Azathioprine-Induced Lung Toxicity and Efficacy of Cyclosporin A in a Young Girl With Type 2 Autoimmune Hepatitis

Abstract

Autoimmune hepatitis is rare in children. Two types are identified according to the presence of different autoantibodies in the serum: the antinuclear and/or smooth muscle antibody (ANA/SMA) characterizes type 1, whereas the liver-kidney microsomal (LKM-1) and/or anti-liver cytosol antibody defines type 2 (1). The treatment of autoimmune hepatitis relies on immunosuppressive drugs, usually prednisone and azathioprine, which are often combined, with azathioprine playing a steroid-sparing role (2,3). However, these drugs may have side effects that can limit or even contraindicate their administration. Among the possible azathioprine-related side effects, interstitial pneumonitis has rarely been reported in adults (4–13), and even more rarely in children (14). The diagnosis must be determined rapidly so that the treatment may be interrupted before irreversible sequelae develop. In such situations, even though corticosteroids are helpful, they may not control the progress of the disease by themselves, making the use of other drugs necessary, among which cyclosporin A appears to be one of the most promising (15). We report on a young girl with autoimmune chronic active hepatitis in whom azathioprine treatment had to be interrupted after the development of interstitial pneumonitis. She was successfully treated with cyclosporin A for 5 years. A relapse occurred a few months later and was controlled with the reintroduction of cyclosporin A. CASE REPORT This 13-year-old girl was admitted to our unit for jaundice and asthenia. She was the first child of healthy unrelated parents and had no particular medical history before admission. Physical examination showed hard hepatomegaly (7 cm below the right costal margin), moderate jaundice, and fever (38.4°C). The spleen was not enlarged. The skin was normal, and pubertal development was nearly complete (Tanner 4). Abdominal ultrasound examination confirmed hepatomegaly but did not reveal any sign of portal hypertension. Laboratory data indicated cytolysis (aspartate aminotransferase [AST]: 1200 IU/l; alanine aminotransferase [ALT]: 1350 IU/l) and cholestasis (γ-glutamyl transferase [GGT]: 150 IU/l; total bilirubin: 185 μmol/l; conjugated bilirubin: 135 μmol/l), whereas factor V and coagulation test results were normal. Serum globulin concentration was 35 g/l (normal: 6–13 g/l). Anti-LKM-1 antibodies were found in the serum at a titer of 1:100,000 (immunodiffusion and identity reaction) whereas other autoantibodies were absent. Liver needle biopsy was performed, and light microscopic examination of the liver tissue revealed chronic active hepatitis with cirrhosis (Fig. 1). Therapy with prednisone (2 mg/kg · day) and azathioprine (1.4 mg/kg · day) was begun. Fever and jaundice slowly disappeared within 3 weeks, and laboratory results improved slowly.FIG. 1.: Liver biopsy specimen obtained when the patient was 13 years of age, showed active hepatitis, piecemeal necrosis, and fibrosis (hematoxylin and eosin; original magnification ×10).Six weeks after the beginning of treatment, the girl was readmitted for fever (39°C) and recurrent cough, despite treatment with amoxicillin and clavulanic acid. Auscultation and percussion showed the chest to be clear, and the liver was enlarged, as previously noted, and hard. Chest radiography showed bilateral interstitial pneumonitis, which was confirmed by tomodensitometry. Laboratory investigations yielded the following results: leukocyte count 2200/mm3 (45% lymphocytes, 40% neutrophils), hemoglobin 7.8 g/dl, platelets 188,000/mm3, C-reactive protein 15 mg/l, AST 39 IU/l, and ALT 109 IU/l. Azathioprine-induced pneumonitis was suspected, and the drug was discontinued immediately, whereas prednisone was continued. Intravenous antibiotherapy was initiated (vancomycin, cefotaxime, amikacin). Bronchial endoscopy was normal and bronchoalveolar lavage showed 380,000 cells/ml with 58% lymphocytes, 31% macrophages, 10% eosinophils, and 1% neutrophils. Among lymphocytes, some were plasmocytes and immunoblasts, and CD4 and CD8 were 46% and 57%, respectively; the CD4-to-CD8 ratio was 0.8 (Fig. 2).FIG. 2.: Bronchoalveolar lavage showing hypercellularity with hyperlymphocytosis (May-Grünwald-Giemsa; original magnification, ×40).Silver-nitrate staining was negative, thus ruling out the diagnosis of Pneumocystis carinii. These findings were suggestive of drug-induced pneumonitis. Bronchial cultures were negative, and the search for viral, bacterial, and fungal infection was negative (repeated blood cultures, urine cultures, and viral serologies). Antibiotics were then discontinued. Fever totally disappeared within 2 weeks, and the cough slowly improved over 3 weeks. At that time, pulmonary function test results were normal, and chest radiography showed regression of interstitial