Small airway dysfunction in bronchiectasis patients with exercise-induced desaturation

Abstract

Background

In non-cystic fibrosis bronchiectasis, small airway airflow limitation is primarily driven by mucus plugging, chronic neutrophilic inflammation, and epithelial injury. These primary mechanisms lead to subsequent airway remodelling and luminal narrowing. Such functional impairment is clinically significant, as a reduced distance–saturation product (DSP) of <280 m% during the 6-minute walk test (6MWT) is associated with increased mortality.

Objective

To evaluate the association between small airway dysfunction (SAD), exercise-induced desaturation (EID), and impaired exercise capacity in patients with bronchiectasis.

Design

A retrospective, cross-sectional study.

Methods

This retrospective study included 114 patients with bronchiectasis (non-desaturators, n = 57; desaturators, n = 57) and 53 non-bronchiectasis controls. Impulse oscillometry (IOS), spirometry, and 6MWT data were analysed. Desaturators were defined as a maximal decline in peripheral oxygen saturation (ΔSpO₂) ≤ −4% during the 6MWT.

Results

Compared with bronchiectasis non-desaturators (n = 57), desaturators (n = 57) exhibited more pronounced IOS-defined SAD, characterised by higher R₅–R₂₀, F_res, and A_X, and lower X₅, as well as spirometry-defined SAD with lower MMEF%pred and FEF50%pred. Desaturators also demonstrated a shorter 6-minute walk distance (6MWD). Greater R₅–R₂₀, F_res, and A_X, and lower X₅, MMEF%pred, and FEF50%pred were associated with a larger reduction in ΔSpO₂, shorter 6MWD, and lower DSP. The multiple linear regression showed age, BRICS, and R₅-R₂₀ are independent predictors for DSP. Among them, R₅-R₂₀(β=-0.345) is a stronger predictor than BRICS (β=-0.285) for DSP.

Conclusion

SAD is independently associated with EID and reduced exercise capacity in bronchiectasis, highlighting its potential role as a functional marker beyond structural disease severity.

Plain language summary

Background Bronchiectasis is a long-term lung condition in which the airways become damaged and inflamed. While damage to the larger airways is well recognised, the smaller airways deep in the lungs are also affected. Scarring and inflammation in these small airways can make breathing more difficult and reduce the lungs’ ability to deliver oxygen during physical activity. A simple walking test, the 6-minute walk test, is often used to assess exercise capacity. A combined measure from this test, called the distance–saturation product (DSP), has been linked to a higher risk of death when it falls below a certain level. Methods We reviewed lung function and exercise test results from people with non-cystic fibrosis bronchiectasis confirmed by lung scans, and compared them with healthy individuals. Lung function was measured using standard breathing tests and a technique called impulse oscillometry, which can assess small airway function during normal breathing. We focused on patients who developed a significant drop in oxygen levels while walking. Results People with bronchiectasis who experienced oxygen drops during walking showed more severe small airway problems than those who did not. They walked shorter distances, had lower DSP values, and had worse results on both standard lung tests and oscillometry. Measures of small airway function were closely linked to the size of the oxygen drop and reduced walking capacity. One oscillometry measure, reflecting reduced lung elasticity, was particularly good at identifying patients with a high-risk DSP value. Conclusion Problems in the small airways were strongly linked to drops in oxygen levels during exercise and to poorer walking ability in people with bronchiectasis.

Introduction

Bronchiectasis is characterised by irreversible dilatation of the bronchi and bronchioles resulting from destruction of elastic and muscular tissue, leading to impaired mucociliary clearance, persistent bacterial colonisation, and chronic airway inflammation. ¹ Beyond structural airway damage, bronchiectasis is increasingly recognised as a heterogeneous inflammatory airway disease associated with significant functional impairment and adverse clinical outcomes. ² Impaired lung function in non-cystic fibrosis bronchiectasis is strongly associated with increased morbidity and mortality. ³ Beyond structural airway damage, bronchiectasis is increasingly recognised as a heterogeneous inflammatory airway disease associated with significant functional impairment and adverse clinical outcomes.

Several studies have demonstrated the presence of exercise-induced desaturation (EID) and reduced exercise capacity in patients with non-cystic fibrosis bronchiectasis.^4,5 Chronic or recurrent hypoxaemia may contribute to pulmonary vascular remodelling and progression of pulmonary hypertension. ⁶ Exercise intolerance in chronic lung diseases is commonly assessed using the 6-minute walk test (6MWT), a simple and reproducible submaximal exercise test in which exertional oxygen desaturation reflects impaired gas exchange during physical activity. In clinical studies, approximately 43% of patients may exhibit desaturation during the 6MWT (American Thoracic Society, 2002). ⁷ Reduced exercise tolerance and oxygen desaturation during the 6MWT have been shown to predict hospitalisation and mortality in bronchiectasis. ⁵ In particular, a distance–saturation product (DSP), calculated as the 6-minute walk distance (6MWD) multiplied by the lowest oxygen saturation during the 6MWT, of less than 280 m% has been associated with a 66.5-fold higher risk of 6-year mortality. ⁸

