Characterisation of severe obliterative bronchiolitis in rheumatoid arthritis

Pulmonary manifestations may account for 10–20% of all rheumatoid arthritis (RA)-related deaths, ranking second after cardiovascular disease 1. These include opportunistic pulmonary infections, drug-induced lung disease, and pulmonary manifestations directly associated with RA, which may affect all anatomic compartments of the lung, with the exception of pulmonary hypertension, which is distinctly rare in RA 2. Pleural disease (with or without effusion), interstitial lung disease (most commonly with a histopathological pattern of usual interstitial pneumonia) and rheumatoid (necrobiotic) lung nodules are the most frequent manifestations, with a prevalence varying according to the investigation employed.

Airway complications include cricoarytenoid arthritis, bronchiectasis (often not clinically significant) and small airways disease, which may cause airflow obstruction; these were first reported in 1977 in six patients, five of whom had RA 3. In the present manuscript, the following terminology is used: “bronchiolitis obliterans” (historically used for the histopathological lesions) to name the pathological feature of reduced lumen of the bronchioles, and “obliterative bronchiolitis” (OB) to name its functional counterpart with airflow obstruction. The estimated prevalence of bronchiolar disease in RA varies widely (from 8% to 65%) between studies 4–8 depending on the criteria used. It may correspond pathologically to either fibrosing (constrictive bronchiolitis obliterans) or cellular bronchiolitis (follicular bronchiolitis or, less commonly, diffuse panbronchiolitis) 2, 4, 5.

Although some of the characteristics of bronchiolar disease have been reported in systematic studies of consecutive patients with RA undergoing pulmonary function tests and/or high-resolution computed tomography (HRCT) of the chest, regardless of pulmonary symptoms 6–8, the small subgroup of RA patients with severe airflow obstruction remains poorly investigated. The current study provides a detailed analysis of a homogenous group of 25 patients with functional evidence of OB (as defined by severe airflow obstruction without clinically relevant tobacco smoking, and with or without pathological diagnosis). Herein it is shown that the clinical radiological presentation of RA-associated severe OB may be nonspecific and resemble chronic obstructive pulmonary disease (COPD) with emphysema, with inconstant direct signs of bronchiolitis at HRCT. Furthermore, it is shown that some features of interstitial lung disease may be present in addition to the airways disease in as many as one third of the cases.

PATIENTS AND METHODS

RESULTS

Chest imaging

The most common findings on chest radiograph were pulmonary hyperinflation (64%), diffuse lung infiltrates (44%) and bronchiectasis (40%). Diffuse alveolar opacities and nodular shadows were reported less frequently (32% and 16% of cases, respectively).

Bronchial wall thickening, centrilobular emphysema (fig. 1⇓), lobular areas of decreased attenuation with mosaic pattern indicative of air trapping, and bronchiectasis (fig. 2⇓) were the most frequent HRCT findings, present in 96%, 56%, 42% and 40%, respectively (table 2⇓). Areas of ground-glass attenuation were found in 44% of cases, and reticular opacities with mild honeycombing in the lower lobes were present in 16% of cases. An association of diffuse emphysema and interstitial opacities of the lower lobes (including ground-glass attenuation) was present in 32% of cases; infiltrative opacities were bilateral and diffuse, without central or peripheral predominance. Pleural effusion was observed in four patients, and rheumatoid lung nodules in nine patients.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

High-resolution computed tomography showing diffuse emphysema lesions in a nonsmoking patient with rheumatoid arthritis and severe airflow obstruction.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

High-resolution computed tomography showing bronchiectasis and centrilobular micronodules in a nonsmoking patient with rheumatoid arthritis and severe obliterative bronchiolitis.

