Non-CF bronchiectasis: does knowing the aetiology lead to changes in management?

Bronchiectasis was first described as “production of fetid sputum along with bronchial dilatation” 1. It is a pathological description of a disease process that has many possible causes. The characteristic features are abnormally dilated thick-walled bronchi that are inflamed and chronically infected by bacteria. It is a chronic condition, which can result in significant physical and social morbidity.

In previous large series, the majority of cases were due to extrinsic factors, especially childhood respiratory infections (severe pneumonia, pertussis, complicated measles and tuberculosis (TB)) as the cause of permanent bronchial damage. Singleton et al. 2 reviewed the case histories of 46 Alaskan native children with bronchiectasis born in the 1970s, and concluded that recurrent pneumonia was the major preceding medical condition leading to bronchial damage. In a study by Eastham et al. 3, previous pneumonia was the most common cause found in 93 cases of non-cystic fibrosis (CF) bronchiectasis. Likewise, nearly 50% of children were found to have developed bronchiectasis after TB or severe pneumonia in a review study by Karakoc et al. 4. Nowadays, with early immunisation and the widespread use of antibiotics in childhood, post-infectious damage is likely to be less relevant. The emphasis has now changed to the investigation of intrinsic defects or noninfective extrinsic insults that predispose to bronchial inflammation or infection. These include congenital defects, aspiration of irritants, immunodeficiency and mucociliary clearance problems. The standard medical treatment of underlying bronchiectasis usually involves airway clearance techniques, antibiotics, anti-inflammatory medication and bronchodilators (where appropriate). Detailed investigations are often carried out to determine the underlying cause of the condition, but there is little point unless they result in a change in management. The expense and inconvenience of an investigation may be considerable, but the value of detailed investigations into the cause of bronchiectasis has, however, not been evaluated.

High-resolution computed tomography (HRCT) has proven to be a reliable and noninvasive method for the assessment of bronchiectasis, and has largely eliminated the need for bronchography 5. HRCT can accurately diagnose bronchiectasis, localise and describe areas of parenchymal abnormality, and identify bronchiolar abnormalities and mucus plugging to the level of fifth- and sixth-order bronchi. It can also identify focal areas of air trapping as an indication of small airway disease 5, 6. It is commonly considered as the technique of choice in the diagnosis of bronchiectasis, but controversy still exists as to whether the distribution of bronchiectasis seen in different diseases is characteristic enough to suggest a specific diagnosis 7–10.

The aim of this study is three-fold: 1) to review the aetiology of cases with non-CF bronchiectasis from two tertiary paediatric respiratory units; 2) to determine how often making a specific diagnosis leads to a change in management; and 3) to assess whether the specific causes of bronchiectasis could be differentiated on the basis of the localisation of abnormalities on HRCT.

METHODS

RESULTS

A total of 136 patients were identified. There were 65 males, and the group median age was 12.1 yrs (range 3.1–18.1) at the time of the study. One hundred cases were referred for investigations by their local hospital paediatricians. An aetiology for the bronchiectasis was identified in 101 patients, whereas no cause could be established in 35. The identified causes of bronchiectasis are shown in table 1⇓. For the group as a whole, at diagnosis, the median (range) FEV₁ was 71% (15–133) and FVC was 77% (14–122). The diagnostic yield of each investigation is shown in table 2⇓. In 77 patients, the cause identified had implications for prognosis and treatment. The patients with combined variable immunodeficiency and others with hypogammaglobulinaemia or agammaglobulinaemia were given Ig replacement therapy. Patients identified to have aspiration were treated with intensive acid-suppression and prokinetic agents; some went on to have Nissen's fundoplication with/without gastrostomy. Patients identified to have primary ciliary dyskinesia (PCD) were seen by a specialist ear, nose and throat (ENT) surgeon in a joint PCD clinic, allowing appropriate management to be instigated 18.

The predominant symptoms at the time of diagnosis are listed in table 3⇓. The most common reason for referral was recurrent “chest infection”, and this was the presenting symptom in over half of the cases.

The distribution of bronchiectasis in relation to the specific aetiology is shown in table 4⇓. In patients with immunodeficiency, nine out of 46 (20%) showed widespread disease, and the commonest sites were the middle (80%) and lower lobes (91%). Bilateral disease was seen in 38 out of 47 (83%) cases. In idiopathic cases, widespread bronchiectasis occurred in 9%, with a predominant middle (71%) and lower (80%) lobe involvement, similar to the immunodeficiency pattern. No PCD patient had widespread disease and, again, there was middle and lower lobe predominance. In cases with bronchiectasis secondary to aspiration, the disease was mainly (80%) seen in the lower lobes. In cases with congenital structural malformations and bronchiectasis secondary to childhood infection, there was no predominant lobar involvement; however, the numbers in these groups were small. Single lobe involvement was more commonly seen in subjects with aspiration and idiopathic bronchiectasis, but the distribution was, once again, not specific to a particular aetiology. There was a great deal of overlap between the underlying aetiology of bronchiectasis and computed tomography (CT) distribution of the abnormalities, and no association between the modalities could be established.

