Sarcoidosis is a systemic inflammatory disease with a predilection for the respiratory system. Although most patients enter remission and have good long-term outcomes, up to 20% develop fibrotic lung disease, whereby granulomatous inflammation evolves to pulmonary fibrosis. There are several radiographic patterns of pulmonary fibrosis in sarcoidosis; bronchial distortion is common, and other patterns, including honeycombing, are variably observed. The development of pulmonary fibrosis is associated with significant morbidity and can be fatal. Dyspnea, cough, and hypoxemia are frequent clinical manifestations. Pulmonary function testing often demonstrates restriction from parenchymal involvement, although airflow obstruction from airway-centric fibrosis is also recognized. Complications of fibrotic pulmonary sarcoidosis include pulmonary hypertension from capillary obliteration and chronic aspergillus disease, with hemoptysis a common and potentially life-threatening manifestation. Immunosuppression is not always indicated in end-stage sarcoidosis. Lung transplantation should be considered for patients with severe fibrotic pulmonary sarcoidosis, as mortality is high in these patients.
Sarcoidosis is a systemic inflammatory disease of unknown etiology. It can affect the eyes, skin, liver, heart, and central nervous system, although the extent of extrathoracic disease is highly variable among patients. In contrast, nearly all patients have some degree of intrathoracic disease. Historically, chest X-ray findings have been used to stage thoracic involvement (Table 1). Most patients with pulmonary sarcoidosis undergo clinical remission with minimal residual organ impairment and favorable long-term outcomes. However, up to 20% develop pulmonary fibrosis as a response to inflammation. Morbidity and mortality are substantially increased for these patients. It is not clear that the risk of developing pulmonary fibrosis, or stage IV disease, varies by ancestry. Although African-American patients were less likely than white patients to present with stage I disease, rates of combined stage III and IV disease were similar between these groups.
Our understanding of the pathophysiology of sarcoidosis is largely derived from studies in acute disease, wherein cell-mediated responses predominate. The oligoclonal expansion of T lymphocytes suggests that sarcoidosis is antigen driven. Epidemiologic data and the high rate of thoracic disease suggest that the antigen, which remains unidentified, is inhaled. Resident antigen-presenting cells traffic to regional lymph nodes and present to circulating CD4+ lymphocytes. Those with a T-cell receptor capable of recognizing antigen become activated and proliferate. Via chemokine signaling, these lymphocytes are recruited to the initial site of inflammation and, polarized to a Th1 phenotype, release cytokines, including IFN-γ, which drive the activation and organization of macrophages. In conjunction with tumor necrosis factor (TNF)-α signaling, these cell-mediated responses result in the formation of granulomas, the histopathologic hallmark of disease. Granulomatous inflammation occurs primarily along lymphatic tracks, which course along bronchovascular bundles and through interlobular septa.
In sarcoidosis, small foci of fibrosis around granulomas can be an expected finding, and small areas of macroscopic fibrosis may occur at sites of healed inflammatory lesions. However, an exuberant fibrotic response is pathologic and leads to frank tissue destruction evident on imaging or gross pathology. We designate this phenotype fibrotic pulmonary sarcoidosis. Even when the extent of fibrosis is anatomically limited and pulmonary function remains normal, a pathologic fibrotic response results in some degree of permanent alteration of pulmonary architecture. Although these alterations meet the definition for radiographic stage IV disease, for the purpose of this review we use the term fibrotic pulmonary sarcoidosis, as it is descriptive, applies also to CT imaging and histopathology, and avoids the implications of “staging.” Although the pathophysiology of fibrotic pulmonary sarcoidosis remains poorly understood, augmented transforming growth factor-β activity, macrophage phenotype switching, and a Th1 to Th2 transition may be important features (11–14). Not all patients with severe or long-standing sarcoid inflammation develop extensive fibrosis. Antigen features, the degree of in vivo antigen containment, and the underlying immunogenotype may drive engagement of a fibrosis program. The exuberance of the response in fibrotic pulmonary sarcoidosis suggests that this program is dysregulated. However, unlike in idiopathic pulmonary fibrosis (IPF), in sarcoidosis there is little evidence to suggest that fibrotic activity, once initiated, continues inexorably.