Radiological features such as nodular bronchiectasis and tree-in-bud patterns suggest predominant involvement of the small airways in bronchiectasis.^9,10 Functional assessment of small airway dysfunction (SAD) and its physiological and clinical implications remain limited. Traditionally, small airway involvement has been inferred from spirometric indices such as maximal mid-expiratory flow (MMEF) and forced expiratory flow at 50% of forced vital capacity (FEF50%). More recently, oscillometry, including impulse oscillometry (IOS), has enabled non-invasive assessment of small airway mechanics during quiet tidal breathing by measuring respiratory system impedance, comprising resistance and reactance. ¹¹ In bronchiectasis, IOS abnormalities have been associated with radiological disease extent and clinical severity, and may be more sensitive than FEV1 in detecting early or mild disease. ¹²

Despite these advances, the relationship between SAD, EID, and submaximal exercise capacity in bronchiectasis remains incompletely understood. We therefore hypothesised that SAD is associated with EID and reduced functional capacity in bronchiectasis, and sought to identify oscillometric and spirometric thresholds associated with a reduced distance–saturation product during the 6MWT.

Methods

Study design and studied subjects

The study retrospectively analysed the parameters of spirometry, modified Borg scale (mBorg) before and after walk and walk distance (6MWD) during 6MWT, peripheral oxygen saturation (SpO₂) at baseline and continuously monitored to obtain the lowest value during the 6MWT, and IOS parameters of patients with CT-confirmed bronchiectasis in Chang Gung Memorial Hospital (Linkou Branch), a university-based tertiary hospital in Northern Taiwan. The presence of bronchiectasis was confirmed according CT reports issued by licensed radiologists, and scored by two independent chest physicians.

This study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. ¹³ (Supplementary material).

Inclusion criteria

All subjects who underwent 6MWT, CT scanning of lungs (Aquilion One Genesis edition, Canon Medical Systems Corporation, Otawara, Japan) and IOS between March 2022 and August 2023 were included. Eligible participants were adults aged 18 years or older with CT-confirmed bronchiectasis. Those without CT-confirmed bronchiectasis—who had been referred for various clinical indications, such as preoperative functional assessment or the evaluation of respiratory symptoms—were enrolled as the non-bronchiectasis control group. No formal sample size or power calculation was performed, as all consecutive eligible subjects during the study period were included in this retrospective analysis. Subjects with bronchiectasis were further categorised into a desaturator group—defined as a maximal change from baseline in peripheral arterial oxygen saturation (ΔSpO₂) ≤-4%(calculated as the lowest SpO₂ recorded during the 6MWT minus the baseline SpO₂)—and a non-desaturator group.^14,15

Exclusion criteria

Patients who were unable to complete the 6MWT (e.g., presenting with an unstable gait) or those with incomplete key measurements were excluded. In the non-bronchiectasis control group, individuals exhibiting emphysema or any of the following spirometric abnormalities were also excluded: (1) FEV1/FVC < 0.7, (2) FVC < 80% of the predicted value, or (3) MMEF < 65% of the predicted value. ¹⁶

Impulse oscillometry (IOS)

The IOS system, Vyntus IOS/v-max 20c option IOS (Vyaire Medical, Mettawa, IL, USA), is routinely calibrated, as suggested by the manufacturer. The examination and analysis were performed prior to bronchodilator administration according to European Respiratory Society technical standards. ¹⁷ Resistance at 5 Hertz (Hz) (R₅) represents total airway resistance, while the resistance at 20 Hz (R₂₀) represents proximal airway resistance. ¹⁷ The difference between R₅ and R₂₀ (R₅-R₂₀) reflects the small airway index. ¹⁸ Reactance at 5 Hz (X₅) relates to the tissue elastance and inertance (the mass-inertive forces of the moving air column). ¹⁷ Resonant frequency (F_res) is a frequency where the capacitance and inertance effects are equal. ¹⁷ The reactance area (A_X) is the area with width of X₅ and height of F_res under the reactance curve, which is related to respiratory capacitance, small airway patency and correlates with R₅–R₂₀.^17,18

Per cent predicted values for IOS parameters (R₅, R₂₀, and X₅) were obtained from device-generated reference equations implemented in the IOS system (Pred. Module: Stand DE#GLI), which incorporate anthropometric factors including age, sex, height, and ethnicity (South East Asian). For R₅–R₂₀, no validated reference equations are available. Therefore, an estimated predicted value was derived by subtracting predicted R₂₀ from predicted R₅ for each individual. However, given the lack of validation and potential for physiologically implausible estimates, R₅–R₂₀ was primarily interpreted using absolute values. Absolute values were also reported for parameters without established reference equations (e.g. A_X and F_res). Z-scores (zR₅, zR₂₀, zX₅, zF_res and zA_X) were derived using Oostveen et al.’s equation. ¹⁹ zR₅-R₂₀ could was not calculated as normative reference data remain unavailable. ²⁰

6-Minute walk test (6MWT)

Participants with bronchiectasis and control subjects performed the 6MWT as far as possible according to American Thoracic Society guidelines. ⁷ Stop and resting during the test were allowed depending on their own physical condition. Peripheral oxygen saturation (SpO₂) and heart rate were recorded every two seconds with a continuous finger-adapted pulse oximeter WristOx2, model 3150 (Nonin Medical, Inc., Plymouth, MN, USA). When extracting the values for statistical analysis, the data series was thoroughly inspected to identify the lowest SpO₂. Any abnormally low values that deviated significantly from the overall trend were identified as artefacts and excluded from the selection. At the beginning and the end of the 6MWT, pulmonary function (by Pneumotrac 6800 spirometer, Vitalograph Ltd., Buckingham, UK) and qualified exertion were assessed by spirometry and modified Borg scale (mBorg). outcomes. Baseline spirometric parameters were utilised to investigate small airway dysfunction prior to the 6MWT. Post-6MWT spirometric measurements were excluded from the current analysis because they reflect acute physiological responses to exertion, rather than the baseline determinants of exercise-induced desaturation or overall exercise performance.