The microbiological flora associated with each diagnosis are shown in table 5⇓. The commonest organisms isolated were: H. influenzae, S. pneumoniae, Pseudomonas aeruginosa, Staphylococcus aureus, and methicillin-resistant S. aureus. Two or more organisms were isolated in only a small proportion of patients, and the commonest pair of organisms isolated was H. influenzae and S. pneumoniae. Fungi (Candida albicans) were only isolated from patients with immunodeficiency. There were no significant differences in spirometry between patients with the same diagnosis but different microbial infection, or between patients with the same microbial infection but different diagnoses.

DISCUSSION

A total of 136 patients with bronchiectasis confirmed on HRCT were reviewed, and their aetiological factors were characterised. A specific cause was identified in 101 patients, but, more importantly, in 77 patients, the cause identified led to a distinct and individualised change in management. The current authors have also demonstrated that the distribution of HRCT abnormalities had no correlation with the underlying cause of bronchiectasis. It should be noted that there was quite marked airflow obstruction already present at the time of diagnosis in some of the patients. Therefore, the exclusion of CF does not mean that respiratory disease was mild.

The most common presentation was recurrent chest infection. It is difficult to know whether these subjects had severe or repeated episodes of respiratory sepsis leading to permanent bronchial damage that were not recalled, or whether they had an unidentified intrinsic impairment to host defence that rendered them more susceptible to normal childhood infections. A proportion of patients with aspiration also presented with recurrent chest infection as their initial complaint. Symptoms of bronchiectasis can be very vague and, on many occasions, they do not point to the underlying aetiology. No aetiological cause was found in 26% of the cases. This percentage of idiopathic cases in the current series is lower compared with previously published data, although it is possible that further diagnostic testing may have uncovered alternative or missed diagnoses. The decision to terminate the investigative process was made by the consultant in charge on an individual basis, in the usual clinical manner. Nikolaizik and Warner 19 reviewed 41 cases of bronchiectasis who presented with chronic productive cough, and no cause was found in 37% of their study group. Forty-eight per cent of studied subjects from a New Zealand cohort, despite extensive investigations, had no known cause for their bronchiectasis 20. This discrepancy may be related to the different referral pattern and study populations of the different series. Another finding from the present study is the number of cases with childhood respiratory infection as the predisposing cause of bronchiectasis. In the current cohort, only five cases were found to have developed bronchiectasis as a result of childhood pneumonia. This is in marked contrast to ∼70% of cases in a series from the 1960s when conditions such as measles, pertussis and TB were still prevalent 21, 22. With the widespread use of antibiotics and early childhood vaccination, intrinsic abnormalities are becoming the more important predisposing factor for non-CF bronchiectasis. Although it is not possible to be definitive without a national survey, it seems likely that post-infective bronchiectasis is less important now than it was previously. Although referral bias is possible (given the special interests of the two centres), children were generally referred with an undiagnosed respiratory problem, rather than with a known diagnosis for tertiary management. Thus, unless by chance all the post-infective patients were retained in the district hospitals, for which there is no evidence, the current authors' conclusion that post-infective bronchiectasis is a less important category seems reasonable (although unproven). In view of this, extensive investigations for a cause of bronchiectasis are definitely warranted, because, in particular, the establishment of a diagnosis has important clinical implications. Chronic cough and recurrent wheeze were common symptoms seen in the present cohort of patients and they were frequently labelled as “difficult to control asthma”. The correct diagnosis allowed the withdrawal of unnecessary and potentially dangerous medications, such as high-dose inhaled corticosteroids 23, and the implementation of airway clearance and antibiotic therapy. Rhinitis in the neonatal period and recurrent otitis media are common presenting symptoms for PCD. In a recently published case series, the mean age at diagnosis of PCD was found to be 4.4 yrs, despite a clear history of early-onset troublesome rhinitis in 76% of the cases 24. The prompt and correct diagnosis of the condition allows early implementation of appropriate therapy, which has been shown to improve prognosis 16, 25, and the introduction of fertility counselling. Furthermore, excessive and unnecessary ENT procedures are frequently performed before the diagnosis is known, leading to considerable morbidity 18. Managing these patients in a specialised PCD clinic leads to a dramatic reduction in the performance of these procedures, with no increased hearing loss or morbidity. In 77 patients from the current study, identifying the cause for bronchiectasis had led to a change in management, targeting treatment and intervention to the specific underlying cause. The current authors' findings are different from those reported by Pasteur et al. 26, who found a causal factor in 47% of adults with bronchiectasis and that, in 15%, identifying the cause had implications for prognosis and treatment. This discrepancy can be explained by different causes for adult and paediatric cases of non-CF bronchiectasis and a possible selection bias of the current cohort of patients.

