The Epidemiology of Chronic Suppurative Lung Disease and Bronchiectasis in Children and Adolescents

In the modern era, the global burden of childhood chronic suppurative lung disease (CSLD) remains poorly captured by the literature. What is clear, however, is that CSLD is essentially a disease of poverty. Disadvantaged children from indigenous and low- and middle-income populations had a substantially higher burden of CSLD, generally infectious in etiology and of a more severe nature, than children in high-income countries. A universal issue was the delay in diagnosis and the inconsistent reporting of clinical features. Importantly, infection-related CSLD is largely preventable. A considerable research and clinical effort is needed to identify modifiable risk factors and socioeconomic determinants of CSLD and provide robust evidence to guide optimal prevention and management strategies. The purpose of this review was to update the international literature on the epidemiology, etiology, and clinical features of pediatric CSLD.

In the modern era, the global burden of childhood chronic suppurative lung disease (CSLD) remains poorly captured by the literature. What is clear, however, is that CSLD is essentially a disease of poverty. Disadvantaged children from indigenous and low-and middle-income populations had a substantially higher burden of CSLD, generally infectious in etiology and of a more severe nature, than children in high-income countries. A universal issue was the delay in diagnosis and the inconsistent reporting of clinical features. Importantly, infection-related CSLD is largely preventable. A considerable research and clinical effort is needed to identify modifiable risk factors and socioeconomic determinants of CSLD and provide robust evidence to guide optimal prevention and management strategies. The purpose of this review was to update the international literature on the epidemiology, etiology, and clinical features of pediatric CSLD.
Keywords: chronic suppurative lung disease, bronchiectasis, children, epidemiology, etiology iNTRODUCTiON Bronchiectasis, a chronic progressive disease of the airways, remains one of the most neglected diseases in respiratory health (1). It is characterized by abnormal dilatation of the bronchi caused by protracted inflammation (2) and by chronic productive or wet cough (3). A definitive diagnosis of bronchiectasis requires a chest high resolution computer tomography (cHRCT) (4), with cases otherwise referred to as having chronic suppurative lung disease (CSLD) (5). Bronchiectasis or CSLD (from hereafter combined and referred to CSLD) has multiple etiologies and are often associated with underlying conditions (e.g., congenital malformation, cystic fibrosis, or immune deficiency) (6,7). However, recurrent acute lower respiratory infections (ALRI) during early childhood, a crucial time for lung growth and development, are arguably the common etiology for CSLD, particularly among socially disadvantaged children (8,9).
Globally, the prevalence of CSLD in high-income countries over the last 50 years has declined with the introduction of antibiotics, immunizations, improved hygiene, nutrition, and access to medical care (10)(11)(12)(13)(14)(15)(16). However, a substantial burden of CSLD persists among socially disadvantaged populations of high-income countries (e.g., Alaskan, Australian, Canadian, Maori, and Pacific Islander children) (8,(16)(17)(18)(19)(20) with the extent of pediatric CSLD in low-and middle-income countries largely unknown. In recent years, there has been a growing awareness of CSLD related to the increased use of cHRCT diagnostics and emerging research into the etiology (6), microbiology, immunology (21), and clinical management (22), yet, robust epidemiological data remain sparse. Our understanding of the complex interplay between the host, pathogens, and the environment is largely superficial, and there have been few clinical intervention trials among children and adolescents. Here, we provide an update on the epidemiology, etiology, and clinical features of pediatric CSLD not associated with cystic fibrosis.

GLOBAL iNCiDeNCe OF CSLD Historical
French physician Rene Laennec first described CSLD in the early nineteenth century (23), with the first surgery (24) and imaging (25) performed around 100 years later. In the 1930s, Roles described the poor prognosis of CSLD patients at the time and highlighted the importance of early diagnosis and the potential of lobectomy for enhanced survival (26). The benefit of surgery over medical management was unclear in several subsequent studies (27)(28)(29)(30), though a surgical case series of CSLD conducted decades later was convincing, particularly for children (13). With the introduction of broad spectrum antibiotics in the 1950s, significant reductions in CSLD incidence were reported. In the United Kingdom (UK) between 1952 and 1960 (12), CSLD admissions fell fivefold from 23.8 to 4.9 per 1,000 total admissions. Similar findings were reported elsewhere in the UK (10) and in the United States of America (USA) (11). Childhood CSLD became confined to disadvantaged populations of high-income countries, earning it the label of an "orphan disease" (14). Alaskan native children <16 years of age appeared vulnerable with a conservatively estimated case prevalence of 41/100,000 population over the decade 1956-1966, largely in the Yukon-Kuskokwim (YK) Delta and with one-third of cases associated with tuberculosis infection (16). In the 1960s, the prevalence of CSLD among Scottish children (<10 years of age) was 10.6/100,000 (27), and in the early 1970s, a substantial burden of CSLD was reported among indigenous children living in Central Australia (31). In this study, 83 indigenous children with CSLD were identified (60% <2 years of age) from an estimated population of around 1,000, despite near complete eradication of tuberculosis, measles, and pertussis, previously considered to be key etiological pathogens (31). In one of the few African studies, 70/1,150 patients consecutively admitted to the University Hospital in Nigeria with respiratory or cardiovascular disease between 1975 and 1979 were diagnosed with CSLD, of which 10 (14%) were children (32).
In the following section, we review the population incidence of pediatric CSLD in the modern era , when immunization and antibiotic interventions control many of the major etiological pathogens of the past, including tuberculosis, pertussis, measles, and pneumococcus.