HRCT scoring system for bronchiectasis

Bronchiectasis Radiologically Indexed CT Score (BRICS) and modified Reiff score were used for the evaluation of bronchiectasis severity (Supplementary Figure S1). BRICS comprise the scoring of bronchial dilatation (absent = 0; mild: lumen just > diameter of adjacent vessels = 1; moderate: lumen 2-3 times > diameter of adjacent vessel =2; severe: lumen >3 times diameter of adjacent vessel = 3) and the number of bronchopulmonary segments with emphysema (none = 0; 1-5 = 1; >5 = 2). The resulting score ranges from 0 to 5(Score 1: mild disease, Score 2-3: moderate disease, Score4-5: severe disease). ¹⁰ The modified Reiff score (ranging from 0 to18) is based on the number of lobes (maximum 6, including lingular lobe) and the severity of bronchial dilatation compared with the adjacent vessel (0 = no bronchiectasis; 1 = tubular, 1-2 times; 2 = varicose, 2-3 times; 3 = cystic, > 3 times). ⁹

Statistical analysis

Statistical analysis was performed using Prism version 10 software package (GraphPad Prism Software Inc, CA, USA) and IBM SPSS Statistics for Windows, version 25 (IBM Corp., Armonk, NY, USA). Results were expressed as median (interquartile range, IQR). Missing data were handled using listwise deletion, where cases with any missing values in the variables included in the models were excluded from the analysis. The differences between non-bronchiectasis controls, bronchiectasis non-desaturators and desaturators were estimated by Kruskal-Wallis test followed by Dunns multiple comparisons. Correlations between parameters were determined by Spearman’s rank correlation in the entire bronchiectasis cohort. Variables that demonstrated both statistically (P<0.10) on univariate analysis and clinically relevance were selected as candidate covariates for the multiple linear regression model to identify independent predictors for DSP. To ensure model parsimony and prevent overfitting, non-significant clinical variables were excluded. Variance inflation factor (VIF) diagnostics were performed to identify and address multicollinearity. Collinear spirometric parameters that exhibited high correlation with R₅-R₂₀ (VIF>5), were omitted to maintain model stability. The predictive efficacy of the IOS and spirometry variables for DSP <280 m% was assessed using the receiver operating characteristic (ROC) curve. Youden index is defined as sensitivity + specificity -1. The point on the ROC curve where the value of Youden index is maximized is selected as the cut-off value. P-values of < 0.05 were accepted as statistically significant.

Results

The medical records of 750 subjects were screened (Figure 1). Of the 129 subjects with CT-confirmed bronchiectasis, 15 were excluded owing to incomplete 6MWT data. A total of 57 bronchiectasis subjects developed EID and were classified as desaturators. The remaining bronchiectasis subjects were classified as non-desaturators (n = 57). Additionally, 53 individuals were included as non-bronchiectasis controls; these subjects had no evidence of bronchiectasis or emphysema, possessed normal lung function (FEV1/FVC ≥ 0.7, FVC ≥ 80% pred, MMEF ≥ 65%pred) and had successfully completed the 6MWT.

Figure 1.

The flow chart of enrolled subjects. IOS: impulse oscillometry; 6MWT: 6-minute walk test; CT computed tomography; ΔSpO₂: a maximal change from baseline in peripheral arterial oxygen saturation during 6MWT (calculated as the lowest SpO₂ recorded during the 6MWT minus the baseline SpO₂); FEV1: forced expiratory volume in 1 second; FVC: forced vital capacity; MMEF: maximal mid-expiratory flow.

Baseline characteristics, comorbidities, and radiographic severity are summarised in Table 1. Compared with non-bronchiectasis controls, bronchiectasis desaturators were older (62 [IQR 49-72] vs 69 [IQR 63–76.5] years; p<0.01), exhibited lower serum eosinophil cationic protein (ECP) levels (5.9[IQR 10.6-21.8] vs 5.4[IQR 2.7-11.9] ug/L; p<0.05), and had a greater cumulative tobacco smoke burden(0[IQR 0-0] vs 0 [IQR 0-20]pack-year; p<0.05). The median body mass index (BMI) was significantly lower in bronchiectasis non-desaturators than in the control group (23.5[IQR20.4-26.5] vs 25.8[IQR 22.7-28.1] kg/m²; p<0.05). No significant differences in comorbidities were observed between the control and bronchiectasis groups, which reached statistical significance (p < 0.05). Regarding potential cardiovascular confounding, no statistically significant differences in tricuspid regurgitation velocity (V_max), tricuspid regurgitation gradient (TRG), or left ventricular ejection fraction (LVEF) were observed between control desaturators, non-desaturators, bronchiectasis desaturators and bronchiectasis desaturators (Supplementary Table S1). Among patients with bronchiectasis, desaturators exhibited significantly higher BRICS (2[IQR 1-3] vs 1 [1-2]; p<0.001) and modified Reiff scores (5 [IQR 3-8] vs 4[IQR 2-6]; p<0.05) than non-desaturators.

