Drug-induced lung disease in rheumatoid arthritis

INTRODUCTION — Drug-induced pulmonary disease is an important consideration in the differential diagnosis of patients with rheumatoid arthritis (RA) who present with respiratory symptoms [1]. Knowledge of the types of lung toxicity that are associated with the individual agents used to treat RA, the patterns of lung disease that are associated with RA (unrelated to medication), and the spectrum of potential comorbid disease processes will help in the diagnosis and management of drug-induced lung disease.

A review of drug-induced lung disease in patients with RA will be presented here. Other aspects of pulmonary disease associated with rheumatoid arthritis are discussed separately. (See "Overview of pleuropulmonary diseases associated with rheumatoid arthritis" and "Interstitial lung disease in rheumatoid arthritis" and "Overview of the systemic and nonarticular manifestations of rheumatoid arthritis".)

ETIOLOGY — A number of drugs used to treat RA can induce alveolar inflammation, interstitial inflammation, and/or interstitial fibrosis (Pneumotox.com), although the exact pathogenesis of the toxicity is unknown [2,3]. The risk and type of lung toxicity varies among the different agents. (See 'Features of individual agents' below.)

In addition to direct lung toxicity, virtually all of the disease modifying antirheumatic drugs (DMARDs) have immunosuppressive effects that increase the risk of bacterial and opportunistic lung infection [4-6]. As an example, in an observational study of 16,788 patients with RA, patients taking prednisone had a higher risk of hospitalization due to pneumonia than patients not taking glucocorticoids (hazard ratio 1.7, 95% CI 1.5-2.0) [7]. The effect was dose dependent. Although methotrexate and tumor necrosis factor (TNF) antagonists were not associated with an increased risk of pneumonia in this study, they have been associated with pneumonia in other studies [8-16]. (See "Tumor necrosis factor-alpha inhibitors: Bacterial, viral, and fungal infections".)

PREVALENCE — The net effect of the disease modifying antirheumatic drugs (DMARDs) on the incidence of drug-induced lung disease in patients with RA is uncertain. Medication side effects may become more common as the clinical use of DMARDs and biological agents such as tumor necrosis factor (TNF) blockers increases [17-20].

Conversely, DMARDs may have a beneficial impact on the natural history of a variety of forms of rheumatoid-associated lung disease, and use of these drugs may influence the frequency or presentation of these problems [21]. In one study of 59 patients with RA who were taking DMARDs but had no pulmonary symptoms, abnormal histology on transbronchial lung biopsy was found in 18 percent of patients on DMARDs, compared to 42 percent of those not taking one of these drugs [22].

CLINICAL MANIFESTATIONS — The clinical manifestations of drug-induced lung disease are variable and nonspecific. Symptoms, when present, can develop days to years into therapy and can progress rapidly or indolently. Drug-induced lung disease is one of the causes of rapid progression of interstitial lung disease (ILD) to respiratory failure. Symptoms include cough, dyspnea, low-grade fever, and occasionally a rash. Lung auscultation may reveal focal or bibasilar crackles, but is often normal.

EVALUATION AND DIAGNOSIS — Drug-induced lung disease is often a diagnosis of exclusion, so the diagnostic approach involves a combination of tests designed to exclude other processes (eg, heart failure, infection, RA-associated interstitial lung disease) and to identify features that are suggestive of a drug-induced process (eg, timing of symptoms relative to drug initiation, eosinophilia in blood or bronchoalveolar lavage fluid). Empiric drug discontinuation is an important diagnostic step. Underlying lung disease due to RA or cigarette smoking may complicate determination of whether current symptoms are caused by a new process or progression/exacerbation of a preexisting process. General approaches to interstitial lung disease and to lung disease in patients with RA are described separately. (See "Interstitial lung disease in rheumatoid arthritis" and "Approach to the adult with interstitial lung disease: Clinical evaluation" and "Approach to the adult with interstitial lung disease: Diagnostic testing" and "Interpretation of lung biopsy results in interstitial lung disease".)