Fibrotic sarcoidosis can be referred to as “burnt-out” disease, as granulomatous inflammation may be absent on histopathology, and treatment does not reliably result in clinical improvement. Even when there is overlap of active inflammation and end-stage fibrosis, the presence of extensive pulmonary fibrosis is a relevant distinction to make among patients with so-called chronic sarcoidosis. Compared with patients with nonfibrotic chronically active disease, the potential complications and long-term outcomes are worse for those with fibrotic sarcoidosis. We herein review the clinical features, outcomes, and management of this relatively uncommon yet potentially devastating phenotype.
Thoracic imaging reveals a variety of findings in fibrotic sarcoidosis. Volume loss and architectural distortion from extensive fibrosis are often easily recognized on chest radiographs. However, patterns of pulmonary fibrosis vary among patients and are best appreciated by computed tomography (CT) imaging. Abehsera and colleagues evaluated the CT scans of 80 patients with fibrotic sarcoidosis and identified bronchial distortion, linear scarring, and honeycombing as main fibrosis patterns (Figures 1 and 2). Bronchial distortion, marked by traction bronchiectasis and airway angulation, was the predominant pattern in 40% of cases; honeycombing was observed as the predominant pattern in 26% and linear fibrosis in 14% of cases. These radiographic patterns may be clinically relevant. The FEV1 was most reduced in bronchial distortion, whereas the total lung capacity and diffusing capacity were most reduced with honeycomb fibrosis (15). However, it is also worth noting that in 20% of cases there was a discrepant read of the predominant pattern by the two observers, and many cases demonstrated an overlap of patterns.
Unlike the rather well-defined features of fibrosis in IPF, the extent and type of fibrosis in sarcoidosis can be highly variable, which likely accounts for the variety of associated clinical findings and outcomes. Although fibrosis is often upper- and middle-lobe predominant (Figure 1A), the distribution of fibrosis also may be somewhat pattern specific; honeycombing was observed more frequently in the upper lobes, and lower lung involvement was more common for diffuse linear fibrosis. Finally, associations between early imaging findings and long-term fibrosis have been identified. Conglomerate peribronchial masses can, on healing, lead to bronchial distortion, and ground-glass opacities can progress to fine honeycombing on serial imaging. Regular follow-up of patients with ground-glass opacities may be warranted, even when there are minimal pulmonary symptoms initially.
The meaning of the term honeycombing has evolved since it was first introduced in the 19th century to describe bronchiectatic and congenital cystic lesions (19). The concept of honeycombing is now more closely associated with the peripheral clusters of small cysts observed in usual interstitial pneumonia (UIP). These cysts are the result of alveolar interstitial fibrosis, which results in parenchymal obliteration and residual irregular airspaces lined by bronchial epithelium. In sarcoidosis, however, honeycombing often is used to describe large and more central cysts (Figure 2). These features are consistent with the loss of terminal airway segments and suggest that the pathophysiology extends beyond alveolar interstitial damage. In sarcoidosis, the prevalence of UIP-type honeycombing versus large cyst honeycombing is not known. Clinical implications may vary accordingly.
Fibrosis in sarcoidosis extends from granulomas and therefore generally occurs in the same lymphatic distribution as active inflammation. This distribution accounts for many of the radiographic and clinical findings. Fibrosis along bronchovascular tracks results in bronchial distortion and large cystic changes, and interlobular septal fibrosis results in linear scarring (21). Less is known about the histopathology of long linear bands and small honeycomb cysts; presumably, granulomas still function as the nidus of fibrosis for these as well. Alveolar interstitial inflammation and fibrosis are uncommon in sarcoidosis.