View this table:

• View inline
• View popup

Table 2—

High-resolution computed tomography findings

Lung pathology

Lung pathology of video-assisted thoracoscopic or open lung biopsy or explanted lung was available in eight and one case, respectively, and analysed by an experienced lung pathologist (FTB). A histopathological pattern of bronchiolitis was found in all cases, occasionally associated with accessory lesions consistent with usual interstitial pneumonia or diffuse alveolar damage (one case each). Bronchiolar changes consisted in chronic constrictive bronchiolitis (six cases), follicular bronchiolitis (one case), and mixed constrictive/follicular bronchiolitis (two cases; fig. 3⇓). Constrictive bronchiolitis was characterised by the presence of concentric fibrosis of the bronchial wall, with severe narrowing of the bronchiolar lumen and/or bronchiolectasis. Inflammatory infiltrates (predominantly lymphocytic) were present within the wall of bronchioles in four cases. Centrilobular emphysema adjacent to bronchiolar changes was further present in five cases.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3—

Histopathological analysis of lung biopsy specimen. a) Lymphocytic constrictive bronchiolitis with nonfollicular lymphocytic infiltrate of the bronchiolar wall. b) Chronic constrictive bronchiolitis, with thickening of the bronchiolar wall by collagen deposition, partial destruction of the muscular fibres, narrowing of the bronchiolar lumen and mild peribronchiolar lymphocytic infiltrates. Scale bars = 40 μm.

DISCUSSION

The present study showed that OB associated with RA is a severe condition, with imaging and functional features resembling those of COPD, with the exception of less impaired D_L,CO. Typical imaging features of bronchiolitis were inconstant at HRCT, and interstitial lung disease was occasionally associated. Respiratory failure was common.

Among the connective tissue diseases, OB has been reported mostly in RA 11, although it may also be encountered in Sjögren disease 12 and, only exceptionally, in systemic sclerosis 13. The association of OB and RA was initially described by Geddes et al. 3 in six patients (five of whom had RA) with rapid progressive dyspnoea and airflow obstruction, with bronchiolitis obliterans demonstrated at necropsy in four cases. Most cases occur <5 yrs after the diagnosis of RA, but OB may be a presenting manifestation of RA in 10–20% of patients 2, as in two of the current patient series.

Since the original description, several studies reported that mild to moderate airways involvement is frequent in patients with RA regardless of pulmonary symptoms 8, 14, 15. Hence, airways obstruction (reduced FEV₁/FVC) was present in 13.6% of consecutive nonsmoking patients undergoing systematic evaluation in one study 16, and some HRCT abnormalities suggestive of airways disease were present in 34% of patients in another study 8. The disease course is usually uneventful in poorly symptomatic patients with only mild airflow obstruction and/or isolated bronchiectasis detected by systematic evaluation 17, a condition distinct from that of symptomatic patients with severe OB (as in the present study).

The patients reported in the present study, which is the largest series of patients with RA-related severe OB, presented with an abrupt or progressive onset of dyspnoea and dry cough, often with squeaks and crackles at auscultation, as previously reported 3. However, the mean±sd delay between respiratory symptoms and diagnosis was 19±32 months, further delaying onset of treatment. The diagnosis may be especially delayed in patients with progressive onset of symptoms, and in those with OB presenting concomitantly or before the diagnosis of RA. Pulmonary function testing revealed airflow obstruction, which seemed to progress more rapidly than in COPD, as previously described 3, 15, 18, with mild to moderate hypoxaemia. In addition to airflow obstruction (with decreased FEV₁/FVC), hyperinflation demonstrated by increased RV and RV/TLC was common. In order to improve early detection of disease, the current authors suggest sequential pulmonary function tests (including measurement of RV and RV/TLC) be performed in any RA patient with progressive dyspnoea on exertion. In addition, lung biopsy should be considered more frequently in this condition, except when contraindicated by the severity of disease or comorbidities, as it might provide access to definitive diagnosis and early therapy. However, the current authors consider that the diagnosis of OB may be obtained without a lung biopsy in a patient with RA and progressive severe airflow obstruction in the absence of significant tobacco smoking and no other cause of airflow obstruction. Treatment with bronchodilators and oral corticosteroids was ineffective in most (84%) of the patients, as previously described 3. Some improvement was obtained in two patients who received off-label anti-TNF-α regimen for OB (one of these has been previously reported 10). Lung function further deteriorated in half of the patients, despite therapy, with chronic respiratory failure in 40%, and death in 16%. The use of high-dose corticosteroids, cyclophosphamide, and of erythromycin, has been suggested 6, 19, 20, but has been poorly evaluated. It is not known whether early diagnosis and treatment with inhaled or high-dose oral corticosteroids, or cyclophosphamide, will decrease progression of OB. Although not formally compared, overall survival seemed lower than that of COPD patients with similar airflow obstruction, and was almost similar to that in post-transplant bronchiolitis obliterans syndrome 21, 22. No treatment recommendation can be provided from published data and the present study. However, it is current experience-based practice, in patients with documented and progressive severe RA-associated OB (and especially those with cellular bronchiolitis at biopsy), to propose combination therapy with bronchodilators, inhaled and oral corticosteroids; macrolides, pulse intravenous cyclophosphamide, or etanercept (with methotrexate) may be considered as second-line therapy.

