Interstitial Pulmonary Fibrosis

Clinical Features

The prevalence of lung disease in SSc depends on the method used for detection. Symptoms or signs of lung disease are common. Dyspnea is present in roughly 55% of patients (range, 21% to 80%). Cough, a less frequently reported symptom, tends to be dry and nonproductive. Hemoptysis is rare but may complicate carcinoma or bronchial telangiectasia. Fine crackles are heard at the lung bases and are of a “Velcro” character. Pleural rub is almost never heard. Pleuritic chest pain is uncommon. Pneumothorax is even less common. Digital clubbing is extremely rare. Secondary pulmonary hypertension with appropriate clinical features of right ventricular strain, raised jugular venous pressure, and ankle edema may be seen during the terminal stages of the disease.

The lung is involved more commonly in patients with diffuse rather than limited cutaneous SSc, but the extent of skin involvement does not correlate with lung function changes. Patients are often without symptoms, even with moderate pulmonary function impairment, but in other cases, exertional dyspnea is present when ILD is trivial or absent, due to pulmonary vascular limitation, cardiac involvement, musculoskeletal problems, general debility, loss of fitness, or a variable combination of these factors. Rarely, scleroderma of the chest wall may cause extrathoracic restriction.

Lung disease may be the first manifestation of SSc. In this setting, a history of Raynaud phenomenon is often helpful. In addition, careful examination of the capillaries in the nail beds reveals the typical feature associated with scleroderma: abnormal loops associated with capillary “dropout” ( Fig. 65-1A ). The presence of autoantibodies, particularly antinuclear antibodies, is helpful in suggesting a CTD etiology ( Table 65-4 ).

Radiographic appearances in patients with systemic sclerosis.

A, Nail bed showing abnormal capillary loops in the cuticle. B, Chest radiograph from a patient with systemic sclerosis. Note the widespread, predominantly peripheral pattern of reticular opacities. The heart borders and the diaphragm are obscured. C, CT scan taken at a level just below the carina in a patient with systemic sclerosis. Note the peripheral reticular pattern that in this case is more prominent anteriorly. In this patient with subtle disease, the dilated esophagus is a clue as to the true cause of the fibrosing lung disease. D, CT scan taken at the level of the diaphragm in a patient with more extensive interstitial pulmonary fibrosis in the context of systemic sclerosis. Note in particular the dilated airways, indicating that the apparent “ground-glass”/consolidation pattern is dense fibrosis. Again, the dilated esophagus indicates that this disease is likely to be part of systemic sclerosis.

( A, Adapted from Iaccarino L, Ghirardello A, Bettio S, et al: The clinical features, diagnosis and classification of dermatomyositis. J Autoimmun 48–49:122–127, 2014, Fig. 1; B–D, Courtesy Michael Gotway, MD.)

Imaging

Chest radiographic abnormalities are present in 25% to 67% of patients, and lung function is impaired in up to 90%, although a large subset of patients has only a minor reduction in gas transfer or diffusing capacity (D l _co ). Chest radiography typically shows a reticular pattern in the lung bases and periphery in early disease. There is obvious loss of volume with more extensive reticular shadowing in more advanced disease. In this situation, honeycombing may be present ( Fig. 65-1B ).

Computed tomography (CT) has revolutionized the interpretation of the pattern and extent of disease. Disease is localized to the lung periphery, appearing earliest at the bases and posteriorly ( Fig. 65-1C and D , eFigs. 65-1, 65-2, and 65-3 ). As disease becomes more advanced, it progresses superiorly, centrally, and anteriorly. Esophageal dilation is common, which can be helpful in the diagnosis if the lung disease is the first manifestation of the systemic disease ( Fig. 65-1C and D ). The pattern may be “ground-glass” (reflecting either fine intralobular fibrosis or a more cellular histopathology) or an overtly fibrotic “reticular” pattern, consisting of intersecting linear abnormalities, often associated with traction bronchiectasis. Honeycomb change is present in up to one third of cases but is usually limited in extent.

CT extent correlates moderately well with lung function variables, particularly measures of D l _co . In the absence of overt pulmonary hypertension, individuals with SSc are less hypoxemic than those with idiopathic pulmonary fibrosis, once the extent of disease on CT has been taken into account, a difference that has been ascribed to a relative absence of new vessel formation in abnormal lung in SSc. The extent of individual patterns (ground-glass vs. reticular) is associated with the type of inflammatory cell found in BAL. More extensive reticular (consistent with more fibrosis) disease is associated with high neutrophil numbers, and this influx appears to be associated primarily with more extensive lung disease.