●Laboratory testing – Laboratory testing is used to determine whether other disease processes are contributing to the patient's respiratory compromise. Complete cell counts and differential are obtained to look for anemia (suggestive of alveolar hemorrhage), neutrophilia (suggestive of infection), or eosinophilia (suggestive of drug hypersensitivity or fungal infection). B-type natriuretic peptide (BNP) can help exclude heart failure as an etiology. Blood cultures, sputum cultures, C-reactive protein (CRP), and serologic studies can also help to identify infectious causes. (See "Approach to the adult with interstitial lung disease: Diagnostic testing", section on 'Laboratory tests'.)

●Imaging – A chest radiograph is obtained to assess the pattern and extent of disease but high resolution computed tomography of the chest is usually needed for full characterization. Various radiographic patterns of drug-induced injury are described, including patchy or diffuse, unilateral or bilateral reticular markings, ground glass opacities, or consolidation, and pulmonary nodules with or without cavitation. Newly appearing or enlarging pulmonary nodules have been associated with methotrexate and leflunomide. Pleural effusions have been associated with methotrexate [3]. (See 'Methotrexate' below and 'Leflunomide' below.)

Hilar lymphadenopathy is an uncommon manifestation of drug induced disease, except in the case of methotrexate-associated lymphoproliferative disease or anti-tumor necrosis factor-alpha-induced lymphadenopathy. (See "Methotrexate-induced lung injury", section on 'Pulmonary lymphoproliferative disease' and 'Granulomatous lung disease' below.)

●Pulmonary Physiology – Frequently, patients are too unwell for complete lung function testing, but if they can be obtained, they can aid in documenting changes in lung function. At a minimum, pulse oxygen saturation (SpO₂) at rest and importantly on exertion, such as during a six-minute walk test (6MWT), is helpful in determining the severity of impairment. Hypoxemia at rest or with exertion is common. (See "Overview of pulmonary function testing in adults".)

While uncommon, a diffusing capacity for carbon monoxide (DLCO) above the predicted range suggests pulmonary hemorrhage. (See "Diffusing capacity for carbon monoxide", section on 'Increased DLCO'.)

●Bronchoscopy – After review of clinical findings, laboratory data, and chest imaging, bronchoscopy with bronchoalveolar lavage (BAL) may be needed to exclude processes such as infection, diffuse alveolar hemorrhage, or lymphangitic spread of tumor (eg, in the presence of fever, widespread or nodular opacities on chest imaging, or rapidly progressive respiratory impairment).

There are no specific findings for drug-induced lung toxicity on bronchoscopy or BAL. BAL fluid cell counts are usually elevated, but the pattern of cellularity is nonspecific. Lymphocytosis, neutrophilia, or eosinophilia may be seen; among these, eosinophilia is more suggestive of a drug-induced process. Thus, the main role of bronchoscopy is to exclude alternative diagnoses. (See "Basic principles and technique of bronchoalveolar lavage" and "Role of bronchoalveolar lavage in diagnosis of interstitial lung disease".)

●Response to therapy – Empiric withdrawal of the implicated drug is a key diagnostic and therapeutic step. In general, noninfectious drug-related lung disease often regresses upon withdrawal of the offending medication. After exclusion of infection, a prompt response to systemic glucocorticoid therapy may be a distinguishing feature of drug-induced lung disease, which often responds better to glucocorticoid therapy than RA-related interstitial lung disease. (See "Overview of pleuropulmonary diseases associated with rheumatoid arthritis" and "Interstitial lung disease in rheumatoid arthritis".)

●Lung biopsy – Lung biopsy is usually not required, and many patients may be too unwell to undergo this procedure. Often, the clinical picture, radiologic findings, and BAL results excluding infection are sufficiently convincing of the diagnosis to make a biopsy unnecessary, particularly in patients who respond quickly to drug discontinuation. In contrast, a lung biopsy is indicated when the patient has acute, progressive or severe disease and the cause of the pneumonitis is uncertain or when lymphoproliferative disease is suspected on the basis of nodular opacities, and when the biopsy findings will change therapy.

Lung biopsy rarely establishes an antirheumatic agent as the definitive source of the lung injury, as there are no pathognomonic findings, and histologic criteria for drug-induced lung disease have not been established [23]. However, when available, lung histopathology can characterize the histopathologic pattern (eg, lymphocytic, granulomatous, eosinophilic, or organizing pneumonia or diffuse alveolar damage), which may help to guide therapy. (See "Role of lung biopsy in the diagnosis of interstitial lung disease" and "Interpretation of lung biopsy results in interstitial lung disease".)