Table 1.

Baseline characteristics and radiographic grading scores of controls and bronchiectasis patients with or without exercise-induced desaturation.

​	Controls (53)	Bronchiectasis non-desaturators (57)	Bronchiectasis desaturators (57)
Male	27(50.9%)	32 (56.1%)	32 (56.1%)
Age (years)	62(49-72)	68(57-74)	69(63-76.5)##
Body height (cm)	163(153.5-167)	162(157-167)	160(152.5-165.5)
Body weight (kg)	66.0(57.0-72.0)	60(52.5-71)	62.0(50.0-72.8)
BMI (kg/m2)	25.8(22.7-28.1)	23.5 (20.4-26.5)#	24.3(21.0-27.0)
Total IgE (IU/mL)	41.7(12.7-137.5)	54.8(14.5-191.0)	51.0(19.4-146.0)
ECP (ug/L)	5.9(10.6-21.8)	8.8(4.3-14.5)	5.4(2.7-11.9)#
Absolute eosinophil count (/uL)	97.6(46-171.5)	109(57-177)	122(62-205.5)
Absolute neutrophil count (/uL)	4160(3122-6677)	4166(2832-5184)	3091(2302-4811)
Smoking pack-year	0(0-0)	0(0-18)	0(0-20)#
Comorbidities
Hypertension	17(32.1%)	15(26.3%)	24(42.1%)
CVA	2(3.8%)	2(3.5%)	3(5.3%)
CAD	2(3.8%)	4(7.0%)	4(7.0%)
Atrial fibrillation	1(1.9%)	2(3.5%)	6(10.5%)
Heart failure	2(3.8%)	2(3.5%)	4(7.0%)
Diabetes mellitus	9(17.0%)	8(14.0%)	13(22.8%)
Dyslipidaemia	11(20.8%)	6(10.5%)	15(26.3%)
Autoimmune diseases	3(5.7%)	3(5.3%)	2(3.5%)
CKD	2(3.8%)	8(14.0%)	4(7.0%)
Emphysema	--	9 (15.8%)	17 (29.8%)
Asthma	14(26.4%)	21(36.8%)	16(28.1%)
Rhinosinusitis	19(35.8%)	17(29.8%)	19(33.3%)
OSA	1(1.9%)	2(3.5%)	5(8.8%)
Reversibility (mL)	6(3.7-26)	120(10-270)	105(18-265)
Reversibility (%)	3(-1-13)	7(1-17)	7(1-18)
BRICS	--	1(1-2)	2(1-3)***
modified Reiff score	--	4(2-6)	5(3-8)*

Patients with computed tomography (CT)–confirmed bronchiectasis were categorised as either desaturators (ΔSpO₂ ≤ −4% during the 6MWT) or non-desaturators (ΔSpO₂ > −4% during the 6MWT). Control subjects exhibited normal lung function (FEV1/FVC ≥ 0.7, FVC ≥ 80%pred, and MMEF ≥ 65%pred) and no CT evidence of bronchiectasis or emphysema. All data are presented as the median (interquartile range).

Abbreviations: BMI, body mass index; IgE, immunoglobulin E; ECP, eosinophil cationic protein; CVA, cerebrovascular accident; CAD, coronary artery disease; CKD, chronic kidney disease; OSA, obstructive sleep apnoea; Reversibility, change in FEV1 from baseline after bronchodilator administration; BRICS, Bronchiectasis Radiologically Indexed CT Score; #p < 0.05, ##p < 0.01, ###p < 0.001 compared with non-bronchiectasis controls; *p < 0.05, **p < 0.01, ***p < 0.001 compared with bronchiectasis non-desaturators.

To further explore the impact of exercise-induced desaturation within the control population, an additional stratified analysis dividing controls into non-desaturators and desaturators is provided in Supplementary Table S1.

Bronchiectasis desaturators exhibit more pronounced SAD than non-desaturators

IOS and 6MWT variables are presented in Table 2, Supplementary Table S2 and Figure S2. IOS-derived lung impedance parameters did not differ significantly between controls and bronchiectasis non-desaturators. In contrast, compared with non-desaturators, bronchiectasis desaturators exhibited increased respiratory impedance, reflected by higher R₅ (0.33[IQR 0.29-0.42] vs 0.43[IQR 0.33-0.50]; p<0.05), F_res 17.69[IQR 15.26-20.09] vs 19.86 [IQR 16.32-23.67]; p<0.05), and A_X values (0.60[IQR 0.375-0.995] vs 1.24 [IQR 0.59-1.90]; p<0.01). Bronchiectasis desaturators also demonstrated significantly poorer lung function, including reduced expiratory airflow in the small airways (lower MMEF%pred and FEF50%pred), accompanied by greater exertional burden (higher post-6MWT mBorg scores and ΔmBorg), reduced exercise capacity (shorter 6MWD), greater oxygen desaturation (ΔSpO₂), and a lower DSP than control and bronchiectasis non-desaturators.

Table 2.

