The etiologies of non-CF bronchiectasis in childhood: a systematic review of 989 subjects

This study followed the Preferred Reporting Items in Systematic Reviews and Meta-Analyses (PRISMA) guidelines [5].

EMBASE, Medline, OVID Cochrane Reviews, Directory of Open Access Journals, Open Science Directory, EPSCO information services, and OAlster were searched from 1966 to March 25, 2014. The bibliographies of all of the selected studies were also reviewed.

The main search term was “non-CF bronchiectasis”. The following filters were used: human, all children and young adult. The searches were performed independently by two of the authors and the results were compared.

Initial evaluation of each article was performed by one author (KSB) and then reviewed by another (MTD). In cases where study populations appeared to overlap, the study with the largest number of subjects was selected. Differences in judgment were resolved first by consensus; ties were adjudicated by the third author (SCA). All studies selected for inclusion were reviewed by the third author.

For each selected study, the following information was recorded: inclusion criteria, number of patients, age range, diagnostic criteria, patient source and country of origin. Categorical and specific etiologies of bronchiectasis were also recorded for each study. Major categories of disease included primary immunodeficiencies, ciliary dyskinesia, infection, aspiration, idiopathic or unknown, congenital malformation, secondary immunodeficiencies, asthma, skeletal disorders, bronchiolitis obliterans, and others. Within each category, specific etiologies were catalogued from those studies that provided specific data.

The categorical and specific etiologies of non-CF bronchiectasis were pooled to provide estimates of the relative prevalence for each disorder. Total sample size estimates for categorical comparisons were calculated from the total number of etiologies identified for the entire review. Total sample size estimates for the comparison of specific etiologies within individual categories were calculated from the total number of etiologies reported in a given category.

Patient sampling by number, locale and institution raised the concern of population homogeneity and possible over- or under-representation of a specific area or ethnic group. Variability in diagnostic evaluation and the identification of multiple etiologies for individual patients were also potential sources of bias in defining etiology. Inconsistencies and vagaries in nomenclature were a potential source of error when studies were combined.

The results of the literature search are shown in Figure 1. Searches of the Medline and EMBASE databases yielded 202 references. An additional 289 citations were found by extensively searching the bibliographies of selected articles (Additional file 1). No additional studies were found by searching the OVID Cochrane Reviews, Directory of Open Access Journals, Open Science Directory, EPSCO information services, or OAlster. From the 491 studies identified, 448 studies were excluded after a cursory review of the title, abstract, and, when necessary, the results section. The full text of the remaining 43 articles was reviewed in detail. Thirty- one of the remaining studies were excluded: 22 reports had overlapping populations with other studies; 3 did not use computed tomography for the diagnosis of bronchiectasis; 1 did not have the minimum number of patients; 2 studies included the same subjects as previous publications; 2 included adult populations that could not be separated from the pediatric subjects and; 1 did not contain any etiology data.

The characteristics of the 12 studies that met the inclusion criteria and comprise the basis of this review are shown in Table 1 [6]-[17]. The reports ranged in size from 22 to 151 participants per study. Non-CF bronchiectasis was defined by computed tomography [18]; in one study, 96% of patients underwent computed tomography [6]. Together, these reports represent a worldwide sample (Australia, Ireland, Turkey, Saudi Arabia, United Kingdom, New Zealand, Alaska, Italy, and Korea). Children were drawn from single centers in 8 studies [6]-[12],[16], multiple centers in 3 studies [13]-[15] and an entire region in 1 study [17]; the total number of patients included in this review is 989.

The underlying conditions associated with non-CF bronchiectasis in children, by study, are shown in Table 2. Banjar, et al. and Eastham et al. reported multiple associations in individual patients [6],[7]. Attempts to contact the authors were unsuccessful. With the inclusion of these studies, the total number of associations (994) exceeds the total number of patients reported (989). In two studies, the most likely association of multiple associations reported for individual patients was the only one counted (e.g. primary immunodeficiency is a more likely association than infection since it was the likely predisposition for the infection) [7],[8]. Singleton et al. described patients with definite and probable bronchiectasis [16]; only definite cases were included.

The categorical disease processes associated with childhood non-CF bronchiectasis are shown in Table 3. Nine hundred and ninety nine associations were identified in 989 patients. No association was found in 366 subjects (40%). Of the identified associations, the most common were infection (173 subjects, 19%), primary immunodeficiency (160 subjects, 18%), aspiration/foreign body (95 subjects, 10%) and ciliary dyskinesia, including Kartagener’s Syndrome (91 subjects, 10%).

