Smoking and Interstitial Lung Disease

Cigarette smoking is now widely accepted as the primary cause of certain ILDs, namely respiratory bronchiolitis–interstitial lung disease (RB-ILD), desquamative interstitial pneumonia (DIP), and pulmonary Langerhans cell histiocytosis (PLCH). Cigarette smoking is also a risk factor for the development of idiopathic pulmonary fibrosis (IPF) and rheumatoid arthritis (RA)–associated ILD, and has been reported to cause some cases of acute eosinophilic pneumonia (AEP) and pulmonary hemorrhage syndromes. Paradoxically, cigarette smoking may confer protection from developing some other ILDs such as hypersensitivity pneumonitis (HP). We have described a classification scheme ( Box 3.1 ) outlining these subgroups and their relationship to smoking. This classification illustrates the highly complex effects of smoking on the lung parenchyma.

The Group 1 diseases ( Box 3.1 ) include the three diffuse lung diseases classically regarded as smoking-related ILDs. This designation is supported by several lines of clinical, epidemiologic, and investigative evidence showing a direct role for cigarette smoking as witnessed in the temporal relationship to disease onset and progression, resolution on smoking cessation, and recurrence or progression on resumption of smoking. Several case series have reported a history of smoking in the overwhelming majority of Group 1 patients, with the prevalence being highest in RB-ILD and PLCH, and less common in DIP ( Table 3.1 ). The reported coexistence of all three lesions in the same patient, the potential for disease remission with smoking cessation, the recurrence of disease in transplanted lungs, and the description of analogous lesions in mice exposed to high doses of cigarette smoke all provide support to the designation of RB-ILD, DIP, and PLCH as smoking-induced ILDs.

For diseases allocated to Group 2 ( Box 3.1 ) the association with cigarette smoking is less robust than for Group 1 diseases. Cigarette smoking, particularly during the relatively early phase after initiation of smoking, seems to be an important precipitating factor in some but not in all cases of Group 2 diseases. The most relevant conditions in this category include AEP and certain pulmonary hemorrhage syndromes. AEP deserves particular attention because a number of recent studies have implicated recent-onset exposure to cigarette smoke as a principal inducer of this disease in some patients diagnosed to have this disorder. In addition, increase in the number of cigarettes smoked per day in chronic smokers has been described to induce AEP. Of particular interest is the response of certain subjects with resolved AEP to a rechallenge with cigarette smoke exposure that triggers peripheral eosinophilia and other associated pathophysiologic abnormalities, suggesting that exposure to cigarette smoke to induce certain responses is relevant to the development of acute diffuse lung disease in susceptible hosts.

Diseases included in Group 3 ( Box 3.1 ) are chronic diffuse lung diseases that are statistically more likely to develop in cigarette smokers. For instance, cigarette smoking is known to increase the relative risk of RA-associated ILD, possibly by triggering RA-specific immune reactions to citrullinated proteins. Similarly, smokers have a higher risk of developing IPF than nonsmokers. The precise significance of these observations has been a topic of substantial debate, but there is limited evidence that smoking itself is directly fibrogenic to the lung. It is not appropriate to consider smoking as an inducer of these diseases, but rather a disease modifier or potentially a cofactor that facilitates the development of profibrotic responses that lead to these diffuse fibrotic lung diseases.

