Clinical, imaging, bronchoscopic characteristics and pulmonary function features in patients with endobronchial tuberculosis: a single-centre cross-sectional study Authors: Cuong Kim Nguyen, The Manh Nguyen, Minh Le Nhat Nguyen, Dai Khac Vu, Phuong Anh Thi Nguyen, Thai Duc Truong, Luong Van Dinh Published: 2026 ABSTRACT AbstractObjectivesEndobronchial tuberculosis (EBTB) is a common manifestation of tuberculosis (TB) that may cause airway obstruction if not promptly diagnosed and treated. This study described clinical, imaging, bronchoscopic characteristics and pulmonary function in patients with EBTB.DesignA single-centre cross-sectional study.SettingThe study was conducted at the National Lung Hospital, a tertiary hospital specialised in TB and lung diseases in Vietnam from May 2023 to May 2024.ParticipantsWe consecutively enrolled 91 patients aged ≥16 years with bacteriologically confirmed EBTB, defined by positive Mycobacterium tuberculosis detection using GeneXpert MTB/RIF assay or liquid culture from bronchial biopsy or aspirate specimens with or without compatible histopathological findings. All participants underwent a standardised diagnostic workup, including chest high resolution CT (HRCT) scan, spirometry, bronchoscopy and histopathology.Outcome measuresThe outcome of our study was the presence of central airway obstruction (CAO), defined as ≥50% narrowing of the trachea, mainstem bronchi, bronchus intermedius, lobar and segmental bronchi visually estimated on bronchoscopy and supported by HRCT findings.ResultsMedian age was 37 years (IQR: 27–54), 64.8% female; 31.9% were post-TB treatment cases and 68.1% newly diagnosed. Overall, 65.9% of patients had ventilatory defects, most commonly the restrictive pattern (66.7%). Lesions were more frequent in the left airway than right (64.8% vs 31.9%, p<0.001). CAO was found in 71.4% of patients, more common in post-TB cases (86.2% vs 64.5%). Multivariate logistic regression analyses showed that three variables were independently associated with the risk of CAO on bronchoscopy, including shorter duration from symptom onset to hospitalisation (<8 weeks) (adjusted OR (aOR) = 3.93, 95% CI 1.18 to 13.03), dyspnoea (aOR = 3.85, 95% CI 1.27 to 11.68) and forced expiratory volume in one second (FEV1) <80% (aOR = 5.45, 95% CI 1.28 to 23.33).ConclusionVentilatory defects and CAO were common among patients with EBTB. A high level of suspicion should be applied to patients presenting with dyspnoea, shorter duration from symptom onset to hospitalisation (<8 weeks) and reduced FEV1 (< 80%). WHAT IS ALREADY KNOWN ON THIS TOPIC Endobronchial tuberculosis (EBTB) can cause significant tracheobronchial stenosis and long-term ventilatory impairment, but its clinical and functional characteristics remain under-reported, especially in low-income and middle-income countries. Endobronchial tuberculosis (EBTB) can cause significant tracheobronchial stenosis and long-term ventilatory impairment, but its clinical and functional characteristics remain under-reported, especially in low-income and middle-income countries. WHAT THIS STUDY ADDS This study demonstrates that central airway obstruction was present in over two-thirds of patients with EBTB, particularly among those with a history of previous tuberculosis (TB) treatment, and identifies short symptom duration, dyspnoea and reduced forced expiratory volume in one second (FEV₁) as independent predictors. This study demonstrates that central airway obstruction was present in over two-thirds of patients with EBTB, particularly among those with a history of previous tuberculosis (TB) treatment, and identifies short symptom duration, dyspnoea and reduced forced expiratory volume in one second (FEV₁) as independent predictors. