Methotrexate-induced lung injury

INTRODUCTION — Methotrexate is an analogue of the vitamin folic acid; it inhibits cellular proliferation by inducing an acute intracellular deficiency of certain folate coenzymes (figure 1) [1-3]. This impairs the intracellular trafficking of single carbon groups and results in impaired synthesis of thymidine, deoxyribonucleic acid (DNA), and ribonucleic acid (RNA) [1,4]. In addition to its antiproliferative effects, methotrexate has antiinflammatory and immunomodulating properties [2,3,5-7]. It is used to treat a variety of malignancies, connective tissue diseases, and also psoriasis. Serious toxicity from methotrexate may affect the lungs, liver, and bone marrow [1,2,8-11].

This topic review will review the pulmonary injury that may result from methotrexate use. Other side effects of methotrexate therapy and an approach to pulmonary toxicity associated with antineoplastic agents are discussed separately. (See "Major side effects of low-dose methotrexate" and "Hepatotoxicity associated with chronic low-dose methotrexate for nonmalignant disease" and "Pulmonary toxicity associated with systemic antineoplastic therapy: Clinical presentation, diagnosis, and treatment" and "Therapeutic use and toxicity of high-dose methotrexate", section on 'Overview of adverse effects' and "Use of methotrexate in the treatment of rheumatoid arthritis", section on 'Mechanism of action'.)

EPIDEMIOLOGY — Lung toxicity most often occurs after weeks to months of low-dose oral methotrexate therapy (as is typically used for non-malignant disease), but can occur following relatively short term use of intravenous or intrathecal administration of higher doses [10,12-14]. In a literature review of 123 cases of methotrexate pneumonitis, about one-half arose in patients receiving therapy for rheumatoid arthritis (range 2.5 to 15 mg weekly), about 20 percent arose during intensification/consolidation treatment for leukemia (doses approximately 20 to 80 mg weekly), and 8 percent were in patients treated for other malignancies (weekly doses ranging from 15 to 1400 mg) (table 1) [10].

The precise frequency with which methotrexate pulmonary toxicity occurs is difficult to assess as some reports have included patients who were receiving other cytotoxic medications, had ongoing infectious diseases, or had underlying disease processes capable of involving the lungs and pleura [1]. In addition, impure preparations of methotrexate may have played a role in some of the earlier reports of toxicity [1]. Many series estimate that acute pulmonary toxicity develops in 1 to 8 percent of patients receiving methotrexate for rheumatologic condition, including rheumatoid arthritis, but some reports suggest an incidence as high as 33 percent [14-22]. On the other hand, others suggest that rates are much lower because not all cases of pneumonitis occurring in patients treated with methotrexate are directly attributable to the drug. The following examples show the frequency of pneumonitis among patients who received methotrexate for rheumatoid arthritis or other inflammatory diseases:

●In a systematic literature review of 3463 patients with rheumatoid arthritis who were receiving methotrexate, 84 patients (2 percent) had some type of lung toxicity, but only 15 were felt to be definitive cases of pneumonitis attributable to methotrexate (0.43 percent) [23]. The mean duration of methotrexate use was 36.5 months and the average dose was 8.8 mg/week.

●In a literature review of studies that compared methotrexate to placebo or another agent in patients with psoriasis, psoriatic arthritis, or inflammatory bowel disease, seven trials (1630 subjects) were identified in which treatment lasted 16 to 52 weeks [24]. Adverse respiratory events were not statistically increased in those taking methotrexate; only one patient developed pneumonitis and that person was in the methotrexate group. While the data are reassuring for patients with these diseases who may take methotrexate, the relatively low total number of patients means that a small but clinically important risk cannot be completely excluded.

The association of methotrexate use with the development of chronic interstitial lung disease in patients with rheumatoid arthritis (RA-ILD) remains controversial. While many pulmonologists and rheumatologists avoid the use of methotrexate in rheumatoid arthritis patients with high resolution chest computed tomography (HRCT)-diagnosed ILD, some studies suggest possible benefit for methotrexate treatment with a decreased risk of developing RA-ILD and a delayed onset, as described separately. (See "Drug-induced lung disease in rheumatoid arthritis", section on 'Methotrexate'.)

PATHOGENESIS AND PATHOLOGY — Pulmonary complications of methotrexate may be classified as inflammatory, infectious, and lymphoproliferative [25].