The fourth and final group consists of diseases that are less prevalent in smokers compared with nonsmokers and includes sarcoidosis and HP. Cigarette smoking seems to provide certain “protective” effects that diminish the potential development of these granulomatous inflammatory lung diseases, possibly by inhibiting certain immunologic responses in the lung that are required for granuloma formation or the development of T-helper 1 (Th1)–polarized immune responses following exposure to inhaled antigens. Epidemiologic studies demonstrate that levels of circulating IgG antibodies to pigeon antigens are higher among nonsmokers than smokers. A similar study in farmers showed that nonsmokers and previous smokers had a higher prevalence of serum precipitin levels to various farmer’s lung antigens, compared with current smokers. Lung macrophages from cigarette smokers also have lower levels of costimulatory molecules than those of controls. Because costimulatory molecules play a critical role in shaping the immune response to inhaled antigens, it is possible that smokers are hyporesponsive to inhaled antigens by virtue of diminished antigen-presenting capacity in the lung. Cigarette smoking and nicotine have also been demonstrated to inhibit the production of the potent Th1-polarizing cytokine interleukin-12 (IL-12). It is conceivable that the diminished capacity of smokers’ macrophages and dendritic cells to generate IL-12 may impede the development of hypersensitivity response to inhaled antigens and granuloma formation in the context of sarcoidosis. The observation that smoking is associated with a lower prevalence of sarcoidosis and HP should not be construed as an indication to promote smoking in patients with these diseases. On the contrary, insight gained from dissecting mechanisms by which smoking suppresses Th1 immunity—an essential driver of the immunopathogenic processes that characterizes these diffuse lung diseases—is also relevant to the pathogenesis of smoking-related lung cancer and airway diseases, diseases that are more prevalent in smokers partly because of impaired Th1 immunity.

The fact that some cases of RB-ILD or DIP may be induced by factors other than cigarette smoke exposure, and that some patients with PLCH are nonsmokers, had been interpreted as implying that these diseases do not necessarily represent specific smoking-induced lung diseases. However, it is well recognized that a number of specific histopathologic entities can be induced by heterogeneous etiologies, potentially a reflection that the lung has only a limited number of ways of responding to various insults. For example, the lesion of usual interstitial pneumonia (UIP) may be induced by asbestos exposure and be seen in patients with chronic HP, as well as in the context of autoimmune diseases such as RA-associated ILD. Cigarette smoking is the most well-defined etiologic factor associated with the development of RB-ILD, DIP, and PLCH; however, the histopathologic lesions of RB, DIP, and PLCH do not exclusively occur in smokers and may occasionally be idiopathic or encountered in the context of other exposures or etiologies.

Defining the relationship between smoking and specific ILDs has important clinical implications. Smoking cessation is imperative for all the diseases listed under Groups 1–3 in Box 3.1 . It is our practice to use aggressive tobacco cessation strategies in these patients and have a low threshold for referral to nicotine dependence counselors. It is our practice to explicitly refer to diseases in Group 1 as “smoking-induced” to underscore the importance of smoking cessation and encourage removal of all tobacco products from the vicinity of the patient, including second-hand tobacco smoke exposure. Similarly, all current smokers with diseases in Groups 2 and 3 should be counseled regarding the emerging and compelling data implicating a direct pathogenic role for cigarette smoke exposure as a potential inducer or cofactor in disease induction and progression. The methods that should be considered in smoking cessation therapy include counseling and behavior therapy, nicotine replacement therapy, and pharmacotherapy, including the use of bupropion, varenicline, and clonidine in selected patients.

Mechanisms by Which Tobacco Smoke May Promote Interstitial Lung Disease

Even in smokers without clinically detectable lung disease, cigarette smoking induces inflammatory cell recruitment, consisting primarily of macrophages, neutrophils, and Langerhans cells (a subtype of the myeloid dendritic cell family expressing surface CD1a receptors), to small airways. Although all smokers have some degree of inflammation in the airways, only a small minority of smokers develop clinically significant diffuse lung disease. The relative rarity of smoking-related ILDs compared with the overall prevalence of cigarette smoking suggests that cigarette smoke is not the only factor responsible for the induction of these diseases and implies that additional factors (endogenous such as genetic factors or exogenous such as infectious pathogens or allergens) are required for the induction of disease.

A characteristic morphologic feature of all Group 1 smoking-related ILDs is prominent bronchiolar inflammation. In addition, Group 1 diseases demonstrate increased macrophages in the interstitium and the alveolar spaces. Pigmented macrophage accumulation in small airways, interstitium, and distal airspaces is a key feature of many smoking-related ILDs. Specific mechanisms by which exaggerated macrophage accumulation occurs in Group 1 diseases are not fully defined, but likely involve exaggerated generation of macrophage recruiting and differentiating factors by airway epithelial cells, enhanced macrophage survival locally, and/or diminished apoptosis of recruited macrophages. In these patients, lung epithelial cells have been demonstrated to aberrantly produce excessive granulocyte-macrophage colony stimulating factor (GM-CSF), a cytokine that provides proliferative and activation signals to both macrophages and dendritic cells. Cigarette smoke extracts have also been shown to induce transforming growth factor beta (TGF-β) production by lung epithelial cells, a cytokine that is involved in Langerhans cell development, immune modulation, and fibrogenic responses in the airways.