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY Early bronchoscopy and spirometric assessment should be integrated into TB care to enable timely detection of airway involvement and guide interventions to prevent irreversible airway damage. Early bronchoscopy and spirometric assessment should be integrated into TB care to enable timely detection of airway involvement and guide interventions to prevent irreversible airway damage. Introduction Tuberculosis (TB) remains the second leading cause of death from infectious diseases worldwide, following COVID-19. According to the Global Tuberculosis Report 2024, an estimated 10.8 million people developed TB and 1.25 million died from the disease in 2023. Vietnam is among the 30 countries with the highest TB burden globally, with approximately 173 000 new cases and 12,000 TB-related deaths reported in 2023.1 Endobronchial tuberculosis (EBTB) is defined as tuberculous infection of the tracheobronchial tree with microbiological and histopathological evidence.2 Although EBTB can involve any part of the bronchial tree, its impact is most severe when affecting the central airways—specifically the trachea, main bronchi and lobar bronchi. However, there are difficulties and delays in diagnosis due to the need for advanced investigations, such as chest CT, preferably high resolution CT (HRCT), or bronchoscopy.3 Both HRCT and bronchoscopy are essential. The standard criterion for a significant central airway obstruction (CAO) is typically defined as a reduction in the airway lumen’s cross-sectional area by more than 50%.4 In a prospective study of 429 patients with newly diagnosed active pulmonary TB who underwent routine bronchoscopy, the prevalence of EBTB was 54.3%, and persistent bronchial stenosis—defined as a narrowing of the airway lumen by more than one third—was observed in 29.6% of patients with tracheal, main bronchial or lobar bronchial involvement.5 EBTB can cause progressive airway narrowing if not recognised and treated promptly. This narrowing is often irreversible, leading to chronic respiratory symptoms, recurrent infections and long-term functional impairment. Early detection and timely initiation of treatment can help prevent severe airway obstruction and preserve lung function. Several studies have shown that delayed diagnosis is associated with a higher risk of fixed airway stenosis and poorer lung function outcomes.6 This highlights the importance of considering EBTB in patients with unexplained respiratory symptoms and performing bronchoscopy early when the disease is suspected. Despite its clinical relevance, EBTB remains poorly described in many high-burden settings, particularly in populations with diverse TB histories and the impact of prior TB treatment on the severity of airway involvement and lung function decline is not well understood. Therefore, this study aimed to (1) describe the clinical, radiological, bronchoscopic and spirometric characteristics of patients with bacteriologically confirmed EBTB, and (2) identify factors associated with CAO. Methods Patient and public involvement Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research. Study design and participants This was a cross-sectional study conducted at the National Lung Hospital in Vietnam from May 2023 to May 2024. The National Lung Hospital is a tertiary referral centre specialising in TB and respiratory diseases, serving as the national reference institution for TB diagnosis, treatment and interventional pulmonology. The hospital receives patients from all regions of Vietnam, including referrals for complicated or treatment-resistant TB. We enrolled all patients aged 16 years and older who were diagnosed with EBTB based on bronchoscopy and histopathological examination. A consecutive sampling technique was applied, whereby all eligible patients who met the inclusion criteria and presented to the National Lung Hospital during the study period (May 2023–May 2024) were enrolled until the sample target was achieved. Inclusion and exclusion criteria Among patients with bacteriologically confirmed pulmonary TB, those who had symptoms or imaging findings suggestive of airway involvement (such as segmental or lobar collapse, atelectasis or bronchial narrowing on HRCT scan) were referred for bronchoscopy. Patients were eligible for inclusion in this study if they met one of the following criteria: (1) histopathology of bronchial mucosal biopsy showed typical granulomatous inflammation consistent with TB (including all components: lymphocytes, epithelioid histiocytes, Langhans giant cells and caseous necrosis) or (2) suggestive histopathology of bronchial mucosal biopsy showed non-typical granulomatous inflammation (presenting at least one of the above-mentioned histological findings) in a patient with a history or current bacteriological confirmation of TB or (3) positive liquid medium culture of bronchial mucosal biopsy specimen for Mycobacterium tuberculosis. We excluded patients with any comorbidities known to be potentially associated with bronchial stenosis, including people living with HIV, those with a history of chronic respiratory diseases, aspergillosis, lung cancer or endotracheal intubation; patients with airway stenosis secondary to mediastinal tumour, individuals with a smoking history of ≥10 pack-years or exposure to biomass smoke for ≥10 years. Data collection All participants underwent standardised evaluations including chest HRCT, spirometry, bronchoscopy and histopathology at the time of enrolment, regardless of their TB treatment status. For newly diagnosed or currently treated patients, assessments were performed before or within the early phase of anti-TB therapy. For those with a history of completed TB treatment, evaluations were performed at the time of presentation for persistent or recurrent respiratory symptoms, regardless of the duration since treatment completion. The exact time interval from the end of anti-TB therapy to spirometry was not systematically recorded. The primary outcomes were CAO, defined as ≥50% occlusion of the airway lumen (including the trachea, mainstem bronchi, bronchus intermedius or lobar bronchi and segmental bronchi) as observed by bronchoscopy, estimated by visual assessment and confirmed with HRCT when available. Pulmonary function testing was performed for all participants using the forced oscillation technique system (Chest MI, Tokyo, Japan) following standard protocols. Ventilatory defects were classified as obstructive, restrictive or mixed according to the American Thoracic Society/European Respiratory Society (ATS/ERS) guidelines.7 Obstructive ventilatory defect was defined as FEV₁/forced vital capacity (FVC) <0.70. Restrictive ventilatory defect was defined as FEV₁/FVC ≥0.70 with FVC <80% of the predicted value. Mixed ventilatory defect was defined as FEV₁/FVC <0.70 and FVC <80% predicted.8 Flow-volume loops were recorded for each participant to evaluate the pattern of airflow limitation. Qualitative assessment of the loops was used to support the clinical diagnosis of airway obstruction. Bronchoscopy was conducted using the Olympus BF-Q170 electronic bronchoscope under local anaesthesia with topical lidocaine. The procedure was performed by experienced pulmonologists in a negative-pressure room with full infection control precautions. Bronchial mucosal biopsies were obtained for histopathological and microbiological analyses. During bronchoscopy, the degree of airway narrowing was estimated visually by using the scope size that can bypass the obstruction, or by comparing the stenotic segment to the nearest normal airway lumen.9 The demographics and clinical data were extracted from the medical records. Variables The dependent variable was the presence of CAO, defined as ≥50% reduction of the airway lumen involving the trachea, mainstem, intermedius, lobar or segmental bronchi, as assessed by bronchoscopy and confirmed by HRCT when available. The independent variables included age, sex, time to diagnosis of EBTB, dyspnoea, FEV1 and history of TB treatment. Age was categorised according to the WHO age classification: <45 years (young adults), 45–59 years (middle-aged adults) and ≥60 years (older adults). Statistical analysis We used descriptive statistics to analyse the study participants. Continuous variables were checked for normality and presented as mean (SD) or median (IQR), as appropriate. Group comparisons were performed using t-test or Mann-Whitney U test, depending on the distribution. Categorical variables were presented as numbers and percentages and compared using χ2 or Fisher’s exact test, as appropriate. To assess the association of demographic characteristics, chest CT findings, history of TB treatment, pulmonary function and the CAO on bronchoscopy, we