Subglottic stenosis and endobronchial disease in granulomatosis with polyangiitis

Abstract

Objectives

To describe tracheobronchial disease in patients with granulomatosis with polyangiitis (GPA) and evaluate the utility of dynamic expiratory CT to detect large-airway disease.

Methods

Demographic and clinical features associated with the presence of subglottic stenosis (SGS) or endobronchial involvement were assessed in a multicentre, observational cohort of patients with GPA. A subset of patients with GPA from a single-centre cohort underwent dynamic chest CT to evaluate the airways.

Results

Among 962 patients with GPA, SGS and endobronchial disease were identified in 95 (10%) and 59 (6%) patients, respectively. Patients with SGS were more likely to be female (72% vs 53%, P < 0.01), younger at time of diagnosis (36 vs 49 years, P < 0.01), and have saddle-nose deformities (28% vs 10%, P < 0.01), but were less likely to have renal involvement (39% vs 62%, P < 0.01). Patients with endobronchial disease were more likely to be PR3-ANCA positive (85% vs 66%, P < 0.01), with more ENT involvement (97% vs 77%, P < 0.01) and less renal involvement (42% vs 62%, P < 0.01). Disease activity in patients with large-airway disease was commonly isolated to the subglottis/upper airway (57%) or bronchi (32%). Seven of 23 patients screened by dynamic chest CT had large-airway pathology, including four patients with chronic, unexplained cough, discovered to have tracheobronchomalacia.

Conclusion

SGS and endobronchial disease occur in 10% and 6% of patients with GPA, respectively, and may occur without disease activity in other organs. Dynamic expiratory chest CT is a potential non-invasive screening test for large-airway involvement in GPA.

Rheumatology key messages

• Subglottic stenosis affects 10% and endobronchial disease affects 6% of patients with granulomatosis with polyangiitis.

• Inflammation of the large airways can occur in absence of other features of disease activity.

• Dynamic expiratory chest CT is a non-invasive test for tracheobronchial disease in granulomatosis with polyangiitis.

Introduction

ANCA-associated vasculitis is a group of disorders characterized by inflammation of small and medium-sized arteries. This subgroup of vasculitis includes granulomatosis with polyangiitis (GPA) and microscopic polyangiitis. The organs most frequently affected in GPA include the upper airway, the lower airway, and the kidneys. Damage to the large airways in the form of tracheobronchial disease, including subglottic stenosis (SGS) and lower tracheal and bronchial stenosis, is a devastating complication of GPA that can lead to upper airway obstruction and potentially life-threatening disease.

SGS is defined as narrowing of the airway immediately below the vocal cords and is the most common manifestation of tracheobronchial GPA, with an estimated frequency of 16–23% [1–3]. In patients with SGS, clinical symptoms of stridor, hoarseness, and dyspnoea require urgent evaluation and may be severe enough to necessitate tracheostomy. Endobronchial inflammation and stenosis occurs less frequently than subglottic disease, but may present with similar clinical symptoms. Images depicting tracheobronchial disease in GPA can be seen in Supplementary Fig. S1, available at Rheumatology online.

There have been few reports from cohorts of GPA focused on tracheobronchial disease. Identification of patient subsets at particular risk for airway disease, and further development of non-invasive screening methods to detect tracheobronchial disease, are unmet needs in GPA. Dynamic expiratory CT has been proposed as a modality for evaluating tracheobronchial disease and could be a possible screening method for detecting airway disease in GPA [4, 5]. A dynamic expiratory CT evaluates the tracheobronchial tree during the full expiratory phase of respiration, and can assess for both upper and lower tracheal stenosis and collapse. This study was conducted in two different phases with the objectives of: (i) describing tracheobronchial disease in a large observational cohort of patients with GPA, and (ii) evaluating the utility of dynamic expiratory CT for detecting tracheobronchial disease in an independent cohort of patients with GPA.

Methods

Study population

Patients with GPA enrolled in the Vasculitis Clinical Research Consortium (VCRC) Longitudinal Protocol for Granulomatosis with Polyangiitis and Microscopic Polyangiitis were included in this study. The VCRC is an international, multicentre research infrastructure, dedicated to conducting clinical research in different forms of vasculitis, from seven referral centres in the USA and two in Canada. Patients with GPA are eligible for inclusion in the VCRC longitudinal cohort if they fulfil the modified ACR criteria for the classification of GPA [6]. All patients with GPA who had SGS and/or endobronchial disease were included in this analysis.