Reported correlations between spirometry and other measurements of the severity of bronchiectasis are conflicting. In adults with bronchiectasis, Wong-You-Cheong et al. 27 showed that CT scores related to FEV₁. A strong correlation was also demonstrated by Edwards et al. 20 between FEV₁ and the extent of bronchiectasis, severity of bronchial wall thickening and air trapping. However, in a study involving 27 children with non-CF bronchiectasis, FEV₁ and other spirometry data were found to relate poorly to CT scores 28. The current authors were not able to comment on the relationship between lung function and severity of bronchiectasis in the present cohort of patients, as formal CT scoring was not carried out. The effects of sputum bacteriology on lung function in adult patients have been reported. Wilson et al. 29 and Hernandez et al. 30 both found that patients infected with P. aeruginosa had a greater extent of lung disease and worse lung function than uncolonised patients. In the current study, no relationships between lung function and microbial infection were demonstrated in the patients.

Whether the distribution of bronchiectasis might be sufficiently characteristic for a specific cause to be diagnosed is still controversial 7–10. Idiopathic bronchiectasis has been reported to be predominantly basal in distribution 31. Bronchiectasis secondary to PCD is often said to involve the middle lobe 32. In hypogammaglobulinaemia, bronchiectasis is shown to involve mainly the lower and middle lobes and the lingula, with no cases of isolated upper lobe involvement 33. Although certain patterns may occur more frequently in each of these examples of bronchiectasis of known aetiology, there is considerable overlap in the distribution of CT abnormalities between these diseases. In the current study, the distribution of bronchiectasis overlapped greatly among the various causes, and it was impossible to predict the specific cause simply by the distribution of the CT abnormalities. In a study by Lee et al. 8, experienced chest radiologists could accurately diagnose the cause of bronchiectasis purely on CT appearances in 45% of the cases, and the level of agreement between three radiologists had a κ-value of 0.20. Thus, the diagnosis of the various causes of bronchiectasis cannot be reliably made on the basis of the CT appearances alone.

There are some limitations to the present study. First, the data were collected at a single time point and no longitudinal findings could be used for comparison; therefore, the current authors were not able to comment on the progression of the radiological abnormalities and disease course. In addition, repeat CT would entail extra radiology, and recent evidence suggested CT abnormalities to be poor markers of exercise capacity and functional status 34. Secondly, the patients were a selected group and there was a possibility of referral bias given the special interests of the two tertiary units: immunology (Great Ormond Street Hospital for Children), and PCD (Royal Brompton Hospital). Some of the patients would not have undergone all of the investigations mentioned; for example, a patient with evidence of hypogammaglobulinaemia was unlikely to be subjected to further extensive testing. In the current series, none of the subjects had a combination of pathologies. The diagnosis of PCD was based on an examination of cilia showing abnormal beat frequency and ultrastructural abnormalities on electron microscopy. CT scanning was not part of the routine work-up, but was performed if bronchiectasis was suspected clinically. Thus, the number from the present cohort might be an underestimation of the prevalence of bronchiectasis in PCD. Likewise, this may also be true for lung damage following childhood infections, the prevalence and importance of which is hard to estimate because of likely inaccurate recall and reporting. It is possible that some of the current cases with idiopathic bronchiectasis actually had lung damage secondary to previous childhood infections. However, with the decline in incidence of these infections in the developed world, it holds true that other extrinsic insults or an underlying predisposition to respiratory infection are likely to be more important in the aetiology of bronchiectasis, and, thus, extensive investigations are warranted. Thirdly, bronchiectasis secondary to CF was not included in this study, so it cannot be determined whether CT was useful in differentiating CF bronchiectasis from bronchiectasis due to other causes. Lastly, the time period between the onset of symptoms and diagnosis was not quantified. CT could have been performed very late into the disease process where a spilling-over effect of the underlying disease may have occurred, thus resulting in the nonsignificant finding between lobar distribution of CT abnormalities and aetiology.

In summary, a large selected population of patients with bronchiectasis was reviewed, and, in 74% of cases, the aetiology could be identified. Compared with previous studies, a decreased prevalence of childhood respiratory infection was demonstrated to be a cause for the bronchiectasis. Immunodeficiency and other intrinsic abnormalities are becoming the more prevalent aetiological factors. In 77 (56%) children, identification of a cause led to a specific change in management. Although high-resolution computed tomography allows delineation of the underlying bronchial damage, the current study suggests that it is not useful in helping to identify the underlying cause. It is proposed that these patients should be thoroughly investigated, since the identification of an underlying cause will have a significant impact on management and prognosis. Non-cystic fibrosis bronchiectasis, though rare, is still an important respiratory condition with significant morbidity. Future research should be directed at providing longitudinal clinical information and serially obtained lung function parameters to ascertain the beneficial effects of correct diagnosis and early implementation of therapy.