Introduction

Bronchiectasis is a chronic respiratory condition characterised by recurrent lower respiratory tract infections and airway inflammation, dilatation of the bronchial airways and impaired mucociliary clearance.¹ The pathogenesis of bronchiectasis is complex and multifactorial, though typically arising via a pathophysiologic combination of recurrent lower respiratory infections, smoking, airway inflammation, impaired ciliary function and other chronic lung diseases such as emphysema.¹ Globally, there has been an observed relationship between socio-economic status and the prevalence of bronchiectasis. Higher socio-economic groups have increased bronchiectasis in the elderly population, whereas socially disadvantaged populations demonstrate a higher prevalence of bronchiectasis that presents at a younger age.² Bronchiectasis is more prevalent among Indigenous populations residing in English-speaking Organisation for Economic Cooperation and Development (OECD) countries compared to non-Indigenous populations, such as the Alaskan and Native Americans, First Nations New Zealand Māori and Pacific Islanders and Aboriginal Australians^{3, 4} (from here on “Indigenous” is used to refer to global First nations populations, while “Aboriginal Australian/people/population” is used to specifically refer to Australia's' First Nations population). Despite known longstanding higher prevalence of bronchiectasis among these groups,⁵ the morphological and radiological characteristics of bronchiectasis have not previously been described.

The exact influence of pulmonary morphological characteristics on bronchiectasis outcomes is a matter of ongoing debate and research. The presence of multiple lobar involvement is included in prognostic tools, such as the Bronchiectasis Severity Index (BSI) or FACED scale, with multi-lobar involvement associated with more exacerbations and hospitalisations.^{6, 7} The influence of bronchiectasis type upon lung function parameters, however, is controversial, with some research showing no significant effect,⁸ and others showing a significant difference between types.⁹ Moreover, micro-organism cultures appear to differ between various lobes affected by bronchiectasis.¹⁰

The lobar distribution and type of bronchiectasis differ across global geography, although, broadly, the lower lobes are the most affected.¹¹ In Israel, 20.7% of patients were identified with right lower lobe (RLL) and 20.4% with left lower lobe (LLL) involvement¹⁰ compared with 52.3% and 64.8%, respectively, in Turkey,⁹ and 60% and 62% respectively in the USA.¹² The reason underlying the relationship between geographical location and lobar involvement is unclear. However, it may be due to environmental factors, micro-organisms, prevalence of comorbidities or other factors.

Currently, chest computed tomography (CT) is the gold standard in the accurate diagnosis of bronchiectasis.¹³ Adult Aboriginal Australians residing in the Top End Health Service (TEHS) regions are reported to have a significantly higher prevalence of bronchiectasis in comparison to non-Aboriginal Australians.¹⁴ A recent study from the Northern Territory (NT) of Australia demonstrated that 23% of Aboriginal adults undergoing chest CT scans showed evidence of bronchiectasis.¹⁵ Other research has shown a disparity in bronchiectasis prevalence and mortality for adult Aboriginals between various health districts within the TEHS region.¹⁶ Given the lack of data in the literature regarding the morphological characteristics, it is timely to describe chest CT scan findings among adult Aboriginal Australians diagnosed with bronchiectasis residing in various TEHS health districts in the NT. Therefore, in this study we examined the chest CT features among an Aboriginal Australian cohort with bronchiectasis over a 10-year period (2011–2020).

Methods

Results

Chest CT data

Bronchiectasis was predominantly recorded in the LLL (73%), followed by the RLL (62%). The upper lobe (LUL) was least commonly affected (22%) (Table 1). Females were significantly more likely to have right middle lobe (RML) involvement (Fig. 1) than males (50 vs. 34%, P = 0.012). Bilateral involvement was common (74%), with no significant difference between the sexes. When the type of bronchiectasis was available to be assessed from the radiologist reports, cylindrical or tubular (50%) and cystic (28.1%) types were the most common. No significant differences in the type of bronchiectasis were noted between the sexes.

