INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: MTX-pneumonitis; lymphocytic alveolitis; CD4/CD8 T-cell ratio

Accurate diagnosis is vital with methotrexate-induced pneumonitis (MTX-pneumonitis), which may be life-threatening, leading to death in about 10% of cases (1), and because withdrawal of MTX may be hazardous, depending on the severity of the underlying disease, i.e., hematologic or gynecologic malignancy and germ-cell tumors. Unfortunately, this diagnosis is difficult to establish. Clinical features, i.e., fever, cough, and dyspnea as well as radiologic findings, i.e., diffuse interstitial pneumonitis are nonspecific. A large number of possible etiologies might be responsible for this clinical and radiologic presentation in these patients, i.e., pulmonary infections (2, 3), underlying disease, pulmonary involvement, pulmonary hemorrhage, and other drug-related pulmonary toxicities or radiation pneumonitis. The pathologic lesions are also nonspecific and require open lung biopsy, an aggressive procedure in such fragile patients.

Nevertheless, fiberoptic bronchoscopy with bronchoalveolar lavage (BAL) has been proposed for the diagnosis of MTX-pneumonitis (4) as used to diagnose opportunistic infection (2) or pulmonary hemorrhage (7) in immunocompromized patients with diffuse interstitial pneumonitis. In addition, results of BAL fluid analysis might also be suggestive of drug-induced pneumonitis when eosinophilic or lymphocytic alveolitis (8, 9), especially of the CD8 subset, is observed. In this context, severe lymphocytic alveolitis has been described in several reports of well-documented MTX-pneumonitis, but the alveolar CD4 to CD8 T-cell ratio (CD4/CD8) was curiously variable (4, 10). To confirm these findings, we retrospectively reviewed the clinical characteristics and BAL cell counts of 14 consecutive cases of MTX-pneumonitis and tried to determine factors contributing to the observed spectrum of alveolar CD4/CD8 T-cell ratios.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical, radiologic, and BAL cell data were retrospectively reviewed in patients hospitalized in our respiratory intensive care unit between 1985 and 2000 for the diagnosis of MTX-pneumonitis. BAL cell counts from patients with MTX-pneumonitis (cases) were compared with those from five patients receiving MTX but who did not have evidence of MTX toxicity (MTX-exposed patients) and 29 healthy subjects (control subjects).

Case Definition

Methotrexate pneumonitis was defined by the following criteria: (1) patients receiving MTX before the onset of pulmonary symptoms; (2) occurrence of new bilateral infiltrates on chest radiograph; (3) exclusion of an infectious etiology for the diffuse interstitial pneumonitis after an extensive microbiologic evaluation, i.e., BAL samples (see below) and protected brush specimens performed during fiberoptic bronchoscopy, sputum and blood samples, and serologic tests; (4) exclusion of congestive heart failure by forced diuresis, cardiac echography, and absence of alveolar hemorrhage on BAL cytologic analysis (see below); (5) exclusion of underlying pulmonary disease by BAL cytologic analysis (see below) and histologic findings obtained by bronchial biopsies; (6) improvement of chest radiographic abnormalities after MTX withdrawal.

Collection of Data Concerning Patients at the Time of the Diagnosis and during the Follow-up

The following data were retrospectively recorded from the medical charts at the time of diagnosis and after a 3-mo follow-up: sex, age, general and pulmonary symptoms, chest radiographic findings, and severity of respiratory failure assessed by Pa_O2 measurement and the need for mechanical ventilation. Particular attention was paid to specific underlying disease leading tot he MTX administration and to the time interval between onset of respiratory symptoms and last MTX administration, as well as to the cumulative dose, route of administration, and duration of MTX treatment. Concomitant medication was also recorded, especially the use of steroids.

Fiberoptic Bronchoscopy and BAL Analysis

Fiberoptic bronchoscopy and BAL were performed as previously described (8). Briefly 4 50-ml aliquots of sterile and warmed isotonic saline solution were infused in the middle or the most radiologically affected lobe in the absence of tumor or bronchial suppuration. The macroscopic appearance of the recovered BAL fluid was noted by operators, especially if hemorrhagic. Appropriate stainings of BAL smears were performed for total and differential cell counts, identification of pathogens, i.e., mycobacteria, viruses, fungi, parasites, and detection of tumor cells and alveolar macrophages containing hemosiderin. Aliquots of BAL fluids were cultured for bacteria, mycobacteria, fungi, and viruses. T-lymphocyte helper and suppressor subsets were determined by immunofluorescence microscopy or flow cytometry (Coulter Electronics Corp., Miami, FL) using anti-CD4- and anti-CD8-specific monoclonal antibodies (Ortho Diagnostics, Raritan, NJ), respectively.

