Abstract
Objective To report on a large series of patients with antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) and bronchiectasis, with a specific focus on the timeline of occurrence of both features.
Methods Retrospective nationwide multicenter study of patients diagnosed with both AAV and bronchiectasis.
Results Sixty-one patients were included, among whom 27 (44.25%) had microscopic polyangiitis (MPA), 27 (44.25%) had granulomatosis with polyangiitis (GPA), and 7 (11.5%) had eosinophilic GPA. Thirty-nine (64%) had myeloperoxidase (MPO)-ANCA and 13 (21%) had proteinase 3–ANCA. The diagnosis of bronchiectasis either preceded (n = 25; median time between both diagnoses: 16 yrs, IQR 4–54 yrs), was concomitant to (n = 12), or followed (n = 24; median time between both diagnoses: 1, IQR 0–6 yrs) that of AAV. Patients in whom bronchiectasis precedes the onset of AAV (B-AAV group) have more frequent mononeuritis multiplex, MPA, MPO-ANCA, and a 5-fold increase of death. The occurrence of an AAV relapse tended to be protective against bronchiectasis worsening (HR 0.6, 95% CI 0.4–0.99, P = 0.049), while a diagnosis of bronchiectasis before AAV (HR 5.8, 95% CI 1.2–28.7, P = 0.03) or MPA (HR 18.1, 95% CI 2.2–146.3, P = 0.01) were associated with shorter survival during AAV follow-up.
Conclusion The association of bronchiectasis with AAV is likely not accidental and is mostly associated with MPO-ANCA. Patients in whom bronchiectasis precedes the onset of AAV tend to have distinct clinical and biological features and could carry a worse prognosis.
Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a group of inflammatory diseases characterized by small- to medium-size vessel necrotizing vasculitis involving various organs and the presence of autoantibodies targeting neutrophil’s cytoplasmic proteolytic enzymes1.
AAV-associated lung involvement encompasses a wide spectrum of manifestations including pulmonary hemorrhage, pulmonary nodules in granulomatosis with polyangiitis (GPA), and eosinophilic-related bronchial hyperresponsiveness in eosinophilic GPA (EGPA)2. Several case series reported the association of interstitial lung disease (ILD; mostly usual interstitial pneumonia) and AAV, occurring predominantly in patients with anti-myeloperoxydase (MPO) autoantibodies3,4,5. Likewise, the association of bronchiectasis and AAV has also been reported in case reports and small case series6,7,8,9,10, and it has been shown that patients with bronchiectasis are at increased risk of GPA onset11,12,13.
We previously reported the case of a patient with lung-limited microscopic polyangiitis (MPA) in whom bronchiectasis (1) positively correlated with both ANCA titers and AAV activity, and (2) dramatically improved with immunosuppressive therapy10. This case suggested that bronchiectasis might be another specific feature of AAV. Moreover, by analogy with rheumatoid arthritis (RA)-associated bronchiectasis (where early-onset bronchiectasis carries a worse prognosis), the time-line of occurrence between both conditions might significantly affect disease prognosis14,15.
The aim of our present study was to report on the clinical and radiological characteristics of a large series of patients with both AAV and bronchiectasis, with a specific focus on the time frame between both disease features.
MATERIALS AND METHODS
Patients
We conducted a retrospective study of adult (> 18 yrs) patients with AAV enrolled in the French Vasculitis Study Group (FVSG) registry prior to May 2015 who were diagnosed with bronchiectasis [according to a chest computed tomography (CT) scan, as defined by Barker, et al]16. Specifically, the database was queried manually using the following keywords: “bronchi-ectasis,” “bronchitis,” “bronchial dilatation,” and “bronchial wall thickening.” Next, a nationwide retrospective survey was performed using complementary multidisciplinary networks (Groupe d’Etudes et de Recherche sur les Maladies Orphelines Pulmonaires, Société Nationale Française de Médecine Interne) to identify additional patients. All patients fulfilled either the 1990 American College of Rheumatology classification criteria17,18 and/or the algorithm of the European Medicines Agency for AAV19. Patients were excluded from the study if they had fibrosis-related traction-bronchiectasis and/or secondary bronchiectasis due to other well-defined conditions (e.g., cystic fibrosis, alpha-1 antitrypsin deficiency)16. Patients previously reported in the study by Néel, et al were not included in the present work9.