Modern era: 1980-2016
We performed a systematic literature review of the PubMed library followed by a bibliography search within all relevant articles. The PubMed search terms: (CSLD OR bronchiectasis) AND (pediatric OR child OR children OR infant) AND (epidemiology) yielded 351 articles, which was reduced to 243 when confined to non-cystic fibrosis-related bronchiectasis articles published since 1990 and performed after 1980, human research, and non-review articles written in English. From these, 13 relevant studies reported CSLD rates: either directly (incidence rate) or indirectly (included a bronchiectasis disease numerator, an estimate of the population denominator, and an observation period). In one New Zealand (NZ) study, age-specific data could not be accurately extracted or obtained from the authors and were, therefore, excluded (33). A government report capturing similar national NZ CSLD data was included in its place (34).
In our defined modern era , the burden of childhood CSLD remains difficult to characterize. In the 1980-1990s, wheezing and persistent cough were often considered clinical features of asthma and not related to CSLD. Thus, due to lack of robust evidence, misclassification was likely and management (e.g., antibiotics) possibly suboptimal. Peer reviewed published burden data were available from only nine countries, diagnostic criteria varied, population denominators were often approximate, and incidence/prevalence data were inconsistently reported. Further, hospital-based studies utilizing International Classification of Diseases coding to identify CSLD (ICD8 518, ICD9 J494, and ICD10 J47) were common. ICD coding does not provide radiological information (though several studies confirmed cHRCT diagnosis via medical records or subsequently performed cHRCT) nor distinguish between primary and repeat admissions (exacerbations), which must be ascertained by diagnosis in conjunction with date of presentation. While rare in most high-income countries, in some settings (e.g., Australian indigenous children) as many as 1 in 136 children have a new case of CSLD detected annually (8). Despite limitations in data accuracy, the trends toward higher rates of CSLD among socially disadvantaged populations were clear though perhaps partially attributed to enhanced surveillance. The collated global epidemiology data are summarized in Table 1.

High-income Countries
In most affluent European populations, pediatric CSLD was relatively uncommon with average annual rates ranging from 0.2/100,000 in the UK (37) to 2.3/100,000 in Ireland (38), however, in the equally affluent, yet non-European United Arab Emirates (UAE; 13.3/100,000) (36) and largely non-European NZ (15.0/100,000) (34), CSLD was more common. The indigenous Arabic children in the UAE children were predominantly from a wealthy social demographic with almost universal immunization coverage, therefore, the surprisingly high rates may be related to an inherited predisposition or to moderate levels of educational attainment. In NZ, the high national rate is a reflection of the high proportion of less affluent Maori and Pacific Islander people (~40%) for which there is a well-established susceptibility to respiratory infection (19,43,44). In this particular report by Simpson (34), CSLD cases were not reported by ethnicity. As clarified in the next section, the annual rate was much lower (1.5/100,000) for NZ people of European heritage (18,20).
In the early 2000s, the average annual incidence of cHRCT confirmed bronchiectasis among indigenous children living in Central Australia was high at 735/100,000 (8). Almost all children (95%) had pneumonia at an early age (median 6 months) and nearly two-thirds had concurrent chronic suppurative otitis media (8). Two further hospital-based studies have investigated CSLD among Australian indigenous children. The first study, a historical cohort examining the entire indigenous population of Australia's Northern Territory (NT) (both Central and Northern Australia from 1999 to 2004) showed an average CSLD incidence of 118/100,000 child-years in the first 12 months of life (39). One in five of these infants were also hospitalized with an ALRI. The early burden of respiratory illness is of particular concern in this region. The second study, a retrospective chart review conducted during 2007-2011 at the Mount Isa Base Hospital in Queensland (Qld), demonstrated an average annual incidence of 410/100,000 in children <2 years of age (40).
Alaskan Natives are also historically at high risk of CSLD. In a conglomeration of CSLD data from the YK Delta region (collated from a state-wide registry, historical and current patients of YK Delta Regional Hospital, or Alaska Native Medical Center), the average annual incidence of CSLD among children born between 1980 and 1990 (the review defined modern era) was 140/100,000 (16); equivalent to the incidence reported among the same population born in the 1940s (16). For a similar population, the Inuit children of Nunavut, Canada, the average annual incidence (1998-2011) of pediatric cHRCT confirmed bronchiectasis ascertained via health record reviews was lower but also substantial at 15.5/100,000 (17).
Nationally in NZ (20), and among indigenous children of Australia (8), Alaska (16), and Canada (17), CSLD was diagnosed at around 5 years of age. In Auckland though, the age was slightly older at 8 years (18) and under 2 years of age for Australian indigenous infants (39,40). Common to disadvantaged populations of high-income countries, was low socioeconomic status, inadequate and overcrowded housing, complex environmental and social issues, limited access to health care, frequent exposure to camp or cooking fires, high smoking rates, and almost a universal history of early infant pneumonia and/or recurrent ALRIs (8, 16-20, 39, 40). In affluent countries, CSLD is more frequently associated with non-infectious etiology (see "Etiology").