Parameters of the 6-minute walk test (6MWT) and impulse oscillometry in controls, bronchiectasis non-desaturators, and desaturators.

​	Controls (53)	Bronchiectasis non-desaturators (57)	Bronchiectasis desaturators (57)
R5 (kPa/L/s)	0.35 (0.29- 0.41)	0.33 (0.29- 0.42)	0.43 (0.33- 0.50)#*
R5%pred	100(87-120)	97(80-120)	117(97-146)#**
zR5	0.30(-0.11-0.90)	0.70 (-0.04 – 1.26)	1.08 (0.46 -1.97) ###
R20 (kPa/L/s)	0.27 (0.24- 0.33)	0.260 (0.225- 0.295)	0.290 (0.240- 0.340)
R20%pred	95(78-108)	85(75-99)	95(82-108.5)
zR20	-0.29(-1.00-0.28)	-0.29 (-1.10-0.31)	-0.03(-0.68- 0.60)
R5-R20 (kPa/L/s)	0.07 (0.05- 0.11)	0.08 (0.05- 0.125)	0.14 (0.07-0.18)##
R5-R20%pred	155(104-213)	170(113-246)	286(141-369) *###
X5 (kPa/L/s)	-0.09 (-0.12- -0.06)	-0.12 (-0.165- -0.07) #	-0.16 (-0.26- -0.10)###
X5%pred	209(83-458)	255(122-422)	319(174-544)
zX5	-1.18(-2.11- -0.37)	-0.10 (-1.28 -1.27) ##	0.36(-0.56- 2.19) ###
Fres (Hz)	16.4 (14.1- 18.2)	17.69 (15.26- 20.09)	19.86 (16.32- 23.67) ###*
zFres	1.10(0.36-1.61)	1.53(1.00-2.04)#	2.03(1.08-2.73) ###
AX (kPa/L)	0.48 (0.26 -0.67)	0.60 (0.375- 0.995)	1.24 (0.59- 1.90) ###**
zAX	0.43(-0.46-0.91)	1.16(-0.07-1.77) ###	1.65(0.93-2.34) *###
FVC%pred	97 (87- 105)	82 (71.0- 94.5)###	70.0 (58.0- 82.5)###*
FEV1%pred	97 (91-106)	75 (59.5- 84.5) ###	53.0 (40.5- 71.5)###**
FEV1/FVC(%)	82 (80- 85)	73 (65- 79) ###	62 (53.0- 75.5)###*
MMEF%pred	80 (71- 92)	42 (26-60.5)###	24 (14.0- 38.5)###**
FEF50%pred	85 (76- 104)	45 (27.5- 64)###	23.0 (14- 38.3) ###**
Pre-6MWT IC %pred	86 (72- 102)	70 (60- 83)###	59 (48.0- 76.0)###*
Post-6MWT IC %pred	82 (73- 96)	67 (58- 81.5)###	58 (49.0- 69.5) ###*
ΔIC%pred	-3.0 (-10- 4)	-2 (-8- 5)	-2.0 (-8.5- 3.5)
Pre-6MWT SpO2 %	97 (96- 98)	96 (95- 97)#	96.0 (94.0- 97.0)#
Lowest SpO2 %	93 (92- 95)	94 (93- 95)	89.0 (84.5- 91.5)###***
ΔSpO2 %	-5 (-3- -2)	-2 (-3- -1)###	-6.0 (-9- -5)###***
Pre-6MWT mBorg	0 (0 -0)	0 (0 -0)	0 (0- 2)
Post-6MWT mBorg	3 (2- 4)	4 (3 -4)	4 (4- 5)###**
ΔmBorg	3 (2-4)	3 (2- 4)	4 (2- 4)#
6MWD (m)	546 (473- 588)	507 (449- 546)	455.0 (377.5- 520.0)###
DSP (m%)	497 (445- 550)	461.8 (416.8- 511.4)	398.6 (317.0- 459.5)###**

Patients with Computed tomography (CT)–confirmed bronchiectasis were categorised as either desaturators (ΔSpO₂ ≤ −4% during the 6MWT) or non-desaturators (ΔSpO₂ > −4% during the 6MWT). Control subjects exhibited normal lung function (FEV1/FVC ≥ 0.7, FVC ≥ 80%pred, and MMEF ≥ 65%pred) and no CT evidence of bronchiectasis or emphysema. All data are presented as the median (interquartile range).

Abbreviations: 6MWD, 6-minute walk distance; 6MWT, 6-minute walk test; A_X, reactance area; DSP, distance–saturation product; FEF50, forced expiratory flow at 50% of vital capacity; FEV1, forced expiratory volume in 1 second; F_res, resonant frequency; FVC, forced vital capacity; IC, inspiratory capacity; Lowest SpO₂%, lowest peripheral arterial oxygen saturation during the 6MWT; MMEF, maximal mid-expiratory flow; Post-6MWT, after the 6MWT; Pre-6MWT, before the 6MWT; Pre-6MWT SpO₂%, baseline peripheral arterial oxygen saturation before the 6MWT; R₅, resistance at 5 Hz; R₂₀, resistance at 20 Hz; X₅, reactance at 5 Hz; Δ, difference before and after exercise; ΔSpO₂, maximal change from baseline in peripheral arterial oxygen saturation during the 6MWT(calculated as the lowest SpO₂ recorded during the 6MWT minus the baseline SpO₂); %pred, percentage of predicted value. #p < 0.05, ##p < 0.01, ###p < 0.001 compared with non-bronchiectasis controls; *p < 0.05, **p < 0.01, ***p < 0.001 compared with bronchiectasis non-desaturators.