constructed multivariable logistic regression models. Both unadjusted and adjusted ORs with 95% CIs were calculated to describe these associations. Statistical analyses were performed using IBM SPSS Statistics for Windows (V.27.0. Armonk, New York: IBM Corp). All statistical testing was two-tailed, with a p value <0.05 considered statistically significant. Inter-rater reliability between bronchoscopic visual estimation and HRCT findings for CAO was assessed using Cohen’s Kappa coefficient. Results Most participants were young to middle-aged adults (median age 37 years), with a predominance of females (64.8%). Nearly one-third of the cohort had a history of previous TB treatment. Among 62 TB incident cases, 19.4% had relapsed TB, and 80.6% were new patients (ie, never treated or had received anti-TB drugs for less than 1 month). Of these, 45.2% patients were on current TB treatment at the time of enrolment, with median treatment duration of 10 weeks (IQR: 4–20 weeks). The demographic characteristics of the participants are shown in detail in table 1. Table 1Clinical characteristics of participantsCharacteristicsAll(n=91)Patients with TB treatment success (n=29)Patients with new episode of TB (n=62)P valueAge (median, IQR) (years)37, 27–5437, 27–5037, 27–540.448c <45 years old58/91 (63.7%)21/29 (72.4%)37/62 (59.7%)0.26a 45–59 years old18/91 (19.8%)3/29 (10.3%)15/62 (24.2%)0.15a ≥60 years old15/91 (16.5%)5/29 (17.2%)10/62 (16.1%)0.92aGender Male (%)32/91 (35.2%)12/29 (41.4%)20/62 (32.3%)0.481a Female (%)59/91 (64.8%)17/29 (58.6%)42/62 (67.7%)0.54aTime to hospitalisation (median, IQR) (weeks)4, 3–144, 2–88, 4–160.039aCough20/91 (22%)7/29 (24.1%)13/62 (21%)0.789aDyspnoea54/91 (59.3%)20/29 (69%)34/62 (54.8%)0.255aWheezing23/91 (25.3%)8/29 (27.6%)15/62 (24.2%)0.798aChest tightness35/91 (38.5%)14/29 (48.3%)21/62 (33.9%)0.248aHaemoptysis7/91 (7.7%)3/29 (10.3)4/62 (6.5%)0.675bBronchoscopic forms of EBTB Actively caseating21/91 (23.1%)2/29 (6.9%)19/62 (30.6%)0.015a Granular11/91 (12.1%)1/29 (3.4%)10/62 (16.1%)0.164a Oedematous-hyperaemic28/91 (30.8%)1/29 (3.4%)27/62 (43.5%)<0.001a Fibrostenotic70/91 (76.9%)28/29 (96.6%)42/62 (67.7%)0.003a Tumorous0 Ulcerative0 Non-specific bronchitic0Central airway obstruction in bronchoscopy65/91 (71.4%)25/29 (86.2%)40/62 (64.5%)0.046aCT images Lower lung volume14/91 (15.4%)6/29 (20.7%)8/62 (12.9%)0.362a Atelectasis30/91 (33.0%)14/29 (48.3%)16/62 (25.8%)0.054a Airway stenosis44/91 (48.4%)18/29 (62.1%)26/62 (41.9%)0.114a Consolidation29/91 (31.9%)6/29 (20.7%)23/62 (37.1%)0.150aP values were calculated using the χ2 testᵃ or Fisher’s exact testᵇ or Mann-Whitney U testc as appropriate.EBTB, endobronchial tuberculosis; TB, tuberculosis. Clinical characteristics of participants P values were calculated using the χ2 testᵃ or Fisher’s exact testᵇ or Mann-Whitney U testc as appropriate. EBTB, endobronchial tuberculosis; TB, tuberculosis. At the time of diagnosis of EBTB, the most common self-reported symptoms of airway obstruction were chronic cough, dyspnoea, wheezing and chest tightness. There were no statistically significant differences in the frequency of these symptoms between patients who had TB treatment completion and those with incident TB. However, the time from symptom onset to EBTB diagnosis was longer in incident TB cases compared with those who had completed TB treatment. Every participant underwent chest HRCT scan, pulmonary function test and bronchoscopy at the time of enrolment. Inter-rater reliability between bronchoscopic visual estimation and HRCT findings for the evaluation of CAO demonstrated a Cohen’s Kappa coefficient of 0.42 (p<0.001). This level of agreement indicated that bronchoscopic assessment aligns reasonably well with HRCT in characterising the severity of CAO. The frequency of these examination findings is shown in tables13 and figure 1. EBTB lesions were more frequently located on the left side of the airway (table 2). Table 2Frequency and distribution of TB lesions in the bronchial treeDistribution of TB lesionsAll(n=91)Patients with TB treatment success (n=29)Patients with new episode of TB (n=62)P valueSites of lesions Lesion only on the right lung29/91 (30.8%)9/29 (31%)20/62 (32.3%)0.896a Lesion only on the left lung59/91 (64.8%)18/29 (62.1%)41/62 (66.1%)0.692a Both sites3/91 (3.3%)2/29 (6.9%)1/62 (1.6%)0.251bTrachea Upper-third3/91 (3.3%)1/29 (3.4%)2/62 (3.2%)1b Middle-third6/91 (6.6%)5/29 (17.2%)1/62 (1.6%)0.023b Lower third7/91 (7.7%)1/29 (3.4%)6/62 (9.7%)0.418b Entire trachea4/91 (4.4%)04/62 (6.5%)0.296bPrimary bronchi Right13/91 (14.3%)5/29 (17.2%)8/62 (12.9%)0.606b Left48/91 (52.7%)16/29 (55.2%)32/62 (51.6%)0.745a Both bronchi1/91 (1.1%)1/29 (3.4%)00.333bSecondary bronchi (lobar bronchi) Superior lobar bronchus27/91 (29.7%)7/29 (24.1%)20/62 (32.3%)0.408a Middle lobar bronchus5/91 (5.5%)05/62 (8.1%)0.161b Inferior lobar bronchus1/91 (1.1%)01/62 (1.6%)1b Intermediate bronchus12/91 (13.2%)4/29 (13.8%)8/62 (12.9%)0.916bTertiary bronchi (segmental bronchi) Segmental bronchus 12/91 (2.2%)02/62 (3.2%)0.541b Segmental bronchus 22/91 (2.2%)02/62 (3.2%)0.541b Segmental bronchus 33/91 (3.3%)03/62 (4.8%)0.278b Segmental bronchus 43/91 (3.3%)1/29 (3.4%)2/62 (3.2%)1b Segmental bronchus 66/91 (7.7%)1/29 (3.4%)5/62 (8.1%)0.651b Segmental bronchus 81/91 (1.1%)01/62 (1.6%)1bP values were calculated using the χ2 testᵃ or Fisher’s exact testᵇ as appropriate.TB, tuberculosis. Frequency and distribution of TB lesions in the bronchial tree P values were calculated using the χ2 testᵃ or Fisher’s exact testᵇ as appropriate. TB, tuberculosis. Table 3Pulmonary function in patients with endobronchial tuberculosisPulmonary function testAll patientsPatients with TB treatment successPatients with new episode of TBP valueFEV1 (<80%)67/91 (73.6%)26/29 (89.7%)41/62 (66.1%)0.02aFVC (<80%)56/91 (61.5%)17/29 (58.6%)39/62 (62.9%)0.69aFEF 25%–75% (<65%)63/91 (69.2%)24/29 (82.8%)39/62 (62.9%)0.06aFEV1/FVC (<70%)20/91 (21.9%)10/29 (34.5%)10/62 (16.1%)0.049aPEF (<80%)63/91 (69.2%)25/29 (86.2%)38/62 (61.3%)0.02aP values were calculated using the χ2 testᵃ. Values are percentage of measured to predicted values (% pred).FEF, Forced Expiratory Flow; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; PEF, peak expiratory flow. Pulmonary function in patients with endobronchial tuberculosis P values were calculated using the χ2 testᵃ. Values are percentage of measured to predicted values (% pred). FEF, Forced Expiratory Flow; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; PEF, peak expiratory flow. Lung function patterns in patients with endobronchial TB. TB, tuberculosis. Lung function impairment was also common, with 65.9% of patients showing a ventilatory defect on spirometry, predominantly a restrictive pattern (66.7%). The spirometrically defined restrictive ventilatory defect was found in 44% and was less common in patients with TB treatment completion (37.9%) compared with those with incident TB (46.8%). The restrictive ventilatory defects were also more frequent in patients with CT scan findings of airway stenosis (52.3%) than in those with other HRCT scan abnormalities (36.2%), although this difference was not statistically significant (p=0.122). Similarly, the restrictive ventilatory defects were also more frequent in patients with bronchoscopic findings of CAO compared with those without (49.2% vs 30.8%; p=0.109). In general, patients with TB treatment completion tended to have worse pulmonary function than incident TB cases (table 3). In our study, the overall prevalence of CAO among patients with EBTB was 71.4%, indicating a substantial burden of airway narrowing in this population. In seven bronchoscopic subtypes of EBTB by Chung’s classification,10 we found 4/7 subtypes in our study participants, namely the fibrostenotic type (76.9%), the actively caseating type (23.1%), the granular type (12.1%) and the oedematous-hyperaemic type (30.8%). Among these, the fibrostenotic subtype was the most common and was significantly more frequent in patients who had completed TB treatment (96.6%) compared with incident TB cases (67.7%) (p value=0.002). Airway stenosis on chest CT was observed in 63.1% with bronchoscopic evidence of CAO, and in 11.5% without bronchoscopic obstruction. The agreement between HRCT scan and bronchoscopic findings in classifying airway obstruction was moderate, with a Cohen’s Kappa of 0.42 (SE = 