INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bronchiectasis usually presents with recurrent lower respiratory tract infection and chronic mucopurulent sputum production. It is a pathologic description of a disease process that has a number of possible causes. The characteristic features are abnormally dilated thick-walled bronchi that are inflamed and colonized by bacteria. Symptoms include chronic cough, mucopurulent sputum production, hemoptysis, breathlessness, and tiredness. The incidence is perceived to have declined over recent decades, but significant numbers of patients continue to present to respiratory physicians (1). Sometimes regarded as a condition in which extensive investigation is unnecessary or unlikely to lead to treatable causes, the aim of this study was to characterize the underlying causative factors in a population of patients with bronchiectasis referred to a specialist clinic.

In the past, the emphasis has been on the role of extrinsic factors, particularly childhood respiratory infections (pneumonia, pertussis, complicated measles, and tuberculosis) as the cause of permanent bronchial damage. Postinfectious damage is likely to be less relevant with early immunization and widespread use of antibiotics in childhood. The emphasis has changed to the investigation of intrinsic defects or noninfectious extrinsic insults that predispose to bronchial inflammation or infection. These include congenital defects, mucociliary clearance defects, aspiration of irritants, and allergic bronchopulmonary aspergillosis (ABPA). In some patients, bronchiectasis is due to a primary immunodeficiency. As well as panhypogammaglobulinemia, more subtle defects of the humoral system such as antibody subclass deficiency and antibody production defects have been associated with pulmonary sepsis (2). Defects of neutrophil adhesion, respiratory burst, and chemotaxis are rare but well-recognized causes of bronchial sepsis and bronchiectasis (3) and may present for the first time in adult life (4, 5). An association has also been noted with conditions that predispose to bronchial inflammation such as rheumatoid arthritis and other connective tissue disorders, ulcerative colitis, and possibly ₁-antitrypsin deficiency. Identifying the underlying cause of bronchiectasis can have major implications for management.

Diagnosing bronchiectasis has become significantly easier with the advent of high resolution computed tomography (HRCT), which has proved to be a highly sensitive noninvasive technique for demonstrating bronchiectatic change in the airways. Using this technique to identify a population of patients with bronchiectasis, we then attempted to identify any underlying causes or associations suspecting that this might be possible in a significant proportion of patients. Particular emphasis was placed on the possible role of inherited, inflammatory, and immunologic factors, including humoral and neutrophil function. Previous studies of immune defects in bronchiectasis have concentrated on the role of antibody deficiency but have not investigated functional humoral responses as thoroughly (6). Disorders of neutrophil function have not previously been investigated in a large population of patients with bronchiectasis, and although severe neutrophil function disorders tend to present in early childhood, it is not known whether less severe defects are a significant cause of bronchiectasis presenting in adult life. Additionally, the prevalence of defects in nonimmunologic aspects of lung defense such as ciliary function have not been reported in a large population.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

One hundred ninety-three consecutive patients, in whom the diagnosis of bronchiectasis was known or suspected on the basis of chronic mucopurulent sputum production, were referred for investigation at the Lung Defence Clinic at Papworth Hospital over a 3-yr period from 1995 to 1997. HRCT identified 150 patients with bronchiectasis, and this group formed the study population. There were 142 referrals from respiratory physicians, four from general practitioners, and four from a lung transplantation unit (prior to transplantation). Three patients with bronchiectasis known to be secondary to panhypogammaglobulinemia at the time of referral were not included in the study (one childhood onset and two secondary to chronic lymphatic leukemia). Approval was obtained from the local ethics committee.

Clinical Assessment

The following were established: age at onset of chronic mucopurulent sputum production, history of persistent purulent rhinosinusitis, smoking history, previous lobectomy or pneumonectomy, history of childhood respiratory infections (pertussis, pneumonia, or measles), or pulmonary tuberculosis (including atypical Mycobacterium infection), predisposition to nonrespiratory infections (indicating possible immune deficiency), atopic or asthma, connective tissue disorders, symptoms of gastroesophageal reflux, infertility, family history of immunodeficiency or pulmonary sepsis, and Young's syndrome (bronchiectasis with sinusitis, obstructive azoospermia, normal ciliary function, and no evidence of other organ system disease that would suggest cystic fibrosis). Risk factors for HIV or other features suggesting this diagnosis (e.g., mucocutaneous candidiasis) were assessed. FEV₁ and FVC were measured on each visit according to American Thoracic Society guidelines (7), and the best results over 1 yr were recorded. For analysis of lung function, patients who had had lung resection were excluded (three men and 22 women).