Oxygen desaturation and exercise intolerance during the 6MWT are associated with SAD in bronchiectasis

We further demonstrated associations between small airway parameters and oxygenation and exercise capacity in patients with bronchiectasis. Increased small airway resistance, indicated by higher R₅–R₂₀ values, greater reactance (more negative X₅, higher F_res, and larger A_X), and reduced expiratory flow in the small airways (lower MMEF%pred and FEF50%pred) were associated with greater reductions in ΔSpO₂ and 6MWD (Supplementary Table S4 and Figure S3). These small airway parameters were not correlated with subjective breathlessness, as assessed by the mBorg scale after the 6MWT. Small airway indices were significantly correlated with modified Reiff and BRICS scores, except for the associations between Fres and the modified Reiff score and between R₅–R₂₀ and BRICS. Notably, X₅, MMEF%pred, and FEF50%pred showed moderate negative correlations with both radiological scores (Supplementary Figure S4).

According to previous studies, a DSP < 280 m% indicates poor outcomes in patients with chronic obstructive pulmonary disease (COPD) and bronchiectasis.^8,21 The multiple linear regression model showed that age, BRICS, and R₅-R₂₀ significantly predicted DSP, accounting for 46.3% of the variance (Adjusted R² = 0.463, F (8, 104) = 12.45, p < 0.001) (Table 3). Although age (β=-0.394) is the strongest predictor of exercise capacity, our study found that the impact of R₅-R₂₀ (β=-0.345) surpassed that of BRICS (β=-0.285).

Table 3.

Multiple linear regression analysis of factors associated with DSP in patients with bronchiectasis.

Variables	B	SE	β	T	P value	95% confidence interval (CI)		VIF
Male	-7.735	23.209	-0.036	-0.333	0.740	-53.776	38.306	2.373
BH	1.982	1.319	0.166	1.503	0.136	-0.634	4.599	2.489
Age	-3.507	0.680	-0.394	-5.156	<0.001	-4.856	-2.158	1.187
Smoking pack-year	-0.272	0.475	-0.045	-0.573	0.568	-1.215	0.671	1.266
BRICS	-27.574	6.980	-0.285	-3.951	<0.001	-41.420	-13.729	1.057
R5-R20*	-40.262	8.933	-0.345	-4.507	<0.001	-57.982	-22.541	1.188
ΔIC%	0.793	0.782	0.072	1.015	0.313	-0.757	2.343	1.030
ΔmBorg	0.051	5.354	0.001	0.010	0.992	-10.569	10.671	1.108

Dependent variable: DSP (distance–saturation product), Adjusted R²: 0.463, F (8, 105) = 12.45, p < 0.001.

*R₅-R₂₀ values were rescaled (multiplied by 10) for the regression analysis to provide more clinically interpretable coefficients.

Abbreviations: B, unstandardized coefficients; SE, standard error; β, standardized coefficients; $\mathrm{T}=\frac{\mathrm{B}}{\text{SE}}$ ; CI, confidence interval; VIF, variance inflation factor; BH, body height; BRICS, Bronchiectasis Radiologically Indexed CT Score; R₅–R₂₀, difference between resistance at 5 Hz and resistance at 20 Hz; Δ, difference before and after exercise; IC, inspiratory capacity; mBorg, modified Borg scale.

Diagnostic efficacy, cut-off values, and sensitivity and specificity of spirometric and IOS variables for predicting DSP < 280 m%

A DSP < 280 m% predicts a 66.5-fold increased risk of 6-year mortality in patients with non-cystic fibrosis bronchiectasis. ⁸ Cut-off values, sensitivities, and specificities for spirometric and IOS parameters are presented in Figure 2 and Supplementary Table S3. ROC curves assessing DSP < 280m% based on Z-score adjusted impulse oscillometry parameters in non-desaturated and desaturated bronchiectasis patients are displayed in Supplementary Figure S5.

Figure 2.

ROC curve of DSP <280 m% with different parameters of pulmonary function test and impulse oscillometry in non-desaturated and desaturated bronchiectasis patients. R₅ and R₂₀: resistance at 5 Hz and 20 Hz; R₅-R₂₀: difference between resistance at 5 Hz and resistance at 20 Hz; X₅: reactance at 5 Hz, F_res: resonant frequency; A_Χ: reactance area; %pred: percentage of predicted value; FVC: forced vital capacity; FEV1: forced expiratory volume in 1 second; MMEF: maximal mid-expiratory flow; FEF50: forced expiratory flow at 50% of vital capacity; AUC: area under receiver operating characteristic curve.

Among all spirometric and IOS indices, A_X demonstrated the highest area under the receiver operating characteristic curve (AUC). The optimal A_X cut-off value for predicting DSP < 280 m% in bronchiectasis was >1.26 kPa/L (AUC 0.86, p = 0.0001), with a sensitivity of 90.91% and a specificity of 72.82%.