BAL total and differential cell counts from five MTX-exposed patients and 29 control subjects were used as comparative groups. To identify MTX-exposed patients we interrogated our computerized medical database, including all patients hospitalized in our chest department for 4 yr with the following key words: "methotrexate and bronchoalveolar lavage, not pneumonitis." Exposed patients were four women and one man with a mean age of 48.2 ± 15.0 yr. Two were current smokers (3.4 and 15 pack/years, respectively). Underlying diseases were psoriasis (n = 2), rheumatoid arthritis (n = 1), myositis (n = 1), and hematologic disorders (n = 1). MTX administration was oral in all cases, with a mean total dose of 850 ± 350 mg. Chest radiographs were normal, and BAL were performed because of recent cough and/or moderate fever that spontaneously disappeared without MTX cessation or specific medication. The control group was extracted from a previous study published by us (8). Control subjects were healthy volunteers enrolled at that time (1987 to 1989) to define normal values of BAL lymphocyte subsets in our laboratory. They were 18 men and 11 women with a mean age of 45 ± 12 yr; nine were smokers (36 ± 10 pack/years) and 20 nonsmokers.

Statistical Analysis

All the results are expressed as mean and standard deviation (SD). Because of the small population, nonparametric tests were used. Statistical comparisons were performed using the Mann-Whitney U-test (for two-group comparisons), ANOVA for Kruskal-Wallis (for three-group comparisons) and Wilcoxon's test for unpaired and paired data, respectively. Statistical correlations were performed using Spearman's Rho Rank Test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Nineteen BAL were performed in 14 consecutive cases of well-documented MTX-induced pneumonitis diagnosed during the study period. Three of these cases were reported earlier (4).

Case Characteristics

The cases were nine women and five men with a mean age of 50.0 ± 1.7 yr. Four of them were current smokers (28 ± 16 pack/years). Underlying diseases and concomitant medication are summarized in Table 1. All patients were symptomatic and complained of dyspnea and/or fever (38.1 ± 1.0 C°); 11 had cough. All patients presented with diffuse interstitial pneumonitis, and one with a mild pleural effusion (Case 14). All were hypoxemic (PO₂: 44.6 ± 9.6 mm Hg; range, 27 to 59), except one (Case 13). The length of treatment by MTX was 8.5 ± 17 mo (range, 2 to 64) and mean total dose was 959.6 ± 1,385.7 mg (range, 45 to 5,010). MTX administration was intravenous in five cases, intrathecal in three, intramuscular in two, and oral in four. Mean time-interval between the onset of respiratory symptoms and the last MTX administration was 5.3 ± 2.5 d (range, 2 to 10). Mean time-interval between last MTX administration and first BAL was 14.5 ± 5.2 d (range, 7 to 24). In all cases MTX was withdrawn, and the final outcome was favorable. One patient (Case 1) required intubation and mechanical ventilation for 5 d. Three patients (Cases 3, 8, and 9) improved without adjunctive steroid therapy. In six patients (Cases 4, 6, 7, 11, 12, 14), steroid therapy was continued at the previous dose (5 to 60 mg/d). In the five remaining patients (Cases 1, 2, 5, 10, and 13), a bolus of intravenous methylprednisolone was prescribed (40 to 1,000 mg/d for 4 to 15 d). Concomitant medication was pursued in 11 patients and interrupted in three.

In patients with MTX-pneumonitis, MTX-exposed patients, and control subjects, mean age, proportion of smokers, and intensity of their tobacco abuse did not differ. Furthermore, length of treatment and mean total dose of MTX administration did not differ in cases and MTX-exposed patients.

Results of BAL Analysis

Bronchoalveolar lavage was performed in all patients with MTX-pneumonitis. Five patients underwent a second BAL. Individual BAL data are given in Table 2. Total and lymphocyte counts expressed as percentage of total cells or as absolute numbers were significantly increased in MTX-pneumonitis compared with MTX-exposed patients and control subjects. Interestingly, lymphocytic alveolitis was present in each case, even in those with a normal total cell count (Table 2). Neutrophil count was significantly increased in cases when compared with control subjects and was higher than 10% in four of them. However, neutrophil count did not differ between cases and MTX-exposed patients. In these latter, neutrophil count was also significantly increased when compared with control subjects (Table 2). By contrast, total macrophage and eosinophil counts did not differ between cases, MTX-exposed patients, and control subjects. However, eosinophil count was slightly raised in two patients with MTX-pneumonitis (6 and 8%).