Overall, the microscopic features of fibrotic sarcoidosis are distinct from those of UIP. It may be that fibroblast foci occur, albeit rarely, in fibrotic sarcoidosis. However, other histopathologic features of UIP, such as prominent bronchiolar honeycombing and fibroblast foci at so-called transition zones, are not typical features of sarcoidosis. Descriptive studies of large cohorts of representative cases will be helpful to further delineate the histopathologic features of fibrotic sarcoidosis.
In sarcoidosis, dyspnea, cough, and wheezing are common in patients with extensive pulmonary fibrosis. While dyspnea and cough can occur in fibrotic and nonfibrotic pulmonary sarcoidosis, wheezing is more common in those with fibrosis, where it is attributed to airway-centric fibrosis. Although bronchial hyperreactivity is not uncommon in sarcoidosis, its prevalence in fibrotic disease, and its role in wheezing, are not known. Chronic purulent sputum production suggestive of bronchiectasis is a variable finding. Exertional hypoxemia and decreased exercise capacity are more common. Hypoxemia may be profound in advanced disease. Respiratory failure is an uncommon but significant finding. However, in contrast to IPF, acute declines from acute exacerbations of disease attributed to diffuse alveolar damage have not been reported in fibrotic pulmonary sarcoidosis.
On pulmonary function testing, restriction and a reduction in diffusing capacity are the most common abnormalities in patients with fibrotic pulmonary sarcoidosis. Obstruction in sarcoidosis can be multifactorial, but it is more likely to occur in fibrotic lung disease. Although granulomatous inflammation of the small airways and air trapping can be observed in sarcoidosis, the physiologic relevance is unclear. In one study, no correlation between air trapping on CT imaging and obstruction was observed. Rather, results of prior studies suggest that fibrosis is likely an important mediator of obstruction. Hansell and colleagues identified the extent of reticulation rather than air trapping as the stronger radiographic predictor of airflow obstruction. Reticulation included honeycombing and was used as an index of fibrosis. In another study of chest X-ray staging, obstruction was significantly associated with stage IV disease. Obstruction in sarcoidosis tends to be irreversible, although this has been primarily assessed in patients with nonfibrotic disease.
The degree to which pulmonary function declines or remains stable over time is not well established for fibrotic pulmonary sarcoidosis. In one study of patients with stage IV disease, pulmonary function remained relatively stable over long-term (1–20 yr) follow-up. In 39% of patients, pulmonary function actually improved. This is consistent with improvement in inflammation after a period of overlapping inflammation and fibrosis. The relationship between pulmonary function and outcomes in sarcoidosis may differ from that in IPF. In a review of patients awaiting transplant, those with sarcoidosis had a significantly lower FVC and FEV1 yet required less supplemental oxygen compared with those with IPF. Notably, functional limitation and survival were similar between the groups.
A variety of mechanisms can cause pulmonary hypertension (PH) in sarcoidosis (Table 2). Among these, capillary obliteration secondary to fibrosis is particularly important and most, though not all, cases of PH in sarcoidosis occur in the setting of fibrotic lung disease. In addition to capillary obliteration, the degree to which parenchymal distortion negatively affects vascular mechanics is not known, but altered flow and shear stress may also play a role in PH associated with fibrotic sarcoidosis. Granulomatous angiitis is another important cause of PH in sarcoidosis. Pulmonary veins are at particular risk, given their location in interlobular septa, and comparisons have been made between granulomatous angiitis and pulmonary venoocclusive disease. In sarcoidosis, the presence of ground-glass opacities may distinguish between fibrotic and nonfibrotic etiologies of PH; in one study these opacities were more common in patients with nonfibrotic disease (21).
PH in sarcoidosis can cause exertional dyspnea, chest pain, and syncope. Many experts recommend assessing for PH in patients with dyspnea out of proportion to underlying lung disease or when a reduction in the diffusing capacity is disproportionate to reductions in other pulmonary function indices. An echocardiogram is often the initial evaluation, although a regurgitant tricuspid jet that is inadequate to characterize pulmonary pressures is not uncommon. Although pulmonary artery pressures by echocardiography and right heart catheterization have been shown in several studies to correlate with each other, over- and underestimations of true pulmonary pressures are common (21, 35). In the setting of vascular inflammation, PH may be at least partially reversible with immunosuppression. Prostacyclin analogs and other vasoactive agents may be helpful but should be instituted by a clinician expert in the evaluation and treatment of PH in sarcoidosis.