Inflammatory and fibrotic lung disease — Hypersensitivity pneumonitis is the most common type of pulmonary toxicity associated with methotrexate and is characterized by a lymphocytic infiltration of the interstitium with epithelial cell hyperplasia, small, poorly formed granulomas, and sometimes eosinophilic infiltration [1-3,9,17]. This pattern is similar to the hypersensitivity pneumonitis associated with inhaled organic antigens. (See "Hypersensitivity pneumonitis (extrinsic allergic alveolitis): Epidemiology, causes, and pathogenesis", section on 'Etiologic agents' and "Hypersensitivity pneumonitis (extrinsic allergic alveolitis): Clinical manifestations and diagnosis", section on 'Surgical lung biopsy'.)

A number of other types of lung injury have been associated with methotrexate use, including [1-3,15]:

●Organizing pneumonia (formerly called bronchiolitis obliterans with organizing pneumonia or BOOP)

●Acute interstitial pneumonia with noncardiogenic pulmonary edema

●Pulmonary fibrosis (which may be rapidly progressive) [9,26]

●Pleural effusion and symptomatic pleuritis occur infrequently [14]

In advanced methotrexate pneumonitis, the lungs show extensive fibrosis and honeycombing, especially at the bases [9]. Approximately 10 percent of patients will demonstrate fibrosis on lung biopsy.

The precise mechanisms by which methotrexate causes pulmonary injury are unknown. Most researchers suggest that methotrexate pneumonitis is a form of hypersensitivity lung disease because there is typically fever, eosinophilia, an increase in CD4+ T cells in bronchoalveolar lavage fluid, and a mononuclear cell infiltration of the lungs with granulomatous inflammation [1,3]. Others suggest that injury may result from a direct toxic effect of methotrexate on lung tissue, while a third theory proposes that impaired host resistance to acquired or latent viral infection (eg, cytomegalovirus or Epstein-Barr virus) underlies the observed pathology [27-30]. Support for direct toxicity comes from a mouse model that demonstrated alveolar epithelial injury, apoptosis, and fibrosis [31]. Finally, activation of mitogen-activated protein (MAP) kinase pathways and altered cytokine expression may contribute to the pulmonary inflammatory response [27,30].

Pulmonary toxicity has occurred following both low and high doses and by a variety of routes of administration, suggesting that some of the pulmonary adverse effects may result from idiosyncratic mechanisms unrelated to folate antagonism. None of the proposed mechanisms account for the observation that pulmonary toxicity may remit despite continued therapy, and may not occur upon rechallenge [1-3]. In summary, the mechanisms of methotrexate pulmonary toxicity are unresolved and may be multiple.

Pulmonary infections — Methotrexate can compromise the immune response and increase the risk for opportunistic infections due to Pneumocystis jirovecii (formerly Pneumocystis carinii), cytomegalovirus, varicella-zoster virus, Nocardia, mycobacteria, or other fungi [32,33]. Pneumocystis has been the most commonly reported agent and has accounted for up to 40 percent of the infectious complications in some series [17,33,34]. (See "Epidemiology of pulmonary infections in immunocompromised patients".)

Pulmonary lymphoproliferative disease — Lymphoproliferative disease (including non-Hodgkin [B cell] and other lymphomas) may appear during methotrexate therapy and regress after methotrexate discontinuation [33,35-40]. Among 48 patients with rheumatoid arthritis receiving methotrexate who developed a lymphoproliferative disease, the primary site was the lung in four [37]. In a separate report, 6 of 28 cases of lymphoproliferative disease involved the lung or pleura, but details were not provided [36].

The reversibility of this disease without specific anti-lymphoma treatment suggests that diminished immune surveillance due to methotrexate may facilitate the development and expansion of malignant lymphoid clones. However, whether methotrexate-immunosuppression has a direct effect on the malignant cells or an indirect effect via recrudescence of Epstein-Barr virus (EBV) infection is not known. A portion of these cases (28 percent) are associated with EBV infection, a finding that is also seen in patients who are immunosuppressed in the setting of organ transplantation or AIDS [37]. (See "Treatment and prevention of post-transplant lymphoproliferative disorders" and "Major side effects of low-dose methotrexate", section on 'Risk of lymphoproliferative disorders'.)