CSLD-ReLATeD MORTALiTY
There are limited data on CSLD mortality in the pediatric population. More deaths occur in adults. In Central Australia, a hospital record review showed that 34.2% (41/120) of indigenous adults diagnosed with CSLD 2001-2007 died during that same period. Most had a history of pneumonia, nearly half were diagnosed in childhood and the median age of death was 42.5 years (56). A study in England and Wales showed that only 0.2% (12/5,745) of total CSLD deaths recorded between 2001 and 2007 occurred among children aged <14 years (57). Whereas, a NZ audit of an electronic CSLD database (1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006), found that 7% (6/91) of children diagnosed with CSLD (median age 7.3 years) died during follow-up (median follow-up 6.3 years) (58). A more recent study in India identified 80 children (mean age 9.6 years) with CSLD via medical record reviews. Among 62 children who were prospectively followed for 1 year, 5 (11%) died. To avoid premature death from CSLD in both children and adults, it is vital to intervene as early as possible.

eTiOLOGY AND CLiNiCAL MANiFeSTATiONS OF PeDiATRiC CSLD
A systematic literature review of the PubMed library using the search terms: (CSLD OR bronchiectasis) AND (pediatric OR child OR children OR infant) and restricted by time (post 1990), type (non-review), language (English only), and availability, yielded 1,163 articles. Further screening to exclude cystic-fibrosis, HIV related, surgical, small (<10 CSLD cases), overlapping, or biased CSLD populations refined the list to 28 articles ( Table 2). From these publications, 26 contributed etiology ( Table 3) and 27 clinical manifestation data ( Table 4).

Demographics
The 26 studies of pediatric CSLD encompassed 2,103 children from 15 countries around the world; 13 high-income (including studies within disadvantaged sub-populations) and 13 low-and middle-income countries ( Table 2). The age of symptom onset varied significantly among the studies (range 0-16 years) but was consistently present many years prior to a diagnosis of CSLD [ Table 2; mean/median differences ranging from 0.8 to 7 years ( Table 2)]. Some of the earliest reported symptoms were among indigenous infants from Australia (0.3years) (8), and Alaska (0.2 years) (9), and New Zealand (0.5 years) emphasizing the early origins of CSLD in high-risk children. Interestingly, Malaysian (low and middle income) and Italian (high income) children had a similar age of symptom onset (0.5 years). Thus, early symptom identification is essential to preventing CSLD progressing and long-term decline in lung function (68)(69)(70). As with studies reporting incidence (several of which overlap), the demographic, etiological, and clinical data varied in their nature relating to differences from study design, study era (e.g., clinical reporting) and the age at diagnosis of children involved (e.g., unable to perform lung function <4 years).

etiology
Chronic suppurative lung disease is the end result of chronic airway inflammation that is driven by persistent infection (2). As such, the natural history of CSLD depends largely on the susceptibility to both acute and chronic infection (71). Historically, congenital malformations, cystic fibrosis, immune deficiency, and aspiration are common antecedents (5), although a primary cause is not always identified. Further, measles, tuberculosis, and pertussis (27,72), once the most common causes of CSLD may be less important today in the wake of advances (5) in vaccinations, antibiotics, and access to health care (67). Reviews of CSLD etiology from pediatric cohorts have been published in 2011 (67) and 2014 (6). Using a systematic search, we further updated the etiology literature, which remains similar as previous reviews and is briefly summarized below (Table 3) (17,38,48,51,58,61,(63)(64)(65)(66).

