Are bronchial Diverticula Indicative of Chronic Cough?
There has been recent interest in bronchial diverticula (BD) and their purported association with cigarette smoking and chronic obstructive pulmonary disease [1-3]. The pathogenesis of BD is unknown, though prior authors have hypothesised that damage from cigarette smoke was important to their development . Our objective was to determine the prevalence of BD in subjects with chronic cough, using a diagnosis of usual interstitial pneumonitis as a surrogate, with or without coincident cigarette smoking or COPD. We also compared the prevalence of BD in subjects with UIP with a historical control group of long-term cigarette smokers .
Materials and Methods
This study received institutional ethics committee approval before commencement.
All CTs of the chest from 1st July 2014 to 31st October 2016 were retrieved from our institutional radiologic information system (Agfa Qdoc™, version 6.2) using the query terms “usual interstitial pneumonitis”, “interstitial pneumonitis”, “pulmonary fibrosis” or “UIP”. Only those scans that fulfilled the CT diagnostic criteria for UIP pattern as cited by the combined American Thoracic Society (ATS) and The European Respiratory Society (ERS) statements were included . Subjects with a history of asbestos exposure and collagen vascular diseases with CT features of UIP were also included. Scans retrieved that mentioned the query terms but did not display convincing evidence for UIP were excluded from the study population. Where there were multiple scans for one patient, only the original scan was included.
All scans had been performed on a Philips 64 slice brilliance™ CT machine during a single breath hold. Imaging parameters were a tube voltage of 120-140 kVp, tube current 40-150 mAs, beam collimation 0.9 mm, gantry rotation time 500 ms, and pitch 1.2, with variations in tube voltage and current dictated by the patient’s body habitus as per the departmental protocol. Images had been reconstructed at 0.9 mm thickness with a 40% overlap in the axial plane on mediastinal and lung algorithms, then stored as such on a long-term digital archive.
The images were reviewed retrospectively by one thoracic radiologist (with ten years of experience post fellowship) and a trainee radiologist who were blinded to the clinical details of the subjects. Consensus was reached as to the presence and number of BD.
The images were read on a diagnostic station utilising 5-megapixel Dome™ monitors and commercially available software tools (Agfa Impax™ version 6.2). The images were reviewed on thin section axial and coronal multiplanar reformats on standard thoracic window levels and width. BD identified on either axial or coronal planes were considered to be present. BD were scored as per a previously published system: grade 0 when there were no BD, grade 1 = one to three BD, and grade 2 = more than three BD.1 BD that arose within 10 mm of the carina were scored as “central” (Figure 1) and those beyond 10mm as “peripheral” (Figure 2).
Figure 1. An axial 0.9 mm thick CT chest image on a subject with central BD (arrow). Note the subpleural reticulations seen in UIP.
Figure 2. A coronal 0.9 mm thick multiplanar reformat of a CT chest on a subject with peripheral BD (arrow).
The medical records of the subjects were reviewed to obtain their smoking history, presence and duration of cough and their pulmonary function tests (PFTs). Subjects who were active smokers within one year of initial diagnostic CT scan were recorded as current smokers. Those who had quit smoking for at least one year were recorded as ex-smokers and those who had never smoked were recorded as never smokers. The current and ex smokers were sub classified as long term (10 packyears or more) or short term (less than 10 pack-years). History of cough was categorised into chronic (lasting eight weeks or longer) and non-chronic (less than eight weeks). All PFTs had been performed within six months of the CT. The forced expiratory volume in one second (FEV1) and forced vital capacity (FVC) expressed as percentages of the predicted value as well as FEV1/FVC ratio were recorded.
Figure 3. A coronal 0.9 mm thick multiplanar reformat of a CT chest on a subject with central BD (black arrow), displayed as small juxtabronchial air cysts. In the periphery, there are bronchial mural corrugations (white arrow), which were not recorded as BD because the neck is broader than the fundus. Note background features of UIP.
A Fisher’s exact test was used to determine the association between the presence of BD and smoking status. In addition, the prevalence of BD was compared with a cohort of long-term smokers reported by Sverzellati et al . A Chi square test was used to determine the relationship between the presence of BD and an obstructive pattern of PFTs. The FEV1/FVC results were dichotomised into a ratio of less than 0.7, in keeping with an obstructive pattern of pulmonary disease, or 0.7 and above . A p-value of less than 0.05 was considered to be statistically significant.
Figure 4. Another example of a central BD (arrow) and changes of UIP on the visualised lung field.
The study population comprised of 140 subjects. BD were identified in 112 subjects (80%), of which 70 were male and 42 female. A total of 344 BD were observed. Of the 112 subjects with BD, 49% had grade 1 and 51% had grade 2. Overall, 212 BD (62%) were peripheral and 132 (38%) were central.
The clinical characteristics of our subjects are summarised in Table 1. The mean age of subjects with BD was 72 years (64% male) and without BD was 73 years (61.5% male). The never-smokers comprised 43% (60) of the study population, the ex-smokers were 44% (62) and the current smokers were 13% (18). BD were displayed in 80% (48/60) of never- smokers, 77% (48/62) of ex smokers and 89% (16/18) of current smokers. All of the current and the ex smokers, except 2 (8 pack-years), were long-term cigarette smokers with a mean of 33 pack-years. All the ex smokers had quit smoking at least 10 years prior. A Fisher’s exact test showed no significant difference in the prevalence of BD amongst the smokers, the ex smokers and the never smokers (Fisher’s exact= 0.497). A Chi square test showed a significant difference (p = 0.000) in the prevalence of BD between the never smokers in our study population and the long-term smokers in the cohort studied by Sverzellati et al, with significantly more subjects with BD in our UIP group (likelihood ratio = 14.22). A history of chronic cough was present in 130 (93%) of subjects. A history of cough was not recorded in the medical records of 10 subjects.