The infections associated with non-CF bronchiectasis are shown in Table 4. Of the 173 patients with an infectious process, 108 (62%) were identified by a specific disease entity. Pneumonia was the most common association (61%) followed by measles (14%), tuberculosis (11%) and pertussis (5%). Varicella, neonatal pneumonia, allergic bronchopulmonary aspergillosis and adenoviral pneumonia were rarely associated with bronchiectasis.

Of the 160 children with bronchiectasis and primary immunodeficiency, 131 (83%) cases were identified by a specific entity (Table 5). B cell disorders accounted for 97 (73.5%) of the primary immunodeficiencies identified: IgG and IgG subclass deficiencies were the most common (66.5%) and IgA deficiency accounted for 6%. A heterogeneous group of combined immunodeficiency disorders accounted for 10% of cases while 7.5% of primary immunodeficiencies resulted from T cell disorders. Of the 29 subjects with secondary immunodeficiencies, 18 (62%) were children who had received chemotherapy for an underlying oncologic process; 6 (20%) of children had HIV/AIDS and 5 (18%) were cardiac transplant recipients.

Of the 95 children with bronchiectasis and aspiration, 18 instances (20%) resulted from aspiration of a foreign body; 14 children (15%) had seizures and recurrent aspiration. Thirty-four children had an underlying congenital malformation and 27 (79%) had a specific entity identified (Table 6). Tracheoesophageal fistulae and cystic lung disease accounted for 52% and 19% of cases, respectively.

The studies included in this review ranged in size from 22 to 151 subjects; the largest study accounted for 17% of the total sample. The countries of origin contributed samples from 9 different countries and multiple continents. Only 1 study drew patients from a large, regional database; the remaining studies represented one or two clinical sites. All of the studies were retrospective in nature and none employed a standardized diagnostic evaluation. Vagaries in nomenclature occurred among patients diagnosed with primary immunodeficiencies deficiencies. Patients with antibody deficiency, antibody dysfunction or IgG deficiency were grouped together as IgG deficiency; those with combined immunodeficiency were grouped together with severe combined immunodeficiency. Identifying patients with “idiopathic” disease is also confusing since this implies a singular, as yet unidentified process. Three studies reported multiple underlying disorders for individual patients with non-CF bronchiectasis [6]-[8]. In two, patients with multiple ascribed etiologies could not be identified [6],[7].

Non-CF bronchiectasis in children usually has an indolent onset and presents with chronic respiratory symptoms [2]-[4]. Cough with daily sputum production is the most common clinical presentation and may be present for years before diagnosis. Hemoptysis, pleuritic chest pain, pulmonary osteoarthropathy, and delayed growth are additional findings associated with non-CF bronchiectasis. The definitive diagnosis of bronchiectasis requires chest imaging usually with high resolution computed tomography [18].

While the pathophysiology of bronchiectasis is well defined, the etiologies of non-CF bronchiectasis are varied [1],[3]. The data presented in this review suggest that 60% of children with this disorder have an underlying etiology. Infections, primary immunodeficiencies, aspiration (both foreign body and recurrent aspiration in neurodevelopmentally challenged patients) and ciliary dyskinesia account for most cases; severe bacterial or viral pneumonias and IgG deficiencies are the most common etiologies encountered. Throughout the literature, patients without an identified etiology are reported as idiopathic disease. A focused medical history and focused laboratory investigation should reveal the etiology of non-CF bronchiectasis in many cases.

Selection bias is a major limitation of this review. Since all of the studies selected for review were retrospective in design, the possibility exists that patients may have been missed. The subjects included in this study represent 9 different countries, lending credence to the notion that the sample is unbiased. Conversely, U.S. children are underrepresented since the only American study was from Alaska. Finally, 8 of the studies selected for review reported patients from a single clinical site and 3 studies used 2 clinical sites; only 1 study employed a regional database [17].

Misidentification and failure to identify an etiology also contribute to the limitation of this review. Identifying multiple etiologies in individual patients occurred in 2 studies, raising the concern of over representation of specific etiologies [6],[7]. The selected studies contained little to no detail regarding the diagnostic approach used to identify the etiology of non-CF bronchiectasis. The absence of a detailed, unified approach to the diagnostic evaluation of these children across studies may have overestimated the number of children without a diagnosis or may have misdiagnosed an unknown number of subjects. Variability in nomenclature compounds this problem. The most common etiology of non-CF bronchiectasis was severe pneumonia, but detail regarding the infectious agent was not available. The second most common etiology was a B cell disorder; unfortunately some subjects were identified with IgG deficiency or an antibody disorder leaving the true diagnosis open to speculation.

This study did not involve any direct contact with primary patient source documents and as such did not require IRB approval or patient consent.