Interstitial Pneumonia

The incidence of mAb-induced interstitial pneumonia is unknown, although three studies made estimates in specific populations. The British Society for Rheumatology Biologics Register (BSRBR) prospectively collects data on all patients with rheumatoid arthritis receiving biological therapy in the United Kingdom. In 2007, Dixon et al¹⁰ studied 9294 patients with rheumatoid arthritis treated with anti-TNF agents and 2454 treated with nonbiological agents from the BSRBR and found a higher prevalence of physician-reported interstitial pneumonia in patients receiving anti-TNF (2.9% vs 1.8%; P = .002), although the possible cause-effect relationship between interstitial pneumonia and initiation of biological therapy was not clear. Two Japanese postmarketing surveillance studies of 7091 patients with rheumatoid arthritis treated with etanercept and 5000 patients with rheumatoid arthritis treated with infliximab, found a prevalence of interstitial pneumonia of 0.6% and 0.5%, respectively.¹¹ Thus, available data suggest a prevalence of interstitial pneumonia of 0.5% to 3% in patients with rheumatoid arthritis treated with anti-TNF agents. However, whether this might indicate more severe rheumatoid arthritis requiring biologic therapies or the effect of exposure to drugs with potential lung toxicity such as methotrexate or leflunomide is not clear.

In 2006, the Study Group on Autoimmune Diseases (GEAS) of the Spanish Society of Internal Medicine created the BIOGEAS project (www.biogeas.org), a multicenter study dedicated to collecting data on the use of biological agents in adult patients with systemic autoimmune diseases. An additional objective of the BIOGEAS project was to collect data on autoimmune diseases secondary to the use of biological agents by a quarterly surveillance of reported cases by a MEDLINE search.^{[3], [6] and [8]} For this review, the baseline MEDLINE search included articles published between January 1990 and January 2010, including the MeSH term “lung diseases, interstitial” combined with “anti-TNF,” “infliximab,” “etanercept,” “adalimumab,” “golimumab,” “certolizumab,” “anakinra,” “abatacept,” “rituximab,” “belimumab,” and “epratuzumab.” Patients receiving biological therapies for neoplasia, pediatric cases, and pulmonary involvement clearly associated with an infectious etiology were excluded. Additional articles were identified through a comprehensive manual search of the references of retrieved articles. According to previous reports,^{[3], [6] and [8]} the etiologic role of mAb in the development of interstitial pneumonia was suggested by at least one of the following factors: disease temporally associated with the initiation of therapy, symptomatology suggestive of an antigen-mediated hypersensitivity process, resolution of disease after discontinuation of therapy, or positive re-challenge phenomenon (reappearance or worsening of symptoms on re-exposure to the agent).

The PubMed search identified 118 cases of mAb-induced new-onset or exacerbation of interstitial pneumonia: 49 come from case series or case reports, 67 from the two Japanese postmarketing studies, and 2 from retrospective studies. However, interstitial pneumonia exacerbation also has been reported in the course of untreated rheumatoid arthritis, and the possibility of reporting bias cannot be totally excluded.

In 90% of patients, biological agents were administered for rheumatoid arthritis and were anti-TNF drugs (etanercept in 55 cases, infliximab in 55, and adalimumab in 3), and rituximab in 5 cases. Epidemiological data in 49 cases showed there were 38 women with a mean age at interstitial pneumonia onset of 61 years (range, 24-80 years). Two-thirds of patients had received methotrexate.

The clinical presentation of interstitial pneumonia was detailed in 49 cases. The disorder appeared a mean of 26 weeks after initiation of biologicals, with 50% of cases appearing within 2 months. The main presenting symptoms included dyspnea in 83% of cases, fever in 40%, and cough in 35%. High-resolution pulmonary computed tomographic scans in all cases suggested interstitial pneumonia; bronchoalveolar lavage in 10 cases showed predominantly lymphocytic alveolitis. Pulmonary function tests in 13 cases showed a restrictive pattern and reduced diffusing capacity of the lung for carbon monoxide. Interstitial pneumonia was confirmed by pulmonary biopsy in 24 cases, mainly usual interstitial pneumonia and nonspecific interstitial pneumonia according to a 2002 interstitial pneumonia consensus classification.¹²