Gallium scans provide no added value. Technetium-99m-labeled diethylenetriaminepentaacetic (DTPA) acid clearance has been used in some medical centers to identify early disease and predict prognosis. A more rapid clearance of this tracer from the air spaces into the circulation is indicative of a loss of epithelial cell integrity. A persistently rapid clearance rate confers an increased risk of subsequent lung function deterioration, and a persistently normal clearance rate predicts lung function stability.

Lung Function Tests

The interstitial pulmonary fibrosis of SSc is characterized by a restrictive ventilatory defect, which results in reduced pulmonary compliance, vital capacity, and total lung capacity. Residual volume is decreased. D l _co is reduced and may be the only abnormality in early disease. Blood gas analysis usually reveals a normal or reduced arterial oxygen pressure (P o ₂ ), reflecting regional vasoconstriction in diseased lung, with a normal or low arterial carbon dioxide pressure (P co ₂ ). In the absence of pulmonary hypertension, hypoxemia is seldom marked until late in the disease process.

There have been a number of studies of the rate of decline of lung function in SSc. In one study of 38 patients with SSc, the mean loss of vital capacity was 100 mL per year (20 to 30 mL/yr being the normal rate of decline). A second study showed that vital capacity loss was greater in patients who had evidence of an active alveolitis on BAL. Lower forced vital capacity (FVC) within 3 years of disease onset predicts decline in pulmonary function. Rates of decline are often higher within the first few years of onset of SSc, emphasizing the importance of identifying lung disease early.

Exercise lung function testing increases ventilation-perfusion mismatching and also increases diffusion abnormalities resulting in hypoxemia and widening of the alveolar-arterial oxygen difference. Minute ventilation increases generally as a consequence of increased respiratory rate rather than tidal volume. Dead space ventilation may increase with exercise, with a rise in the dead space–to-tidal volume ratio. For a given degree of lung involvement defined by CT, abnormalities on exercise gas exchange are more severe in patients with idiopathic pulmonary fibrosis than in those with interstitial pulmonary fibrosis in association with SSc.

Bronchoalveolar Lavage

BAL may identify alveolitis in SSc before the onset of pulmonary symptoms. A neutrophil alveolitis has been described by some authors to predict more progressive disease. However, it has been shown that an increase in neutrophils reflects an increase in the extent of disease on CT, particularly of the reticular pattern; thus, this increase is likely a marker of more extensive disease rather than an independent index of progressive disease. BAL should not be used alone to determine whether treatment should be started. Many patients with apparently normal BAL findings may exhibit progressive disease. BAL is, therefore, not recommended routinely for the diagnosis or monitoring of SSc-related interstitial disease.

Biopsy

Surgical lung biopsy is virtually never required in the diagnosis of SSc-related interstitial disease unless clinical and CT findings are atypical for ILD in SSc and an alternative diagnosis is suspected. No useful additional information is provided by transbronchial biopsy. Pathologically, the most prevalent pattern is nonspecific interstitial pneumonia, with thickening of the alveolar walls with inflammatory cells (mononuclear cells, granulocytes, and plasma cells); connective tissue matrix cells; and proteins combined with intra-alveolar inflammation (predominantly by macrophages), type II pneumocyte proliferation, and vascular obliteration. The distribution is subpleural and basal and is maximal in posterior segments, with the macroscopic findings of the lung surface taking on a fine nodular, “cirrhotic” appearance in the early stages and honeycombing with more advanced disease. This pattern differs from usual interstitial pneumonia in exhibiting a homogeneous pattern of pathology ( eFig. 65-4 ). Much less commonly, the usual interstitial pneumonia pattern of histopathology is seen. One of the hallmarks of usual interstitial pneumonia is a heterogeneous appearance, with normal alveoli seen in the same section as areas of extensive alveolar remodeling (areas of fibroblastic proliferation and dense fibrosis). Crucially, outcome does not differ materially. In idiopathic disease, the usual interstitial pneumonia pattern of injury is associated with a worse outcome compared with the nonspecific interstitial pneumonia pattern. However, in SSc, outcome is not related to the histologic pattern in the largest series, although there is evidence in a smaller series that some patients with usual interstitial pneumonia pattern have a progressive course. Overall, surgical biopsy data do not provide sufficient prognostic information to justify the procedure in SSc.