DIFFERENTIAL DIAGNOSIS — The differentiation between a drug reaction, underlying rheumatoid-associated lung disease, infection, and heart failure may be difficult, since there is significant overlap in the clinical syndromes. In addition, many of the pulmonary reactions to drugs used for the treatment of rheumatoid arthritis are rare and are published as case reports (table 1). An online repository of drug-induced lung disease is available to help identify potential culprit medications (Pneumotox.com) [2]. When patients present with fever, rapidly progressive respiratory insufficiency, or widespread or nodular opacities on chest imaging, bacterial and opportunistic lung infections are high on the differential and should be pursued vigorously. (See 'Evaluation and diagnosis' above and "Approach to the immunocompromised patient with fever and pulmonary infiltrates".)

FEATURES OF INDIVIDUAL AGENTS

Methotrexate — Methotrexate (MTX) is the most commonly used disease modifying antirheumatic drug (DMARD) in RA. Pulmonary complications of methotrexate are not associated with folate deficiency and may infrequently occur with the relatively low doses (<20 mg per week) that are used in patients with RA [24-26]. The risk factors, clinical manifestations, diagnosis, and treatment of MTX-induced pneumonitis are discussed in greater detail separately. (See "Major side effects of low-dose methotrexate" and "Methotrexate-induced lung injury".)

Acute or subacute interstitial pneumonitis typically presents with nonproductive cough, dyspnea, and sometimes fever or chest pain after weeks to months of oral therapy. Among patients with a subacute onset of MTX-induced lung disease, up to 50 percent demonstrate peripheral blood eosinophilia, which strongly supports the diagnosis, when present. Early drug discontinuation at the onset of respiratory symptoms may obviate the need for invasive testing.

Less commonly, interstitial fibrosis, accelerated rheumatoid lung nodulosis, asthma, and air trapping have been reported during MTX treatment. In many cases, it is not clear whether these less common abnormalities are drug-related or are due to underlying rheumatoid disease [27,28]. Furthermore, data are conflicting about the development and progression of chronic fibrotic lung disease during treatment with methotrexate, and more recent studies have failed to document an association [25,29,30].

●A systematic review and meta-analysis (22 studies with 8584 participants) reported a small increase in respiratory adverse events with methotrexate treatment (RR 1.10, 95% CI 1.02-1.19), but the definition of ILD was not consistent across studies [25].

●In contrast, support for the long-term safety of MTX comes from a multicenter, multiethnic case control study with 410 patients with RA and chronic fibrotic ILD and 673 patients with RA and no lung disease [29]. The study had discovery and replication steps, and the combined results showed that MTX use was not associated with an increased risk of ILD (odds ratio 0.43, 95% CI 0.26-0.69); ILD was detected later in RA patients taking MTX, than in those without MTX treatment.

As noted above, the differential diagnosis of MTX-induced lung disease includes infectious complications, which must be excluded prior to the initiation of immunosuppressive therapy to treat a possible drug reaction. Infections reported in the setting of methotrexate therapy include Pneumocystis jirovecii pneumonia, cryptococcal pneumonia, invasive pulmonary aspergillosis, disseminated histoplasmosis, pulmonary Nocardia infection, and viral pneumonia (eg, parainfluenza and cytomegalovirus) [6,8-13]. (See 'Evaluation and diagnosis' above.)

Leflunomide — Leflunomide is a DMARD used in RA that blocks a key enzyme of pyrimidine synthesis in activated lymphocytes. ILD and cases of new or accelerated pulmonary nodule formation have been reported, although the exact pathogenesis is not known [3,31-35]. As with other immunosuppressive agents, leflunomide is associated with an increased risk of infection in some [36], but not all studies [37]. (See "Pharmacology, dosing, and adverse effects of leflunomide in the treatment of rheumatoid arthritis".)