Our study complied with Good Clinical Practice guidelines and the Declaration of Helsinki. Owing to the retrospective design of this study, and in compliance with the French legislation in force at the time of the study (Loi Huriet-Sérusclat 88-1138, December 20, 1988, and its subsequent amendments), an observational study that does not change routine management of patients does not need to be declared or submitted for the review of a research ethics board.
Baseline measurements
All medical records were reviewed regarding demographic, clinical, biological, radiological, pulmonary function, and histological findings. Studied biological variables included proteinuria (> 0.4 g/24 h), hematuria (> 10 red blood cells per high-power field), serum creatinine level, C-reactive protein level, and the results of ANCA testing by immunofluorescence and/or ELISA at AAV diagnosis20. AAV disease activity was assessed at the time of AAV diagnosis using the Birmingham Vasculitis Activity Score (BVAS) version 3 (scores range from 0 to 63, with higher scores indicating more active disease)21. The modified BVAS was used for GPA22, and AAV disease prognosis was assessed by the 1996 Five-Factor Score (FFS; with severe disease corresponding to a FFS ≥ 1: i.e., presence of cardiac, central nervous system or gastrointestinal involvement, peak creatinine level > 140 μmol/l, and/or proteinuria > 1 g/d)23.
All diagnoses of bronchiectasis were reassessed by the investigators (RL, MC, and CT). Available chest CT scans were retrospectively reviewed by an expert thoracic radiologist (PYB) blinded to clinical and histopathological data. The severity of bronchiectasis was gauged using the FACED score [FEV1, age, chronic Pseudomonas aeruginosa colonization, extension, and dyspnea; a tool validated for patients with noncystic fibrosis bronchiectasis leading to 3 levels of severity: mild bronchiectasis (0–2 points), moderate bronchiectasis (3–4 points), or severe bronchiectasis (5–7 points)], and in all patients with available high-resolution CT scan, using the Bhalla scoring system24,25. Briefly, the Bhalla score (ranging from 0 to 25 points, the latter corresponding to the most severely involved patients) depicts various aspects of the pulmonary involvement (e.g., the maximum diameter of bronchiectasis, peribronchial thickening, and the presence of sacculation and/or abscesses). For patients with more than 1 available CT scan, Bhalla score was assessed yearly (or more in case of disease progression).
An obstructive ventilatory defect was defined by a forced expiratory volume in 1 s (FEV1)/forced vital capacity ratio of < 70%, while a restrictive ventilatory defect was defined as a total lung capacity of 80% or less than the predicted value26.
Long-term outcome measures
During follow-up, several long-term events were recorded: the evolution of bronchiectasis, respiratory function, the occurrence of chronic respiratory failure26, severe lower respiratory tract infections, death, and AAV relapse. Specifically, bronchiectasis worsening and AAV relapse were defined, respectively, as (1) a decline of ≥ 1 point(s) of the Bhalla score; and (2) the reappearance or worsening of AAV features with BVAS > 0 and requiring treatment escalation27.
Treatment procedures
Drugs used both to induce and maintain remission were recorded, as well as those used as antibiotic prophylaxis (i.e., cotrimoxazole and/or azithromycin).
Statistical analysis
Patient characteristics are reported as the number and percentage for categorical variables and as the median and IQR for continuous variables. Patient subsets were differentiated based on the type of AAV and on the first disease manifestation. Hence, patients in whom bronchiectasis preceded the diagnosis of AAV were classified as the “B-AAV” subgroup, while patients in whom bronchiectasis was diagnosed concomitantly or after the diagnosis of AAV were grouped into the “AAV-B” subgroup. For these subsets, quantitative variables were compared using the Wilcoxon-Mann-Whitney U test (or the Kruskal-Wallis test when > 2 independent groups), and categorical variables were compared using Fisher’s exact test.