The causal relationship between lymphoma and methotrexate use has been difficult to prove due to the increased rate of lymphoma among patients with rheumatoid arthritis. Large database studies suggest that it is unlikely that long-term oral methotrexate therapy increases the risk of lymphoma [23,41-43]. As an example, in a prospective study of 19,562 patients (89,710 person-years of follow-up), with rheumatoid arthritis, similar risks of lymphoma were observed among the 68 percent of patients who received methotrexate versus those who did not [43]. However, there was no adjustment for the severity of the arthritis, a known factor in the risk of lymphoma in these patients [23]. Thus, the relationship between methotrexate use and lymphoma in patients with conditions such as rheumatoid arthritis remains uncertain. However, there are also well-documented cases of regression of lymphoma when methotrexate is held. (See "Disease outcome and functional capacity in rheumatoid arthritis", section on 'Lymphoproliferative disorders'.)

RISK FACTORS — A number of studies have identified risk factors for methotrexate pulmonary toxicity. One multicenter case-control study involving 36 patients with methotrexate-induced lung injury following treatment for rheumatoid arthritis and 82 matched controls identified the following risk factors [44]:

●Age greater than 60 years

●Rheumatoid pleuropulmonary involvement

●Previous use of disease-modifying antirheumatic drugs

●Hypoalbuminemia (either before or during therapy)

●Diabetes mellitus

The mechanism by which some of these factors may confer excess risk is unclear, although there is evidence that hyperinsulinemia, which may occur with treatment for diabetes mellitus, is associated with increased polyglutamation of methotrexate [45]. The previous use of disease-modifying antirheumatic drugs may be a marker for more severe rheumatoid arthritis, and hypoalbuminemia could potentially result in a lower degree of protein binding and higher free levels of methotrexate.

Additional risk factors for methotrexate pneumonitis have been suggested by others, including:

●Higher weekly doses of methotrexate [46]

●Daily, rather than weekly drug administration [1]

●Preexisting lung disease [29,47] (see 'Prevention' below)

●Abnormal pulmonary function tests prior to therapy [47]

●Decreased elimination of methotrexate (eg, as seen in renal insufficiency or with the presence of third-space fluid collections such as ascites) [9,48]

●Inherited polymorphisms in some gene alleles related to methotrexate transport and metabolism may be associated with increased risk of toxicity, but further studies are needed to validate these associations [49]

CLINICAL MANIFESTATIONS — Methotrexate pulmonary toxicity may present in an acute, subacute, or chronic form. Subacute presentations are most common. The majority of patients who develop methotrexate pulmonary toxicity during chronic oral low-dose therapy do so within the first year of treatment (mean nine months) [14], although cases have been reported as early as 12 days and as late as 18 years after the drug was initiated [14,17-20,28,29,50].

●The clinical presentation of acute methotrexate pneumonitis is generally nonspecific, with symptoms (fever, chills, malaise, nonproductive cough, dyspnea, chest pain) that are progressive over several days [10,14,18-20,28,29]. Rapid progression to respiratory failure may occur [1,3].

●Subacute pneumonitis manifests with the insidious onset of dyspnea (82 percent), nonproductive cough (81 percent), fever (76 percent), crackles (50 percent), and cyanosis over a few weeks [10,28,51]. Approximately 17 percent of patients also have cutaneous manifestations of methotrexate toxicity [28]. Progression to pulmonary fibrosis is observed in approximately 10 percent of patients.

●Rarely, methotrexate is associated with a more chronic presentation of dyspnea, cough, and coarse crackles that progress to respiratory failure over weeks to months [10,26,52].  

The clinical manifestations of pulmonary involvement in methotrexate-associated lymphoproliferative disorder are poorly described [36,37]. In a series of 48 patients with rheumatoid arthritis receiving methotrexate who developed a lymphoproliferative disease, four had a primary site in the lung but the exact pattern of lung involvement was not delineated [37].

The extrapulmonary toxicities of methotrexate are listed in the table and are discussed separately (table 2). (See "Major side effects of low-dose methotrexate".)