Radiology

In 135 patients, HRCT scans were performed by one of two specialist respiratory radiologists (CDF, RAC) at Addenbrooke's and Papworth Hospitals, and in 15 cases HRCT was performed at other centers. Images were obtained using 1-mm collimation at 10 mm intervals in full inspiration. Bronchiectasis was deemed to be present if there was evidence of bronchial dilatation (internal bronchial diameter greater than the diameter of the accompanying pulmonary artery) and a lack of bronchial tapering on sequential slices. The bronchi were evaluated on a lobar basis (regarding the lingula as a separate lobe) and the severity of bronchiectasis graded from 1 to 3 according to established criteria (CDF) (8). The maximum possible score for the six lobes was 18.

Microbiology

Sputum specimens were obtained on the patient's first visit and at three-monthly intervals (while not receiving antibiotic treatment whenever possible). Bacteria were cultured on blood and heated blood ("chocolate") agar, mannitol salt agar, CLED, and Sabouraud's agar plates according to standard procedures (9). Sputum was cultured for mycobacteria only if there was no response to standard antibiotic treatment. A patient was considered colonized with a species if it had been cultured on two or more occasions, at least 3 mo apart, in a 1-yr period.

Inherited Factors: CF Genetics, ₁-Antitrypsin Level and Phenotype, and Ciliary Function

Cystic fibrosis (CF) gene mutation analysis was performed by the Molecular Genetics Laboratory, Addenbrooke's Hospital, Cambridge, after counseling and obtaining consent. Mutations screened for were: F508, G85E, R117H, 621+1G>T, G551D, R553X, G542X, N1303X, and I507. These account for 86% of mutant alleles found locally. The finding of one or more CF gene mutations led to measurement of equivalent sweat sodium chloride concentration using the Sweat-Chek sweat conductivity analyzer (Wescor, Inc., Utah) (greater than 80 mmol/L defined as abnormal). Serum ₁-antitrypsin level and phenotype were determined. The function of cilia on nasal epithelial brushings was assessed using a light microscopy technique (10). In the absence of rapid, coordinated ciliary beating, the test was repeated on a second occasion, after treating rhinitis if necessary. If still abnormal, a brushing was sent for electron microscopy.

ABPA and Sensitization to Aspergillus fumigatus

Specific Aspergillus fumegatus (Asp f ) IgE levels were measured using the Pharmacia CAP system and were assigned to Class 0 to 6 accordingly (Pharmacia, Uppsala, Sweden). In sensitized subjects (Class 1 or above), assessment was made of total serum IgE, eosinophil count, Asp f skin prick test (SPT), Asp f IgG precipitins, history of atopy or asthma, the presence of Asp f in the sputum, and proximal and/or upper lobe HRCT changes. IgE greater than 170 kU/L and eosinophils greater than 0.4 × 10⁹/ml were considered raised. These investigations were used to classify patients as either probable ABPA or simply sensitized to Asp f.

Immune Function

Humoral immunity (immunoglobulin levels and integrity of antibody production) and neutrophil function (adhesion molecule expression, respiratory burst, and chemotaxis) were assessed. Serum total IgM, IgA, and IgG and IgG subclass levels (IgG₁, IgG₂, IgG₃, IgG₄) were measured in all patients using a radioimmunodiffusion assay (Binding Site, Birmingham, UK). The IgG₄ assay did not, however, allow detection of IgG₄ deficiency as the lower limit of the normal range was below the limit of detection of the assay. Antibody production was assessed by measuring the total serum IgG antibody response to the Pneumovax II vaccine (Pasteur Mérieux, Berkshire, UK) using an ELISA based on previous methods (11). The standard curve was derived from pooled heat-inactivated serum from 72 healthy subjects. Baseline serum levels of total IgG against the vaccine were measured for each patient at first clinic visit. For patients who had antibody levels below the 25th percentile of the normal range, test immunization using 0.5 ml of Pneumovax II intramuscularly was performed, and postimmunization serum obtained between 2 and 4 wk later. A twofold rise in antibody titer was judged to be a satisfactory response (11).