0.08; 95% CI 0.25 to 0.58; p <0.001). To identify factors associated with CAO, we compared the clinical, radiological and functional characteristics between patients with CAO (n=65) and those without CAO (n=26). As shown in table 4, several variables showed significant differences between the two groups, including dyspnoea (p=0.01), history of TB treatment success (p=0.033) and FEV1 <80% (p=0.007). Table 4Comparison of clinical, radiological and functional characteristics between patients with EBTB with and without CAOCharacteristicsAll(n=91)Patients with CAO (n=65)Patients without CAO (n=26)P valueAge <45 years old58/91 (63.7%)41/65 (63.1%)17/26 (65.4%)0.764a 45–59 years old18/91 (19.8%)14/65 (21.5%)4/26 (15.4%) ≥60 years old15/91 (16.5%)10/65 (15.4%)5/26 (19.2%)Gender Male32/91 (35.2%)20/65 (30.8%)12/26 (46.2%)0.165a Female59/91 (64.8%)45/65 (69.2%)14/26 (53.8%)Time to hospitalisation <8 weeks50/91 (54.9%)39/65 (60%)11/26 (42.3%)0.125aCough20/91 (22%)12/65 (18.5%)8/26 (30.8%)0.2aDyspnoea54/91 (59.3%)44/65 (67.7%)10/26 (38.5%)0.01aWheezing29/91 (31.9%)25/65 (38.5%)4/26 (15.4%)0.033aChest tightness35/91 (38.5%)27/65 (41.5%)8/26 (30.8%)0.34aHistory of TB treatment success29/91 (31.9%)25/65 (38.5%)4/26 (15.4%)0.033aCT images Lower lung volume14/91 (15.4%)11/65 (16.9%)3/26 (11.5%)0.75b Atelectasis30/91 (33%)23/65 (35.4%)7/26 (26.9%)0.438a Airway stenosis44/91 (48.4%)41/65 (63.1%)3/26 (11.5%)<0.001a Consolidation29/91 (31.9%)18/65 (27.7%)11/26 (42.3%)0.176aFEV1 <80%67/91 (73.6%)53/65 (81.5%)14/26 (53.8%)0.007aFVC (<80%)56/91 (61.5%)43/65 (66.2%)13/26 (50%)0.152aFEF 25%–75% (<65%)63/91 (69.2%)50/65 (76.9%)13/26 (50%)0.012aFEV1/FVC (<70%)20/91 (21.9%)14/65 (21.5%)6/26 (23.1%)0.873aPEF (<80%)63/91 (69.2%)49/65 (75.4%)14/26 (53.8%)0.044aP values were calculated using the χ2 testᵃ or Fisher’s exact testᵇ.CAO, central airway obstruction; EBTB, endobronchial tuberculosis; FEF, Forced Expiratory Flow; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; PEF, Peak Expiratory Flow; TB, tuberculosis. Comparison of clinical, radiological and functional characteristics between patients with EBTB with and without CAO P values were calculated using the χ2 testᵃ or Fisher’s exact testᵇ. CAO, central airway obstruction; EBTB, endobronchial tuberculosis; FEF, Forced Expiratory Flow; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; PEF, Peak Expiratory Flow; TB, tuberculosis. Variables associated with CAO on bronchoscopy in univariate analyses (with p value <0.05) included TB diagnosis, dyspnoea and FEV1 (table 5). After incorporating these variables into the multivariable logistic regression model, we found that shorter duration from symptom onset to hospitalisation (<8 weeks) (aOR=3.93, 95% CI 1.18 to 13.03), dyspnoea (aOR=3.85, 95% CI 1.27 to 11.68) and FEV1 <80% (aOR=5.45, 95% CI 1.28 to 23.33) were independent risk factors associated with an obstruction of at least 50% of the airway lumen on bronchoscopy. Table 5Regression analysis of risk factors related to central airway obstruction in bronchoscopyFactorCrude OR (95% CI)P valueAdjusted OR (95% CI)P valueSex Male11 Female1.93 (0.76 to 4.91)0.171.65 (0.46 to 4.86)0.40TB diagnosis New TB diagnosis11 History of TB treatment completion3.44 (1.06 to 11.15)0.042.63 (0.66 to 10.61)0.18Age <45 years old11 45–59 years old1.45 (0.42 to 5.00)0.551.52 (0.36 to 6.47)0.57 ≥60 years old0.83 (0.25 to 2.80)0.770.37 (0.06 to 2.21)0.28Time to hospitalisation ≥8 weeks11 <8 weeks2.05 (0.81 to 5.15)0.133.93 (1.18 to 13.03)0.03Dyspnoea No11 Yes3.35 (1.30 to 8.63)0.013.85 (1.27 to 11.68)0.02FEV1 FEV1 ≥80%11 FEV1 <80%3.79 (1.40 to 10.22)0.015.45 (1.28 to 23.33)0.02Atelectasis on Chest CT No11 Yes1.47 (0.54 to 4.06)0.440.60 (0.16 to 2.18)0.44FEV1, forced expiratory volume in one second; TB, tuberculosis. Regression analysis of risk factors related to central airway obstruction in bronchoscopy FEV1, forced expiratory volume in one second; TB, tuberculosis. Among 65 patients with CAO, 60% received interventional bronchoscopy, including mechanical dilation with rigid bronchoscopy (66.7%) and stent placement (33.3%). Most procedures were performed once (median 1, range 1–3 sessions), depending on the severity and