Cigarette smoking induces several abnormalities in immune and other lung cells that are likely relevant to the pathogenesis of smoking-related ILDs. Certain constituents in cigarette smoke are known to activate epithelial cells, macrophages, neutrophils, and dendritic cells in vitro, promoting generation of chemokines and cytokines that lead to inflammation by promoting immune cell recruitment. It is reasonable to speculate that smokers in whom ILD develops have an amplified inflammatory cascade associated with activation of multiple immune cell types that promote a vicious cycle of inflammatory cell recruitment. Whether failure of endogenous antiinflammatory mechanisms or additional exogenous insults such as viral infections have a role in promoting smoking-related ILDs is unknown but should be an important area of future research.

Respiratory Bronchiolitis-Associated Interstitial Lung Disease

Niewoehner described RB as a histopathologic finding of pigmented macrophage accumulation centering on respiratory bronchioles and neighboring alveoli, a finding that was ubiquitous in cigarette smokers at autopsy. Subsequent case series described similar findings on lung biopsy specimens from cigarette smokers. RB can thus be considered a histologic marker of smoking and must be distinguished from RB-ILD, a term coined by Myers and colleagues to recognize the clinicopathologic ILD occurring in cigarette smokers in whom surgical lung biopsy revealed only RB. In patients with RB-ILD, the lesion of RB is not felt to be a mere indicator of exposure to smoking, but rather constitutes the primary and only histopathologic lesion accountable for the observed ILD. Following the original description by Myers and colleagues, other reports described in greater detail clinical and radiologic features of RB-ILD as a specific interstitial and bronchiolar process occurring in smokers and defined by the presence of RB as the only definable pathologic abnormality present on lung biopsy.

The true prevalence of RB-ILD is difficult to estimate as many patients with this disorder may be asymptomatic. The duration of exposure to cigarette smoke need not be lengthy or severe, although many have substantial cumulative tobacco exposures. Most patients present in the fourth and fifth decade of life, and there is no gender predilection. A clinicopathologic syndrome indistinguishable from RB-ILD can occasionally be encountered following exposure to solder fumes, diesel smoke, and fiberglass.

RB-ILD usually presents in a nonspecific fashion with chronic cough and exertional dyspnea; rarely, acute presentation may occur ( Table 3.1 ). The physical examination reveals inspiratory crackles in approximately half of patients, but digital clubbing is infrequent. Pulmonary function testing yields various patterns, including normal, obstructive, restrictive, or mixed abnormalities. The severity of physiologic impairment, if present, is usually mild to moderate.

Chest radiography reveals bilateral, fine reticular, or reticulonodular opacities in about 60–70% of patients but may appear normal in some patients. The main findings on chest high-resolution CT (HRCT) include bronchial wall thickening, fine centrilobular nodules, and patchy areas of ground-glass attenuation. The ground-glass changes are typically bilateral and affect both upper and lower lung fields ( Fig. 3.1 ). Coexisting emphysematous changes are frequently noted, but honeycombing, traction bronchiectasis, and parenchymal fibrosis are not.

High-resolution CT of the chest showing patchy areas of ground-glass attenuation in upper lung fields in a smoker with respiratory bronchiolitis–associated interstitial lung disease.

The differential diagnosis of RB-ILD includes not only consideration of other bronchiolar diseases, including infectious bronchiolitis, follicular bronchiolitis, and diffuse aspiration bronchiolitis, but also ILDs characterized by ground-glass opacities, particularly HP and nonspecific interstitial pneumonia (NSIP). Although surgical lung biopsy is often required for a definitive diagnosis, in clinical practice a provisional diagnosis may be established in many patients on the basis of epidemiologic, clinical, and radiologic features, and reasonable exclusion of other potential diagnoses. Bronchoscopic lung biopsy has a low yield, and bronchoalveolar lavage (BAL) findings are nonspecific in RB-ILD, but may be diagnostically helpful in distinguishing RB-ILD from other conditions such as HP that are associated with more specific features.