To evaluate the utility of dynamic expiratory CT, patients with GPA were recruited into an independent, prospective, observational cohort at the National Institutes of Health (NIH) in Bethesda, MD, USA. Patients could be enrolled at the time of diagnosis or later during the disease course. Patients also fulfilled the modified ACR criteria for the classification of GPA [6]. All study participants provided written informed consent, and local ethics approval was obtained at each participating study site.

Data collection

Using standardized data collection forms, baseline demographic information was recorded, including age at diagnosis, age at symptom onset, sex, and race. ANCA status/type was recorded as immunofluorescence pattern [cytoplasmic (C)-ANCA or perinuclear (P)-ANCA] and ELISA testing (anti-myeloperoxidase or anti-proteinase-3) or ANCA-negative. Historical presence of constitutional symptoms, musculoskeletal symptoms, cutaneous symptoms, eye symptoms, cardiovascular symptoms, renal involvement, and nervous system involvement was queried. The presence ever of ENT symptoms was recorded including: (i) nasal septal perforation, (ii) saddle-nose deformity, (iii) SGS and (iv) any other upper airway involvement. The presence ever of pulmonary involvement was recorded including: (i) endobronchial disease and (ii) any other pulmonary involvement (e.g. lung nodules, alveolar haemorrhage).

Disease activity in the large airways was defined based upon the clinical judgement of the treating physician and recorded as part of BVAS [7]. Pattern of organ involvement at time of active subglottic inflammation and active endobronchial disease was also recorded, including constitutional symptoms, pulmonary involvement, mucous membrane involvement, cutaneous involvement, cardiovascular disease, gastrointestinal symptoms, renal involvement and nervous system involvement.

Dynamic expiratory CT

Patients with GPA in the NIH cohort were randomly chosen to undergo a dynamic expiratory CT. All images were assessed by a thoracic radiologist (A.S.) who was blinded to clinical information. Each patient was evaluated on a 320-row multi-detector CT unit (AquilionONE, Canon Medical Systems, Otawara, Japan). First, a routine thoracic CT was obtained to assess the tracheal anatomy, followed by the dynamic expiratory portion of the exam. The field of view extended from the trachea to the level of the bronchus intermedius. All patients received respiratory coaching prior to the exam. Patients were instructed to take in a deep breath, forcefully exhale, and then repeat this process until told to stop. Volumetric CT data acquisition was acquired over several respiratory cycles. Images were evaluated on a PACS workstation (CareStream, Version 12.1.6.1005, Carestream Health, NY) and processed on a 3 D software programme (Vitrea, Version 7.10.1.20, Vital Images Inc, Minnetonka, MN). The area of the lumen was manually planimetered at full inspiration and during exhalation. The amount of collapse was determined by the following equation: (area during full inspiration – area during exhalation)/area during full inspiration. More than 50% decrease of area during exhalation was considered abnormal.

Statistical analysis

Baseline demographic features, ANCA type, and clinical features for patients with GPA with and without SGS were compared. These same parameters were compared among patients with GPA with and without endobronchial disease. Fisher’s exact test was used for categorical variables expressed as a number (or percentage). For continuous variables, means with standard deviations were compared using a two-tailed t-test. Multivariable logistic regression analysis was performed to evaluate the association of variables that were significant in univariable analyses in association with SGS and endobronchial disease. A value of P < 0.05 was considered significant. All analyses were performed using JMP 13.0 (2016 SAS Institute, Cary, NC, USA).

Results

Study population

There were a total of 962 patients with GPA evaluated from nine VCRC referral centres. The mean age of the cohort was 52 (16) years, and 55% of the cohort was female. The majority of patients were Caucasian (92%), followed by Asian (4%), African American (2%), and unknown race (2%). Sixty-seven percent of patients were C-ANCA/PR3 positive (644 patients), 25% patients were P-ANCA/MPO positive (239 patients), 5% were ANCA negative (50 patients), and 3% had unknown ANCA status (29 patients).