Interstitial pneumonitis — Data on the risk of interstitial pneumonitis due to leflunomide are conflicting [3,33-38]. Overall, the risk appears low, except possibly in patients with underlying ILD or a history of MTX-induced lung toxicity. We generally avoid leflunomide in such patients, realizing the limitations of the data. Decisions need to be individualized, however, as the benefits of leflunomide may be substantial.

Studies showing the range of findings with leflunomide include the following:

●A systematic review and meta-analysis that included eight clinical trials with 4579 participants suggested a decreased risk of noninfectious respiratory adverse events (RR 0.64, 95% CI 0.41-0.97) with leflunomide compared with MTX or placebo [37]. This analysis supports the contention that "channelling bias" may explain the results of the above observational studies. However, clinical trials would have excluded patients with underlying ILD. Thus, the safety of leflunomide in these patients requires further study.

●In a review that used linked prescribing and administrative databases for more than 235,000 patients with RA, the relative risk of ILD among those treated with leflunomide was 1.9 compared to those treated with other DMARDs [38]. However, there was no significant increase in risk among patients who had no prior diagnosis of ILD and no prior MTX use.

●Similar findings were reported in an observational study of 5054 patients who were treated with leflunomide; 1.2 percent developed new or worsening ILD [39]. Risk factors included preexisting lung disease (most important factor) with an odds ratio of 8.17 (95% CI 4.63-14.4), smoking, low body weight, and use of a loading dose.

Pneumonitis due to leflunomide usually presents within the first 20 weeks of therapy and may occur after cessation of the medication [34,40-42]. Typical symptoms are fever, cough, and dyspnea. The main findings on high resolution computed tomography (HRCT) are ground glass opacities, bilateral reticular opacities, and honeycombing, although areas of consolidation can also be seen [34,43]. Eosinophilia in the bronchoalveolar lavage (BAL) fluid has been described [42]. Lung biopsies show interstitial pneumonitis (sometimes with eosinophilia), organizing pneumonia, or diffuse alveolar damage.

In one report, leflunomide-induced pneumonitis was fatal for 11 of 29 patients [32]. In another report, there were no fatalities; the authors attributed this to prompt recognition and treatment with glucocorticoids and cholestyramine [33].

Leflunomide has a long half-life. Because of the hepatobiliary circulation of leflunomide, cholestyramine resin (eg, 8 g/day for three days) can be used to hasten elimination [32,44]. (See "Pharmacology, dosing, and adverse effects of leflunomide in the treatment of rheumatoid arthritis", section on 'Pregnancy and lactation'.)

Rheumatoid pulmonary nodules — Leflunomide is also associated with appearance or accelerated progression of rheumatoid (necrobiotic) pulmonary nodules, which can occasionally lead to pneumothorax [31,45,46]. Development of pulmonary rheumatoid nodules may be associated with cough and low grade fever [31,46]. On imaging studies, the nodules may be cavitary [31]. Cessation of leflunomide usually leads to improvement or resolution of the nodules. (See "Rheumatoid nodules" and "Overview of pleuropulmonary diseases associated with rheumatoid arthritis", section on 'Rheumatoid lung nodules'.)

Biological agents — Despite broad use of biologic agents to treat RA, cases of substantial pulmonary toxicity are infrequently reported [3,21,47,48]. In general, the biologic agents, tumor necrosis factor (TNF)-alpha blockers (soluble p75 TNF receptor fusion protein [etanercept], dimeric anti-TNF-alpha antibody [infliximab], anti-TNF-alpha monoclonal antibody [adalimumab]), interleukin (IL)-1 blockers (anakinra), anti-B-cell monoclonal antibody (rituximab), and a selective costimulation modulator which prevents T cell CD28 binding (abatacept), have been shown to improve symptoms and joint disease, and possibly also lung disease in patients with RA [49-51]. Case series have also reported patients with other pulmonary disorders, such as nodules and infection, in association with biologic therapies for RA. (See "Overview of biologic agents in the rheumatic diseases".)