extent of airway narrowing. Symptomatic improvement, particularly in dyspnoea, was reported in over two-thirds of cases, and no major procedural complications occurred. Airway stents were placed in 33.3%, primarily in those with severe (>75%) or recurrent stenosis after dilation. Silicone stents were used in all cases to maintain airway patency. All patients showed improved airway calibre on follow-up bronchoscopy and reported subjective relief of respiratory symptoms. Stent migration occurred in one case and was managed by repositioning. There were 26 patients who did not receive interventional bronchoscopy due to either ineligibility for the procedure (eg, significant coagulopathy, severe underlying comorbidities, and so on) or lack of consensus between the patient and healthcare providers regarding the intervention. Discussion Our study provided a comprehensive overview of patients with EBTB in a Vietnamese tertiary care setting, integrating clinical, imaging, bronchoscopic and pulmonary function data. Our findings highlighted the substantial burden of ventilatory defects and CAO in patients with EBTB, especially in those with a history of completed TB treatment. In a large study of 1221 cases of EBTB in China using the flexible bronchoscopy, the lesions were more common in the right lung (757 cases, 62.0%) and the left main bronchus (270, 22.1%) compared with the right main bronchus (247 cases, 20.2%). The most common endoscopic subtypes were necrotising ulcerative and granulation proliferative types.11 In a retrospective study of 55 patients with EBTB in Turkey, follow-up bronchoscopy performed at the end of treatment revealed that exudative, ulcerative and granular lesions were healed without complicated sequelae, while caseous and tumorous lesions of advanced disease were more likely to result in bronchostenosis associated with bronchiectasis.12 When comparing patients with completed TB treatment and those with newly diagnosed or active disease, distinct clinical and bronchoscopic patterns were observed. Individuals who had completed therapy more frequently exhibited fibrostenotic lesions, fixed airway narrowing and restrictive ventilatory defects, reflecting sequelae of healed inflammation and fibrosis. Importantly, microbiological cure does not equate to anatomical recovery. In our cohort, patients with completed TB therapy had the highest rates of fibrostenotic CAO. This suggests that while anti-TB drugs eradicate the pathogen, they cannot reverse the structural damage or exuberant fibrotic remodelling established during the active phase. Thus, severe obstruction in these patients represents permanent anatomical sequelae of EBTB rather than active disease. In contrast, newly diagnosed or currently treated patients tended to show edematous, granular or caseating lesions suggestive of active endobronchial inflammation. These findings support the dynamic nature of EBTB, which evolves from an active inflammatory phase to irreversible fibrotic scarring if not recognised and managed early. Furthermore, the integration of mucosal biopsy in our protocol was paramount for differential diagnosis, especially in a tertiary setting where complex pathologies are common. Visual bronchoscopic findings of EBTB can often be non-specific and may mimic other serious conditions, such as bronchogenic malignancy or endobronchial fungal infections. Histopathological verification ensures diagnostic accuracy and prevents the risk of misdiagnosing treatable non-tuberculous diseases as simple TB sequelae. A recent systematic review and meta-analysis of 75 960 patients (of whom 7447 with a history of pulmonary TB) showed that previous TB diagnosis had a negative impact on all spirometric parameters compared with healthy controls. Specifically, the pooled mean differences were FEV1 −0.41 L (95% CI −0.51 to −0.32, I2=90.4%), FVC −0.25 L (−0.33 to −0.17, I2=80.6%) and FEV1/FVC ratio −0.37 (−0.54 to −0.19, I2=92.0%).13 The impact of TB on lung function has been well documented, with several large-scale studies demonstrating long-term ventilatory impairment following pulmonary TB. Although these studies primarily focused on parenchymal disease rather