Sarcoid-like Disorder

Recently, the mAbs have been used to treat refractory/severe sarcoidosis.¹³ Paradoxically, by August 2010, 40 cases of mAb-induced sarcoid-like disease were reported in patients with rheumatologic autoimmune diseases, including 39 from case series or case reports and 1 from a randomized controlled trial. Recently, Daien et al¹⁴ reported 10 cases of sarcoid-like granulomatosis after anti-TNF administration (4 for rheumatoid arthritis and 6 for spondyloarthropathies) and estimated a prevalence of 0.04%, fourfold higher than for idiopathic sarcoidosis.¹⁵

Agents included anti-TNF drugs (mainly etanercept in 24 cases) and anakinra in 1 case.^{[14], [15], [16] and [17]} Epidemiological data in 38 cases showed there were 25 women with a mean age at sarcoid-like disease onset of 50 years (range, 27-73 years). Seventeen patients had received methotrexate.¹⁸

The clinical presentation was detailed in 39 cases, and included mainly respiratory, cutaneous, and general symptoms (Table 2). Imaging studies disclosed pulmonary involvement in 92% of cases, mainly pulmonary infiltrates and hilar adenopathies. According to the classification of radiographic stages of sarcoidosis,¹⁵ most cases had 2 stages (50%). Thirteen patients (33%) presented cutaneous involvement.

Laboratory tests showed raised angiotensin converting enzyme levels in 48% of patients and a positive tuberculin test in 19% cases. Polymerase chain reaction for mycobacterial infection was negative in the 31 patients tested. Noncaseating granulomas were demonstrated histologically in all but 1 case.¹⁹

Pulmonary Vasculitis

Only 3 cases of mAb-induced pulmonary vasculitis are reported^{[20], [21] and [22]}: 1 patient with rheumatoid arthritis with necrotizing pulmonary nodules and vasculitis after etanercept therapy, and another patient with rheumatoid arthritis who developed alveolar capillaritis, crescentic glomerulonephritis, and (peri) nuclear antineutrophil cytoplasmic antibody after etanercept initiation.

Other Pulmonary Diseases

Patients treated with mAb often present acute or delayed hypersensitivity reactions, including mild respiratory symptoms.^{[23] and [24]} Severe respiratory manifestations, including severe bronchospasm,^{[25] and [26]} hypersensitivity pneumonitis,²⁷ and acute respiratory distress syndrome caused by eosinophilic pneumonia²⁸ also are reported. A 2006 systematic review of randomized trials of patients with rheumatoid arthritis treated with anti-TNF agents, suggested a dose-dependent increased risk of malignancies, including some cases of lung cancer.²⁹ However, a recent study using BSRBR data found no increase in incident malignancy in patients with rheumatoid arthritis with prior malignancy receiving anti-TNF therapy after a mean follow-up of 3 years.³⁰

Outcome of Monoclonal Antibody Induced Pulmonary Diseases

Interstitial Pneumonia

Data on therapeutic management were available in 50 cases and included withdrawal of biological agents in all cases but 1.³¹ Corticosteroids were used in 84% cases, with methylprednisolone pulses administered in one-third and immunosuppressive agents added in 8 cases. The mAb therapy was reinitiated in 1 case, with infliximab changed to etanercept to control the underlying disease, without further complications.³²

Interstitial pneumonia outcomes were detailed in 48 cases, with complete resolution in 32%, improvement or partial resolution in 28%, and no resolution in 40%. Fifteen patients (32%) died during follow-up, 70% within 5 weeks of mAb initiation. The cause of death was progressive pulmonary disease in all cases but 1. Postmortem studies in 3 cases confirmed pulmonary fibrosis as the cause of death. Patients with pre-existing interstitial lung disease had a worse prognosis than newly diagnosed patients, with a mortality rate of 67% vs 32%, respectively. BSRBR 2003 data show that, after any anti-TNF therapy, the mortality rate was sixfold higher in patients with pre-existing rheumatoid arthritis-associated interstitial lung disease (n = 184) than those without (n = 6061), (90 per 1000 person-years of follow-up vs 14 per 1000). After adjusting for age and sex, patients with baseline rheumatoid arthritis-associated interstitial lung disease treated with biologics had a mortality rate 4.4 times higher (95% confidence interval, 1.8-10.7) than patients without pulmonary disease treated with biologic agents (Biologics Register Newsletter, May to August 2004, British Society for Rheumatology).