Subsets of T-lymphocytes were analyzed in 13 BAL from 11 cases. CD4 and CD8 cell counts were both significantly higher in cases than in MTX-exposed patients and control subjects (Table 3). By contrast, CD4 and CD8 lymphocyte percentages did not differ between the three groups (Table 3). These findings resulted in the large spectrum of CD4/CD8 ratios, ranging from low (< 1.5) to high (> 2.5). Five patients with MTX-pneumonitis underwent a second BAL 3 to 15 d later. In each case, the total and lymphocyte cell counts decreased, reaching a statistical significant difference for total lymphocytes (p = 0.04) (Figure 1). Among the second BAL, the T-lymphocyte subset was available in one patient (Case 2) and showed a decrease in CD4/CD8 ratio from 4.1 to 0.6. 

View larger version (14K): [in this window] [in a new window]   Figure 1.   Significant reduction of lymphocytic alveolitis after methotrexate withdrawal. Each closed circle represents individual data. First BAL (1st) was performed at the time of diagnosis and second BAL (2nd) was performed 3 to 15 d after MTX withdrawal.

Finally, results of BAL data did not significantly differ in cases, between patients receiving (n = 8) or not receiving (n = 6) cyclophosphamide as concomitant medication and with regard to specific underlying disease leading to the MTX administration, i.e., hematologic disorder (n = 6), cancer (n = 4), or rheumatologic disease (n = 4).

Factors Associated with CD4/CD8 Values

The analysis of the possible factors leading to the variation in CD4/CD8 T-cell ratios is given in Table 4. The alveolar CD4/ CD8 ratios correlated positively with total cell and lymphocyte counts (Table 4). By contrast, they correlated negatively with the interval between the last MTX administration and BAL (Figure 2, left panel) and with the steroid cumulative dose calculated from the last MTX administration (Figure 2, right panel). On the other hand, no correlation was found between CD4/CD8 ratio and MTX cumulative dose or Pa_O2.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Negative correlation between alveolar CD4/CD8 ratio and time elapsed between last MTX administration and date of BAL (p < 0.01, Rho = 0.73) (left panel ) and steroid cumulative dose given after last MTX administration (p < 0.01, Rho = 0.78) (right panel ). Each closed circle represents individual data.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the series reported herein, the clinical and radiologic presentation of MTX-induced pneumonitis was very homogeneous. It presented as subacute diffuse interstitial pneumonitis with high-grade fever and mild to severe hypoxemia (Pa_O2 range: 27 to 82 mm Hg) occurring 2 to 10 d after the last MTX administration. BAL cell counts showed severe lymphocytic alveolitis. Patients all recovered in a few weeks after MTX discontinuation and/or steroid adjuvant therapy introduction. These features were similar to those previously reported by us (4) and others (1, 5, 6, 10, 11). However, MTX-induced pneumonitis has also been reported as an acute respiratory distress syndrome (12, 13) and as a chronic fibrosing alveolitis (14).

Subacute occurrence of diffuse interstitial pneumonitis in patients receiving MTX for various underlying diseases should first suggest a pulmonary infection. The presence of severe lymphocytic alveolitis pointed to Pneumocystis carinii pneumonia and miliary tuberculosis (15, 16). P. carinii pneumonia was easily excluded in our patients by BAL examination. Miliary tuberculosis was also excluded since microbiologic evaluations were negative for Mycobacterium tuberculosis and since they all improved without any antituberculosis therapy. Specific pulmonary involvement of underlying neoplastic disease was completely eliminated by the outcome: total recovery from the pneumonitis. Lastly, the role of rheumatoid arthritis in pulmonary manifestations in Cases 11, 13, and 14 could not be eliminated, but interstitial pneumonitis resolved after MTX withdrawal without an increase in the low dose of steroid treatment (5 to 7 mg daily).