Consistent with previous studies 3, 11, 15, 13 out of 25 patients had received previous therapy with d-penicillamine or tiopronine, and 10 out of 25 had received gold salts, raising the concern of drug-induced bronchiolar disease. In addition, FEV₁ was significantly lower in patients who had ever been treated with d-penicillamine compared with those who had not, with a similar trend of more severe impairment of lung function in patients ever treated with methotrexate or gold salts. A majority of patients with RA and OB are females 20, contrasting with the increased prevalence of interstitial lung disease associated with RA in males 11. Whether associated Sjögren disease affects the development of OB is controversial 8, 16, 23. In addition, one study reported a higher frequency of OB in patients with human leukocyte antigen-B40 and -DR4 24. The risk of OB was not related to the severity of rheumatic disease, as RA was controlled by treatment in two thirds of the present patients.

Only eight out of 25 patients in the current series were ex-smokers, with a mean smoking history of only 10 pack-yrs. It is, therefore, the conviction of the present authors that progressive and severe airflow obstruction could be explained in them neither by tobacco smoking, nor by any known cause other than RA 11. In fact, pulmonary function was less severely impaired in ex-smokers than in never-smokers. Most patients had isolated, unexplained, and progressive obstructive lung disease, somewhat similar to COPD (although usually with more rapid functional worsening and less impaired D_L,CO). Hence, the Global Initiative for Obstructive Lung Disease functional definition of moderate, severe, and very severe COPD 25 was fulfilled in 20%, 56% and 24% of the 25 patients, respectively. Interestingly, cigarette smoking is associated with an increased risk of developing RA 26 and with RA activity, an association probably due to post-translational protein modifications induced by tobacco-derived products 27. The relationship between tobacco smoking and the development of lung disease in RA is unclear 2. Although severe OB may develop without tobacco smoking in patients with RA, it is conceivable that smoking might have contributed to pre-existing small airways disease in ex-smokers, and all RA patients should be discouraged from smoking. Lymphocytic activation within bronchus-associated lymphoid tissue (where the immunopathology of RA mostly takes place within the lung) also probably contributes to bronchiolar disease in RA 28, with involvement of both B-lymphocytes and CD4+ T-lymphocytes 28, 29. Whether the onset or outcome of OB can be affected by the presence and treatment of gastro-oesophageal reflux in patients with RA is unknown and would require specific investigation.

HRCT features consisted mostly of indirect signs of bronchiolitis, with a mosaic pattern reflecting air trapping in 44% of cases, compared with 20–32% in studies conducted in RA patients regardless of symptoms 7, 8. The frequency of mosaic pattern was probably underestimated, however, since expiratory images were not systematically obtained. Part of the imaging was performed using single-detector row scanners. Direct signs of bronchiolitis (centrilobular micronodules and tree-in-bud opacities) and thin-walled cysts were also reported in previous imaging studies of consecutive patients with RA, with centrilobular opacities reported in 6% 8, 28% 7, and 71% 30 of patients.