Pulmonary outcomes were assessed by comparing respiratory findings at diagnosis of bronchiectasis (whether it was assessed prior to or after that of AAV) and at last follow-up. Patient’s survival after onset of AAV was analyzed using the Kaplan-Meier method and compared using log-rank tests. Results are presented as HR with 95% CI. The Cox proportional hazard model was used to estimate HR using variables with P < 0.2 in univariable analysis. P values < 0.05 were considered significant. All statistical analyses were conducted using R 3.3.3 for Windows (R Foundation for Statistical Computing).
RESULTS
Patient characteristics at diagnosis of AAV and treatments
Of the 2035 AAV patients included in the FVSG database in May 2015, 43 (2%) were reported to have bronchiectasis. Forty-six additional patients with both AAV and bronchiectasis were included from a national survey. After exclusion of 28 patients (ILD-related traction bronchiectasis: n = 2; insufficient data: n = 26), 61 patients (16 males, 45 females) from 19 centers were included in our present study. Their baseline characteristics are reported in both Table 1, and Supplementary Table 1 (available with the online version of this article).
Patients’ characteristics and treatment at diagnosis of antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) according to whether bronchiectasis occurred before AAV.
Median age at AAV diagnosis was 64 (IQR 55–75) years. AAV consisted either of MPA (n = 27), GPA (n = 27), or EGPA (n = 7). Thirty-nine patients (64%) had anti–MPO-ANCA (including 2 patients with concomitant antibactericidal/permeability-increasing protein ANCA), and 13 patients (21%) had anti-proteinase 3-ANCA, while ANCA lacked specificity (n = 6) or were absent (n = 5) in the remaining cases (consisting of ANCA-negative EGPA patients with biopsy-proven and/or clinical surrogates of vasculitis)25. The diagnosis of bronchiectasis preceded that of AAV in 25 cases (41%), with the median time between both diagnoses 16 years (IQR 4–54). Both diagnoses were simultaneous in 12 cases (20%), while bronchiectasis occurred after the diagnosis of AAV in the last 24 remaining cases (39%; median time between both diagnoses: 1 yr, IQR 0–6). Patients belonging to the B-AAV subgroup had more frequent neurological involvement (64% vs 28%, P = 0.01), MPO-ANCA (80% vs 53%, P = 0.03), and a trend toward more frequent MPA (60% vs 33%, P = 0.07) than their counterparts from the AAV-B subgroup. At the diagnosis of AAV, the median BVAS was 13 (IQR 7–18) and the median FFS was 0 (0–1), without difference between patient subgroups.
Glucocorticoids were the cornerstone of therapy (n = 60, 98%), alone or associated with another immunosuppressive therapy (n = 53, 87%), consisting either of cyclophosphamide (CYC; n = 45, 74%), azathioprine (n = 32, 52%), mycophenolate mofetil (n = 12, 20%), rituximab (RTX; n = 15, 25%), or methotrexate (n = 14, 23%). Long-term antibiotics with cotrimoxazole (n = 33, 54%) and/or azithromycin (n = 12, 20%) were also frequently prescribed (Table 1).
Pulmonary characteristics at diagnosis of bronchiectasis
Table 2, and Supplementary Table 2 (available with the online version of this article) summarize patients’ pulmonary characteristics at diagnosis of bronchiectasis. Respiratory symptoms were present in 45 cases (74%): cough (n = 32, 52%), dyspnea (n = 29, 48%), and hemoptysis (n = 16, 26%). Among the 61 patients, 8 had low-resolution chest CT scans that, although showing bronchiectasis, lacked sufficient resolution to accurately assess the Bhalla score. Among the 53 remaining patients, the median Bhalla score was 9 (IQR 6–12; Supplementary Table 3, available with the online version of this article). Bronchiectasis was usually widespread [affecting > 6/20 bronchopulmonary segments in 27 cases (51%)] and presented up to the fifth generation of bronchial division in 46 cases (87%). The dilatation of bronchi was usually moderate to severe [diameter of the lumen ≥ 3× of normal features in 30 (56.5%) cases], while peribronchial thickening tended to be mild to moderate [bronchi wall thickness being superior to twice the diameter of adjacent vessel in 44 cases (83%); Supplementary Figure 1, available with the online version of this article]. Besides bronchiectasis, other findings on the initial high-resolution chest CT scan included nodules (n = 19, 36%), nodule excavations (n = 2, 4%), ILD (n = 1, 2%; in a patient with focal bronchiectasis unrelated with ILD), and ground-glass opacities suggestive of alveolar hemorrhage (n = 1, 2%), without differences between patient subgroups.