The histopathologic findings required for the diagnosis of RB-ILD are those of RB and include the presence of yellow-brown–pigmented macrophages in the lumens of respiratory bronchioles, alveolar ducts, and peribronchiolar alveolar spaces without significant associated interstitial pneumonia. At low power, these features are patchy and are generally confined to peribronchiolar regions (bronchiolocentric distribution). Mild peribronchiolar fibrosis can be seen, but honeycombing is unusual.

As in all of the Group 1 diseases, smoking cessation is a key component of RB-ILD management. Smoking cessation may lead to improvement in radiologic abnormalities and lung function. The degree of improvement following smoking cessation appears to be limited in some patients, and abnormalities may persist for years. For patients with significant lung impairment, corticosteroids or other immunosuppressive medications have been used in an attempt to limit progression of lung disease; however, evidence of their effectiveness is lacking. Most patients with RB-ILD have a relatively good prognosis, and mortality from RB-ILD is uncommon. Although smoking cessation may lead to disease remission in some patients with RB-ILD, longitudinal studies have shown that some patients remain symptomatic for years after smoking cessation.

Desquamative Interstitial Pneumonia

DIP was originally believed to be a diffuse parenchymal lung disease resulting from desquamation of alveolar epithelial cells into the alveolar space but later was recognized as a process of alveolar filling from macrophage accumulation. DIP is associated with cigarette smoking in at least two-thirds of cases but can also be seen in nonsmokers, particularly in the context of autoimmune diseases, some infections, and drug exposures. It has been reported to occur in children as well as adults.

The clinical presentation of DIP is nonspecific with dyspnea and cough, and physical examination reveals inspiratory crackles in ∼60% and digital clubbing in 25–50% of patients ( Table 3.1 ). Pulmonary function testing reveals restriction in one-third of cases, normal findings in 10–20%, and a mixed defect in the remainder.

Chest radiography typically reveals patchy haziness or interstitial patterns with lower zone predominance. The striking abnormality on HRCT is ground-glass opacities predominantly in the lower lung zones and often in a peripheral distribution ( Fig. 3.2 ). Irregular linear opacities are frequently present; however, honeycombing and significant architectural distortion are uncommon. In some instances, patients with DIP have been reported to develop HRCT findings suggestive of fibrotic NSIP (irregular linear opacities) on longitudinal follow-up. Small parenchymal cysts and apical emphysematous changes may also be seen.

High-resolution CT of the chest showing more extensive ground-glass opacities in a smoker with desquamative interstitial pneumonia.

On light microscopy, lung biopsies show characteristic filling of alveolar spaces with pigment-laden alveolar macrophages. While both RB-ILD and DIP are associated with the accumulation of pigment-laden macrophages in alveolar spaces, the distribution of abnormality is more bronchiolocentric and patchy in RB, whereas in DIP it tends to be more diffuse. The extent of interstitial fibrosis, lymphoid follicles, and eosinophilic infiltration has been reported to be more prevalent in DIP than in RB-ILD. Fibroblast foci are not seen, and the DIP lesion appears temporally uniform. A definitive diagnosis of DIP usually requires surgical lung biopsy, as it may be difficult to reliably differentiate DIP from NSIP or RB-ILD by clinical, radiologic, and bronchoscopic biopsy criteria.

For those DIP patients who are smokers, smoking cessation is an essential component of therapy. Prolonged remission of DIP after smoking cessation has been described, but similar to all other smoking-related ILDs, the effect of smoking cessation on the natural history of DIP remains poorly characterized. Although most DIP patients have a relatively good prognosis with a better than 90% 5-year survival, some patients progress to respiratory failure and premature death within 5 to 10 years following the diagnosis. Patients with DIP are frequently treated with corticosteroids, but the effectiveness of steroid therapy is variable and has not been evaluated in a prospective study. Other immunosuppressants such as azathioprine and methotrexate have been used in anecdotal cases. Lung transplantation is an option for patients with progressive disease, but DIP can recur in the transplanted lung.