In 2010, however, by examining the influence of anti-TNF therapy on mortality in patients with pre-existing rheumatoid arthritis-associated interstitial lung disease, Dixon et al³³ concluded that mortality is not higher in patients receiving anti-TNF therapy compared with those treated with traditional disease modifying antirheumatic drugs, although no recommendations on patient screening, selection, or monitoring were made.

Available evidence suggests that pre-existing pulmonary disease, especially interstitial pneumonia, may have prognostic value in patients receiving mAb, and that anti-TNF agents should be used with extreme caution in patients with pre-existing or underlying severe lung disease. Heightened surveillance of pulmonary complications during postmarketing studies of anti-TNF therapies is recommended.

Sarcoid-like Disease

Idiopathic sarcoidosis generally has a good prognosis. Many patients are asymptomatic and at least 60% of cases resolve spontaneously.¹⁵ However, in patients with chronic progressive disease, reported mortality is 1% to 6%.¹⁵ Of the 36 reported cases of mAb-induced sarcoid-like disease with detailed outcomes, 67% resolved completely, there was improvement or partial resolution in 22%, and no resolution in 11%; no patients died. The mAb therapy was withdrawn in all cases except 4,^{[14], [34] and [35]} and corticosteroids were used in 16 patients (40%). The mAb therapy was continued in 4 cases (with recurrence of sarcoid-like disease in 2 cases) and reinitiated in 12 cases to control the underlying disease. In 11 cases, the anti-TNF agent was changed, with sarcoid-like disease reappearing in 1 case. The limited evidence suggests that continuation or re-initiation with the same agent is more frequently associated with recurrence of sarcoid-like disease compared to switching agents. Whether mAb therapy unmasks and accelerates idiopathic sarcoidosis or induces a condition that resembles sarcoidosis is unclear.

Pathophysiological Mechanisms

Of the 161 cases reviewed, 97% were associated with TNFα blockers. TNFα, a cytokine that is an essential component of the innate immunity, has a very complex network of interactions. TNFα is associated with two etiopathogenic processes closely related to idiopathic interstitial pneumonia and sarcoidosis: it is a pivotal cytokine in the pathophysiology of pulmonary fibrosis^{[36] and [37]} and plays an essential role in the formation and maintenance of granulomas, as shown by studies of the physiopathological mechanisms of mycobacterial infection.^{[38] and [39]}

TNFα is postulated as a key cytokine in the pathogenesis of interstitial pneumonia. Paradoxically, anti-TNF agents are postulated as a potential therapy for idiopathic interstitial pneumonia, while there are increasing reports of anti-TNF-induced interstitial pneumonia, suggesting TNFα may have both profibrotic and antifibrotic effects. TNFα promotes pulmonary tissue repair, eliminating inflammatory cells by apoptosis.^{[40] and [41]} Kuroki et al⁴² demonstrated that TNFα plays an important role in limiting pulmonary inflammation in a TNFα −/− mouse model developing an accelerated form of bleomycin-induced pulmonary fibrosis. Other studies describe synergistic inhibition of fibroblast proliferation when TNFα was combined with interleukin-1 or interferon.⁴³ TNFα may play a profibrotic role by upregulating pulmonary transforming growth factor-beta 1 (TGF-β1) expression through mechanisms including extracellular regulated kinase-specific pathway activation in fibroblasts,⁴⁴ increased expression and DNA binding of the activator protein-1,⁴⁴ and regulation of p21, a cyclin-dependent kinase inhibitor, a key regulator of cell cycle progression, DNA repair, and apoptosis. Yamasaki et al⁴⁵ found that p21 is a negative regulator of TGF-β1-induced pulmonary inflammation and fibrosis, and regulates TGF-β1-induced tissue responses via TNFα-signaling pathways.