Respiratory characteristics at diagnosis of bronchiectasis according to whether bronchiectasis occurred before AAV.
Respiratory function tests were performed in 35 patients (57%) at the diagnosis of bronchiectasis. Twenty patients (57%) had airway obstruction [the latter being more frequent in EGPA (100%) than in MPA (35%) and GPA (5%); P = 0.01]. Five patients (14%) exhibited airway restriction (without differences between patient subgroups), while the latter tests were unremarkable in the remaining 10 patients.
Long-term outcomes
Table 3, and Supplementary Table 4 (available with the online version of this article) summarize patient outcomes. After a median AAV follow-up of 7 (4–13) years, the overall survival was 100% at 1 year, 97% at 5 years, and 89% at 10 years but tended to be worse in the B-AAV subgroup (overall survival at 10 yrs of 62% vs 93%, P = 0.03; Figure 1A). The leading causes of death consisted of pneumonia (n = 2, 22%), acute renal failure (n = 2), massive hemoptysis due to uncontrolled AAV (n = 1), myocardial infarction (n = 2, 22%), chronic respiratory failure due to bronchiectasis (n = 1, 11%), and sudden death of unknown origin (n = 2). Vasculitis relapses occurred in 26 patients (43%) and chronic respiratory failure in 1 case (2%), without differences between patient subgroups. In the B-AAV subgroup, there were no significant differences regarding patients’ survival between those treated with RTX and/or CYC and those who received neither of these drugs (HR 0.14, 95% CI 0.01–1.7, P = 0.1; Figure 1B).
Kaplan-Meier survival curves of patients with AAV-associated bronchiectasis according to (A) the first disease manifestation: AAV (“AAV-B,” thick line) or bronchiectasis (“B-AAV,” dashed line); and (B) the treatment regimen during follow-up: glucocorticoids alone (“CS alone,” thick line) or glucocorticoids in combination with cyclophosphamide and/or rituximab (“CS + CYC/RTX”, dashed line). AAV: antineutrophil cytoplasmic antibody-associated vasculitis; B: bronchiectasis; CS: corticosteroids; CYC: cyclophosphamide; IS: immunosuppressive agent CYC and/or RTX; RTX: rituximab.
Patients’ long-term outcomes according to whether bronchiectasis occurred before AAV.
As for respiratory outcomes, 32 patients (52%) presented at least 1 pulmonary infection with a median rate of annual severe infection of 0.1 per year (IQR 0–0.4). P. aeruginosa colonization occurred in 18 cases (30%), and infection due to nontuberculous mycobacteria in 3 cases (5%). Of the 37 patients with more than 1 available CT scan, 6 B-AAV patients (17%) improved their Bhalla score while under immunosuppressants versus 5 (22%) AAV-B patients (P = 0.3); the median improvement being 1 (IQR 1–2) point. Among the 17 patients who exhibited pulmonary nodules and underwent follow-up chest CT, such nodules resolved (or significantly improved, with only mild sequelae) in 12 patients (71%). Follow-up evaluations of the respiratory function tests were available in 16 cases (26%), with a median FEV1 decrease of 4.5% (IQR −14 to 4.5) between the first and last follow-up visits. Although nonsignificant, the FEV1 tended to be lower [−10%, (IQR −18 to 2) vs 1% (IQR −6 to 5), P = 0.2] in the AAV-B subgroup.