Electron microscopy shows early endothelial and epithelial cell injury, even without abnormalities on light microscopy. Autopsy studies show diffuse lung disease in up to 80% of cases and pulmonary vascular disease in up to 30% ( eFig. 65-4C ). In some patients with a more accelerated course of disease, histopathologic examination has shown a diffuse alveolar damage pattern.

Serologic Investigations

Although SSc is traditionally defined on the basis of the extent of the skin disease, as outlined previously, the pattern of autoantibodies appears to be a much stronger determinant of the associated internal organ involvement. Antinuclear antibodies are found in 90% to 100% of patients with SSc (roughly 30% of normal individuals have antinuclear antibodies at a titer of 1 : 40). Three major autoantibodies include anti-centromere, seen in 57% of patients with limited cutaneous disease; anti-topoisomerase (Scl-70), seen in 40% of patients with diffuse disease; and PM-Scl, seen in a small fraction of cases in association with the polymyositis overlap syndrome. It is rare to have both Scl-70 and anti-centromere antibodies. Diffuse lung disease is rare in the presence of anti-centromere antibodies, and a protective role for this autoantibody has been argued but not substantiated. Interstitial pulmonary fibrosis is strongly associated with the Scl-70 antibody and with diffuse scleroderma. Limited cutaneous disease associates with vascular disease and the anti-centromere antibody. Both forms of lung disease may progress to pulmonary hypertension. Other autoantibody studies have shown associations with organ involvement, including a nucleolar pattern associated with diffuse lung disease and pulmonary hypertension ; the antibody against B23, a nucleolar phosphoprotein, associated with pulmonary hypertension and the presence of the anti-fibrillarin antibody; and anti-Th/To antibodies associated with pulmonary hypertension and diffuse lung disease.

Prognosis

Although crude mortality rates are 3.9%/yr for men and 2.6%/yr for women, lung disease remains the most common cause of death in patients with SSc. In one series, lung disease accounted for 21% of all deaths. There is also an increased risk of lung cancer in SSc. With the widespread use of high-resolution computed tomography (HRCT), mild or trivial ILD (see eFigs. 65-1 , 65-3A ) is detected in many SSc patients, leaving the clinician with difficult decisions as to whether to introduce therapy or to observe carefully without immediate intervention. This dilemma is most frequently confronted in SSc but is also increasingly encountered in other CTDs. In overtly severe ILD, the decision to treat is straightforward. However, the majority of SSc patients have milder lung involvement, and a reliable means of discriminating between stable and progressive ILD would be useful in management. On the basis of clinical series and accumulated experience, prognostic evaluation should focus on three main considerations.

The decision to treat patients with SSc is most strongly influenced by the extent of disease at presentation, as judged by lung function tests (particularly FVC and D l _co ), and the extent of disease on CT. A simple staging system using CT extent (above and below 20%) and FVC at presentation (above and below 70%), the United Kingdom Raynaud’s and Scleroderma Association (UKRSA) staging system, is highly discriminatory for both progression and death ( Fig. 65-2 ). The prognostic value of the UKRSA staging system has been confirmed, and the severity thresholds were virtually identical to those found to identify the likelihood of a treatment effect in the first controlled study of oral cyclophosphamide in SSc.

Staging algorithm for interstitial lung disease due to systemic sclerosis.

A, Flow diagram of limited/extensive staging system with use of high-resolution CT (HRCT) scores and pulmonary function test data. Identification of overtly minimal/limited disease is based on disease extent threshold of 20% on HRCT versus severe/extensive disease that is identified by HRCT disease extent greater than 20%. In cases with an indeterminate extent of disease on HRCT, a forced vital capacity (FVC) threshold of 70% is used to separate the cases into limited or extensive disease. The HRCT scans are scored at five levels for total disease extent, extent of reticulation, proportion of ground glass, and coarseness of reticulation. B, Survival compared between patient subgroups with ( Ba ) FVC levels above and below a threshold value of 70%; ( Bb ) HRCT disease extent above and below a threshold value of 20%; and ( Bc ) limited disease (HRCT extent ≤ 10%, or when HRCT extent was 10% to 30%, FVC ≥ 70%) versus extensive disease (HRCT > 30%, or when HRCT extent was 10% to 30%, FVC < 70%).