SGS

SGS was identified in 95 (10%) patients (Table 1). Compared with patients without SGS, patients with SGS were more likely to be female (72% vs 53%, P < 0.01), younger at time of diagnosis (35.7 vs 48.7 years, P < 0.01) and younger at symptom onset (33.8 vs 47.6 years, P < 0.01). They were also more likely to have nasal septal perforations (29% vs 11%, P < 0.01), saddle-nose deformities (28% vs 10%, P < 0.01) and endobronchial involvement (27% vs 4%, P < 0.01). Patients with SGS were less likely to report constitutional symptoms, cardiovascular involvement, renal disease, or nervous system involvement. Most patients with SGS were C-ANCA/PR3 positive (67%), but 23% were P-ANCA/MPO positive. Of the seven patients with SGS who were ANCA negative, 5 patients (71%) had disease limited only to the upper and lower airways. There were no differences in the prevalence of each ANCA subtype in patients with SGS compared with those without SGS. In multivariable logistic regression models, female sex (ß = 0.25, s.e. 0.12, P = 0.04), younger age at diagnosis (ß =−0.04, s.e. 0.01, P < 0.01) and no prior history of kidney involvement (ß = −0.39, s.e. 0.12, P < 0.01) remained significant independent predictors of SGS in patients with GPA. A total of 63 patients underwent interventions for SGS (e.g. dilation, surgery), with 9 patients requiring tracheostomy.

Table 1.

Clinical features of patients with and without SGS

	SGS	SGS	P-value
	Yes	No
	n = 95	n = 867
Sex (female, %)	68 (72%)	457 (53%)	<0.01
Age, mean (s.d.), years
At diagnosis	35.7 (15)	48.7 (17)	<0.01
At symptom onset	33.8 (14)	47.6 (17)	<0.01
Race (Caucasian, %)	90 (95%)	788 (92%)	0.31
ANCA type (n, %)
C-ANCA/ PR3	63 (67%)	581 (67%)	0.91
P-ANCA/ MPO	22 (23%)	217 (25%)	0.80
ANCA-negative	7 (7%)	43 (5%)	0.33
Unknown status	3 (3%)	26 (3%)	N/A
Clinical features
Constitutional	66 (70%)	690 (81%)	0.02
Musculoskeletal	53 (56%)	544 (63%)	0.17
Cutaneous	23 (24%)	251 (29%)	0.32
Eye	29 (31%)	232 (27%)	0.49
Ear, nose, and throat	95 (100%)	641 (75%)	<0.01
Nasal perforation	28 (29%)	67 (11%)	<0.01
Saddle-nose deformity	27 (28%)	63 (10%)	<0.01
Cardiovascular	0 (0%)	35 (4%)	0.04
Gastrointestinal	2 (2%)	27 (3%)	0.57
Pulmonary
Any manifestation	69 (72%)	574 (67%)	0.29
Endobronchial	25 (27%)	34 (4%)	<0.01
Renal	37 (39%)	540 (62%)	<0.01
Nervous system	12 (13%)	202 (24%)	0.02

SGS: subglottic stenosis; C-ANCA: cytoplasmic-ANCA; P-ANCA: perinuclear-ANCA.

There were 87 patients with SGS clinically evaluated during the time of active subglottic inflammation. Fifty (57%) patients had subglottic inflammation with isolated upper airway disease. Thirty-seven of these patients had additional upper-airway manifestations (e.g. bloody nasal discharge, sinus involvement) along with subglottic inflammation, and 13 patients presented with only subglottic inflammation and no other disease features. Overall, sinus involvement was seen in 63 patients (73%), along with subglottic inflammation, with 26 patients having additional organ involvement. Of those patients with active subglottic inflammation and additional disease manifestations outside of the upper airway, pulmonary disease was most frequently noted, involving 24 patients (28%), including 11 (13%) with endobronchial involvement. Other organ involvement is detailed in Fig. 1. Constitutional symptoms were noted in 13 (15%) patients, whereas mucous membrane involvement, renal disease, and nervous system involvement were infrequently detected, and no patients had associated cutaneous, cardiovascular, or GI disease flares.

Fig. 1.

Organ involvement in patients with active subglottic disease

There were 87 patients with known GPA who were evaluated during time of active subglottic inflammation. Patients were divided on the basis of isolated upper airway disease or additional disease outside of the upper airway. Additional features of each group are displayed. GPA: granulomatosis with polyangiitis.