CLiNiCAL FeATUReS AND OUTCOMeS
In this update of the literature surrounding clinical features of CSLD, importantly, we have included the largest prospective international multicenter study for CSLD in children (9,75). The most comprehensive clinical data in this section are recorded among indigenous populations and low-and middle-income countries (summarized in Table 4).
Clinical features of CSLD differ depending on the extent of the disease. In CSLD, recurrent wet or productive cough is probably the earliest and most important symptom to recognize and investigate, although it should be noted that young children do not readily expectorate (3). Despite the importance of cough in respiratory health, this was only specifically reported in 59%  (9)] reported high rates of cough. Under-appreciating the importance of cough may delay diagnosis and treatment, which can lead to a decline in lung function and poorer long-term prognosis (76). Likewise, cough is an important clinical feature of CSLD exacerbations (1).
In low-to middle-income countries, the proportion of CSLD children presenting with wheeze ranged from 20% in Tunisia (48) and South Korea (63) to 66% (49) in Saudi Arabia, though it was unclear whether all 66% of the Saudi children had wheeze as it co-reported with other symptoms. The only high-income country reporting wheeze was the UK where 10% of children were identified. Interestingly, in the multicenter study of indigenous children (9), Alaskan natives had a much higher percentage of wheezing (41%) than either indigenous children of Australia (7%) or Pacific Islander and Maori children of NZ (17%). This correlated with an earlier study that showed both Alaskan native children with (65%) and without (37%) CSLD had higher rates of wheeze (77). It is likely that there are different underlying susceptibilities to and clinical manifestations of CSLD across population groups, but further exploration is needed to further understand these relationships.
Inadequate nutrition (macro and micro) is another factor that can increase the risk of developing and/or worsening CSLD by compromising innate and adaptive immunity (78), which can have long-term consequences (67). As an example, sufficient vitamin D is known to be important for regulating immune responses to respiratory pathogens (79). Among indigenous Australian children, deficiency is associated with an increased risk of hospitalization for respiratory infection (80). A typical feature of suboptimal energy and protein intake is failure to thrive (FTT), which was reported just over one-third of studies.
In more advanced CSLD, clinical features include chest wall deformity, digital clubbing, hemoptysis, and dyspnea. Again, these markers were common to indigenous populations and low-and middle-income countries ( Table 4). A further feature, chest pain, was reported in two studies from Seoul (3%) (63) and India (43%) (64). It is possible that chest pain relates to a delayed diagnosis of CSLD (median age 7.6 years in Korea and 9.6 years in India) again emphasizing the need for earlier symptom surveillance and diagnosis to improve clinical outcomes for children with CSLD. In the longer term, complications associated with CSLD extend beyond the respiratory system including cardiovascular disease an area requiring further investigation (84).
While there are limited long-term clinical data, these studies indicate that children with CSLD have different clinical manifestations across populations. These data also highlights that children from socially disadvantaged groups (e.g., indigenous populations and low-and middle-income countries) appear to have more extensive clinical features and greater propensity for poorer longterm outcomes than those from high-income countries.

PReveNTiON AND MANAGeMeNT
Despite the high global burden of respiratory diseases, only a small percentage of research and development funding and resources are devoted to these conditions (71,85). To prevent CSLD, a multidisciplinary approach is required beginning in the antennal period continuing throughout childhood. Key strategies include immunizations, improved hygiene, nutrition, and education (that is culturally appropriate) to improve health-related outcomes for children with CSLD (9,86). Other fundamental factors to consider are exposure to tobacco, camp fire, biomass fuels, and other environmental air pollutants, overcrowding, housing quality, and access to running water. Evidence-based clinical guidelines have been developed in several settings to guide primary and secondary health care (22,78,87,88), yet, more robust and consistent evidence from populations across the globe is essential. Other papers in this series will focus on the management of CSLD in more detail.

FUTURe ReSeARCH
There is a need for more epidemiological research globally to clarify the prevalence of CSLD, in particular, for at-risk populations (e.g., indigenous and low-and middle-income countries). Importantly, there are also no long-term prospective studies in children to guide clinical care and management into adolescence and adulthood. Clinical trials across multiple countries are essential to further improve interventional and clinical management of CSLD. This can only be achieved with substantial investment and support from governments and funding bodies.

CONCLUSiON
Once considered an "orphans disease" among high-income populations, this review reinforces that CSLD remains a disease of poverty, common among children from indigenous populations and low-and middle-income countries. Delayed diagnosis was common as was severe disease. Despite the fact that the burden of CSLD in these settings was primarily related to early preventable infections, this condition is neglected globally in terms of research priority and funding. To improve the respiratory health of disadvantaged children, a concerted international effort is needed to determine and understand the burden of pediatric CSLD and to provide a solid evidence base for future clinical care and management.