Lymphocytic alveolitis has also been largely documented in diffuse interstitial pneumonitis related to various drugs (8). Of drugs received by our patients, doxorubicin, adriamycin, and vincristine as well as ramipril, omeprazole, and amitriptyline were not or rarely reported to induce interstitial pneumonitis. Cytosine arabinoside caused noncardiogenic pulmonary edema (17) and diclofenac eosinophilic pneumonitis (18). Cyclophosphamide and bleomycin were responsible for more insidious pneumonitis associated with neutrophil and/or eosinophil alveolitis, often leading to severe pulmonary fibrosis with a fatal outcome (19, 20). Finally, the fact that all the patients recovered upon MTX withdrawal while other drugs were pursued in most cases strongly suggested that these drugs were not implicated in the respiratory manifestations observed.

The severe lymphocytic alveolitis observed in MTX-induced pneumonitis has already been described by us and others (4, 10). Rare cases of neutrophil or eosinophil alveolitis have also been reported (8, 10, 21, 22). In this context, increases in alveolar neutrophil and eosinophil counts were observed in three and in two cases, respectively (Table 2). The mechanism supposed to be implicated in MTX-induced pneumonitis, as for several other drugs associated with lymphocytic alveolitis, i.e., amiodarone, propanolol, gold salts, or nitulamide (8, 23), is an immune lung reaction, which is supported by clinical and pathologic findings. First, onset of pulmonary symptoms usually occurs a couple of days after the last administration of MTX, i.e., 2 to 10 d in our series with no relationship with dose or route of administration (4, 10). Second, the marked lymphocytic alveolitis as well as the lung pathologic findings in cases of MTX-induced pneumonitis recently reviewed (11) were similar to those described in cases of hypersensitivity pneumonitis (26). Third, lymphocytic alveolitis rapidly decreased after MTX withdrawal, as demonstrated in some of our patients (Figure 1). The report of rare cases in which MTX has been rechallenged as these patients did not develop the pneumonitis, did not deserve the immune lung reaction hypothesis (10, 27). However, diagnosis of MTX-pneumonitis in some of these reports was based only on clinical criteria because BAL or lung pathologic analysis were not performed (27). Furthermore, other medications were withdrawn concomitantly to MTX cessation and were not secondarly reintroduced (29), and the role of associated pulmonary infection was finally suggested in some cases (27, 29).

The few case reports describing lymphocyte subset analysis in BAL fluid from MTX-induced pneumonitis remain controversial (4, 5, 6, 10). Likewise, we have clearly shown that a wide range of values of alveolar CD4/CD8 ratio was observed in MTX-induced pneumonitis. This phenomenon was not observed in patients receiving MTX but who did not have evidence of MTX toxicity and were not related to specific underlying disease leading to the MTX administration (10). Nevertheless, these values correlated positively with the degree of lymphocytic alveolitis (Table 4) but negatively with the time lapse between the last MTX administration and the date of BAL (Figure 2, right panel). These findings may reflect the kinetics of the pulmonary immune response and inflammatory cell recruitment after MTX challenge. Indeed, Leduc and colleagues (22) described early neutrophil alveolitis followed by lymphocytic interstitial pneumonitis with granuloma shortly after MTX reintroduction in two patients with rheumatoid arthritis. In this context, we also observed that chronically MTX-exposed patients presented latent neutrophil alveolitis compared with healthy subjects. Furthermore, we described lymphocytic alveolitis decreasing over time after MTX withdrawal in five patients (Figure 1). In one of these patients (Case 2), the alveolar CD4/CD8 ratio varied from 4.1 to 0.6. Thus, alveolar lymphocyte recruitment may involve a first wave of lymphocytes comprising CD4 helper cells, followed by a second wave of CD8 suppressive cells, thus explaining both the decrease in total lymphocyte count and the between-patient discrepancies in CD4/CD8 T-cell ratios. Bronchoalveolar lavage cell profile might also be modified by concomitant medication, i.e., cyclophosphamide and steroids. The responsibility of cyclophosphamide was unlikely in our series for results of BAL data did not differ between MTX patients receiving or not receiving cyclophosphamide as concomitant medication. By contrast, steroids might be implicated as suggested by the negative correlation between CD4/CD8 ratio and the steroid cumulative dose given after the last MTX administration (Figure 2, left panel). Although steroid-induced apoptosis is a well-recognized regulator of T-cell number and function in humans (30), our small study population constitutes a potential limitation to such conclusions.

The clinical and radiologic presentation of the MTX-induced pneumonitis in our series of patients was very homogeneous, and was also characterized by severe lymphocytic alveolitis. However, a large range of CD4/CD8 ratios was observed and reflected the variation in time elapsed between the last administration of MTX and BAL evaluation and the use of steroids as adjunctive therapy.