Endobronchial disease

Endobronchial disease was identified in 59 (6%) patients. Data on presence or absence of endobronchial disease was missing on 17 (1.8%) out of 962 patients, and those patients were excluded from analysis. As shown in Table 2, patients with endobronchial disease were more likely to be younger at diagnosis (39.8 vs 47.7 years, P < 0.01) and at symptom onset (39.2 vs 46.5 years, P < 0.01), compared with the entire cohort. Compared with patients without endobronchial disease, they were also more likely to have ENT involvement (97% vs 77%, P < 0.01) and C-ANCA/PR3 positivity (85% vs 66%, P < 0.01), and less likely to have renal involvement (42% vs 62%, P < 0.01) and P-ANCA/MPO positivity (13% vs 25%, P = 0.04) or be ANCA negative (0% vs 6%, P = 0.04). There were no significant differences in female prevalence between patients with and without endobronchial involvement (56% vs 55%, P = 0.86). There were 25 patients (2.6%) in the cohort who had both SGS and endobronchial involvement. Concomitant SGS and endobronchial involvement was not associated with sex (60% vs 55% female, P = 0.60).

Table 2.

Clinical features of patients with and without endobronchial disease

	Endobronchial – Yes	Endobronchial – No	P-value
	n = 59	n = 886
Sex (female, %)	33 (56%)	485 (55%)	0.86
Age, mean (s.d.), years
At diagnosis	39.8 (15)	47.7 (17)	<0.01
At symptom onset	39.2 (15)	46.5 (17)	<0.01
Race (Caucasian,%)	55 (93%)	813 (92%)	0.69
ANCA type (n, %)
C-ANCA/anti-PR3	50 (85%)	588 (66%)	<0.01
P-ANCA/anti-MPO	8 (13%)	225 (25%)	0.04
ANCA-negative	0 (0%)	54 (6%)	0.04
Unknown status	1 (2%)	19 (2%)	N/A
Clinical features
Constitutional	52 (88%)	694 (79%)	0.09
Musculoskeletal	36 (61%)	556 (63%)	0.75
Cutaneous	16 (27%)	257 (29%)	0.73
Eye	14 (24%)	247 (28%)	0.47
Ear, nose, and throat	57 (97%)	674 (77%)	<0.01
SGS	25 (42%)	66 (8%)	<0.01
Nasal perforation	10 (19%)	84 (13%)	0.24
Saddle-nose deformity	12 (21%)	78 (12%)	0.04
Cardiovascular	1 (2%)	34 (4%)	0.39
Gastrointestinal	1 (2%)	28 (3%)	0.52
Pulmonary(additional)	44 (75%)	577 (65%)	0.15
Renal	24 (42%)	546 (62%)	<0.01
Nervous system	12 (21%)	197 (23%)	0.76

SGS: subglottic stenosis; C-ANCA: cytoplasmic-ANCA; P-ANCA: perinuclear-ANCA; pulmonary (additional): pulmonary manifestation other than endobronchial involvement (e.g. lung nodules, alveolar haemorrhage).

There were 53 patients with endobronchial disease who were evaluated during a time of active endobronchial disease. Seventeen (32%) patients had isolated endobronchial disease without any additional pulmonary manifestations (i.e. pulmonary nodules, alveolar haemorrhage). Thirty-six patients (68%) had additional disease features as detailed in Fig. 2. ENT disease was most frequently seen in patients who had additional disease at the time of active endobronchial disease, with 29 patients having ENT manifestations, including 11 with subglottic inflammation.

Fig. 2.

Organ involvement in patients with active endobronchial disease

There were 53 patients evaluated during time of active endobronchial disease. Patients were divided on the basis of isolated endobronchial disease and those who had additional disease features.

Dynamic expiratory CT

A total of 23 patients with GPA underwent screening with a dynamic expiratory chest CT. Four patients had known SGS confirmed by bronchoscopy prior to imaging, and no patient had known tracheobronchomalacia (TBM). There were 7 patients with large-airway pathology on imaging, as shown in Table 3. Dynamic CT confirmed SGS in all 4 patients with previously identified disease. TBM was discovered in 4 patients, including one male patient who was previously thought to have isolated SGS. A total of 8 patients who underwent dynamic CT chest had chronic airway complaints, and 4 of these patients had TBM, highlighting the possibility of patients with GPA and chronic airway complaints having underappreciated underlying airway disease. Additional characteristics of the 23 patients with GPA who underwent dynamic CT are provided in Supplementary Table S1, available at Rheumatology online. Representative images of a dynamic CT in a patient with TBM are shown in Fig. 3.

Table 3.

Patients with abnormal findings on dynamic expiratory CT

Patient	Sex	Age at disease onset (years)	Known SGS	Known TBM	Respiratory symptoms	SGS on CT	TBM on CT
1	M	43	Yes	No	Hoarseness	Yes	No
2	F	56	Yes	No	None	Yes	No
3	F	15	Yes	No	Hoarseness	Yes	No
4	M	49	Yes	No	Chronic cough	Yes	Yes
5	M	61	No	No	Chronic cough	No	Yes
6	F	54	No	No	Chronic cough	No	Yes
7	F	58	No	No	Chronic cough	No	Yes

SGS: subglottic stenosis; TBM: tracheobronchomalacia.