Table 1. CT scan results detailing lobar involvement, type of bronchiectasis and laterality

	Total (n = 459)	Female (n = 254)	Male (n = 205)	P-value
CT- Lobar location of bronchiectasis
RLL	285 (62.1%)	148 (58.3%)	137 (66.8%)	0.427
RML	196 (42.7%)	126 (49.6%)	70 (34.1%)	0.008*
RUL	126 (27.5%)	67 (26.4%)	59 (28.8%)	0.955
LLL	337 (73.4%)	184 (72.4%)	153 (74.6%)	0.955
Lingula	130 (28.3%)	73 (28.7%)	57 (27.8%)	0.955
LUL	100 (21.8%)	52 (20.5%)	48 (23.4%)	0.932
Type
Cylindrical or tubular	80 (50%)	51 (54.3%)	29 (43.9%)	0.775
Cystic	45 (28.1%)	28 (29.8%)	17 (25.8%)	0.955
Traction	35 (21.9%)	16 (17%)	19 (28.8%)	0.502
Type not stated	299 (65.1%)	160 (63%)	139 (67.8%)	0.852
Laterality
Bilateral	339 (73.9%)	181 (71.3%)	158 (77.1%)	0.712

• P-values obtained via logistic regression, adjusted for multiple hypotheses testing via Romano-Wolf method. LLL, left lower lobe; LUL, left upper lobe; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe.
• *Statistically Significant.

There were minor differences noted between health district regions in the CT lobar involvement and for the type of bronchiectasis (Table S1). RLL involvement was the most common in DU (67.6%), while LLL involvement was most common in each of DR (78.8%), EA (62.7%) and KR (76.3%) regions. The LLL and Lingula involvement were significantly more common in DR compared to the DU region (uncorrected P = 0.032 and P = 0.044, respectively). Among those patients where a type of bronchiectasis was reported, traction bronchiectasis was the most common in DU region (33.3%), cylindrical bronchiectasis in DR (38.4%) and KR (33.3%) regions, and both traction and cystic bronchiectasis in EA (34.5% each). Bilateral involvement did not significantly differ between regions. In multivariate analysis, the prior noted differences in lobar involvement between regions were attenuated.

Pairwise correlations between lobar location of bronchiectasis

For 337 (73.4%) patients, bronchiectasis affected more than one lobe. The strongest correlation was noted between the right upper lobe (RUL) (Fig. 2) and LUL, R² of 0.539 (P < 0.001) (Table 2). However, moderate correlations (R² > 0.4) were also noted between the Lingula and the LUL (R² 0.418) and between the Lingula and the RML (R² 0.415). The RML and LLL showed a significant positive correlation with cylindrical or tubular type bronchiectasis (R² 0.265 and 0.260, respectively) and a negative correlation with traction type bronchiectasis (R² –0.247 and −0.253, respectively). LUL involvement showed a significant positive correlation with cystic type bronchiectasis (R² 0.270) (Table 3).

Table 2. Pairwise correlations between lobar location of bronchiectasis

CT- Lobal location	RLL (n = 285)	RML (n = 196)	RUL (n = 126)	LLL (n = 337)	Lingula (n = 130)	LUL (n = 100)
RLL
RML	129 (45.3%)
	0.066 (1.000)
RUL	90 (31.6%)	78 (39.8%)
	0.118 (0.167)	0.239 (<0.001)
LLL	248 (87.0%)	138 (70.4%)	86 (68.3%)
	0.394 (<0.001)	−0.059 (1.000)	−0.072 (1.000)
Lingula	87 (30.5%)	98 (50%)	60 (47.6%)	103 (30.6%)
	0.063 (1.000)	0.415 (<0.001)	0.263 (<0.001)	0.083 (1.000)
LUL	67 (23.5%)	63 (32.1%)	73 (57.9%)	80 (23.7%)	64 (49.2%)
	0.053 (1.000)	0.217 (<0.001)	0.539 (<0.001)	0.079 (1.000)	0.418 (<0.001)

• Data displayed as number (%) followed by Pearsons R² (Bonferroni adjusted P-value). Bolded cells indicate correlations with P < 0.05. LLL, left lower lobe; LUL, left upper lobe; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe.