Discussion

We showed increased IOS-defined airway resistance (R₅, R₅-R₂₀) and lung reactance (X₅, Fres and A_X), as well as reduced spirometry-defined lung function (FVC%pred, FEV1%pred, FEV1/FVC, MMEF%pred, FEF50%pred, IC%pred before and after 6MWT), in bronchiectasis desaturators. IOS data were inversely correlated with spirometry data. After adjusting for confounding factors, age, BRICS, and R₅-R₂₀ remained independent predictors of DSP. Notably, for every 0.1 unit increase in R₅-R₂₀, a reduction of approximately 40.26 m% in DSP was observed (B = -40.262, p < 0.001). While age remains the primary determinant of exercise tolerance, our findings reveal that R₅-R₂₀ exerts a greater influence than BRICS. This indicates that small airway resistance is a more decisive factor for DSP than visible radiographic changes. The ΔSpO₂ were correlated with airway reactance (R₅%pred, R₅, R₂₀, R₅-R₂₀), lung reactance (X₅, Fres, A_X) and spirometry-defined lung function.

Area of Reactance (A_X), defined as the integral of the reactance curve from the lowest frequency to the Fres, captures the elastic properties of the lung dominated by capacitance. Compared to other IOS parameters—namely R₅-R₂₀, X₅, and Fres—A_X provides a more comprehensive assessment of pulmonary mechanics. ²² A_X also demonstrates the greatest percentage change among these parameters following bronchodilator administration in patients with asthma, alongside consistently noted baseline elevations in the previous studies.^16,23 These parameters, particularly R₅-R₂₀ and A_X, serve as surrogate markers for peripheral airway resistance and alveolar compliance.^24,25 By reflecting both small-airway obstruction and parenchymal damage, they are intrinsically linked to impaired gas exchange and reduced exercise capacity. This cumulative physiological decline is clinically captured by the DSP. Our findings demonstrate that R₅-R₂₀ and A_X are closely associated with the DSP <280 m%, suggesting their utility as non-invasive early indicators for patients at risk of functional decline, especially in clinical settings where the 6-minute walk test (6MWT) is impractical or contraindicated. To the best of our knowledge, this is the first report revealing the of role of IOS in prediction of O₂ desaturation and exercise intolerance in bronchiectasis. Importantly, reduced oxygenation and exercise capacity have been shown to predict mortality in bronchiectasis.

Due to a limited pool of eligible subjects, the non-bronchiectasis control group was not matched for age or body size and was generally younger. However, this imbalance has a negligible impact on the main findings, as the regression analyses focused on differences within the bronchiectasis cohort (desaturators vs non-desaturators) rather than comparisons with the control group. Furthermore, while height often influences walking performance, it was not an independent predictor in the final model, as its advantages tend to diminish with ageing. ²⁶

Bronchiectasis is a chronic lung disease characterised by persistent inflammation, progressive lung destruction, and impaired pulmonary function, leading to reduced gas exchange and EID. ²⁷ Desaturation during the 6MWT has been associated with more severe lung disease, increased dyspnoea, reduced muscle strength, and limited daily physical activity across chronic respiratory diseases. ²⁸ In our study, abnormalities in oscillometry-derived small airway parameters were also associated with EID and reduced 6MWD.

Although normal and restrictive spirometry patterns can also be observed, the typical ventilatory pattern of bronchiectasis is progressive airway obstruction. ²⁹ Compared with bronchiectasis non-desaturators, desaturators had lower FEV₁ but not a higher prevalence of CT-confirmed emphysema, suggesting that emphysematous air sac destruction may not be the main factor contributing to EID in bronchiectasis. COPD accounted for 10% of cases of bronchiectasis in Taiwan, ³⁰ and COPD-related structural lung changes may coexist in some patients. However, our findings suggest that emphysema is unlikely to be the primary driver of EID in this cohort.

In bronchiectasis, repeated infections lead to structural changes in the smaller airways, including fibrosis, enlargement of lymphoid follicles in the airway walls, airway wall thickening, and mucus retention.^31–33 The decline in expiratory flow rate reflects the development of small airway constrictive bronchiolitis, which may also be associated with mosaic attenuation on expiratory phase of computed tomographic scanning. ³⁴ These alterations contribute to airflow limitation and air trapping due to loss of elastic recoil and airways closure. ³⁵ These mechanisms may help explain the association between small airway dysfunction and EID observed in our study. Guan et al. demonstrated that elevated IOS parameters were associated with higher Bronchiectasis Severity Index (BSI) and HRCT scores (modified Reiff score), an increased number of bronchiectatic lobes in bronchiectasis. ¹² All IOS parameters but X₅ were positively correlated with the duration of bronchiectasis symptoms, number of bronchiectatic lobes, HRCT total scores, and BSI (all p < 0.05), but not sputum bacterial density (p > 0.05). Compared with FEV1, any single IOS parameter being abnormal was more common in mild bronchiectasis, particularly in patients with HRCT score of 5 or lower. IOS parameters, particularly X₅, which indicates reduced lung elasticity, have demonstrated predictive value for assessing bronchiectasis severity as measured by the FACED score. ³⁶ Our study demonstrated a moderate negative correlation of MMEF%pred and FEF50%pred with the moderate Reiff score and BRICS, as well as X₅ with the moderate Reiff score (p<0.05). Other small airway parameters showed either low correlation (p<0.05) or no correlation (Supplementary Figure S4). In this study, IOS parameters were not more correlated BSI/HRCT scores than FEV1, MMEF%pred, FEF50%pred.