Due to the infrequent occurrence of clinically significant ILD, the likelihood that these agents are used in patients with more severe RA, and the risk of reporting bias, it is difficult to ascertain the exact risk of drug-induced ILD or worsening of preexisting ILD [3]. The British Society for Rheumatology Biologics Register (BSRBR) prospectively collects data on all patients in the UK receiving biologic agents (>8000 patients) [52]. An early report in the organization's newsletter suggested that the odds ratio for mortality was 4.4 times higher (95% CI 1.8-10.7) for those RA patients with preexisting pulmonary disease who were treated with biologics compared to those without pulmonary disease. However, further analysis of the BSRBR included a prospective study of 367 patients with preexisting ILD (299 on anti-TNF; 68 on DMARDS) [53]. The adjusted (overall) mortality rate ratio comparing anti-TNF-alpha agents with DMARDS was 0.81 (95% CI 0.38-1.73), suggesting that biologics do not increase overall mortality. On the other hand, ILD appeared to be a more common cause of death with biologic agents compared with DMARDs (age and sex adjusted mortality rate ratio 2.63, 95% CI 0.60-11.45), although the numbers were small. A systematic review of published reports of drug-induced ILD in RA estimated the overall the risk with biologic agents at around 1 percent, but the mortality associated with ILD due to anti-TNFa-agents was high at 35 percent, compared with estimates of 18 and 13 percent for leflunomide and MTX, respectively [3].

The British Society of Rheumatology performed a systematic review of non-TNF-alpha agents and, based on the limited available data, concluded that abatacept and rituximab appear less likely to exacerbate ILD than other biologic agents and may be preferred over anti-TNF agents in patients with RA requiring therapy with a biologic agent [3,26,48]. The importance of these findings is in balancing the great overall benefit of biologic agents in controlling RA with the relative rarity, but potential severity, of drug-induced ILD. Additionally, as new agents are introduced, clinicians should maintain a heightened awareness of the potential for any drug to cause lung complications.  

Inflammatory pneumonitis — Drug-induced ILD has been reported with the TNF-alpha inhibitors, rituximab, tocilizumab, and anakinra, but not abatacept [2,54,55]. Development of new or worsening cough, dyspnea, and radiographic abnormalities should alert the clinician to the possibility of drug-induced ILD.

The frequency of drug-induced ILD and the possibility that underlying ILD can potentiate this process are illustrated by the following studies, and are summarized in a review [21]:

●Etanercept and infliximab (TNF-alpha inhibitors) – Drug-induced ILD is uncommon, but potentially severe in patients treated with etanercept or infliximab, particularly patients aged 65 or older and those with preexisting ILD [20,48]. Among 108 patients with RA who developed new-onset or worsening of preexisting ILD while taking these agents, the majority of cases developed after the first six months of therapy. Drug-discontinuation and glucocorticoids resulted in improvement or resolution in approximately 65 percent, but progressive and death in 12 percent. In an observational study of 7091 patients with RA who were treated with etanercept, ILD developed in 42 (0.6 percent) [56]. A number of case reports have described lung toxicity as well. A patient with a prior history of MTX therapy developed lung injury eight weeks after initiation of etanercept, but responded to cessation of etanercept and treatment with oral prednisone 40 mg/day [57]. In two small case series, 10 patients with RA developed progressive usual interstitial pneumonitis (UIP) temporally related to initiation of infliximab or etanercept, and one developed organizing pneumonia; eight patients died from progressive UIP [58,59]. Among the fatal cases, all had preexisting ILD with a UIP pattern and three were concomitantly taking azathioprine and glucocorticoids. In these patients, MTX had been avoided because of the preexisting lung disease. Three cases of acute interstitial pneumonitis have been reported shortly after addition of infliximab in patients taking a stable dose of MTX [60]. All patients improved with discontinuation of MTX and infliximab and treatment with methylprednisolone.

●Adalimumab (TNF-alpha inhibitor) – One patient with RA on a stable dose of MTX developed rapidly worsening ILD after initiation of adalimumab [61]. Respiratory improvement followed discontinuation of MTX and adalimumab, and treatment with oral prednisolone.

●Certolizumab (TNF-alpha inhibitor) – In a patient with refractory RA despite MTX and leflunomide, certolizumab was added with subsequent development of dry cough, breathlessness, and basilar crackles [62]. Ground glass and reticular opacities were noted on computed tomography (CT). Despite cessation of certolizumab and administration of systemic glucocorticoids, the patient experienced progressive respiratory failure. Among 4049 patients with RA treated with certolizumab, no instances of drug-induced ILD were reported [63].