Pulmonary Langerhans Cell Histiocytosis

PLCH (also referred to as pulmonary Langerhans granulomatosis or pulmonary eosinophilic granuloma) is induced by cigarette smoke exposure in the majority of adult patients diagnosed to have this disorder and is characterized by accumulation of CD1a-expressing Langerhans cells in the lung, and occasionally in other organ systems. Adult PLCH forms part of the spectrum of histiocytic diseases, which range from relatively benign processes such as unifocal LCH involving bone to disseminated multiorgan forms (more commonly in children) associated with significant morbidity and mortality. Contrary to DIP and RB-ILD, which exclusively affect the lungs, ∼15% of adult PLCH patients may have disease outside the thoracic cavity. PLCH represents ∼5% of the total number of diffuse parenchymal lung diseases diagnosed by lung biopsy. PLCH tends to affect younger adults in their third and fourth decades. PLCH appears to affect both men and women equally.

Approximately 95% of adults with PLCH are active or former smokers or have been exposed to substantial second-hand cigarette smoke. While the pathogenesis remains poorly understood, it is likely that cigarette smoke constituents activate epithelial cells and other cell types in the airways to produce cytokines that promote recruitment, activation, and retention of Langerhans cells in the subepithelial regions of the airways. Cigarette smoke also induces the production of cytokines with profibrotic functions, such as TGF-β; in turn, TGF-β and other cytokines such as GM-CSF may further promote local expansion of Langerhans cells and facilitate the development of tissue remodeling and fibrosis as is evident in more advanced PLCH cases. It is possible that certain cigarette smoke constituents are taken up by immune or other cells and result in direct immune cell activation in peribronchiolar regions. Activated Langerhans cells and macrophages in peribronchiolar regions are likely to then promote secondary recruitment of T cells, plasma cells, and eosinophils, resulting in the formation of eosinophilic granulomatous inflammation from which the descriptive term “eosinophilic granuloma” is derived.

In recent years, somatic mutations leading to constitutive activation of the mitogen-activated protein kinase pathway, a key regulator of many cellular functions, including cell growth and proliferation, have been identified in LCH. Similar mutations have been identified in a subset of cases of adult PLCH and include BRAF (usually BRAF-V600E), MAP2K1, and NRAS mutations. Cumulatively, these data suggest that some cases of adult PLCH represent myeloid neoplastic processes rather than a reactive process induced by exposure to tobacco smoke. This observation raises the possibility of targeted therapy, such as vemurafenib in BRAF-V600E-positive cases, for patients who exhibit progressive disease despite smoking cessation, the cornerstone of management in adult PLCH.

As in other smoking-related ILDs, the clinical presentation tends to be nonspecific and includes dry cough and shortness of breath ( Table 3.1 ). About one-third of patients are asymptomatic. Constitutional symptoms occur in approximately 20–30% of patients, while few patients (around 10–15%) may present with a spontaneous pneumothorax that can be recurrent. Rarely, patients may present with symptoms related to extrapulmonary manifestation, such as skin, lymph node, or bony involvement.

Pulmonary function testing demonstrates variable results and may show obstructive, restrictive, mixed, or nonspecific abnormalities; pulmonary function testing may at times be completely normal. Physiologic studies reveal limitations in the exercise capacity that can occur even with relatively normal resting ventilatory function. Exercise limitation correlates with markers of pulmonary vascular dysfunction, implying vascular involvement as an important cause of exercise limitation in these patients.

The chest radiograph is usually abnormal and shows reticulonodular opacities more prominent in the middle and upper lung zones. The HRCT of the chest often reveals characteristic abnormalities that include nodules and cysts in varying combinations bilaterally with relative sparing of the lung bases ( Fig. 3.3 ). Nodules with or without cavitation predominate in early disease, whereas cystic changes predominate in more advanced disease.

High-resolution CT of the chest demonstrating a combination of nodular and cystic lesions in the upper lung fields and relative sparing of the lung bases in a smoker with pulmonary Langerhans cell histiocytosis.