Neutrophil function tests were performed and repeated if necessary to obtain values during chronic stable state. All tests on patients were run concurrently with a sample from a healthy volunteer to ensure the validity of each assay. Neutrophil surface adhesion molecule immunophenotyping was performed using flow cytometry of cells labeled with monoclonal antibodies against CD11b (Harlan Seralab, Sussex, UK) and Sialyl Lewis x (Becton Dickinson, Oxford, UK) and using an Epics XL flow cytometer (Coulter, Luton, UK). A normal result was defined as a level of expression greater than or equal to 80% of that on neutrophils from a healthy person measured on the same day and within the normal range for the laboratory. Neutrophils for respiratory burst and chemotaxis assays were separated from whole blood (12). Measurement of the respiratory burst was with a chemiluminescence-based assay (13): Neutrophils (0.5 × 10⁶/ml) were stimulated with 0.2 µM fMLP (N-formyl-Met-Leu-Phe) (Sigma Chemical, Poole, Dorset, UK) in the presence of 0.2 mM luminol (Sigma) and integral luminescence measured over a 40-min period on a Lumicon chemiluminometer (Hamilton Bonaduz AG, Chur, Switzerland). A normal range for the laboratory was determined using a population of healthy control subjects. Chemotaxis was measured by migration under sterile agarose in response to 10⁷ M fMLP (14). Neutrophils (1 × 10⁶/ml) were incubated at 37° C for 2 h in a well in an agarose gel, and the numbers of cells migrating toward a well with 10⁷ M fMLP counted using phase contrast microscopy. On the basis of studies of a group of 50 healthy control subjects, a count of 50 neutrophils per high power field was deemed a satisfactory chemotactic response.

Statistical Analysis

Groups were compared using the unpaired t test when a normal distribution of values was found within the study population. For HRCT scores, age at onset and duration of symptoms and the Mann Whitney U test were used as data were not normally distributed. Frequencies of CF mutations in patients with ABPA and ₁-antitrypsin phenotypes were analyzed using the chi-square test. SPSS 6.1 (SPSS, Chicago, IL) was used throughout.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Demographic Data, Lung Function, Radiology, and Microbiology

The sex, age at presentation, age at onset of symptoms, HRCT scores, and lung function are shown in Table 1 and Figure 1. Female patients outnumbered male patients and were slightly older at presentation; 98% of patients were white, reflecting the East Anglian population. There were two sibling pairs, and three other patients had a history of bronchiectasis affecting a sibling. The age of onset of symptoms (median, 14 yr) showed a wide and skewed distribution (Figure 1) but no significant sex difference. No patient was known to be HIV positive, and it was not felt necessary to test any patient based on clinical assessment of risk. Lung function tests showed mild or moderate airflow obstruction in most cases. HRCT scores showed a median bronchial dilatation score of 4 (range, 1 to 17) with male and female patients affected similarly (21 patients had changes in one lobe only). No patient had features suggestive of atypical mycobacterial infection (cavitation, nodules, or consolidation). Sputum culture results are shown in Table 2. Hemophilus influenzae and Pseudomonas aeruginosa were the most frequently isolated organisms. In 23% of patients, no organism was isolated despite repeated cultures while not receiving antibiotics. Staphylococcus aureus was isolated from sputum of three of the four patients with CF. No mycobacteria were isolated. Persistent colonization was also most commonly with P. aeruginosa (24% of total patients) and H. influenzae (17%). Patients colonized with P. aeruginosa had a lower FEV₁ (mean, 60.4% predicted) compared with those not colonized (78.1%) (p = 0.001). There was no significant age difference between the two groups, but there was a greater duration of symptoms in those colonized with P. aeruginosa (median, 45.5 yr) compared with noncolonized (26.0 yr) (p = 0.001).

View larger version (15K): [in this window] [in a new window]   Figure 1.   The age at onset of mucopurulent sputum production is shown. Patients are grouped within decades according to when symptoms began. The x-axis labels are the middle age of each decade.