Factors associated with the worsening of bronchiectasis
Factors associated with the worsening of bronchiectasis are reported in Table 4. Thirty-seven patients (61%) had more than 1 chest CT scan that could be further analyzed. Age at diagnosis of AAV and arthralgia were predictors of bronchiectasis worsening. Conversely, AAV relapse requiring treatment escalation tended to be protective (HR = 0.6, 95% CI 0.4–0.99, P = 0.049), with at the end of follow-up, a mean Bhalla score variation of 2.1 (SD ± 3) points in patients who underwent AAV relapse (vs +2.5, SD ± 3.7 points variance in nonrelapsing patients). Results persisted in multivariate analysis.
Factors associated with bronchiectasis worsening.
Factors associated with death
Factors associated with death are reported in Table 5. Those that were statistically significant included age at diagnosis of AAV (HR 1.2, 95% CI 1.1–1.3; P = 0.003), MPA (HR 18.1, 95% CI 2.2–146.3; P = 0.01) and belonging to the B-AAV subgroup (HR 5.8, 95% CI 1.2–28.7; P = 0.03), while ENT involvement was associated with lower mortality (HR 0.1, 95% CI 0.01–0.6; P = 0.01). Statistical significance disappeared in multivariate analysis but within the same trends as for HR (data not shown).
Factors associated with death during patient follow-up.
DISCUSSION
We report on patients with both AAV and bronchiectasis from a large nationwide multicenter series set up through complementary multidisciplinary networks. While bronchiectasis can both precede the onset of AAV or occur during vasculitis follow-up, each scenario seems to correlate with a specific patient phenotype: patients with bronchiectasis preceding AAV (B-AAV) tend to have more frequent neurological symptoms (especially mononeuritis multiplex), MPA, MPO-ANCA, and are at a 5-fold increased risk of death compared to patients with AAV preceding (AAV-B). Moreover, bronchiectasis can also be reported in a substantial proportion of ANCA (mainly MPO)-positive patients with initially no other systemic manifestation of vasculitis. This is an important finding, highlighting that ANCA serology should be part of the initial diagnostic investigation of newly diagnosed patients with bronchiectasis and that, in the context of ANCA-associated bronchiectasis, pulmonologists should look for nonrespiratory manifestations of AAV28,29,30,31.
Various underlying pathophysiological processes might account for the highlighted differences in both disease presentation and ANCA subtypes. On the one hand, up to two-thirds of the patients developed bronchiectasis during the follow-up of AAV, often shortly after AAV diagnosis, suggesting an active immune-related process leading to the onset and worsening of bronchiectasis. Hence, in AAV-B patients, analogous to chronic graft-versus-host disease-related bronchiectasis32, bronchiectasis seems to be a dynamic inflammatory process that can, at least partially, be reversed. Because neither cartilage nor small vessels are found in the wall of small bronchioles, the disease’s pathophysiology is likely not related to genuine vasculitic manifestations. Yet, a plausible explanation could be the activation of neutrophils by ANCA in the lumen of bronchi, thus triggering bronchial inflammation33. On the other hand, in B-AAV patients, analogous to chronic Staphylococcus aureus nasal carriage in GPA, chronic bronchial suppuration might be a contributing factor to the development of autoimmunity and AAV onset. Strikingly, and in line with this hypothesis, Pearce, et al demonstrated in a case-control study involving 757 GPA patients and 7546 controls that GPA patients were 5 times more likely than controls to have a previous diagnosis of bronchiectasis (OR 5.1, 95% CI 2.7–9.4; P < 0.0001), even when patients were diagnosed with bronchiectasis more than 5 years before vasculitis onset11. Unfortunately, given the retrospective design of our present study, cytobacteriological examination of the sputum was not performed sequentially, and further longitudinal microbiological studies would be of paramount interest to better decipher the interplay between bacteriological findings and bronchiectasis onset and/or worsening.
The rates of bronchiectasis reported in patients with AAV are highly variable from one study to another. In our present study, a high-resolution CT scan was not mandatory at enrollment in the FVSG database, and it was not a queryable menu item — the most likely explanations for the low rate reported herein, which is likely to be considerably under estimated. Moreover, in another large series of patients with AAV, bronchiectasis was found in 21% of patients with AAV and considered severe in 2%8. More recently, in a monocentric retrospective study of French AAV patients with available chest CT scans (and that did not overlap with the present work), the prevalence of ANCA-associated bronchiectasis was reported as high as 38% (n = 22/58) of AAV patients, although the severity of bronchiectasis was not specified9. Further large-scale studies with systematic implementation of chest CT are warranted to better assess the epidemiology and true burden of AAV-associated bronchiectasis.