DIFFERENTIAL DIAGNOSIS — The typical symptoms and signs of methotrexate-induced lung injury (eg, dyspnea, cough, diffuse radiographic opacities and fever) are nonspecific and may be caused by intercurrent infection (eg, Pneumocystis, Cryptococcus, cytomegalovirus, herpes zoster, Nocardia, mycobacteria), the underlying disease (eg, rheumatoid arthritis), malignancy (eg, lymphangitic tumor, lymphoproliferative disease), or a concomitant medication. Bacterial, fungal, and viral infections are more likely to have an acute presentation of fever, dyspnea, and cough with or without sputum production, while mycobacterial disease is usually more insidious in onset. (See "Approach to the immunocompromised patient with fever and pulmonary infiltrates", section on 'Initial evaluation' and 'Pulmonary infections' above and 'Pulmonary lymphoproliferative disease' above and "Overview of pleuropulmonary diseases associated with rheumatoid arthritis".)

As both usual and opportunistic lung infections have been described in patients taking methotrexate, it is imperative to exclude the possibility of pulmonary infection early in the course of diagnosis and prior to initiating treatment with glucocorticoids [34]. (See 'Pulmonary infections' above and "Approach to the immunocompromised patient with fever and pulmonary infiltrates", section on 'Etiology of pulmonary infiltrates'.)

DIAGNOSTIC EVALUATION — While no test is absolutely diagnostic of methotrexate-induced pneumonitis, some tests can help in the inclusion or exclusion of other processes. Radiographic evaluation, a trial of drug cessation, bronchoalveolar lavage, and lung biopsy are the primary ways to narrow the differential diagnosis when methotrexate-induced lung disease is suspected. However, bronchoscopy and lung biopsy are not necessary in most patients. A beneficial response to methotrexate withdrawal in the appropriate clinical setting may be sufficient in the patient who is not acutely or severely ill.

Laboratory testing — Laboratory testing (eg, complete cell counts, coagulation tests, B-type natriuretic peptide [BNP], blood cultures, sputum cultures, viral serology) is performed in most patients to determine whether other disease processes are contributing to the patient's respiratory compromise.

Up to 50 percent of patients with subacute onset of methotrexate lung toxicity demonstrate peripheral blood eosinophilia, which strongly supports the diagnosis, when present [28].

Imaging — Diffuse lung parenchymal opacities are usually the earliest radiographic finding in methotrexate lung toxicity, and they may progress rapidly to patchy acinar consolidation [28,53,54]. Nodular or ill-defined acinar opacities may be diffusely scattered throughout the lung fields. Less typical chest radiographic abnormalities include hilar lymphadenopathy, atelectasis, and pleural effusions [4,52,54]. The chest film is rarely normal when the disease is fully established, but may be normal in the early stages of dry cough. Patients who develop a nonproductive cough on methotrexate, but do not have changes on radiographic studies, often improve with methotrexate discontinuation, supporting the view that methotrexate lung toxicity can be radiographically silent.

In general, high resolution computed tomographic (HRCT) scanning is superior to conventional chest radiographs for characterizing the pattern, extent, and location of lung involvement and is performed in most patients [55,56]. (See "Pulmonary toxicity associated with systemic antineoplastic therapy: Clinical presentation, diagnosis, and treatment", section on 'Imaging'.)

A variety of HRCT patterns have been described with methotrexate-induced toxicity, all of which can be seen with infection.

●Patchy areas of ground glass attenuation with small area of consolidation [55,57]

●Diffuse ground glass opacities with or without reticulation or consolidation [58]

●Ill-defined centrilobular nodules [58]

●Bilateral, symmetric, lower lung zone, irregular, reticular opacities, traction bronchiectasis and architectural distortion with or without consolidation or honeycombing [26,57]

The degree of HRCT abnormality correlates quantitatively with lung volume reductions on pulmonary function testing [55,59].

The radiographic manifestations of pulmonary lymphoproliferative disease associated with rheumatoid arthritis and methotrexate are poorly described; pleural effusion has been reported in a few cases in addition to single or multiple nodular opacities [14,33,35-38,60].

Pulmonary function testing — Pulmonary function tests (PFTs) are helpful for characterizing the pattern (eg obstructive, restrictive) and severity of respiratory impairment in patients with respiratory symptoms. Among patients suspected of having methotrexate-related lung injury, PFTs are primarily performed in those who present with an indolent onset of dyspnea or cough during methotrexate therapy, rather than a more fulminant presentation or respiratory compromise.

PFTs in patients with methotrexate pneumonitis typically reveal a restrictive pattern with a decrease in the diffusing capacity for carbon monoxide (DLCO), hypoxemia, and an increased alveolar-arterial (A-a) gradient (calculator 1) [1,3]. Some reports have noted worsening of airflow obstruction or development of a restrictive pattern with long-term, low-dose methotrexate [61-63], but other reports have not found a clinically significant decrease in pulmonary function in the majority of patients [50,61,64].