Clinical Assessment

Sixty-four patients (43%) had a history of pneumonia (40, 27%), pertussis (32, 21%), measles (20, 13%), or tuberculosis (3, 2%) which predated the onset of lower respiratory tract infections. Twenty of these had a history of chronic rhinosinusitis, indicating a generalized susceptibility to respiratory tract infection, and lower respiratory tract infections are unlikely to be the primary cause of bronchiectasis in these cases. Early childhood infection was a possible causative factor in the remaining 44 (29%). In 21 of these, another cause for bronchiectasis was also identified. Five patients (3%) were shown to have Young's syndrome (15) (ciliary function was normal in all five cases and none had CF gene mutations). This syndrome has been linked to mercury poisoning (Pink disease) (16), present in the history of one patient. Aspiration was felt to be relevant in five patients (3%). One had a history of inhalation of an industrial solvent, one had aspiration of gastric contents postanaesthetic, and three had chronic gastroesophageal reflux predating infections and with improvement of symptoms after receiving antireflux treatment. One patient had diffuse panbronchiolitis (on open lung biopsy) with associated bronchiectasis. One had tracheobronchomegaly (Mounier Kuhn syndrome) (17), a congenital deficiency of cartilage. One had primary lymphoedema with bilateral pleural effusions but without dystrophic nails. Four patients had rheumatoid arthritis (three seronegative), and respiratory symptoms began in childhood in all four. Two patients had ulcerative colitis. Seventy-one patients (47%) had a history of rhinosinusitis, 31 (21%) of whom had required upper respiratory tract surgery (in the 13 who could recollect details this was endoscopic sinus surgery in 12 and a Caldwell-Luc procedure in one).

Inherited Factors

One hundred forty-six patients consented to CF genetic analysis, and 11 were found to have one or more CF mutations, four of whom were felt likely to have CF (Table 3). One patient had two CF gene mutations and a positive sweat test, one patient had one mutation with a positive sweat test, and two had one mutation with a borderline raised sweat test and phenotypic features consistent with CF (upper lobe bronchiectasis in one and infertility in another). All four were pancreatic-sufficient on clinical assessment and have not needed pancreatic enzyme supplementation. The seven remaining patients we presume are carriers as they have none of the other phenotypic features of CF and (except in the two cases in which it was not determined) have a normal sweat test. In two cases, it was not possible to perform a satisfactory sweat test because an insufficient volume of sweat was obtained despite repeated attempts. The three R117H carriers were fertile.

The frequency of ₁-antitrypsin phenotypes is shown in Table 3. There were no Pi-ZZ patients. The frequency of phenotypes did not vary significantly from those expected in the UK population (18) (p = 0.17). In all but two patients, the serum AAT level was above the lower limit of the normal range of the testing laboratory (0.9 to 1.8 g/L). Two Pi-MZ patients had levels of 0.7 and 0.85 g/L. Three patients were identified with ciliary defects. One had no ciliary movement, bronchiectasis, sinusitis, or situs invertus (Kartagener's syndrome). Two had absent or uncoordinated ciliary movement with ultrastructural abnormalities on electron microscopy. One of these and two other patients had a history of childhood mercury poisoning (Pink disease), known to affect ciliary function.

ABPA

Twenty-eight (19%) of patients had evidence of sensitization to Asp f. In 11 patients it was felt ABPA was likely (Table 4) as they were atopic with SPT positive for Asp f and either very high total IgE and Asp-f-specific IgE (Class 4 or above) or lower levels of Asp-f-specific IgE (Classes 1 to 3) with other supportive evidence such as proximal or predominantly upper lobe bronchiectasis on HRCT, positive IgG antibodies, or sputum culture. Contrary to previous reports, which have suggested serum IgG precipitins are positive in 60% of cases (19), there were only two patients with positive results. Seventeen patients had evidence of sensitization to Asp f without other evidence of ABPA. Thirteen were atopic (seven asthmatic), one had CF, and three had no evidence of atopy. Two patients with ABPA were CF gene mutation carriers (one F508 and one R117H), but this frequency was not significantly different from that expected in the general population (p = 0.2).