than endobronchial involvement, the mechanisms of lung function loss—such as airway fibrosis, residual inflammation and destruction of the bronchial wall—are also relevant to patients with EBTB. In our study, the high prevalence of restrictive ventilatory patterns among patients with EBTB may reflect similar structural changes secondary to airway scarring and fibrosis. Understanding these mechanisms is essential for long-term respiratory care and rehabilitation planning in this population. In our multivariable analysis, three factors were independently associated with CAO: shorter duration from symptom onset to hospitalisation, presence of dyspnoea and reduced FEV₁ (<80% predicted). A shorter time to hospitalisation (<8 weeks) was likely a surrogate marker of more rapid symptom progression, suggesting that patients with aggressive endobronchial inflammation or oedema may present earlier but already have significant luminal narrowing. Dyspnoea, one of the strongest clinical indicators, directly reflects impaired airflow due to central airway narrowing and can serve as a simple but valuable screening symptom for EBTB in clinical practice. The association between decreased FEV₁ and CAO is expected, as FEV1 reduction indicates mechanical limitation of expiratory flow. This finding aligns with previous research showing that airflow limitation on spirometry correlates with the extent of tracheobronchial stenosis assessed bronchoscopically. Together, these parameters may help clinicians identify patients at high risk for significant airway obstruction who would benefit from early bronchoscopy and potential intervention. The strength of our study is its integration of different diagnostic modalities, including histopathology, microbiology, bronchoscopy and chest CT to describe the characteristics of EBTB and the preliminary results of interventional bronchoscopy in the management of the CAOs in this population. However, our study had several limitations. First, as a single-centre study conducted at the National Lung Hospital—a national reference institution for TB—our findings may be subject to selection bias towards more complex or treatment-resistant cases. This may limit the generalisability of our results to patients with EBTB in community or primary hospital settings. Second, this study did not include bronchodilator reversibility testing in spirometry, which may have helped differentiate fixed from reversible airway obstruction. However, as our primary objective was to characterise ventilatory patterns in EBTB rather than to diagnose obstructive airway diseases such as asthma or chronic obstructive pulmonary disease (COPD), baseline spirometric measurements were considered sufficient for analysis. Furthermore, we did not evaluate the relationship between the duration of anti-TB treatment and pulmonary function abnormalities. As the interval from treatment completion to spirometry varied among post-treatment patients, residual airway obstruction might reflect both previous disease severity and the degree of fibrotic healing. Third, this study only described a short-term treatment outcome without information for the long-term follow-up duration after stent implantation, stent indwelling time and monitoring protocols for late complications. Additionally, the observational design precludes causal inference, and the moderate agreement between CT and bronchoscopy suggests that reliance on imaging alone may miss significant airway pathology. Future studies should aim to validate these findings in multicentre cohorts and include patients in all levels of care to better understand the full spectrum of EBTB. Longitudinal studies are needed to assess the progression of airway obstruction, the optimal stent indwelling duration and the risk of late complications, as well as the impact of interventional bronchoscopy on long-term pulmonary function. Conclusion CAO was found in a high proportion (71.4%) of patients with EBTB, particularly among those with a history of prior TB treatment. Early clinical suspicion is warranted in patients presenting with dyspnoea, short symptom duration and reduced FEV1 to prevent irreversible airway damage.