Although Aspergillus spores are ubiquitous in the environment, pulmonary disease develops primarily in the setting of damaged lungs or compromised immune function. Aspergillomas and progressive aspergillosis syndromes have been observed in fibrotic pulmonary sarcoidosis (Table 2). Simple aspergillomas are often clinically stable over time, although hemoptysis is an important complication (38). Locally progressive aspergillosis has been variably called chronic necrotizing disease or chronic cavitary pulmonary aspergillosis. It has also been referred to as semiinvasive disease, although reports documenting parenchymal invasion are scant. More important than the terminology is the fact that these conditions often go unrecognized and place patients at risk for poor outcomes, including life-threatening hemoptysis, a wasting syndrome, and progressive respiratory insufficiency. They also affect considerations for lung transplantation, as described below. In general, a compromised immune system is a recognized risk factor for progressive aspergillosis. For sarcoidosis in particular, the role of systemic corticosteroids, where use is common among patients who develop aspergillosis, is not known.
The diagnosis of a simple aspergilloma can often be made on imaging. Progressive aspergillosis is diagnosed by the combination of radiographic and clinical findings. The diagnosis is supported by recovery of fungus from lung lavage or biopsy specimens or by repeatedly positive sputum cultures. Testing for Aspergillus antibodies is frequently positive. The erythrocyte sedimentation rate and C-reactive protein are often quite elevated.
Treatment recommendations for pulmonary aspergillus disease are derived from case reports and expert opinion statements . For simple aspergillomas, local installation of antifungal antibiotics may be effective in symptomatic cases (46). Systemic antifungal agents are not often indicated. In contrast, for progressive aspergillosis, systemic antifungal therapy can help control disease, although cure is not often achieved (Figure 3).
Bronchial artery embolization can terminate life-threatening bleeding; when indicated, it should be performed by an experienced radiologist. Surgical resection may be required in refractory hemoptysis, although poor lung function and progressive aspergillosis are associated with postoperative hemorrhage, bronchopleural fistulas, pleural infection, and prolonged respiratory insufficiency. Treatment with an oral azole antifungal agent before surgery may decrease the incidence of complications, including fungal contamination of the chest cavity. In our opinion, caution is indicated regarding immunosuppression for sarcoidosis in the setting of progressive aspergillosis.
There are few published data on the natural history of fibrotic sarcoidosis, including rates of exacerbations. Exacerbations of sarcoidosis are due to an acute-on-chronic increase in disease activity or to the recurrence of disease in patients previously in remission or with stable end-stage disease. In a recent review, it was noted that exacerbations of sarcoidosis are not uncommon; however, for long-standing fibrotic disease in particular, exacerbation rates remain largely undefined (51).
Subjective reports and objective changes in lung function contribute to the assessment, and changes of at least 10% in the FVC or FEV1 have been proposed as criteria for an exacerbation of sarcoidosis. Chest radiograph profusion scores were not associated with exacerbations in a study of patients with various stages of sarcoidosis; whether CT imaging is helpful to diagnose exacerbations in fibrotic disease has not been evaluated. When angiotensin-converting enzyme levels or other inflammatory markers correlate with disease activity, they can help identify the recurrence of active sarcoidosis. For the majority of patients who lack trending biomarkers, we rely on the overall clinical context, with particular attention to the cardiac status, to guide our assessment of increased pulmonary symptoms.