Since infliximab binds to transmembrane TNFα expressed on the surface of macrophages and CD4+T-cells, leading to cellular lysis,⁴⁶ increased local release of macrophage-derived proteolytic enzymes might enhance the potential pulmonary toxicity of methotrexate, which might be potentiated by anti-TNF therapy through deficient apoptosis of infiltrating inflammatory cells. Therefore, anti-TNF combined with methotrexate might contribute to interstitial pneumonitis, while the longer half-life of some agents could explain slow pulmonary infiltrate clearance. Anti-TNF therapy also may result in systemic and/or pulmonary shifts toward anti-inflammatory cytokines, such as TGF-β1, contributing to profibrotic states,⁴⁷ which would explain the exacerbation of pulmonary fibrosis in patients with pre-existing interstitial lung disease and the enhanced risk of pulmonary fibrosis in patients taking methotrexate concomitantly.

Although the etiology of rituximab-induced interstitial pneumonia is unclear, rituximab is associated with complement activation, B-lymphocyte cytolysis, and TNFα release,⁴⁸ which may contribute to pulmonary damage.

Anti-TNF-induced sarcoid-like disease overwhelmingly occurs in the lungs and skin, which are in direct contact with potential environmental pathogens. Granulomas are organized aggregates of macrophages that recruit B and T lymphocytes through antigen-driven Th1-type immune responses, providing a local environment where these cells may interact and suppress the pathogen. The unique organization of granulomas also prevents the spread of infection. Because granuloma formation is defective in murine models lacking TNFα activity, it is postulated that TNFα orchestrates the initial formation of granulomas in mycobacterial infections.⁴⁹ Although the etiopathogenic mechanisms involved in mAb-induced sarcoid-like disease are unknown, the demonstrated high risk of mycobacterial infection after anti-TNF administration^{[38] and [39]} suggests their involvement in granuloma formation. Studies show fewer granulomas, lacking their characteristic architecture and organization, are formed in the absence of TNFα signaling.^{[50] and [51]}

Studies in Mycobacterium tuberculosis-infected mice found that TNFα blockade resulted in failure to maintain the granuloma structure and led to diffuse pulmonary infiltration,^{[52] and [53]} which might contribute to noninfectious sarcoid-like granulomas and interstitial lung damage. Flory et al⁵⁴ described a key regulatory role of TNFα and interleukin-1β in monocyte chemoattractant protein-1-mediated pulmonary granuloma formation in rats, while Shinozaki et al⁵⁵ recently examined healing of cutaneous wounds in a TNFα-deficient murine model and found that the absence of TNFα promoted granulation tissue formation.

Two-thirds of reported cases of anti-TNF-induced sarcoid-like disease are associated with etanercept. Studies in rheumatic patients suggest that infliximab and etanercept block TNFα through different etiopathogenic mechanisms. Infliximab causes in vitro antibody-dependent and complement-dependent lysis of cells expressing the transmembrane TNFα form,⁵⁶ inhibits interferon-γ production,⁵⁷ and reduces T cells expressing TNFα and interferon-γ,⁵⁸ while etanercept seems not to influence interferon-γ pathways. This difference may be related to the TNFα type targeted by infliximab and etanercept (transmembrane and soluble vs soluble, respectively). While transmembrane TNFα may mediate the response to microbes, soluble TNFα may be involved in less specific inflammatory mechanisms involved in autoimmune disease.⁵⁹ The different effects of anti-TNFα agents on the regulation of interferon-γ production by T-cells¹⁶ might account for the higher number of reports of sarcoid-like disease induced by etanercept.

Patient Management

Although the evidence is based mainly on case reports and small series, and patients had other risk factors for interstitial lung disease such as rheumatoid arthritis and methotrexate, we suggest some recommendations for the management of patients with mAb-induced pulmonary diseases (Figure 1).

