which would have been inhaled if the test had continued. The proportion of subjects who do not cough will obviously vary between different cough challenge methodologies. In a recent as yet unpublished study the proportion of non-coughing healthy volunteers in our laboratory was approximately 10% with a Mefar dosimeter based protocol3 and a maximum inhaled concentration of 3 M citric acid. Pounsford and Saunders4 found that over one quarter of women tested did not cough during their citric acid cough challenge and that these subjects had the ability to smoke cigarettes pleasurably on an intermittent occasional basis. In more recently published studies5,6 over half of the large control group of healthy individuals did not achieve C5 thresholds to their maximum challenge of 500 μM capsaicin. The discrepancy between the relative prevalence of the two findings raises concerns that recurrent pneumonias can be purely attributed to low cough reflex sensitivity per se.
This brings us to our second area of concern and offers a possible alternative explanation for their cases. Niimi et al state that they excluded patients in an immunocompromised state (diabetes mellitus, corticosteroid therapy, active malignancy, AIDS) and found that immunoglobulin levels including IgG subclasses were normal. They do not, however, seem to have excluded functional immunoglobulin deficiencies which may be diagnosed by measuring patients’ immune responses to tetanus toxoid and pneumococcal vaccination.7 This well recognised condition is a cause of recurrent bronchial sepsis which is associated with normal immunoglobulin/complement levels, as well as normal neutrophil/lymphocyte counts. It is possible that the patients described may have been suffering from undiagnosed immunodeficiency as an alternative cause for their recurrent pneumonias, and we believe that functional immunoglobulin deficiency should be looked for in all such patients.
We thank Barber et al for their comments on our paper and agree that unexplained recurrent pneumonia is not a very common condition. In fact, we could recruit only seven cases over a 5 year period at a university hospital department which specialises in respiratory infection.1
They claim that insensitivity to tussive stimuli is a relatively common finding in normal healthy volunteers on the basis of a number of previous publications2,3 and their own experience, and suggest that the recurrent pneumonias may not be entirely the result of a low cough sensitivity per se. This argument against the conclusion of our study does not, however, seem to be relevant. We used a 15 second tidal breathing method to test capsaicin cough sensitivity4 which, although it has been used at only a few institutions, has established reproducibility.5 In our study the capsaicin cough threshold that caused five or more coughs (C5) in 21 healthy controls (18 men, mean (SD) age 61 (15) years) ranged from 2.44 to 156.3 μM with a median of 19.5 μM. The highest concentration of capsaicin prepared was 625 μM. In another group of 57 younger healthy volunteers from our laboratory (50 men, mean (SD) age 28 (6) years) C5 titres ranged from 0.61 to 156.3 μM with a median of 19.5 μM. Only two subjects had a C5 titre of 156.3 μM (unpublished data). Similarly, in 160 normal volunteers investigated by Fujimura et al6 using the 15 second tidal breathing method of capsaicin sensitivity testing, C5 titres ranged from 0.49 to 1000 μM (the highest dose) with a median of 15.6 μM. Only one subject had a C5 titre of 1000 μM. We are therefore convinced that, when the 15 second tidal breathing method is used, “healthy” subjects rarely fail to respond to the highest concentrations of capsaicin. The different prevalence of non-responders in our laboratory from that of Doherty et al,2,3 who also used capsaicin, may be because they adopted a single breath method using a dosimeter.7 However, a study in 100 healthy volunteers using capsaicin and a dosimeter8 showed a distribution of C5 similar to ours. We would like to emphasise that, when the results of cough sensitivity tests are analysed, they should be cautiously compared with the results from appropriate controls obtained using exactly the same method, preferably in the same laboratory.
We did not exclude functional immunoglobulin deficiencies which might have been present in our patients. If this had been the case, however, impairment of the cough reflex might have played an additional role in the development of recurrent pneumonia.