A bronchoscopic or surgically obtained lung biopsy is recommended to confirm the diagnosis but is not always necessary. Bronchoscopy is diagnostically useful if an elevated percentage of CD1a-positive cells is identified in the BAL fluid, with ≥5% being virtually diagnostic of PLCH.

Histologic features of early PLCH include loosely formed nodules of mixed inflammatory cells centered on small airways in a bronchiolocentric pattern. These bronchiolocentric lesions of pulmonary LCH typically form stellate lesions with central scarring. Langerhans cells are abundant in early lesions and may be identified by immunohistochemical staining for the CD1a or Langerin (CD207) cell surface antigens or by the identification of intracellular Birbeck granules (pentalaminar rod-shaped intracellular structures) by electron microscopy. Eosinophilic infiltration is often encountered and may be quite extensive earlier in the course of the disease. Varying degrees of parenchymal infiltration with macrophages, lymphocytes, and eosinophils are noted, and in rare cases, extensive alveolar macrophage infiltration causes a “pseudo-DIP” reaction. Some cases are associated with extensive vascular infiltration of inflammatory cells, resulting in a proliferative vasculopathy involving both arteries and veins.

A critical component in the management of PLCH is smoking cessation. Smoking cessation often leads to stabilization of symptoms and radiologic abnormalities. However, some individuals may show disease progression leading to respiratory failure despite smoking cessation. There is no biomarker to predict which patient will improve and who will continue to get worse despite smoking cessation. For patients with severe disease, systemic pharmacotherapy is often considered in addition to smoking cessation. Corticosteroid therapy in the form of oral prednisone 40–60 mg daily with slow tapering over months has historically been used to treat patients with severe or progressive disease, but the data on therapeutic benefit of corticosteroids are limited. Because of the perceived lack of effectiveness of corticosteroids, a number of other immunosuppressive agents, namely vinblastine, chlorodeoxyadenosine (also known as 2-CDA), cyclophosphamide, and methotrexate, have been used to treat progressive PLCH. Chlorodeoxyadenosine has been successfully used in the management of multisystem LCH involving the bone and skin, but its utility in the management of smoking-related PLCH is not well defined. Whether immunosuppressive therapy is effective in the management of patients with progressive disease who continue to smoke is currently not known. As noted earlier, identification of somatic mutations in a subset of patients with PLCH raises the possibility of targeted therapy such as vemurafenib, a selective oral inhibitor of BRAF V600 kinase, in the treatment of BRAF V600E mutation-positive cases.

Management of PLCH also includes treating associated complications and sequelae such as pneumothorax, pulmonary hypertension, and respiratory failure. Pneumothorax is generally managed initially by chest tube drainage. Pleurodesis should be considered for most patients with spontaneous pneumothorax associated with PLCH because the recurrence rate of pneumothorax with conservative management only is ∼60%. Pulmonary hypertension is a complication that can be seen even in the absence of severe ventilatory impairment or hypoxemia in patients with PLCH and is present in nearly all patients with advanced disease. The presence of pulmonary hypertension portends a poor prognosis. We routinely perform a two-dimensional echocardiogram on patients with PLCH at the time of diagnosis and later in the clinical course if dyspnea or the degree of hypoxemia seems out of proportion to the severity of ventilatory impairment on pulmonary function testing. If the patient has echocardiographic evidence of pulmonary hypertension, right-sided heart catheterization should be performed to confirm the presence, determine the severity, and assess response to vasomodulator therapy. The use of vasomodulators such as the endothelin antagonist bosentan and the phosphodiesterase inhibitor sildenafil should be considered in patients with moderate to severe pulmonary hypertension.

Overall, most patients with PLCH have a relatively good prognosis, particularly if complete smoking cessation is achieved. The overall median survival from time of diagnosis is ∼13 years, with 5-year and 10-year survival rates of 75% and 64%, respectively. Some individuals may progress to extensive pulmonary scarring and cystic changes leading to respiratory failure. Lung transplantation is an option for patients with advanced PLCH. The overall survival of PLCH patients with lung transplants is comparable to that of individuals with other indications for lung transplantation. Recurrence of PLCH in the transplanted lung, even after smoking cessation, has been described in a few cases.