† >10 pack-years
Table 1. Characteristics of patients in our cohort according and their association with the bronchial diverticula (BD).
Of the 140 subjects, 112 (80%) had contemporaneous pulmonary function test results. Of these 112, 20 (18%) had an obstructive pattern of airways disease (defined as FEV1/FVC of less than 0.7 of the predicted as per clinical routine) . There was no relationship between obstructive airways disease and bronchial diverticula on chi square testing (p = 0.792). Our numbers were too small for a subset analysis for difference between grade 1 and grade 2 BD.
In 1976, in a paper describing the electron microscopic findings of BD, Wang and Ying observed BD in subjects with and without COPD . They described their findings of BD in stages of development from submicroscopic depressions and dilatations of bronchial mucus glands and ducts, incorporation of these depressions into diverticula, and herniations of BD through the submucosal muscular bundles. Attention has also been given to BD in the recent imaging literature [1-3].
Sverzellati et al, in their study of 503 long-term cigarette smokers recruited for the Multicentric Italian Lung Detection (MILD) project, found that BD were a frequent finding in long-term smokers and suggested an essential role of cigarette smoking in their pathogenesis . Miyara et al showed that although BD were also observed in lifelong nonsmokers, the group with BD had a higher percentage of smokers than the group without BD . In support of this, older bronchographic studies suggested that BD were seen in chronic bronchitis secondary to smoking, and presumed to represent pseudodiverticula secondary to smooth muscle hypertrophy of the airways . In contradiction, a recent publication by Higuchi et al showed that BD were not related to airflow limitation . The prevalence of BD in our subjects with airflow limitation on PFTs was no different to those without airways obstruction. In addition, the 60 never smokers with UIP in our group had significantly more BD than the long-term smokers in the population studied by Sverzellati et al .
Our results indicate that BD form in the absence of airflow limitation or cigarette smoking. Indeed, we have found a high prevalence of BD in chronic restrictive lung disease and no association with airflow limitation, acknowledging relatively small numbers with obstruction.
A history of chronic cough was recorded in two-thirds of the subjects with BD (90/130). Cough is a cardinal symptom of UIP, being present early in the disease process and almost invariably present in late disease . The cough associated with UIP is typically a dry, non-productive cough [4,10]. Colonic diverticulosis is a similar pathologic entity to BD. The pathogenesis of colonic diverticula involves intermittent abnormal elevation of intracolonic pressure . We suggest that this is analogous to the intermittent significant fluctuations in intrabronchial pressure produced by chronic cough.
Inflammatory airways disease is not a component of UIP, though remodelling of peripheral airways with traction bronchiectasis (in the absence of inflammatory change) is a late pathological feature . Regardless, BD are central airways abnormality and we did not observe BD peripheral to the segmental bronchial divisions. As such, we are doubtful that BD are secondary to the UIP disease process per se. Given the likely commonality of a high prevalence of chronic cough in our subjects and previously published cohorts, we propose that cough is an important (probably the most important) aetiological factor in the development of BD. Confirmation of our hypothesis would require further studies with higher numbers of never smokers.
The chief limitation of our study is the small sample size. In particular, the assessment of differences between current, ex and never smokers with UIP required a Fischer’s exact test and therefore the intergroup comparisons should be interpreted with some caution. The large number of controls in the historical MILD group allowed a high level of confidence in our finding that BD were more prevalent in our subjects with UIP than in previously published long-term cigarette smokers. We do not know how many of the MILD subjects had UIP, though given the low community prevalence of UIP  this is not likely to be a significant confounding factor. We could not retrieve the pulmonary function tests for all our subjects. This further limited our analysis of the association between bronchial diverticula and airways obstruction. UIP is a restrictive lung disease and we did not expect many subjects with airways obstruction. Nonetheless our results are supportive of those published by Higushi et al .
There may have been a degree of rater error as there is a gradation between serration of the bronchial wall and the presence of BD (figure 3). We recorded a diverticulum if we observed the appearance of a neck and a broader fundus, or a juxtabronchial air cyst with no discernible neck. Ultimately, the observation of BD (figure 4) was subjective and it may be that we were more sensitive than prior authors.
In summary, our study indicates that BD were present in majority of subjects (80%) with UIP and almost all (93%) of subjects with UIP had a history of chronic cough, as expected. BD were not associated with airflow limitation or smoking status, as has previously been suggested. We propose that the most important factor in the aetiology of BD is chronic cough.
1.Saccente M, Woods GL. Clinical and laboratory update on blastomycosis. Clin Microbiol Rev. 2010, 23(2): 367-381.
7.Bariola JR, Perry P, Pappas PG, Proia L, Shealey W et al. Blastomycosis of the central nervous system: a multicenter review of diagnosis and treatment in the modern era. Clin Infect Dis. 2010, 50(6): 797-804.
9.Chapman SW, Dismukes WE, Proia LA, Bradsher RW,Pappas PG et al. Clinical practice guidelines for the management of blastomycosis: 2008 update by the Infectious Diseases Society of America. Clin Infect Dis. 2008, 46: 1801-1812.