The utility of screening pulmonary function tests for the prevention or early detection of methotrexate pneumonitis is not well established.

●As interstitial lung disease is a risk factor for methotrexate pneumonitis, some experts recommend baseline pulmonary function testing prior to initiating chronic methotrexate to identify underlying interstitial lung disease and to use as a baseline in case of a later onset of symptoms [65]. We typically perform baseline pulmonary function tests prior to initiation of methotrexate, if the patient reports dyspnea on exertion or if the pretreatment chest radiograph is abnormal. We avoid use of methotrexate in patients with more than mild respiratory impairment at baseline (eg, DLCO <70 percent predicted) [65].

●The value of routine screening for methotrexate lung toxicity was assessed in a prospective study of 124 patients with rheumatologic diseases receiving long-term oral methotrexate. Surveillance spirometry, lung volumes, and DLCO were not helpful in the preclinical diagnosis of methotrexate pneumonitis [66].

Bronchoscopy — Bronchoscopy with bronchoalveolar lavage (BAL) is performed in patients with new onset lung disease complicating methotrexate therapy, when it is necessary to exclude processes such as infection or malignancy (eg, in the presence of fever, widespread or nodular opacities on chest imaging, or rapidly progressive respiratory impairment). BAL is more helpful in ruling out processes such as infection and lymphangitic tumor spread, than in making a definitive diagnosis of methotrexate pneumonitis.

BAL samples are sent for bacterial, fungal, and mycobacterial smear, special stains, and culture. Samples are also sent for cytologic examination for viral inclusion bodies and malignant cells. Samples are sent for viral culture and immunohistochemistry, where available. (See "Pulmonary toxicity associated with systemic antineoplastic therapy: Clinical presentation, diagnosis, and treatment", section on 'Bronchoscopy' and "Basic principles and technique of bronchoalveolar lavage".)

In the BAL fluid of patients with methotrexate pneumonitis, cell counts reveal lymphocytic predominance (eg, 33 to 68 percent) with an increase in the number of CD4 lymphocytes and the CD4/CD8 ratio [67,68]. These findings are not specific and can be seen in a variety of inflammatory lung processes, including sarcoidosis, berylliosis, rheumatoid arthritis, and miliary tuberculosis; other forms of hypersensitivity pneumonitis may yield similar results, although the CD4/CD8 ratio is usually decreased [67,68]. (See "Role of bronchoalveolar lavage in diagnosis of interstitial lung disease".)

Atypical epithelial cells are sometimes seen in BAL fluid from patients with pulmonary fibrosis due to methotrexate [1,3,9]. This manifestation may be detected earlier in the course of the disease than histologic evidence of fibrosis on lung biopsy.

In patients without specific contraindications, transbronchial lung biopsy may improve the likelihood of identifying lymphangitic spread of tumor or an invasive infection over BAL alone. In addition, biopsy allows distinction between colonization and invasion in fungal (or viral) infection. (See "Approach to the immunocompromised patient with fever and pulmonary infiltrates", section on 'Lung sampling' and "Flexible bronchoscopy in adults: Indications and contraindications", section on 'Contraindications' and "Flexible bronchoscopy in adults: Overview" and "Flexible bronchoscopy in adults: Associated diagnostic and therapeutic procedures", section on 'Transbronchial biopsy'.)

Lung biopsy — Pulmonary histopathology may be helpful in establishing the diagnosis of methotrexate-induced pulmonary toxicity, but is not always required. Patients who respond quickly to drug discontinuation generally do not need a lung biopsy. 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. Surgical lung biopsy can be obtained by video-assisted thoracoscopic surgery or open thoracotomy, depending on the expertise and preference of the surgeon. (See "Role of lung biopsy in the diagnosis of interstitial lung disease", section on 'Surgical lung biopsy'.)