lesions. Despite treatment with prednisone, laboratory data indicated cytolysis (AST: 450 IU/l; ALT: 500 IU/l) and cholestasis (GGT: 135 IU/l) with hypergammaglobulinemia (23 g/l), and anti-LKM-1 autoantibodies were still present in the serum (titer: 1:1000). Because prednisone did not control the disease, oral cyclosporin A was introduced (5 mg/kg · day) after checking the normality of renal function (blood pressure: 105/65 mmHg, creatinine clearance: 125 ml/min · 1.73 m2). Within 2 months, cytolysis and cholestasis disappeared, and anti-LKM-1 autoantibodies were no longer detected in the serum. Prednisone was discontinued, and the patient was treated with cyclosporin A alone. During the next 3 months, trough serum levels of cyclosporin A were between 195 and 245 ng/ml. Once sustained remission was obtained, cyclosporin A dosage was tailored so that serum trough levels remained between 100 and 150 ng/ml for 2 years. The girl was regularly followed-up. Growth and blood pressure were normal and, except for mild development of hirsutism, no side effects of cyclosporin A were noted. After 2 years of complete remission under this regimen, cyclosporin A doses were slowly decreased under laboratory control over 3 years, and the treatment was discontinued. At that time, renal function was unchanged (creatinine clearance: 115 ml/min · 1.73 m2), and there were no extrahepatic manifestations of the disease. Six-and-half months later, whereas physical examination was normal, laboratory data indicated mild cytolysis (AST: 86 IU/l, ALT: 126 IU/l) and anti-LKM-1 autoantibodies were present at a titer of 1:640 in the serum. Oral cyclosporin was reintroduced so that serum trough levels remained at approximately 100 ng/ml. Within 6 weeks, serum transaminases concentrations were normalized. The patient will continue with this regimen for several years. DISCUSSION We report on a girl with type 2 autoimmune hepatitis who was initially treated with a conventional combination of prednisone and azathioprine. The latter had to be discontinued and was successfully replaced by cyclosporin A, and prednisone was discontinued 2 months later. Cyclosporin A alone then controlled the course of the disease. Azathioprine-associated pulmonary toxicity occurs rarely (review in Israel-Biet et al. [4]). In our patient, even though pulmonary function tests were not initially performed, this diagnosis appears likely in view of the results of bronchoalveolar lavage (especially the CD4/CD8 ratio), the negativity of repeated cultures and viral serologies, the absence of Pneumocystis carinii, and the regression of clinical and radiologic manifestations within the weeks after the withdrawal of azathioprine therapy. Drug-associated pulmonary toxicity may be difficult to prove, but the possibility of an iatrogenic complication should always be considered in patients with pulmonary symptoms. Clinical and radiologic manifestations are not specific. Pulmonary function tests usually reveal hypoxemia with hypocapnia and decreased diffusion capacity of the lung for carbon dioxide Dlco. Moreover, bronchoalveolar lavage is mandatory and usually shows hypercellularity with hyperlymphocytosis, and an inversion of the CD4-to-CD8 ratio (4,11,12). Finally, viral and bacteriologic searches are negative, and all symptoms regress once the drug has been discontinued. However, among drugs known to be potentially toxic to lungs, azathioprine is not one of the major culprits, especially when compared with gold salts, amiodarone, or cytotoxic agents such as bleomycin or methotrexate (4,11). In our patient, the absence of portal hypertension makes pulmonary involvement due to liver disease unlikely and argues for the direct toxicity of azathioprine. The conventional treatment of autoimmune hepatitis in children comprises prednisone and azathioprine (2,3). However, this therapy may fail to achieve complete remission. Both prednisone and azathioprine have side effects that may limit their use, and, if relapses occur, the risk of drug intolerance increases. The efficacy of cyclosporin A in children with type 2 autoimmune hepatitis had been suggested (16–18) and has recently been demonstrated (15). Furthermore, cyclosporin A can induce both clinical and laboratory remission even when given as first-line therapy. However, as pointed out by Debray et al. (15), the possible side effects of cyclosporin A have to be borne in mind, especially the potential long-term nephrotoxicity, and close monitoring of cyclosporin A blood levels are mandatory to avoid toxicity while maintaining complete remission. Even though a relapse occurred in the current case, it seems that cyclosporin A may be a useful alternative treatment in children with autoimmune hepatitis, especially when either azathioprine (as in our case) or prednisone is not tolerated and has to be discontinued. Acknowledgments: The authors thank Alan Strickland for editing the English.