Assessment of Immunologic Function

Immunoglobulin and IgG subclass deficiencies together with antibody responses to Pneumovax II are shown in Table 5. No patients with panhypogammaglobulinemia were identified. One patient was found to have common variable immunodeficiency (CVID) with low IgG (2.6 g/L) and IgA (< 0.3 g/L) levels, low IgG₁ (1.1 g/L) and IgG₂ (0.34 g/L) levels and normal IgM, IgG₃ and IgG₄ levels. One patient had IgM deficiency (< 0.03 g/L), a well recognized immunodeficiency (20). Six patients had isolated IgG subclass deficiency. One of these had severe IgG₂ deficiency (< 0.3 g/L) with evidence of defective antibody production in addition. The remaining five patients had borderline low levels (IgG₁, 2.87 g/L; IgG₂, 1.18 and 1.08 g/L, IgG₃, 0.15 and 0.12 g/L), but still within the range found in a survey of healthy subjects without bronchial sepsis (21). Three of these had evidence of defective antibody production and were classified as immunodeficient. Three selective IgA-deficient patients (< 0.5 g/L) were identified. This occurs in 0.1 to 0.2% of the population (22) and is not necessarily indicative of immunodeficiency. One was judged to have a deficit of humoral function as there was evidence of defective antibody production. Of the 139 with normal levels of immunoglobulins, 35 had baseline low antipneumococcal titers (below the 25th percentile). Of the 29 who consented to test immunization, six (4%) had evidence of isolated defective antibody production. All six, in fact, had baseline levels below the 10th percentile. A total 12 (8%) patients had evidence of humoral immune deficiency.

Neutrophil function was assessed in 148 of the 150 patients (one patient died, one lost to follow-up). Neutrophil adhesion markers (Sialyl Lewis X and CD11b) were normal in all 148 patients, making a defect in neutrophil adhesion in this group of bronchiectasis patients unlikely. Two patients were found to have a persistently depressed neutrophil respiratory burst. One was atopic, and the level of depression was only slightly below that of control neutrophils. Atopy is associated with depressed function of neutrophils in vitro, and so a primary defect of neutrophil function is unlikely. The second patient had defective neutrophil respiratory burst (NRB) on repeated testing in the absence of atopy or active infection. A nitroblue tetrazolium test was normal, ruling out chronic granulomatous disease and leaving the possibility of a subtle defect in the NRB. Three patients had defective neutrophil chemotaxis, in two atopic patients only slightly below the normal range and in one there was no secondary cause.

Summary of Causes of Bronchiectasis

A summary of investigations into the cause of bronchiectasis is shown in Table 6. One or more cause was identified in 70 patients (47%), whereas no cause could be established in 80 (53%). In 22 patients (15%), the cause identified had implications for prognosis and treatment. The patients with CF were referred to a specialist clinic, the patient with CVID was commenced on immunoglobulin replacement therapy, and the patients with ABPA were given a trial of oral steroid treatment, and aspiration was treated with intensive acid-suppression. No correlation was found between age at diagnosis or age at which symptoms started with a treatable cause of bronchiectasis.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have reviewed 150 patients with bronchiectasis confirmed on HRCT and characterized the manifestations of disease and causative factors. The patients studied were a selected group with an established or suspected diagnosis of bronchiectasis referred largely by respiratory physicians and excluded patients already known to have panhypogammaglobulinemia. Therefore, this study does not provide data on the prevalence of bronchiectasis. The patients were drawn from a population base that is northern European, predominantly white, and has a low incidence of HIV and tuberculosis, which must be considered when extrapolating the results to other groups. In keeping with most (1, 16, 23, 24), but not all (25), previous studies in bronchiectasis, we found a predominance of female subjects. The similar age distribution of male and female subjects together with the fact that HRCT scores and measures of lung function were almost identical in the two sexes, suggests that women have an increased predisposition to develop bronchiectasis, but that lung damage progresses at the same rate in both sexes. Respiratory tract sepsis began at a young age in many40% were younger than 10 yr of age at the onset of symptoms. No cause was identified in 60% of these, and it is difficult to know whether these subjects had unrecollected severe or recurrent episodes of respiratory sepsis leading to permanent bronchial damage or whether they represent a group with an intrinsic impairment of host defense that render them more susceptible to the normal childhood and adult infections.