Fig. 3.

Dynamic expiratory CT chest in a patient with tracheobronchomalacia

A 67-year-old male with GPA underwent dynamic expiratory CT. He had a 7-year history of GPA diagnosed on the basis of presence of PR3-ANCA, sinusitis, arthralgias, constitutional symptoms, and diffuse alveolar haemorrhage. He was in long-term clinical remission on repeated courses of rituximab, with persistent, unexplained dry cough. Dynamic CT chest showed normal appearance of trachea on inspiration (A) and tracheal collapse of anterior and lateral tracheal walls (62%) on expiration (B) consistent with tracheobronchomalacia. GPA: granulomatosis with polyangiitis.

Discussion

SGS and endobronchial disease can be devastating complications of GPA. SGS has previously been reported in older cohorts, with an estimated frequency of 16–23% in patients with GPA [1–3], while endobronchial disease has been reported less frequently than SGS, but without a well-defined incidence in GPA [8–10]. In this cohort of 962 patients with GPA, the frequency of SGS was 10%, while endobronchial disease was noted in 59 patients (6%). Concomitant SGS and endobronchial disease was seen in 25 patients (2.6%).

The demographic features associated with SGS differ from the typical demographic characteristics of GPA. SGS is more commonly seen in young, female patients with GPA. Typically, GPA is equally common in both sexes and occurs most frequently in middle age [11, 12], as exemplified in the overall population in this study, in which both sexes were nearly equally affected (55% female and 45% male), and the average age at diagnosis of GPA was 52 years. Seventy-two per cent of patients with SGS in this study were female (68 female patients and 27 male patients), and the average age at diagnosis and symptom onset was 36 years and 34 years, respectively, which is consistent with prior reports [1, 2, 9, 13, 14]. Patients with idiopathic SGS are also more frequently female, although the average age tends to be slightly older, between 30 and 50 years of age [15, 16]. There has been suggestion of a hormonal cause for SGS, as it is mainly seen in women, but prior studies have failed to demonstrate the presence of oestrogen receptors in tissue biopsies of patients with SGS [17].

Unlike SGS, the demographic features associated with endobronchial disease more closely mirror the typical demographic pattern seen in patients with GPA. There was no imbalance of sex among patients with endobronchial disease, with a total of 33 female patients and 26 male patients having endobronchial involvement. The average age at diagnosis was 40 years, which was still younger in comparison to the overall cohort, but older than the average age at diagnosis for patients with SGS. The few previous case series evaluating endobronchial disease in GPA are limited by small sample sizes, making the current study valuable for characterizing the clinical features associated with this relatively rare disease manifestation [9, 18].

There were unique clinical associations found in patients with SGS and endobronchial disease. Both groups were less likely to have renal disease, which is consistent with findings from prior studies [2, 9, 14]. This is in contrast to what is typically seen in GPA, where >75% of patients with GPA have been reported to have renal involvement [11]. Patients with SGS were more likely to have associated endobronchial disease compared with the overall cohort, and more likely to have destructive sinonasal disease, including nasal septal perforations and saddle-nose deformities. Prior studies have also noted that patients with SGS more frequently have other ENT manifestations [2, 9].

There has been conflicting data about ANCA status in patients with SGS, and little reported data about ANCA status in endobronchial disease. Prior studies have shown variability in ANCA status, with some studies concluding that patients with SGS are more likely to be C-ANCA/PR3 positive vs P-ANCA/MPO positive [2, 19], whereas other studies have not noted this pattern and have found many patients with SGS to also be P-ANCA/MPO positive or ANCA negative [9, 20, 21]. In the current study there were no significant differences among ANCA type (anti-PR3, anti-MPO, or ANCA negative) between patients with and without SGS, while patients with endobronchial disease were more likely to be C-ANCA/PR3 positive than those without endobronchial disease. Importantly, approximately one-third of the patients with SGS were not C-ANCA/anti-PR3 positive. Many patients were P-ANCA/anti-MPO positive (23%), and some were ANCA negative (7%).