Evaluation of Pre-existing Lung Disease

Before mAb initiation (especially anti-TNF), a careful clinical evaluation of previous pulmonary disease is mandatory, especially in patients with rheumatoid arthritis, in whom interstitial pneumonia is a frequent extra-articular manifestation. Additional factors associated with a high risk of interstitial pneumonia, including smoking,⁶⁰ autoimmune diseases such as Sjögren's syndrome and systemic lupus erythematosus,⁶¹ and, especially, concomitant therapies (methotrexate, azathioprine, cyclophosphamide, and leflunomide) should be evaluated.

Pre-therapy Studies

Chest radiographs and immunological studies are recommended before initiating mAb therapy.⁶² Due to the increased risk of development of auto-antibodies (antinuclear and anti-DNA antibodies and rheumatoid factor) after mAb therapy,⁶² it is advisable to have, or perform if not available, an immunological study before initiating biological therapy, in order to better evaluate a posterior autoimmune disease triggered after the initiation of therapy. Pulmonary function tests (and pulmonary computed tomography scan if necessary) can discard subclinical interstitial lung disease, especially in patients at high risk of pulmonary disease (Figure 1).

A recent review by Lioté et al⁶³ identified 45 patients with rituximab-interstitial lung disease, although 44 were hematological patients, with only 1 case of rheumatoid arthritis. The most common presentation was acute/subacute hypoxemic organizing pneumonia (n = 37), starting 2 weeks after the last infusion and resolving, in most cases, provided glucocorticoid therapy was given early. Acute respiratory distress syndrome occurred in 5 patients within a few hours, usually after the first infusion. Eight patients died. The authors recommended clinical support and possibly radiological monitoring between infusions, particularly in patients with a history of reversible respiratory symptoms.

Pointers to Early Diagnosis

A typical patient with mAb-induced interstitial pneumonia is a 60-year-old woman with rheumatoid arthritis receiving anti-TNF therapy presenting progressive dyspnea and cough during the first weeks of therapy. In sarcoid-like disease, patients have rheumatoid arthritis or seronegative spondyloarthropathy, a mean age of 50 years old, a female:male ratio of 2:1, and present with dyspnea, fatigue, and/or cutaneous nodules ([Figure 2] and [Figure 3]).

Mycobacterial infection should be exhaustively investigated. Dixon et al⁶⁴ found that infliximab and adalimumab are associated with a threefold or fourfold higher rate of tuberculosis (62% extrapulmonary), compared with etanercept in patients with rheumatoid arthritis. Nearly half the cases of disseminated tuberculosis occurred after cessation of adalimumab, indicating the need for vigilance after discontinuation. Therefore, tuberculosis prophylaxis should be considered in eligible patients before beginning anti-TNF therapy.⁶⁵ Interferon-γ Elispot assays have been reported to perform better than the tuberculin skin test in recognizing rheumatic patients with latent tuberculosis infection who are candidates for anti-TNF treatment.⁶⁶

Tailored Therapy

Therapy should be tailored according to the severity and organ involvement. In patients with mild features (general symptoms, arthralgias, or isolated cutaneous lesions), mAb cessation may suffice, although closely monitored, switching to another biological agent to control the severe underlying disease might be considered after resolution of the autoimmune process. In patients with severe interstitial lung disease, mAb cessation is mandatory, with initiation of oral corticosteroids and addition of immunosuppressive agents as necessary: mAb re-initiation (especially the same agent or family) is not recommended.

Conclusion

The mAbs have been used for rheumatologic autoimmune diseases but there are increasing reports of the paradoxical induction of autoimmune processes due to these agents, overwhelmingly anti-TNF. To date, over 150 cases of drug-induced pulmonary disease have been reported, mostly from uncontrolled studies, meaning solid conclusions are not possible. Concerns include the variable incidence of induced disease according to ethnic group, underlying disease and activity, type of study, possible dose-related effects, concomitant medications, and differences in the pathogenic mechanism according to the molecule type (eg, chimeric vs humanized). De novo granulomatous disease seems to be mostly reversible, whereas pulmonary fibrosis has the worst prognosis, especially in the setting of prior lung fibrosis. Management recommendations are shown in Figure 1. Large, prospective, postmarketing studies including nonbiological agents as controls may provide an accurate risk-benefit analysis of pulmonary disease in patients with rheumatologic autoimmune diseases receiving mAb.