●Rituximab (CD-20 antibody) – Several case reports have described ILD associated with rituximab therapy for hematologic malignancies, but only a few have described ILD in patients with RA [64-67]. In a report of patients with underlying RA-associated ILD who were treated with rituximab, one experienced further progression and another developed acute respiratory distress syndrome due to possible pneumonia [66]. A separate report described development of organizing pneumonia in a patient with RA treated with rituximab and MTX [64]. (See "Pulmonary toxicity associated with antineoplastic therapy: Molecularly targeted agents", section on 'Rituximab'.)

●Tocilizumab (interleukin [IL]-6 receptor antibody) – Isolated cases of combined pulmonary fibrosis with emphysema, ILD, acute pneumonitis, and idiopathic pulmonary fibrosis have been reported with tocilizumab (monoclonal antibody to interleukin-6) [68-72].

●Anakinra (IL-1 inhibitor) – Although infrequently used in the treatment of RA, anakinra has been associated with ILD in a small number of patients [73-75].

Treatment of drug-induced ILD due to the various biologic agents requires cessation of the drug; concomitantly administered DMARDs (eg, MTX, leflunomide) will also need to be stopped. If it is unclear which agent caused the lung toxicity, cautious reintroduction of one of the agents may be possible after resolution of symptoms and radiographic changes. Alternatively, a different biologic agent or DMARD may need to be substituted.

The efficacy of systemic glucocorticoids in this setting is not known. For patients who have stable or improving pneumonitis after cessation of the drug, glucocorticoids are generally withheld and the patient observed, as resolution of pulmonary toxicity often accompanies drug discontinuation. In contrast, empiric glucocorticoid therapy is usually initiated in a patient who has rapidly progressive or more severe pulmonary toxicity, although scientific evidence to support this practice is lacking.

As noted, these reports describe the potential for serious, adverse effects with biologic agents, particularly among patients with preexisting RA-induced ILD [20]. Predictors of which patients with RA-ILD are at greatest risk of a severe lung reaction have not been determined, but caution is advised for older adults, a UIP pattern, or severe disease [76]. However, biologic agents carry the potential for significant improvement in overall RA disease activity, so the balance of all factors should be considered when deciding to administer or withhold TNF-alpha blockers in patients with preexisting RA-induced ILD.

Granulomatous lung disease — Lung disease characterized by granuloma formation (both noncaseating and necrotizing) without evidence of mycobacterial or fungal infection has been reported in a number of case reports [77-82]. In the largest series, six patients were on etanercept, two infliximab, and three adalimumab; some were also taking leflunomide or MTX [82]. Symptoms of cough, dyspnea, chest pain, or asthenia were reported by five of the patients. Imaging revealed single, multiple, or cavitary nodules; one patient had hilar adenopathy. Anti-TNF-alpha therapy was discontinued in six patients, but maintained in the others. Two patients were treated with rituximab after discontinuation of anti-TNF-alpha therapy, with resolution of the nodules.

Infections — A variety of serious infections have been described with use of biologic agents to treat RA, especially TNF-alpha inhibitors. The presence of underlying chronic obstructive pulmonary disease (COPD) and the combination of biologic agents with glucocorticoids or other immunomodulatory agents may further increase the risk of infection [15]. Presenting symptoms may be subtle, and a high degree of clinical suspicion for infection should be applied in the assessment of patients treated with these agents.

In one study of 5326 patients, a four-fold increase in hospitalization due to infection was attributed to use of TNF-alpha inhibitors, compared with a two-fold increase that was associated with MTX [16]. The most common infection was bacterial pneumonia. In a separate observational study of 7091 patients taking etanercept, pneumonia occurred in 0.8 percent [56]. In a retrospective study in 146 older adults receiving biological agents for rheumatoid arthritis, the most common severe adverse reaction was infection, which occurred in 32 patients (22 percent) [83]. A variety of opportunistic infections (eg, Listeria, Mycobacteria tuberculosis, Coccidioides, Histoplasma and other fungal species, and cytomegalovirus) are noted in case reports and series. (See "Tumor necrosis factor-alpha inhibitors: Bacterial, viral, and fungal infections" and "Risk of mycobacterial infection associated with biologic agents and JAK inhibitors".)