Sputum microbiology revealed a spectrum of organisms that are found commonly in bronchiectasis, with P. aeruginosa and H. influenzae predominating. The finding of S. aureus in patients with bronchiectasis should alert the clinician to the possibility of CF and lead to sweat testing and/or CF genetic analysis. Colonization with P. aeruginosa correlated strongly with duration of symptoms and worse lung function, in keeping with the finding that P. aeruginosa occurs in patients with poorer lung function, as well as leading to a more rapid decline in spirometry (26). We did not routinely screen for HIV or atypical mycobacteria, but both these possibilities should always be considered, especially if clinical features or typical HRCT changes are present as specific treatment may halt the progression of bronchiectasis.

The importance of lung damage occurring after pneumonia, pertussis, measles, or tuberculosis as a cause of bronchiectasis, though often cited (16, 24, 27, 28), is hard to estimate because of inaccurate recall and reporting and the ability of most people to make a full recovery without lasting pulmonary symptoms after these infections (29). We have made the assumption for the purposes of definition that chronic upper respiratory tract symptoms (rhinosinusitis) reflect an underlying airway or immunologic disorder. If previous severe chest infection was the cause of bronchiectasis, this cannot explain concomitant upper respiratory tract sepsis. Acute lower respiratory tract infection is unlikely to be the primary reason for the development of bronchiectasis and sinusitis in these patients, and they were excluded from the estimate of postinfectious bronchiectasis. Chronic rhinosinusitis was present in 31% of those with a history of early pneumonia, measles, or pertussis, suggesting that these infections do not lead to bronchiectasis in more than roughly two thirds of cases. The figure of 29% is still likely to be an overestimate as many of these will not have suffered permanent damage (29). With the decline in incidence of these infections, other extrinsic insults or an underlying predisposition to respiratory infection are likely to be more important in the etiology of bronchiectasis.

Late presentation of CF has been described, particularly with mutations associated with mild lung disease and pancreatic sufficiency (30). Making the diagnosis is important as it has prognostic and treatment implications, and genetic counseling of the patient or their family may be needed. Two other European studies have investigated CF in patients with bronchiectasis. A study of 38 cases revealed a 5% incidence of CF based on elevated sweat chloride levels (31) and another of 32 cases found 6% homozygous and 15% heterozygous for CF gene mutations (32). To our knowledge, our study is the largest survey of CF gene mutations in bronchiectasis with four cases (3%) identified though the high incidence of CF mutations in the United Kingdom compared with nonwhite and some other white populations means the frequency may not be so high in other population groups. Although sweat testing would have identified all cases in our study, genetic testing may be necessary to be certain that rare mutations associated with a normal sweat test are not missed. Of less clear-cut importance is the role of ₁-antitrypsin deficiency as a cause of bronchiectasis. Case reports have shown that bronchiectasis occurs in some persons with the Pi-ZZ phenotype (severe deficiency), but there were alternative causes for bronchiectasis in many of these cases, and thorough exclusion of other causes was not always performed (33). Our findings of no Pi-ZZ persons and no significant increase in the frequency of partial deficiency phenotypes (Pi-MS and Pi-MZ) fail to support any role for ₁-antitrypsin deficiency in the etiology of bronchiectasis and is in keeping with some previous studies (31, 34).

Recurrent respiratory tract infections are the most common presentation of primary immunodeficiency, and in milder humoral and neutrophil disorders this may be the only feature (3, 35, 36). Humoral immunodeficiency (antibody deficiency and/or a defective antibody response) was detected in 8% of cases. The identification of one case of severe IgG and IgM deficiency (CVID) led to treatment with immunoglobulin replacement therapy (35). The relevance of isolated IgG subclass deficiency is less clear-cut as, although linked to respiratory sepsis (36), normal ranges for each subclass are difficult to define, and severe or even complete absence of a subclass has been found in healthy subjects (36, 37). We only ascribed severe IgG subclass deficiency to values that fell outside the entire range found in healthy persons, and the low incidence of significant IgG subclass 1, 2, or 3 deficiency in our study (1 patient, < 1%) is similar to that found in others using the same criteria (2%) (21). Another study using less stringent control data reported a higher incidence of IgG subclass deficiency (48%) (38). Patients found to have only borderline low IgG subclass levels were not classified as immunodeficient unless they had evidence of a functional defect as well, determined by a defective response to immunization. The inability to produce a satisfactory response to immunization with polysaccharide antigen in the presence of normal serum immunoglobulin levels of all isotypes is another indicator of primary immunodeficiency (36, 39). Studies show that 90 to 100% of healthy subjects mount a twofold or greater response to immunization with the 23 valent Pneumovax II (11). We assessed the response to Pneumovax II in those with baseline levels below the 25th percentile on the basis that levels above this indicate a satisfactory response to pneumococcal antigen at some point previously. Defective antibody production accounted for nearly half those with primary immunodeficiency, emphasizing the need to include functional studies, as well as serum immunoglobulin and IgG subclass levels in the investigation of defects in humoral immunity.