(Data from Goh NS, Desai SR, Veeraraghavan S, et al: Interstitial lung disease in systemic sclerosis: a simple staging system. Am J Respir Crit Care Med 177:1248–1254, 2008.)

The selection of patients in need of treatment is also influenced by the duration of systemic disease. The greatest risk of ILD progression is during the first 4 years of SSc. Early declines in FVC are strongly predictive of subsequent severe ILD, which most often develops in the first 4 years. Put simply, in SSc and other CTDs, the early development of mild to moderate ILD is a marker of likely rapid disease progression, which should reduce the threshold for initiating therapy.

Finally, evidence of recent disease progression is an important justification for treatment. It should be acknowledged that the long-term prognostic value of observed disease progression has not been quantified in SSc or, indeed, in other CTDs. However, declines in lung function variables, especially FVC, have consistently been predictive of long-term mortality in other forms of pulmonary fibrosis and these observations can reasonably be extrapolated to CTDs.

Although no overall treatment algorithm of ILD has been devised to integrate the duration of systemic disease, evidence of recent progression, and the severity of disease, consideration of all three factors provides a basis for rational treatment decisions. More recently, attention has focused on refining the selection of appropriate SSc patients with pulmonary fibrosis to participate in controlled treatment trials. In order for treatment effects to be demonstrated, it is necessary to enroll patients with progressive disease: the UKRSA staging system has recently been endorsed for this purpose by an expert group. Biomarker data might, in principle, be used, both in clinical prognostication and to select patients for trial involvement, but at present, despite promising preliminary data, no single biomarker is currently fit for routine use.

Treatment

A wide variety of treatments for interstitial pulmonary fibrosis in SSc have been explored. Most of these include immunosuppression. A number of uncontrolled reports suggested an advantage of treatment with cyclophosphamide, together with prednisolone, to improve lung function and prognosis. The Scleroderma Lung Study demonstrated a small but significant benefit of oral cyclophosphamide compared with placebo for FVC, skin scores, and quality of life and dyspnea scores at 12 months. The oral cyclophosphamide treatment regimen used most commonly is 2 mg/kg/day orally up to a maximum of 150 mg with moderately low dosages of prednisolone at 10 mg/day or 20 mg every other day. Close scrutiny with full blood counts, liver function tests, and urine testing for evidence of hemorrhagic cystitis is necessary.

Intravenous cyclophosphamide may have a better safety profile. Studies of intravenous cyclophosphamide (750 to 1000 mg/m ² ) given at 2- to 4-week intervals for 6 to 12 months tended toward significance, as judged by CT and pulmonary function data. The similarity in the amplitude of treatment effects in the Scleroderma Lung Study and a smaller placebo-controlled trial of intravenous cyclophosphamide prompted the European League Against Rheumatism to state that “in view of the results from two high-quality RCTs and despite its known toxicity, cyclophosphamide should be considered for treatment of SSc-ILD.” However, it cannot be emphasized too strongly that in SSc and in other CTDs alike, careful selection of patients with severe or progressive lung disease is required, given the toxicity associated with intense immunosuppressive therapy.

In controlled trials, treatment effects largely consisted of the prevention or reduction in progression of lung disease, as judged by serial lung function trends, with the greatest benefit seen in patients with extensive fibrotic disease, endorsing the important principle of intervention in irreversible fibrotic lung disease, with the primary aim of achieving stabilization, as opposed to regression of disease. However, in the Scleroderma Lung Study, after the cessation of active treatment, therapeutic benefits were transient. Longer-term treatment approaches have not been evaluated in controlled studies. In SSc and in other CTDs, oral immunosuppressive agents that are less toxic than cyclophosphamide have been widely used, including azathioprine, methotrexate, and mycophenolate mofetil, on the basis of accumulated clinical experience and small retrospective case series. The most convincing data were generated in a recent retrospective series of 100 CTD patients, in which treatment with mycophenolate mofetil was well tolerated and was associated, on average, with stabilization of pulmonary fibrosis over a number of years.