It can be difficult to distinguish a recurrence of sarcoid activity from a complication of underlying fibrosis or a superimposed condition. In our experience, left-sided congestive heart failure is not uncommon in patients with end-stage sarcoidosis and can be an important contributor to acute respiratory declines. Coronary artery disease and hypertensive cardiomyopathy are highly prevalent, and therapies for these conditions, if present, often result in clinical improvement. A low threshold of suspicion to evaluate for cardiac sarcoidosis in patients with signs or symptoms of cardiac dysfunction is prudent. Right heart dysfunction from PH can also cause clinical decline. In addition, new data suggest that the incidence of pulmonary embolism may be increased in sarcoidosis. These conditions and other events should be excluded in patients with a declining pulmonary status and are summarized in Table 3.
As noted above, of the serious complications of fibrotic pulmonary sarcoidosis, parenchymal fibrosis with restrictive physiology, airway-centered fibrosis with airflow obstruction, and PH are the most common and reflect the predilection for disease along the bronchovascular and other lymphatic regions (Table 2). Less common complications include central airway involvement with upper airway obstruction, bullous emphysema, pneumothorax, and pleural thickening. Pulmonary complications in sarcoidosis unrelated to fibrosis include pleural effusion, chylothorax, pulmonary embolism, and PH from mechanisms other than capillary obliteration.
Although mortality rates are low among all patients with sarcoidosis, mortality is higher for those with pulmonary fibrosis. In a recent retrospective review from France, a 10-year mortality rate of 16% in patients with fibrotic sarcoidosis was significantly higher than that projected for the general population of a similar age. The extent and stability of fibrosis and the development of secondary complications are likely important determinants of survival in fibrotic sarcoidosis. Once patients are sick enough to be listed for lung transplant, mortality rates are high and similar to those for patients listed for IPF.
Important questions regarding immunosuppression and pulmonary fibrosis remain unanswered. One of the most important is whether immunosuppression prevents fibrosis. In numerous trials, the use of systemic corticosteroids in sarcoidosis—to treat symptoms or used empirically in asymptomatic disease—has not been shown to prevent the development or halt the progression of pulmonary fibrosis. Data on nonsteroidal immunosuppressives for fibrotic pulmonary sarcoidosis are largely lacking. However, in a phase 2 study of infliximab, a TNF-α inhibitor, no change in the fibrosis score in patients with chronic pulmonary sarcoidosis was observed at follow-up. Finally, although there are important differences between IPF and fibrotic pulmonary sarcoidosis, the potential exists for antifibrotic agents currently under investigation for IPF to be evaluated in sarcoidosis.
Pulmonary fibrosis is an irreversible event. However, in at least some patients, fibrosis coexists with active inflammation. In these cases, immunosuppression can lead to some measurable improvement in clinical status. Although treatment is not known to prevent ongoing fibrosis, a reduction in superimposed inflammation may be clinically meaningful. Unfortunately, we lack a valid assessment model for disease activity to determine which patients may benefit from treatment. As reviewed, when serum angiotensin-converting enzyme or other inflammatory markers track with disease activity, their levels can aid in therapy decisions.
In established fibrosis, CT imaging can be helpful to assess for coexisting active inflammation. Clusters of small nodules and interstitial or peribronchial opacities without architectural distortion suggest active sarcoidosis. As a complement to CT imaging, positron emission tomography (PET) scanning may help qualify active inflammation in fibrotic disease. Mostard and colleagues recently reported increased uptake on PET scanning in patients with underlying fibrosis. A correlation of these findings with those of concomitant CT imaging was not reported. Further studies of PET scanning in patients who lack signs of active inflammation on CT scan will provide insight into the presence of occult inflammation in those who appear to have only burnt-out disease.
We favor immunosuppression for those with biomarker or radiographic findings suggestive of active inflammation (Table 3). Often, however, a progressive clinical decline alone prompts a trial of therapy. Having specific treatment endpoints in mind assists decision making at follow-up regarding continuation of an empiric trial of immunosuppression. The risks associated with immunosuppression are not trivial, particularly in patients who have substantial comorbidities (64). In addition to ongoing surveillance of medication tolerability, several clinical parameters require monitoring (Table 4).