Table 3. Pairwise correlations between lobar locations and bronchiectasis type

Bronchiectasis type	RLL (n = 100)	RML (n = 69)	RUL (n = 61)	LLL (n = 120)	Lingula (n = 48)	LUL (n = 41)
Cylindrical	59 (59%)	45 (65.2%)	25 (41%)	69 (57.5%)	25 (52.1%)	17 (41.5%)
	0.232 (0.056)	0.265 (0.013)	−0.142 (1.000)	0.260 (0.016)	0.027 (1.000)	−0.100 (1.000)
Cystic	29 (29%)	23 (33.3%)	23 (37.7%)	35 (29.2%)	21 (43.8%)	20 (48.8%)
	0.025 (1.000)	0.101 (1.000)	0.167 (0.621)	0.040 (1.000)	0.228 (0.068)	0.270 (0.011)
Traction	17 (17%)	7 (10.1%)	17 (27.9%)	19 (15.8%)	9 (18.8%)	7 (17.1%)
	−0.152 (0.983)	−0.247 (0.029)	0.114 (1.000)	−0.253 (0.022)	−0.050 (1.000)	−0.068 (1.000)

• Data displayed as number (%) followed by Pearsons R² (Bonferroni adjusted P-value). Bolded cells indicate correlations with P < 0.05. LLL, left lower lobe; LUL, left upper lobe; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe.

Chest CT location of bronchiectasis to lung function parameters and sputum microbiology

Neither lobar location nor bronchiectasis type showed any significant association with lung function parameters either for FEV₁, FVC or FEV₁/FVC in univariate or multivariate models (Figure S1). However, RML, Lingula and LUL involvement were associated with significantly greater percent predicted values of DLCO in univariate regression (16.2 (95% CI, 3.76, 28.64), P = 0.023, 16 (95% CI 0.29, 31.71), P = 0.023 and 16.1 (95% CI −4.47, 36.67), P = 0.023 respectively), though the associations were attenuated on multivariate regression (14.08 (95% CI 1.72, 26.44), P = 0.092, 12.75 (95% CI −0.45, 25.94), P = 0.095 and 15.1 (95% CI −2.61, 32.82), P = 0.067, respectively) (Fig. 3). No CT parameters showed any significant association with TLC.

Sputum microbiological cultures were available for 147 (91.9%) patients for whom the type of bronchiectasis was also reported. In multivariate modelling, presence of non-Aspergillus fungus in sputum was associated with significantly reduced odds for traction type bronchiectasis (RW P = 0.0319). There were no other significant associations between sputum culture and type of bronchiectasis (Table S2) nor between sputum culture and lobar location of bronchiectasis (Table S3).

Discussion

To the best of the authors knowledge, this is the first comprehensive study assessing chest CT data in an adult Indigenous population with bronchiectasis, especially from the Top End, NT of Australia. In line with previous studies, lower lobes are predominantly affected, and bilateral involvement is common. Although data pertaining to the subtype of bronchiectasis was limited, the available data demonstrated that cylindrical or tubular types of bronchiectasis are most common in this population. We identified that RML involvement is significantly more common in females and that there are differences in chest CT findings based on geographic location. Furthermore, we identified better lung function values for DLCO with RML, Lingula and LUL involvement.

Currently there is sparse literature detailing chest radiology data for Indigenous populations globally,^{17, 18} and for the Australian Aboriginal population in particular.^19-24 In other global non-Indigenous populations, chest CT findings of bronchiectasis patients and their association with other clinical parameters such as potential aetiology, sputum microbiology and lung function parameters have been well described.¹¹ Chest CT characteristics of bronchiectasis secondary to tuberculosis (TB) are documented in previous reports from geographic regions with high incidences of TB.^{21, 22} However, the incidence of TB in Australia is much lower (including among Aboriginal Australians) as compared to certain global geographic regions.²³ Hence, it is unlikely that the vast majority of the CT findings of bronchiectasis observed among our study participants are sequalae secondary to pulmonary TB. Similarly, chest CT characteristics of non-tuberculosis mycobacterial (NTM) infections are recognised to be associated with RML involvement.¹¹ The incidence of NTM infections, however, are lower in Aboriginal Australians in comparison to their non-Aboriginal counterparts.²⁴ In our study, a significantly higher proportion of females were noted to have bronchiectasis affecting the RML. However, in multivariate regression, there was no significant association between NTM and RML involvement. It is unclear at this stage the reasons for heightened RML involvement among Aboriginal Australian females in our study cohort. Therefore, it is reasonable to speculate that previously unexplored or unknown etiologic factors may be contributing to bronchiectasis in global Indigenous populations and in Aboriginal Australians.