There are no guidelines available for the management of patients with both bronchiectasis and AAV. Interestingly, AAV relapse had a protective effect toward the deterioration of bronchiectasis, but treatment with either CYC or RTX was not associated with improved survival, and the same conclusions were drawn when specifically studying highest-risk patients (i.e., B-AAV subgroup). Hence, to date there is no rationale for AAV treatment escalation based only on the presence of bronchiectasis. Conversely, given the low annual infection rates (that tend to be the same as those of patients with AAV but without bronchiectasis) reported in both the present study and that of Néel, et al9, the presence of bronchiectasis (even with bronchorrhea or mucoid impaction on chest CT) should not prevent the use of potent immunosuppressive drugs, when required. Yet, a study of 192 patients with AAV treated with RTX suggested that the presence of bronchiectasis was associated with a 6-fold increased risk of severe respiratory tract infections34. Moreover, life-threatening nontuberculous mycobacterial infections have also been reported in this setting35. Hence, we recommend that all patients receive influenza and pneumococcal vaccinations36, ideally at a time when they are treated with neither CYC nor RTX, which reduce the vaccination’s immunogenicity37. Because of their antimicrobial and antiinflammatory properties, both cotrimoxazole and azithromycin deserve to be investigated in this setting, particularly in patients with frequent acute exacerbations30,38,39. Last, although ciprofloxacin was used in combination with wide-spectrum beta-lactams in selected cases of patients with acute infection related to P. aeruginosa, fluoroquinolones were never prescribed in the long term, to prevent antimicrobial drug resistance.
We acknowledge some limitations of our study, such as its retrospective design, the small sample of patients, and the limited number of events, which did not enable multivariate analysis in some cases. Last, because both vasculitis and bronchial disease were diagnosed concomitantly in 12 AAV-B patients, we cannot rule out that some of the patients with concomitant diagnosis already had long-lasting mild bronchiectasis and thus were misclassified. Yet bronchiectasis is seldom asymptomatic, and the comparison of 3 subgroups (i.e., B-AAV vs AAV-B vs patients with concomitant diagnoses of both bronchiectasis and AAV) yielded the same results, with MPA, neurologic involvement, and MPO positivity being more frequent in the B-AAV subgroup (the latter patients also presenting higher mortality rates; data not shown). Moreover, such dichotomization tends to reflect daily practice, with bronchiectasis being diagnosed either by pulmonologists (prior to AAV onset), or by AAV-treating physicians during vasculitis follow-up, in patients previously free from respiratory symptoms. For the sake of completeness, crude data of patients according to both vasculitides’ phenotypes and ANCA specificity are also presented in Supplementary Tables 1 and 2 (available with the online version of this article).
Our findings suggest that the association of bronchiectasis with ANCA and vasculitis may not be accidental, especially among patients with MPO-ANCA and/or MPA. Two clinical phenotypes seem to coexist, according to the timeline between both bronchiectasis and AAV diagnoses. Bronchiectasis is usually a mild disease, but patients in which bronchiectasis precedes the diagnosis of AAV tend to have a poorer prognosis. Last, ANCA serology should be part of the diagnostic investigation of patients with newly diagnosed bronchiectasis, and because systemic vasculitis is likely to occur even decades after the diagnosis of bronchial disease, pulmonologists should look for nonrespiratory (especially neurological) manifestations of AAV.
ONLINE SUPPLEMENT
Supplementary material accompanies the online version of this article.
Acknowledgment
We thank the members of both the French Vasculitis Study Group and the Groupe d’Etudes et de Recherche sur les Maladies Orphelines Pulmonaires for their help in data collection.
- Accepted for publication November 8, 2019.
REFERENCES
In this issue