Controlled treatment data in SSc (and in other CTDs) have been confined to immunomodulatory strategies. Novel antifibrotic agents with documented treatment effects in idiopathic pulmonary fibrosis (e.g., antioxidant therapy, pirfenidone ) have not been evaluated in CTD. A well-conducted placebo-controlled trial of bosentan in SSc pulmonary fibrosis was definitively negative. A study of interferon-α showed a greater deterioration in lung function at 1 year than with placebo. With bone marrow transplantation in severe progressive SSc, significant improvements in lung function variables are seen in some cases. The most promising pilot data apply to the use of rituximab in SSc-ILD and, more convincingly, in polymyositis-dermatomyositis, especially when given as rescue therapy in CTD patients with severe progressive disease despite high-dose corticosteroid and immunosuppressive therapy. However, the use of rituximab in lung disease in SSc and other CTDs is currently unclear.

The question of Pneumocystis prophylaxis is not resolved. Some medical centers use cotrimoxazole three times a week if immunosuppressive agents are being given.

It has been observed that steroid therapy is associated with scleroderma renal crisis, both with moderate to high-dose therapy (≥15 mg/day prednisone or equivalent) and, more recently, with prednisolone doses of less than 10 mg daily. However, confounding by severity cannot be excluded, as patients with more aggressive systemic disease are more likely to receive corticosteroids, although occasional cases of renal crisis are undoubtedly linked to high-dose corticosteroid therapy. Thus, low-dose steroid therapy remains justified as an invaluable adjunct to the treatment of lung disease, although renal function should be monitored.

End-stage lung disease has been treated with single-lung transplantation provided there is no evidence of disease activity in other organs and no major esophageal dysfunction. With careful selection of patients, outcome with transplantation may differ little from that in individuals with idiopathic diffuse lung disease, although it should be understood that perceived contraindications to lung transplantation due to systemic disease activity vary widely between transplant units. In terminal disease, consideration must be given to oxygen therapy, treatment of supervening heart failure, and infection and advanced care planning, with palliative care specialists often having an invaluable input.

Interstitial Pulmonary Fibrosis

The interstitial pulmonary fibrosis of RA has a male predominance (male-to-female ratio, 3 : 1). High titers of rheumatoid factor and the presence of rheumatoid nodules are associated with an increased prevalence of pulmonary fibrosis in RA. Smoking is a risk factor for the development of overt pulmonary fibrosis in RA and has also been associated with subclinical disease. In the presence of HLA-DR susceptibility genes, smoking is strongly associated with development of anti-citrulline-positive RA. It has been suggested that the seropositive RA actually starts in the lungs.

In early disease, a lymphocytic interstitial infiltrate is often the predominant abnormality, and prominent peribronchiolar follicles, containing aggregates of lymphocytes with germinal centers, are often seen. In long-standing disease, fibrosis predominates, often resulting in cystic changes or honeycombing. It is difficult to estimate the exact prevalence of subcategories of interstitial pulmonary fibrosis because, in historical series, histologic descriptions of “interstitial fibrosis” have not been detailed. Both nonspecific interstitial pneumonia and usual interstitial pneumonia are present in significant proportions of patients with interstitial fibrosis ( eFig. 65-5A ). Usual interstitial pneumonia is associated with a better outcome in CTD than in an idiopathic setting. However, in RA, usual interstitial pneumonia is associated with a worse outcome than nonspecific interstitial pneumonia. Furthermore, usual interstitial pneumonia has a worse outcome in RA than in other CTDs.

However, even taking this fact into account, the outcome in RA patients with usual interstitial pneumonitis in the above series was somewhat better than the outcome in idiopathic pulmonary fibrosis, in keeping with trends observed in the histologic series of Park. However, it is also clear that a subset of RA patients with usual interstitial pneumonia have a progressive course similar to that in idiopathic pulmonary fibrosis, especially when HRCT appearances are similar to those in classical idiopathic pulmonary fibrosis (see Figs. 65-3A and 18-23 ), with predominantly basal subpleural honeycomb change and little or no ground-glass attenuation.

In general, surgical lung biopsies are not performed for prognostic indications in CTD. In pulmonary fibrosis in RA, it is not clear that knowledge of the histologic pattern adds usefully to HRCT assessment in prognostic evaluation, although surgical biopsies are performed in some centers. Minor pulmonary fibrosis is common in RA; in one open-lung biopsy series performed in the previous century in volunteers with RA (some without clinical evidence of ILD), pulmonary fibrosis was seen in 60% of patients.