One unexpected finding of the present study was the high frequency of diffuse emphysema lesions on HRCT (56% of patients, five of whom were smokers) and on histopathology (five out of nine cases with biopsy, including three smokers), a prevalence higher than in previous studies (i.e. 4–39% 7, 8, 31, 32), which may be partly due to the inclusion criteria, which included air trapping. Centrilobular emphysema was not influenced by a history of moderate tobacco smoking in 32% of patients; it may conceivably develop as a late result of chronic bronchiolitis, as in smokers 33, and/or be related to air trapping. Interstitial lung disease was also present in more than half of the patients with OB, consistent with previous reports 5, 28; interstitial lung disease was mild in most patients but evolved to severe pulmonary fibrosis and death in one patient. In about one third of patients, both emphysema and infiltrative opacities were present, and was somewhat reminiscent of the syndrome of combined pulmonary fibrosis and emphysema identified in smokers 34, 35. Taken together, the findings suggest that severe OB occurring in RA present significant similarity with COPD, especially with emphysema on chest imaging, and subtle changes suggestive of bronchiolar disease may be overlooked.

Bronchiectasis is present in 8–30% of patients with RA 7, 16, 31, 36 and has been associated with a higher risk of airflow obstruction, pulmonary infections and death 37. Bronchiectasis, bronchial wall thickening, cough and bronchorrhea in the patients from the present study with OB suggest comprehensive airways disease, involving both bronchioles and bronchi of all size. Therefore, airway disease with OB occurring in RA compares to that complicating bone marrow transplantation 21, 38 and lung transplantation 21.

Pathological analysis of the lung biopsy available in nine out of 25 cases (a proportion possibly greater than that in current clinical practice) demonstrated bronchiolitis in all cases, predominantly with constrictive fibrotic bronchiolitis obliterans and peribronchiolar fibrosis (so-called bronchiolocentric fibrosis 39), and occasionally follicular or mixed constrictive/follicular bronchiolitis, in a similar manner to previous studies 5, 6. Both pathological patterns often coexist in patients with RA 6. Although constrictive and follicular bronchiolitis share common clinical, functional and imaging characteristics, follicular bronchiolitis might theoretically respond better to therapy.

The limitations of the present study are its retrospective and multicentric nature; the management of patients was not uniform, limiting interpretation of effect of therapy. Expiratory chest HRCT was not performed in all patients. Referral bias probably occurred, with the most severe cases being referred to participating tertiary centres. Although inclusion bias may have occurred, as in any retrospective study, the current authors consider that it was limited by including patients in whom lung biopsy was not performed (with potentially more severe disease), therefore the series reflects current clinical practice. The design of the study did not allow an evaluation of the epidemiology of disease (and especially its actual prevalence) and evaluation of the spectrum of severity of disease (and especially patients with mild or moderate disease were not included), but rather focused specifically on describing the subpopulation of patients with severe OB, in order to improve the knowledge regarding the manifestations, chest imaging, outcome and, possibly, therapy of this orphan condition, and to further justify and promote an earlier detection. While only severe cases were included, it cannot be ruled out that inclusion may be skewed toward cases with less progressive disease and better outcome (with ensuing longer follow-up facilitating recollection by the physician). As patients included were diagnosed over a period of 20 yrs, image acquisition was not standardised. The limited number of patients did not allow multivariate statistical analysis.

In conclusion, the current authors describe a homogenous cohort of 25 patients with severe obliterative bronchiolitis attributable to rheumatoid arthritis, characterised by severe dyspnoea, lung hyperinflation, bronchial wall thickening, frequent bronchiectasis, and centrilobular emphysema on chest imaging. Disease presentation was somewhat similar to that of chronic obstructive pulmonary disease in tobacco smokers. Severity of disease was variable. Response to oral corticosteroids was poor, frequently leading to chronic respiratory failure and possible death.

Statement of interest

None declared.