Although commonly presenting in childhood if the defect is severe, defective neutrophil function manifests at varying ages and with varying severity according to the nature of the genetic defect responsible (40). In vitro tests of peripheral blood neutrophils are sensitive in detecting defects in adhesion markers, NRB, and chemotaxis, but caution is needed in their interpretation as a suboptimal response may be due not only to a primary defect but also to the effects of circulating inflammatory mediators released in infected or atopic patients (41). There were no patients with persistent defects in expression of CD11b and SLe^x, making deficiency of these markers unlikely to be a cause of bronchiectasis in this population. The NRB was defective in two patients, but in one it was borderline and could have been explained by severe asthma and ABPA. In one 15-yr-old male subject, the NRB was persistently defective. A severe defect such as chronic granulomatous disease was excluded, raising the possibility of a previously unrecognized defect of NRB that had predisposed to the development of bronchiectasis. Neutrophil chemotaxis was defective in one subject with possible causal relevance.

ABPA occurs in asthmatic or atopic persons, accounting for around 10% of cases of bronchiectasis (42) and may also complicate CF. It is important to identify as treatment with corticosteroids may significantly improve lung function and symptoms. Evidence of sensitization to Asp f with either positive skin prick test or raised specific IgE is a prerequisite for the diagnosis and was present in 19% of the study population. The diagnostic difficulty is distinguishing those who are simply sensitized (as is the case in 5 to 10% of asthmatics) (43) from those in whom the allergic response to Asp f is contributing to bronchial damage. The titer of total IgE and specific anti-Asp f IgE is higher in patients with ABPA than in sensitised asthmatics (44). Pronounced eosinophilia, anti-Asp f IgG antibodies and positive sputum cultures are supporting features. However, the eosinophil count may be normal in advanced "burnt out" disease (45), IgG antibodies have been reported in only 60% of cases, and it may not always be possible to culture Asp f from sputum (19). Characteristic HRCT features are proximal bronchiectasis with involvement of the upper lobes in particular (46), though distal rather than proximal bronchiectasis and involvement of just the lower lobes have been described (8). We have classified patients as having ABPA if there was evidence of significant sensitization to Asp f and past history of atopic illness with either upper lobe bronchiectasis, greatly raised total IgE, positive IgG antibodies, eosinophilia or positive sputum cultures. The history of atopic illness (usually asthma) in these patients points to ABPA being the cause of bronchiectasis rather than sensitization occurring secondary to colonization of damaged airways. The low frequency of positive IgG antibodies in our experience makes this test insensitive in predicting those with ABPA, and it should not be used in isolation. Unlike Miller and colleagues (47), we failed to find an association between ABPA and the F508 CF gene mutation.

We have investigated a large selected population of patients with bronchiectasis and found that in 49% of cases one or more potential cause may be found. Compared with previous reviews, we have shown a decreased prevalence of childhood respiratory infection and increased number of patients with humoral immune disorders. In particular, functional antibody response testing to exclude defective antibody production, as well as assessment of immunoglobulin levels is indicated. In our experience, screening patients for disorders of neutrophil respiratory burst, chemotaxis, or adhesion is unlikely to be helpful. ABPA constitutes an important treatable condition, and a number of less common causes such as ciliary defects, rheumatoid arthritis, and congenital defects continue to occur in lower numbers. It is important to consider the possibility of CF in every patient with bronchiectasis, even in older adults, as this contributed a small but important number of cases. Patients with bronchiectasis deserve thorough investigation as this may impact on management.