Lung biopsy can help to distinguish methotrexate-induced lung disease from other processes, such as infection, neoplasia, and lung involvement by the underlying disease (eg, rheumatoid arthritis), although this is not always possible. In a patient with rheumatoid arthritis, for example, poorly formed granulomas with scattered eosinophils, consistent with hypersensitivity pneumonitis due to methotrexate, are distinct from the lung injury usually associated with rheumatoid arthritis. The latter is typically characterized by a usual interstitial pneumonia (idiopathic pulmonary fibrosis) or nonspecific interstitial pneumonia pattern (NSIP). On the other hand, acute interstitial pneumonitis (AIP) and organizing pneumonia (OP) have been described with methotrexate alone and with rheumatoid arthritis in the absence of methotrexate; thus, the finding of AIP or OP does not differentiate these processes. (See 'Inflammatory and fibrotic lung disease' above and "Interstitial lung disease in rheumatoid arthritis", section on 'Lung biopsy'.)

Making the diagnosis — The diagnosis of methotrexate-induced pulmonary toxicity is based on the combination of the appropriate clinical setting, clinical manifestations, radiographic abnormalities, bronchoalveolar lavage findings, lung histology (when available), and response to drug discontinuation.

Given the absence of a definitive diagnostic test, a set of diagnostic criteria has been proposed to establish a diagnosis of methotrexate-induced lung injury [4,14].

In patients receiving methotrexate, the major criteria are:

●Hypersensitivity pneumonitis by histopathology without evidence of pathogenic organisms

●Radiographic evidence of patchy or diffuse pulmonary ground glass or consolidative opacities

●Blood cultures (if febrile) and initial sputum cultures (if sputum is produced) that are negative for pathogenic organisms.

The minor criteria are:

●Shortness of breath for less than eight weeks

●Nonproductive cough

●Oxygen saturation ≤90 percent on room air at the time of initial evaluation

●DLCO ≤70 percent of predicted for age

●Leukocyte count ≤15,000 cells/mm³

Methotrexate pneumonitis is characterized as "definite" if major criteria 1 or 2 and major criterion 3 are present in conjunction with three of the five minor criteria. "Probable" methotrexate pneumonitis is present if major criteria 2 and 3 plus two of the five minor criteria are present.

While this combination of factors can support a diagnosis of methotrexate-induced lung injury, the scoring system is predominantly useful in assuring comparable patient populations in different clinical trials. The clinical utility of the scoring system has not been adequately validated, and it should not be strictly relied upon to establish the diagnosis in a given patient [4,14].

TREATMENT — The optimal therapy for methotrexate-induced lung toxicity has not been established and no prospective trials of therapy have been performed. The treatments for opportunistic infections in the lung and lymphoproliferative disease are discussed separately. (See "Approach to the immunocompromised patient with fever and pulmonary infiltrates", section on 'Selection of initial therapy' and "Major side effects of low-dose methotrexate", section on 'Risk of lymphoproliferative disorders'.)

Initial management

●It is mandatory to exclude an infectious etiology of the pulmonary findings [12]. Empiric antimicrobial therapy directed at likely pathogens may be indicated while definitive procedures and cultures are performed.

●Discontinuation of methotrexate is the initial step in management. In most cases, clinical improvement occurs within days of stopping the drug, followed by radiographic improvement over several weeks [4]. While there are reports of successful rechallenge with methotrexate and even regression of pulmonary toxicity despite continued therapy, neither of these approaches is recommended [1,3,5,9,14,20,53].

●Supportive care includes supplemental oxygen as indicated by pulse oximetry and mechanical ventilation, if clinically indicated.

●Systemic glucocorticoids are used in selected patients.

Glucocorticoids — The clinical and histological similarities of methotrexate-induced pneumonitis to hypersensitivity pneumonitis imply a role for glucocorticoids. No prospective trials of glucocorticoid therapy have been performed, but clinical experience and case reports suggest that use of glucocorticoids accelerates the recovery of pneumonitis [10,28,69-71], although fatalities may still occur [71]. The efficacy of glucocorticoid therapy is illustrated by the following reports:

●In one report of nine patients with otherwise unexplained symptoms who experienced a clinical course consistent with a drug-induced lung reaction to methotrexate, three had a moderate severity of findings and fully recovered with discontinuation of methotrexate [70]. All six patients who had marked respiratory compromise recovered fully with methotrexate discontinuation and high-dose glucocorticoids.

●In a second series of nine patients with presumed methotrexate pneumonitis, methotrexate was discontinued in all, and glucocorticoids given to eight [10]. One patient's illness resolved completely after 5 days, five improved steadily over the course of 3 to 20 weeks; one continued to have respiratory insufficiency requiring periodic oxygen supplementation at 19 months; and two died of respiratory failure.