Severe cough, when present, impairs quality of life and is often refractory to inhalers and immunosuppressive therapies. Cough regimens containing codeine may be helpful. Bronchiectasis in sarcoidosis can be associated with recurrent infection, hemoptysis, and chronic sputum production (28, 65). Antibiotic use aimed at airway microbes demonstrated on sputum culture has led to clinical improvement in patients who have both bronchiectasis and a decline in pulmonary status (28). In addition, although not formally evaluated in sarcoidosis, airway clearance maneuvers, including scheduled nebulizer treatments and use of an airway clearance device, may be helpful for those with chronic sputum and associated airway symptoms.
Lung transplantation is an important therapeutic option for patients with fibrotic pulmonary sarcoidosis given the increased mortality, with 75% of deaths related to respiratory causes. Lung transplantation should be considered in patients with severe or progressive respiratory limitation from fibrotic pulmonary or pulmonary vascular disease. The optimal timing of transplantation can be difficult to gauge, as models predicting mortality in patients with sarcoidosis have not been validated, algorithms that predict mortality in other forms of pulmonary fibrosis do not apply, and decrements in lung function may not correlate directly with risk of death.
Historically, mortality among patients with sarcoidosis listed for lung transplant has been high. A retrospective analysis of the United Network for Organ Sharing transplant database in the late 1990s showed that mortality was similar to patients with IPF. Yet patients with sarcoidosis were less likely to receive lung transplantation despite long-term outcomes comparable to those for patients undergoing transplantation for other pulmonary diseases.
In separate studies using the same database, factors associated with mortality in patients with sarcoidosis awaiting transplant included African American race, PH, and oxygen use (67, 68). A single-center retrospective study of 43 patients with sarcoidosis listed for lung transplantation showed that hypoxemia, PH, decreased cardiac index, and elevated right atrial pressure were associated with death while on the waiting list. After multivariate analysis, only right atrial pressure greater than or equal to 15 mm Hg (relative risk, 5.2; 95% confidence interval, 1.6–16.7) remained an independent prognostic variable. Importantly, most of the data on lung transplantation in sarcoidosis are retrospective and were generated before implementation of the lung allocation scoring system in 2005. Although we await studies of outcomes using the new scoring system, patients with fibrotic pulmonary sarcoidosis should be considered for a lung transplant evaluation if they require oxygen, have PH, or have a progressive decline in lung function. Beyond a patient’s respiratory status, other situations unique to sarcoidosis need to be considered before listing. Extrapulmonary involvement, such as serious neurosarcoidosis, may preclude lung transplantation, whereas cardiac or liver involvement may require consideration of an additional organ transplant. The presence of suppurative bronchiectasis may be an indication for double lung transplant, as is done in patients with cystic fibrosis. Aspergillus pulmonary disease should prompt consideration of double lung transplant, as involvement may be bilateral. Use of antifungal therapy pre- and post-transplant may improve outcomes. Successful lung transplant of a patient with bilateral mycetomas has been reported.
Among patients who undergo lung transplant, there does not seem to be an increased risk of bronchiolitis obliterans syndrome for those with sarcoidosis. However, sarcoidosis was found to be a risk factor for primary graft dysfunction in a recently published large multicenter study of data prospectively collected from 2002 to 2010 using a fully adjusted multivariate model (odds ratio, 2.5; 95% confidence interval, 1.1–5.6). Recurrence of sarcoidosis post-transplant is described but does not affect outcomes.
Mechanisms of fibrosis in sarcoidosis are largely unknown and understudied. In particular, we do not know if a fibrosis cascade is activated at disease onset in predisposed individuals or if it evolves as a wound-healing response to poorly controlled inflammation. Knowing this has therapeutic implications. Advancing our understanding of the pathophysiology of fibrosis in sarcoidosis is therefore essential and will complement what we know about the general mechanisms of disease. Fibrotic sarcoidosis is not a common phenotype, but it is an important one. Although life expectancy is longer in fibrotic sarcoidosis compared with IPF, once patients are sick enough to be listed for transplant, outcomes for both are similar: end-stage lung disease is end-stage lung disease.