In the current study, we found that chronic cough diminished partially or completely in 59% of the patients who used inhaled β2 agonist followed by ICS. The percentage of BRC was higher than that in the previous literature [11, 14], probably because we used the bronchodilator with a higher dose. CVA was the major cause of chronic cough in our study, as previously reported in Japan [15, 16] although this observation is quite different from those in Western countries . In the current study, on the other hand, there was a distinctly different group of patients who responded to beta agonist but did not have airway hyperresponsiveness measured by the methacholine challenge test. We named this condition NHBRC, and it consisted of a significantly large part of BRC (30%) with a higher prevalence in women (Table 1). Fujimoto et al. reported a similar phenomenon in a limited number of patients who responded to a beta agonist without airway hyperresponsiveness . Although the airway reversibility test estimated by the change in FEV1 was not positive in all BRC patients, there was a difference in the flow volume parameters among BA, CVA and NHBRC. A small but significant increase was observed in PEFR after inhalation of salbutamol in BA and CVA, but not in NHBRC (Table 4). There was a significant difference in the magnitude of increase in PEFR between BA and NHBRC (Figure 2). These results suggest that bronchoconstriction or increased smooth muscle tone seem to occur less in NHBRC than in the other BRCs. On the other hand, FEFR50% and FEFR25%, which are used for estimating flow limitation in the peripheral airways, were significantly increased after salbutamol inhalation in all BRC, and the magnitude of increase among BA, CVA and NHBRC was comparable. Since cough diminished after procaterol inhalation in these patients with BRC, bronchoconstriction of peripheral airways may be the common cause of chronic cough. Although bronchoconstriction does not have a direct effect on the sensitivity of the cough receptor in healthy subjects , bronchoconstriction or increased bronchial tone itself may stimulate the afferent tussive nerve, possibly by deformation of the airway epithelium in the peripheral airways . Airway remodeling in BA and CVA has been documented , and similar remodeling could be involved in NHBRC inducing exaggerated airway wall deformation during bronchoconstriction so as to cause chronic cough.
Takemura et al., who compared classical BA and CVA, reported that 15% of CVA patients developed BA after two year observation and that these patients had a higher IgE level than CVA patients who did not develop BA . Their results indicated that the severity of atopic status may be associated with the development of classical asthma with wheezing. Since patients with NHBRC had a significantly higher IgE level as with BA and CVA, it is possible that they could develop CVA or even BA in the future.
As shown in Table 1, there was no combination of diagnoses between BRC and AC. Since ICS was used for BRC in step 1 treatment, and ICS was also effective for AC preventing this combination. Even so, the number of AC patients in our study was far less than was reported by Fujimura et al. . According to the diagnostic criteria of AC , patients with AC do not respond to beta agonist. The dose of beta agonist for diagnosis, however, is not defined. In the current study, twenty out of 33 NHBRC patients were atopic. If the patients with atopic NHBRC in our study were undertreated with β2 agonists, they could well have been diagnosed as AC. Then, the percentage of AC would have become approximately 27%, which is close to the percentage of AC in the previous study by Fujimura et al. . Non-asthmatic eosinophilic bronchitis (NAEB), which shares similar clinical characteristics with AC, is one of the common causes of chronic cough . NAEB is also similar to NHBRC with respect to the effectiveness of ICS and the lack of airway hyperresponsiveness. Since we did not measure the eosinophil counts, it is difficult to speculate whether eosinophilic airway inflammation was involved in NHBRC. On the other hand, the response to beta-agonist has not systematically been investigated in patients with NAEB. Therefore, it seems to be very important to investigate whether there are overlaps between NAEB, AC and NHBRC, since a part of patients with NAEB were reported to develop more serious conditions such as BA or airflow limitation during the few years of follow up .
Postnasal drip syndrome has been recognized as one of the major causes of chronic cough . The guideline issued by the American College of Chest Physicians suggested the use of the term “Upper airway cough syndrome (UACS)” instead of PND . In the current study, we did not categorize the patients as UACS since most of the cough was diagnosed otherwise by the systematic treatment cascade. Although PND was observed in 17-32% of BA, CVA, NHBRC, AC, and GERD patients (Table 3), the cough diminished without H1-antagonist in all groups except for AC. Five out of the 17 ACs had PND (Table 3) and could be diagnosed as having UACS, since both UACS and AC respond to the H1-antagonist, suggesting that there could be an overlap between UACS and AC. Moreover, it is not clear that PND itself is the independent cause of chronic cough since PND from allergic rhinitis or rhinosinusitis is frequently associated with airway disorders such as asthma [27, 28], and the cough diminishes after treatment of asthma. O’Hara and Jones suggested that PND due to rhinosinusitis without a coexistent chest disease is not a predominant cause of chronic cough . In the current study, all the patients with SBS had PND, as well as a productive cough with phlegm, and responded well to the treatment using clarithromycin and carbocysteine suggesting lower airway involvement. Kohno et al. classified the lower airway involvement in SBS into three categories; chronic bronchitis, bronchiectasis, and diffuse panbronchiolitis . Although it is not known why there are more SBS patients in Japan compared with Western countries, the genetic factors may partially contribute to the discrepancy .
GERD is one of the most common causes of chronic cough and is reported to cause up to 41% of it . However, Fujimura et al. reported that only 2% of chronic cough was caused by GERD in Japan . In the current study, 6.5% of chronic cough was caused by GERD alone, and 10.9% of patients had GERD in combination with other disorders (Tables 1 and 2). This result shows that the percentage of GERD is altogether similar to the results in the literature from Western countries . The prevalence of GERD in Japan is increasing since the end of the 1990s [31, 32], possibly due to the change in dietary style and to the decrease in Helicobacter pylori infection brought about by antibiotics treatment. Therefore, the discrepancy between the previous study by Fujimura et al.  and ours could be due to the increasing occurrence of GERD, since our study period was several years later than theirs.