Reductions in D l _co are found in 40% of unselected RA patients. However, radiologically overt pulmonary fibrosis is found in only 1% to 5% of RA patients (based on three large chest radiographic series). In the largest prospective series reported to date, 150 consecutive patients were screened for pulmonary disease. Of these, 19% had CT evidence of interstitial fibrosis, which was often subtle; 43% of those with interstitial abnormalities also had emphysematous bullae. Chest radiography identified abnormalities in only 14% of the whole cohort, but physiologic abnormalities were seen in 82% (gas transfer) and 14% (restrictive pattern of ventilatory defect). Disease severity as judged by CT generally reflects functional impairment.

The most frequent symptom is exertional dyspnea, although this may be masked by a general loss of mobility due to systemic disease. The clinical picture is usually identical to that of idiopathic pulmonary fibrosis, with bilateral, predominantly basal crackles and tachypnea, cyanosis, and right heart failure in advanced disease. Finger clubbing, which is occasionally striking, is more prevalent than in other CTDs. Severe progressive disease requiring hospitalization may be associated with cor pulmonale and respiratory failure and has a poor prognosis, with a 5-year survival rate of less than 50%. A subgroup of patients has more indolent disease that progresses little during prolonged follow-up. A strong predictor of clinical decline is a gas transfer less than 55% of predicted at presentation.

Organizing Pneumonia

Organizing pneumonia is characterized by plugs of granulation tissue in the air spaces distal to and including the terminal bronchioles, associated with lymphocytic infiltration within well-preserved bronchiolar walls and the surrounding lung interstitium. Organizing pneumonia has a very different profile from the entity of bronchiolitis obliterans (also found in RA), with a clinical presentation of pneumonia (as opposed to airflow obstruction), multifocal consolidation on chest radiograph and CT (see Figure 65-3B ), a restrictive functional defect, and a much higher chance of responsiveness to corticosteroids than bronchiolitis obliterans (with a good outcome in 15 of the first 17 reported cases). Organizing pneumonia is more common in RA than in other CTDs (with the exception of inflammatory myopathy and mixed connective tissue disease). In a series of 40 patients with RA undergoing open-lung biopsy, organizing pneumonia (6 cases) had a prevalence similar to that of interstitial fibrosis (5 cases), although organizing pneumonia might have been overrepresented owing to an acute presentation (and thus a perceived need to reach a definitive histologic diagnosis). The good prognosis generally seen in organizing pneumonia has been emphasized in the medical literature, but a minority of RA patients with organizing pneumonia progress to respiratory failure and death despite treatment.

Bronchiolitis Obliterans

Bronchiolitis obliterans (BO) in RA has now been described in numerous case reports and small series. BO (synonymous with obliterative bronchiolitis and constrictive bronchiolitis) is characterized histologically by destruction of the bronchiolar wall by granulation tissue, effacement of the lumen, and eventual replacement of the bronchiole by fibrous tissue. There is circumstantial evidence to suggest that, in some cases, BO may be preceded by an inflammatory exudate; prominent bronchiolar inflammation may be found in patients with the shortest symptomatic course. The expression of HLA antigens B40 and DR1 is increased in BO associated with RA (but not in isolated BO).

In early descriptions, the hallmark of BO was a rapidly progressive, often fatal, course; however, because clinician awareness of the disease was then low, patients with advanced and progressive disease were undoubtedly overrepresented. There is great heterogeneity in the speed of progression, with some patients having indolent disease ; the use of CT in RA patients with suspected pulmonary complications has now identified a subgroup with occult bronchiolitis (see Fig. 65-3C ), often admixed with ILD. The prevalence of unsuspected BO in unselected RA patients remains uncertain, with unexplained airflow obstruction identified in a significant minority (including many nonsmokers) in one study but no increase in the prevalence of pulmonary function abnormalities suggestive of isolated small airway disease in two subsequent controlled series.

Two major associations with the development of BO in RA have been reported. It is likely that secondary Sjögren syndrome is an important predisposing factor, being associated with BO in five of six RA patients in one series ; the spectrum of histologic abnormalities in these cases, ranging from peribronchiolar lymphocytic infiltration to small airway destruction, was analogous to changes seen in the parotid gland in Sjögren syndrome. As the presence or absence of Sjögren syndrome is not documented in many case reports of BO in RA, the etiologic importance of Sjögren syndrome in this context remains uncertain.