Rituximab has also been associated with serious infections, including bacterial pneumonia and Pneumocystis pneumonia [15,84-86]. It is not clear if the incidence of most of these infections is greater than that associated with other DMARDs [87].

The incidence of serious infections is approximately doubled when abatacept is added to a TNF-alpha inhibitor [50].

The use of anakinra at high dose may increase the risk of serious infection, although further data are needed [86].

The guidelines from the British Society for Rheumatology assessed the risk of infection with TNF and non-TNF biological agents and found a lower risk of infection with etanercept or abatacept compared with rituximab in patients at high risk of infection [48].

Tuberculosis screening — TNF-alpha inhibitors are associated with an increased risk of reactivation tuberculosis (TB). Appropriate screening before the start of treatment and vigilance for the occurrence of active TB during therapy are essential, although screening does not eliminate the risk of development of active TB. These issues are reviewed separately. (See "Risk of mycobacterial infection associated with biologic agents and JAK inhibitors", section on 'Screening'.)

Sulfasalazine — Sulfasalazine, which is sometimes used in combination therapy regimens for RA, has been associated with pneumonitis, commonly in conjunction with fever and rash [3,88]. Nearly half of affected patients present with the clinical syndrome of pulmonary infiltrates with eosinophilia [88]. Drug reaction with eosinophilia and systemic symptoms (DRESS) is also reported [89,90]. Cough and crackles on lung examination are commonly present. Eosinophil counts in the peripheral blood range from 432 to 7500/mm³. (See "Drug reaction with eosinophilia and systemic symptoms (DRESS)".)

Clinical improvement typically follows cessation of sulfasalazine, while progressive respiratory failure and death occurred in two of three patients who continued the medication. The role of systemic glucocorticoid therapy is not well-studied, as most patients improve with sulfasalazine withdrawal. Among 20 patients treated with systemic glucocorticoids, 13 had complete symptom resolution, 3 improved, 1 had persistent disease, and 3 died of progressive respiratory failure. Rechallenge is not recommended.

Other pulmonary disorders associated with sulfasalazine include nonspecific interstitial pneumonia, organizing pneumonia (formerly known as bronchiolitis obliterans organizing pneumonia), granulomatous lung disease, bronchiolitis obliterans, and rarely pleural effusion [88,91-93].

Nonsteroidal anti-inflammatory drugs — Nonsteroidal anti-inflammatory drugs (NSAIDs) have both analgesic and antiinflammatory properties, but do not alter articular outcomes in RA. Several NSAIDs (eg, ibuprofen, naproxen, diflunisal) have been associated with pulmonary infiltrates with eosinophilia, but not all of these reports were in patients with RA [94-98]. It is not known whether this is a class effect or a unique feature of certain NSAIDs.

Cessation of the implicated agent is indicated, as resolution of pulmonary toxicity often accompanies drug discontinuation. Systemic glucocorticoids are generally withheld as long as pneumonitis is stable or improving. In contrast, empiric glucocorticoid therapy is usually initiated in a patient who has rapidly progressive or more severe pulmonary toxicity, although data to support this practice are limited to case reports. (See "Nonselective NSAIDs: Overview of adverse effects", section on 'Pulmonary infiltrates with eosinophilia'.)

Rarely used medications

Gold — Pneumonitis due to gold therapy is well recognized but uncommon, particularly now that more effective treatments for RA have supplanted the use of gold. Several forms of pulmonary disease occur among patients treated with gold, including nonspecific interstitial pneumonitis, organizing pneumonia, and bronchiolitis obliterans [99]. Of these, nonspecific interstitial pneumonitis is the most common.

Features of gold-induced pneumonitis that may allow differentiation from underlying RA-induced ILD include the presence of fever (50 percent), absence of clubbing (1 percent), BAL lymphocytosis (70 percent) rather than neutrophilia, and extrapulmonary signs of gold toxicity, such as a skin rash (36 percent), eosinophilia (36 percent), liver dysfunction (19 percent), and proteinuria (19 percent) [99]. Gold-induced pneumonitis typically begins within the first six months of therapy.