●Among six patients who developed "life-threatening" methotrexate-induced pneumonitis, drug discontinuation and administration of high-dose glucocorticoids were associated with improvement in five, but death in one [16].

●In a third report, one of three patients with methotrexate-induced pneumonitis died despite institution of high-dose methylprednisolone; the others recovered completely [28].

Thus, although systemic glucocorticoid therapy may accelerate recovery, it does not appear to prevent fatalities. We typically reserve glucocorticoids for patients with more severe initial pulmonary toxicity (eg, dyspnea at rest, a decrease in oxygen saturation below 90 percent or more than a 4 percent decrease from baseline) or worsening clinical status despite methotrexate discontinuation.

Systemic glucocorticoid therapy is typically initiated with prednisone 1 mg/kg per day orally or the equivalent (table 3). If the patient is severely ill, intravenous glucocorticoids (eg, methylprednisolone) 1 mg/kg once or twice a day are administered and transitioned to oral prednisone as the patient stabilizes. Tapering of glucocorticoid therapy is dependent upon the clinical response, as assessed by improvement in the clinical, physiologic (pulmonary function tests), and radiographic parameters, but generally occurs over a few weeks [5,53].

For patients with advanced pulmonary fibrosis (eg, extensive reticular opacities and honeycombing on HRCT or fibrosis on biopsy) and a slowly progressive course, we are cautious in administering systemic glucocorticoids due to their known side effects and unlikely benefit.

The potential adverse effects of glucocorticoid therapy are reviewed separately. (See "Major adverse effects of systemic glucocorticoids".)

Prognosis — Most patients recover from methotrexate pneumonitis, although some physiologic abnormalities may persist [1,10]. Thus, aggressive supportive care for acute respiratory failure is reasonable while glucocorticoids are administered. A systematic review reported a mortality rate of 13 percent, although other series have reported lower rates [1,10,52]. The cause of the discrepancy in reported fatality rates is unclear.

Rechallenge — A high rate of relapse of lung disease has been reported with methotrexate rechallenge, although individual patients who tolerate rechallenge have been reported [14]. In addition, several cases of fatal lung toxicity have been reported with rechallenge. Thus, rechallenge is avoided, and an alternate form of therapy employed, if possible.

PREVENTION — There is no established method for preventing methotrexate lung toxicity. Lung toxicity is less common with lower doses of methotrexate, but still occurs. To reduce the risk of pulmonary toxicity, dose adjustments of methotrexate may be needed in patients who have renal or hepatic dysfunction, a third space fluid collection (ie, ascites or pleural effusion), or are elderly [48]. (See 'Risk factors' above and "Nephrotoxicity of chemotherapy and other cytotoxic agents".)

●Folic acid and leucovorin – Many adverse reactions to methotrexate (eg, stomatitis, myelosuppression) can be alleviated or prevented by the addition of either folic acid or leucovorin (also called folinic acid). However, the use of these agents does not reduce the risk for methotrexate pulmonary toxicity. (See "Major side effects of low-dose methotrexate", section on 'Prevention of side effects with folate' and "Therapeutic use and toxicity of high-dose methotrexate", section on 'Rationale for leucovorin rescue'.)  

●Screening – Given that underlying pulmonary fibrosis may be a risk factor for developing methotrexate lung toxicity, most experts obtain a chest radiograph prior to initiating long-term methotrexate therapy [14,63,65,72]. If the chest radiograph suggests underlying interstitial lung disease, further evaluation with pulmonary function testing and high resolution computed tomography may be indicated. (See 'Pulmonary function testing' above and "Interstitial lung disease in rheumatoid arthritis", section on 'Evaluation'.)

An alternate agent instead of methotrexate is preferred for patients who have evidence of underlying interstitial lung disease. The baseline radiograph can also be useful for future comparison should a patient develop dyspnea while on methotrexate.

SUMMARY AND RECOMMENDATIONS

●Epidemiology and risk factors – Methotrexate-induced lung toxicity most often develops after chronic low-dose therapy, but can occur following short-term administration of high-dose methotrexate. Risk factors include age over 60, rheumatoid pleuropulmonary involvement, previous use of disease-modifying antirheumatic drugs, hypoalbuminemia, higher weekly doses of methotrexate, pre-existing lung disease, and decreased elimination of methotrexate. (See 'Epidemiology' above and 'Risk factors' above.)