More contentious is the reported association between BO and the use of penicillamine, first reported in the late 1970s. After a number of case reports and small series, a significantly higher prevalence of BO was identified in RA patients who took penicillamine (3 of 133) than in other RA patients (0 of 469) in a large cohort. It is possible that the development of BO and the use of penicillamine are both markers of more aggressive RA and are linked for this reason; however, BO developed less than 1 year after penicillamine was begun in 19 of the first 20 cases, and thus the association is unlikely to be entirely spurious. Because BO has been reported in many RA patients not taking penicillamine, it is likely that an underlying predisposition to BO in RA is unmasked by penicillamine (which disrupts collagen linkage and thus interferes with tissue repair). A relationship between BO and gold therapy has been suggested but is not endorsed by recent clinical experience.

Follicular Bronchiolitis

Follicular bronchiolitis (FB) is characterized by external compression of bronchioles by hyperplastic lymphoid follicles, with variable lymphocytic infiltration of the bronchiolar wall ( eFig. 65-5B ). FB is associated more commonly with RA than with other CTDs and is often found incidentally at lung biopsy in RA patients with interstitial pulmonary fibrosis. No causative mechanism has been identified, and it is unclear whether FB predisposes to the subsequent development of BO. When found in isolation, FB simulates ILD, with reticular or nodular opacities on chest radiography and a pattern of functional impairment that may be restrictive or obstructive. Clinically significant isolated FB is rare in RA, but its recognition is important because a response to corticosteroid therapy, although not the rule, is much more likely than in BO; in six of the first nine reported cases, disease stabilized or regressed with treatment. In a CT study of patients with histopathologically proven FB, the most prominent features were centrilobular (see Fig. 65-3D and E ) and peribronchial nodules together with patchy ground-glass increases in attenuation in a bronchocentric distribution.

Bronchiectasis

The prevalence of bronchiectasis is higher in RA than in other CTDs. One literature review identified 289 patients with bronchiectasis associated with RA reported since 1928; however, because respiratory symptoms preceded the systemic manifestations of RA in 90%, it is likely that chance association accounts for a high proportion of early reported cases. Although associated with long-standing RA in one study, on prospective evaluation of 50 RA patients, bronchiectasis was present on 30% of CT scans (see Fig. 65-3F and G ). Bronchiectasis in RA is not associated with more aggressive systemic disease and is often clinically silent, with little or no sputum production and a less progressive and disabling course than in patients with idiopathic bronchiectasis.

Pulmonary Vasculitis

It is surprising that pulmonary vasculitis is reported only rarely in RA, given the relatively high prevalence of systemic vasculitis in the disease. Pulmonary hypertension resulting from pulmonary vascular disease (as opposed to extensive pulmonary fibrosis) is uncommon, although occasional cases of pulmonary vasculitis have been found at autopsy. Similarly, diffuse alveolar hemorrhage is a rare complication of RA.

Pulmonary Rheumatoid Nodules

Pulmonary rheumatoid nodules may be single or multiple, are found on chest radiography in less than 1% of RA patients, and are usually associated with rheumatoid nodules elsewhere in the body. Occasionally pulmonary rheumatoid nodules precede the development of systemic disease. Nodules are circumscribed, with central necrotic material contained by palisading epithelioid cells and surrounded by fibrosis and lymphocytic infiltration ( eFig. 65-5C ). Nodule cavitation occasionally causes hemoptysis, and pneumothorax may result from the rupture of subpleural nodules. Diffuse infiltration by small nodules leading to respiratory failure has been reported. However, nodules generally present as asymptomatic abnormalities on imaging (see Fig. 65-3H-J ) and may vary in size according to underlying rheumatoid activity; thus, when solitary, their growth as judged by chest radiography may simulate malignancy. Caplan syndrome consists of the association of single or multiple nodules with coal workers’ pneumoconiosis, which is often trivial, in keeping with the suggestion that nodule formation results from a hypersensitivity reaction to inhaled coal dust perhaps amplified by immunologic overactivity.

Pleural Disease

Pleural disease is seen at autopsy in approximately 50% of patients, and 20% give a history of pleuritic chest pain ( eFig. 65-5D ). However, pleural effusions are found in less than 5%, usually in men, and are frequently asymptomatic, often being identified on routine chest radiography. In a minority, pleuritic pain and fever are prominent, and the exclusion of empyema (which may be more prevalent in RA) is required. Occasionally, effusions may develop acutely in association with pericarditis or exacerbations of arthritis; more typically, radiographic abnormalities are chronic, often remaining unchanged for years. The fluid is an exudate, with a low glucose level (correlating poorly with serum glucose), a low pH, and usually, a predominant lymphocytosis (although a neutrophilia is occasionally found). Pleural fluid rheumatoid factor levels tend to mirror serum levels and have little independent diagnostic value.