We observed that cylindrical or tubular (50%) and cystic (28%) bronchiectasis were the most common types among Aboriginal Australian patients. Variations in the lobar involvement and the type of bronchiectasis among patients residing in differing health districts/urban versus rural and Aboriginal Australian communities were also noted. The reason for the observed geographical variations in the chest CT characteristic was not explored in this current study, as that was beyond the scope of this study. However, future studies may be useful to explore the reasoning behind geographic variation. Previous studies have demonstrated that the presence of cystic bronchiectasis is associated with a higher incidence of pseudomonas infections.²⁵ We found no such association in the current study; however, non-Aspergillus fungus was positively associated with cylindrical types (though the P-value was attenuated following Romano-Wolf correction) and negatively associated with traction-type bronchiectasis.

One important observation we noted in this study is that a significant proportion of the study patients demonstrated extensive bronchiectasis in association with significant parenchymal lung architectural distortion (Fig. 4). This suggests that the pathological damage to the pulmonary parenchyma affects the smaller airways and alveoli as well as larger airways that could give rise to substantial respiratory physical impairment. Indeed, shortness of breath and chronic cough with mucous production are noted to be the most common clinical manifestations among adult Aboriginal patients presenting with respiratory disorders in day-to-day clinical practice.^{4, 14} Nonetheless, functional radio-pathological correlation will require further research in this area.

Lung function parameters in relation to chest CT findings among patients with bronchiectasis are documented in the literature for non-Indigenous patients, showing a correlation of bronchial wall thickening against FEV₁ values,²⁵ and greater lung function impairment among patients with cystic bronchiectasis.^{9, 26} Moreover, significant impairment in DLCO is reported to occur among patients with bronchiectasis.²⁷ However, there is limited data on the correlation of specific lobar involvement against lung function impairment, more specifically for DLCO. Previous studies have shown significant impairment in lung function parameters among adult Australian Aboriginal patients with chronic airway disease, including among patients with bronchiectasis.⁴ Moreover, DLCO values are noted to be lower for Aboriginal Australian patients in comparison to Caucasian counterparts.²⁸ In this study, we noted that the presence of bronchiectasis in the RML, lingula or LUL was associated with a higher DLCO, although these associations were attenuated to borderline significant in multivariate models. The RML, Lingula and LUL also demonstrated significant pairwise correlations, although their strength was weak to moderate (R² 0.21–0.41). Exactly what this means for patients in clinical sense, and the potential mechanisms underlying these findings require further exploration.

This study has demonstrated radiological manifestations of bronchiectasis in an adult Aboriginal Australian cohort and the results of study may be useful in the future to develop predictive models in bronchiectasis severity classifications in global Indigenous and among Aboriginal Australians. The burden of chronic respiratory disorders is significantly higher in Aboriginal Australians, contributing to higher overall morbidity and mortality.^{4, 29} Therefore, it is imperative that future prospective studies investigate other various clinical conundrums³⁰ associated with chest radiology findings and clinical outcomes in Indigenous/Aboriginal peoples, including innovative strategies to address chronic respiratory burden in regional and remote Aboriginal communities.

Funding

This research received the Thoracic Society of Australia and New Zealand (TSANZ) - Robert Pierce Grant-In-Aid for Indigenous Lung Health. The TSANZ did not have any role in the study design, data collection, analysis or interpretation.

Author contributions

All authors contributed to the conception and design, interpretation of the data, drafting of the paper, revising it critically for intellectual content and final approval of the version to be published and agree to be accountable for all aspects of the work.

Ethics statement

This study was approved by the local Human Research Ethics Committee.