Gold therapy should be permanently discontinued in all cases. Although data are very limited, systemic glucocorticoid therapy (prednisone 40 to 60 mg per day, followed by a two to six month taper) is usually initiated (after exclusion of infectious etiologies) when respiratory impairment is severe or when drug discontinuation does not lead to improvement [100]. Clinical remission (ie, improved symptoms and normalization of the chest radiograph or pulmonary function tests) occurs in most patients.

Penicillamine — Pulmonary complications related to penicillamine are rare, particularly given that more effective treatments for RA are preferred [1,101]. Original reports of bronchiolitis obliterans in patients with RA suggested that penicillamine was the causative agent, but bronchiolitis obliterans also occurs in the absence of penicillamine in patients with RA, and this association has not been seen in other patients using penicillamine (eg, in Wilson disease).

Bronchiolitis obliterans in patients with RA on penicillamine has been described between 3 and 14 months after the initiation of therapy [1,102]. Patients present with typical clinical, radiographic, and pathologic features of bronchiolitis obliterans, including dyspnea, hyperinflation, airway obstruction, and air trapping. Additional features and treatment of bronchiolitis obliterans are discussed separately. (See "Overview of bronchiolar disorders in adults", section on 'Bronchiolitis obliterans'.)

Other rare complications of penicillamine include membranous nephropathy, drug-induced lupus erythematosus, pulmonary hemorrhage, and interstitial fibrosis [17]. (See "Drug-induced lupus" and "Membranous nephropathy: Pathogenesis and etiology", section on 'Drugs'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Rheumatoid arthritis" and "Society guideline links: Interstitial lung disease".)

SUMMARY AND RECOMMENDATIONS

●Most of the drugs used to treat rheumatoid arthritis (RA) have been reported to cause pneumonitis, including methotrexate, leflunomide, biologic agents, sulfasalazine, gold, penicillamine, and nonsteroidal antiinflammatory drugs (NSAIDs). Many of the pulmonary reactions to drugs used for the treatment of rheumatoid arthritis are rare and are published as case reports, but have a high mortality. (See 'Prevalence' above.)

●Pulmonary disease may infrequently occur with the relatively low doses of methotrexate (<20 mg per week) that are used in patients with RA. There is no clear evidence that methotrexate contributes to chronic progressive fibrosing pneumonitis. (See 'Methotrexate' above.)

●In patients with methotrexate-induced pneumonitis, the chest radiograph typically demonstrates diffuse bilateral reticular opacities or mixed reticular and ground glass patterns (40 percent). Unilateral or nodular changes, effusions, and bilateral hilar lymphadenopathy are uncommon. Early drug discontinuation at the onset of respiratory symptoms may obviate the need for invasive testing. (See 'Methotrexate' above and "Methotrexate-induced lung injury".)

●Interstitial pneumonitis and new and accelerated pulmonary rheumatoid nodule formation have been described with leflunomide, although the risk appears low. We generally avoid leflunomide in patients with underlying lung disease or a history of methotrexate-induced lung toxicity, but gathering evidence suggests this may not be necessary. (See 'Leflunomide' above.)

●New onset or progressive pneumonitis may uncommonly occur in patients treated with the tumor necrosis factor-alpha (TNF-alpha) blockers. A growing number of other biologic agents are being used to treat RA; further study is needed to determine whether they also cause or contribute to RA-associated interstitial pneumonitis. (See 'Inflammatory pneumonitis' above.)

●TNF-alpha inhibitors are associated with an increased risk of reactivation tuberculosis (TB) and other opportunistic infections. Appropriate screening before the start of treatment and vigilance for the occurrence of active TB during therapy are essential. (See 'Tuberculosis screening' above.)

●Granulomatous lung disease and hilar adenopathy are also associated with TNF-alpha inhibitors in the absence of demonstrable mycobacterial or fungal infection. Beyond careful exclusion of infection, the optimal management is not known. (See 'Granulomatous lung disease' above.)

●The rarely used medications, gold and penicillamine, have been associated with interstitial lung disease and bronchiolitis, respectively. (See 'Rarely used medications' above.)