This study has several limitations. First, although per cent predicted values were reported for IOS parameters with available reference equations, certain indices (particularly X₅ and R₅-R₂₀) showed values deviating from expected ranges in the control group. This likely reflects limitations of currently available reference equations, which are not population-specific and may not adequately represent our study population. In addition, R₅-R₂₀ lacks validated reference equations, and derived predicted values may not be physiologically meaningful. Therefore, absolute IOS values may provide a more reliable representation of small airway dysfunction in this cohort. Notably, $\mathrm{z}‐\text{score}$ -derived IOS indices (zR₅, z R₂₀, zR₅-R₂₀, zX₅, zF_res, and zA_X) demonstrated lower discriminatory performance than their absolute counterparts. This discrepancy likely arises because exercise capacity depends on both respiratory impedance and anthropometric traits—the latter being preserved in absolute values but eliminated in $\mathrm{z}‐\text{scores}$ . Nonetheless, absolute cut-off thresholds from a single-centre retrospective study should be applied cautiously, as they may lack generalisability to other populations. Although spirometry was performed before and after the 6MWT, oscillometric measurements were only available prior to exercise; therefore, we could not determine whether changes in resistance and reactance occurred dynamically in parallel with exertional symptoms, oxygen desaturation, or reduced 6MWD. The cross-sectional design also precludes inference regarding the directionality of the association between hypoxaemia and SAD. Another limitation is the lack of systematic screening for pulmonary hypertension or heart failure in all consecutive participants because of the retrospective design. Although this could theoretically confound the findings in control desaturators, our available echocardiographic findings indicated no substantial burden of underlying cardiovascular disease or pulmonary hypertension within this subgroup. The lack of grading for emphysema severity, combined with the absence of diffusing capacity of the lungs for carbon monoxide ( $\text{DLCO}$ ) measurements, precludes a full determination of how alveolar destruction and impaired gas exchange contribute to exercise-induced desaturation. Furthermore, our analysis focused primarily on radiological severity, as the inclusion criteria were based on CT diagnosis. Due to the retrospective nature of the records, multidimensional indices such as the Bronchiectasis Severity Index (BSI), FACED score, patient-reported symptoms and long-term clinical outcomes were not available owing to the retrospective nature of the records, limiting assessment of the prognostic significance of oscillometric abnormalities. Future studies incorporating serial oscillometric measurements during exercise, as well as longitudinal follow-up, may help clarify the temporal and mechanistic relationships between SAD, EID, and functional limitation in bronchiectasis.

Conclusions

Our findings highlight a close association between SAD and EID in non-cystic fibrosis bronchiectasis. Reduced mid-expiratory flow on spirometry and increased resistance and reactance on oscillometry were associated with greater oxygen desaturation and reduced exercise capacity. Among small airway indices, A_X demonstrated the strongest association with a reduced distance–saturation product, suggesting that oscillometric measures may capture physiologically relevant abnormalities not fully reflected by conventional spirometry. These results support the concept that small airway mechanics contribute to exercise-related gas exchange vulnerability in bronchiectasis and may represent an important physiological domain for further investigation across chronic airway diseases.

Appendix.

Abbreviations

Δ

difference before and after exercise

ΔSpO2

a maximal change from baseline in peripheral arterial oxygen saturation during 6MWT (calculated as the lowest SpO2 recorded during the 6MWT minus the baseline SpO2)%pred: percentage of predicted value

6MWD

6-minute walk distance

6MWT

6-minute walk distance

AUC

area under receiver operating characteristic curve

AX

reactance area

BRICS

Bronchiectasis Radiologically Indexed CT Score

BMI

body mass index

DSP

distance-saturation product

ECP

eosinophil cation protein

EID

exercise-induced desaturation

FEF50

forced expiratory flow at 50% of vital capacity

FEV1

forced expiratory volume in 1 second

Fres

resonant frequency

FVC

forced vital capacity

IC

inspiratory capacity

IgE

immunoglobulin E

IOS

impulse oscillometry

Lowest SpO2

lowest peripheral arterial oxygen saturation during 6MWT

mBorg

modified Borg scale

MMEF

maximal mid-expiratory flow

Pre-6MWT

before 6MWT

Post-6MWT

after 6MWT

Pre-6MWT SpO2

baseline peripheral arterial oxygen saturation before 6MWT

R5

resistance at 5Hertz(Hz)

R20

resistance at 20Hz

ROC curve

receiver operating characteristic curve

SAD

small airway dysfunction

X5

reactance at 5 Hz.

ORCID iDs

Chiung-Hsin Chang https://orcid.org/0000-0003-2495-6464

Chun-Yu Lin https://orcid.org/0000-0003-4165-3588

Chun-Yu Lo https://orcid.org/0000-0003-3864-8702

Ethical considerations

The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of Chang Gung Memorial Hospital (approval number: 202301380B0).

Consent to participate

The requirement for informed consent was waived by the Institutional Review Board owing to the retrospective nature of the study.

Data Availability Statement

The data sets analysed during the current study are available from the corresponding author upon reasonable request.*