Other Pulmonary Complications

Other pulmonary complications of RA are rare. Lymphocytic interstitial pneumonia is an occasional finding at lung biopsy and responds variably to corticosteroid therapy. Apical fibrosis mimicking the lung disease of ankylosing spondylitis has been reported. Extensive apical cavitation in the absence of nodules or other causes of fibrocavitary disease (including tuberculosis) has been described in a handful of cases and may follow a fulminant course. Pulmonary amyloidosis has been reported occasionally. The incidence of secondary pulmonary hypertension increases with time from diagnosis. Lower respiratory tract infection is increased in frequency in RA; bronchopneumonia is a common terminal event, accounting for 15% to 20% of deaths in RA patients. An increase in lung cancer has been reported in RA patients.

Drug-Induced Pulmonary Disease

Drug-induced pulmonary disease is a particular problem in RA because of the widespread use of methotrexate in routine clinical practice, and lung disease has been reported in 3% to 18% of RA patients treated with methotrexate. Methotrexate pneumonitis is potentially life threatening. It presents with cough, dyspnea, fever, widespread crackles, and pulmonary opacities on chest radiography and CT (see eFig. 71-4 ), which may be focal or diffuse. Although the presentation is sometimes explosive, more often, the onset is subacute (with symptoms evolving for up to 2 months before diagnosis) ; 50% of cases are diagnosed within 4 months of initiation of methotrexate therapy. Because the clinical and radiographic features are nonspecific, methotrexate-induced lung disease should always be suspected in the treated patient presenting with progressive lung disease. Unfortunately, histologic findings are nonspecific, although finding a prominent lymphocytic infiltration increases the likelihood of methotrexate pneumonitis. There is conflicting evidence on whether preexisting lung disease predisposes to lung injury caused by methotrexate, but the published evidence does not suggest that functional impairment is an absolute contraindication to its use (although particular caution is warranted when pulmonary reserve is grossly compromised, and patients with previous methotrexate toxicity should not be retreated). Pulmonary methotrexate toxicity is associated with a mortality rate of 15% to 20% ; because the lymphocytic component of disease is wholly or partially reversible, immediate withdrawal of methotrexate and the early institution of steroid therapy are warranted.

Pulmonary toxicity in RA has also been documented with sulfasalazine, gold therapy, and penicillamine (discussed earlier; see Chapter 71 ). Pulmonary opacities (due to organizing pneumonia) associated with sulfasalazine most commonly present in the upper lobes; this side effect is rare in RA, with only a handful of cases reported (most cases were reported in ulcerative colitis). Pulmonary disease induced by gold takes the form of alveolar opacities adjacent to bronchovascular bundles, best demonstrated by high-resolution CT, and often associated with fever or skin rash, relatively low rheumatoid factor titers, and a BAL lymphocytosis. In most patients with sulfasalazine or gold toxicity, lung disease largely regresses with withdrawal of the agent and corticosteroid therapy.

There have been a number of reports of rapidly progressive ILD associated with anti-TNF-α therapies, including etanercept and infliximab, although there are divergent views on whether these reports are truly indicative of drug toxicity or reflect the selective use of these therapies in more aggressive RA. The range of reported toxicity ranges from infection (with particular focus on reactivation of tuberculosis), undoubtedly a true association, to interstitial disease and vasculitis. Leflunomide has also been associated with pulmonary rheumatoid nodules and interstitial disease.

Acute Exacerbations of Interstitial Lung Disease

Acute exacerbations of ILD, representing the development of diffuse alveolar damage, are less prevalent in CTD than in idiopathic pulmonary fibrosis. In both settings, presenting features include worsening dyspnea over 2 to 4 weeks, new ground-glass opacities on HRCT ( eFig. 65-6 ), and the absence of infection or other overt causes of decline. Acute exacerbations are more common in RA than in other CTDs (see eFig. 65-6 ), with a high early mortality. Triggers of acute exacerbations have not been identified in RA and there is currently no proven therapy, although high-dose corticosteroid therapy is usually given.