The present study should inform clinicians about which patients with GPA are at highest risk for developing SGS or endobronchial disease. One should consider SGS in young, female patients with GPA, particularly those who have severe, destructive, sinonasal disease but no associated renal involvement. Endobronchial disease is likely under-recognized in GPA, and should also be considered in younger patients, both male and female, especially those who are PR3-ANCA positive, and who have a history of ENT involvement or SGS, but no history of renal disease.

This study highlights the fact that recognition of active subglottic disease can be challenging in patients with GPA because it may occur in isolation without evidence of disease elsewhere. During a time of active subglottic inflammation, >50% of patients in this cohort had disease flares limited only to the upper airway. Isolated airway disease has also been reported in other cohorts [1, 2, 9, 10, 13, 22–24]. Of those patients who had additional disease outside of the upper airway, pulmonary manifestations and constitutional symptoms were most frequently associated with subglottic inflammation, whereas nervous system and renal involvement were infrequently associated with subglottic inflammation. Therefore, patients with GPA who develop symptoms raising concern regarding SGS (i.e. stridor, hoarseness, dyspnea) should undergo urgent evaluation for subglottic inflammation, even in the absence of other symptoms raising concern regarding disease flare. In contrast, during time of active endobronchial disease, patients were more likely to have additional disease features. Over two-thirds of patients with active endobronchial disease had evidence of active disease elsewhere, most commonly ENT manifestations or additional pulmonary features.

Dynamic expiratory chest CT is a non-invasive screening test to evaluate tracheobronchial disease in GPA. It has previously been proposed as a modality to evaluate for TBM [4, 5], as it has been shown to better evaluate degree of collapse in patients with TBM compared with end-expiratory CT [4]. In this study, dynamic CT confirmed the presence of SGS in 4 patients with known SGS, and identified 4 patients with TBM, all of whom had chronic unexplained airway complaints, but no prior diagnosis of TBM. Clinicians can consider obtaining a dynamic expiratory CT in patients who have chronic airway complaints when there is a concern about possible TBM. Previously, magnetic resonance imaging has been assessed as a potential imaging tool for detecting SGS in patients with GPA [25, 26], and it correlates with laryngoscopy and pulmonary function tests. The advantage of dynamic expiratory CT over MRI is the ability to evaluate dynamic pathological changes in the airways during respiration in addition to detecting fixed stenoses. Patients who have known SGS but persistent airway complaints, should also be evaluated for more extensive tracheobronchial disease, particularly male patients, because they are less likely to have isolated SGS. TBM is likely an under recognized complication of GPA, and early identification may prompt earlier referral to an interventional pulmonologist for diagnostic confirmation of TBM and clinical management of airway-related issues. Identification of TBM is also important when assessing overall disease damage in patients with GPA and may play a role in decision-making about future maintenance therapy.

Several limitations of this study should be noted. The evaluation of patients with SGS and endobronchial disease in the VCRC cohort was done by retrospective analysis and relied on identification and reporting of patients having disease by the physician-investigator at each centre. Details of the procedures performed to detect SGS or endobronchial disease were not systematically recorded; however, all patients were evaluated at major academic centres with access to otolaryngology and pulmonary specialists. Additionally, patients did not undergo routine endoscopic evaluation of the upper airway per a standardized protocol, so patients with mild SGS may have been missed. Differentiating active subglottic and endobronchial inflammation from symptoms related to damage in the large airways can be clinically challenging. In this study, active airway disease was defined based upon the clinical judgement of the treating physician-investigator. There was limited information about clinical outcomes related to specific medical or surgical interventions in this cohort. Patients with TBM identified by dynamic expiratory chest CT did not undergo routine follow-up bronchoscopy to confirm TBM; however, prior studies have shown that dynamic chest CT findings correlate well with bronchoscopy [27]. In addition, there are discrepant reports about the threshold to use to define pathological airway collapse on dynamic CT. A reduction in cross-sectional area of the airway of ⩾50% at mid-expiration is typically considered diagnostic for TBM and was used in this study [4], although some studies have proposed more conservative threshold values of ⩾70% collapse [28], and others have used relatively low thresholds of >18–28% collapse [29].

In conclusion, SGS and endobronchial disease can occur in up to 10% of patients with GPA. SGS is more commonly seen in female patients with GPA, whereas bronchial involvement is not associated with sex. A high index of suspicion in relation to evaluating airway disease in GPA is warranted, especially in younger patients and those with sinonasal disease. Dynamic expiratory chest CT is a potential screening test to evaluate for SGS and TBM, and should be considered in patients with GPA who have unexplained airway complaints or clinical features suggestive of large-airway disease.