●Clinical presentation – Subacute presentations are most common. The acute presentation typically includes fever, chills, malaise, nonproductive cough, dyspnea, and chest pain; the subacute presentation is characterized by a more insidious onset of dyspnea, cough, and fever. The majority of patients who develop methotrexate pulmonary toxicity do so within the first year of therapy. (See 'Clinical manifestations' above.)

●Evaluation – While no test is absolutely diagnostic of methotrexate-induced pneumonitis, some tests can help in the inclusion or exclusion of other processes.  

•Laboratory testing (eg, complete cell counts and differential, coagulation tests, brain natriuretic peptide [BNP], blood cultures, sputum cultures, viral serology) is used to determine whether other disease processes are contributing to the patient's respiratory compromise. Up to 50 percent of patients with subacute methotrexate lung toxicity have peripheral blood eosinophilia. (See 'Laboratory testing' above.)

•Radiographic patterns on high resolution computed tomography (HRCT) vary, and may include patchy or diffuse ground glass attenuation, ill-defined centrilobular nodules, and bilateral, reticular opacities associated with traction bronchiectasis and architectural distortion. (See 'Imaging' above.)

•Bronchoscopy with bronchoalveolar lavage (BAL) may be indicated in patients whose symptoms and pattern of onset suggest possible infection or malignancy (eg, acute onset, fever, widespread or nodular radiographic opacities, moderate to severe respiratory impairment). BAL is more helpful in ruling out an infectious or neoplastic etiology than in making a definitive diagnosis of methotrexate pneumonitis. (See 'Bronchoscopy' above.)

•In patients without specific contraindications, transbronchial lung biopsy may improve the likelihood of identifying lymphangitic tumor spread or an invasive infection over BAL alone. Surgical lung biopsy is indicated when the patient has progressive or severe respiratory impairment and the cause of the pneumonitis is uncertain, or if methotrexate-induced lymphoproliferative disease is suspected. (See 'Bronchoscopy' above and 'Lung biopsy' above.)

●Diagnosis – The diagnosis of methotrexate-induced pulmonary toxicity is typically based on the combination of the appropriate clinical setting, clinical manifestations, radiographic abnormalities, and either the response to drug withdrawal or the results of BAL and lung histology. A beneficial response to methotrexate withdrawal in the appropriate clinical setting may be sufficient in the patient who is not acutely or severely ill. (See 'Diagnostic evaluation' above.)

●Differential diagnosis – The differential diagnosis of methotrexate-induced lung injury includes opportunistic infection (eg, Pneumocystis, Cryptococcus, cytomegalovirus, herpes zoster, Nocardia), the underlying disease (eg, rheumatoid arthritis), malignancy (eg, lymphangitic tumor, lymphoproliferative disease), and lung toxicity from a concomitant medication. (See 'Differential diagnosis' above.)

●Treatment – The management of suspected methotrexate lung injury starts with discontinuation of the drug. In most cases, clinical improvement occurs within days to weeks; radiographic improvement may take several weeks. (See 'Treatment' above.)

•For most patients with an acute or subacute symptom presentation and moderate to severe respiratory impairment (eg, a decrease in room air oxygen saturation below 90 percent or more than a 4 percent decrease from baseline), we suggest systemic glucocorticoid therapy, rather than observation and supportive care alone (Grade 2C). Among the factors to consider in the decision to initiate glucocorticoid therapy are the rapidity of clinical deterioration worsened respiratory status despite discontinuation of methotrexate, and the presence of serious comorbid conditions that would increase the risk of serious adverse effects from glucocorticoids (eg, infection, myelosuppression, uncontrolled diabetes mellitus).

The usual initial dose of glucocorticoid is prednisone 1 mg/kg per day orally (or the equivalent); for impending respiratory failure, methylprednisolone 1 mg/kg is administered once or twice a day intravenously (table 3). Tapering is guided by the clinical response. (See 'Treatment' above and "Major adverse effects of systemic glucocorticoids".)

•For patients with evidence of advanced pulmonary fibrosis (eg, extensive honeycombing on HRCT or biopsy), we avoid systemic glucocorticoids because of the known side effects and uncertain benefit. (See 'Treatment' above.)

●Avoidance of rechallenge – Methotrexate rechallenge should be avoided, due to the high risk of recurrent disease. (See 'Rechallenge' above.)