Disseminated tuberculosis after anti-TNFα treatment

Abstract

A 74-year-old white woman with rheumatoid arthritis since 1993 presented to us in September, 2005, with a 2-month history of progressive shortness of breath, fatigue, and dry cough which had worsened over the preceding 3 weeks, with no associated fever, night sweats, or weight loss. Her medical history included ischaemic heart disease, hypertension, chronic renal impairment, and diabetes mellitus. She was taking meloxicam, aspirin, gliclazide, rosuvastatin, lansoprazole, furosemide, insulin, prednisolone (10 mg/day), and adalimumab. She had tried various disease modifying drugs, including hydroxychloroquine, sulfasalazine, methotrexate (15 mg/week for over a year) and lefl unomide (20 mg/day) for her arthritis; all had to be stopped because of side-eff ects or ineffi cacy. In October, 2003, her disease activity score (DAS) was 9·3 (25 swollen joints, 28 tender joints, ESR 106 mm/h, visual analogue scale [VAS] 100). She was assessed according to the British Society for Rheumatology guidelines to receive anti-TNFα therapy. She had no history of tuberculosis, although her sister did. The screening chest radiograph was normal and tuberculin skin test was negative. She was started on adalimumab in November, 2003, but did not receive concomitant methotrexate because it had been stopped previously owing to shortness of breath. Her arthritis responded well with a DAS of 3·52 in March, 2004 (1 swollen joint, 5 tender joints, ESR 17 mm/h, VAS 0). On admission, she was apyrexial, blood pressure 131/53 mm Hg, pulse 97 beats per min, respiratory rate 20 breaths per min, and oxygen saturation 96% on air. She had a right-sided pleural eff usion confi rmed on chest radiograph. Full blood count, urea and electrolytes, and liver function were normal. Pleural fl uid for acid-fast bacilli, tuberculosis culture and PCR, and cytology was negative. CT of the thorax and abdomen confi rmed right pleural eff usion, lower lobe collapse, mediastinal lymphadeno pathy, and lung infi ltrates with numerous masses in the liver. While awaiting liver biopsy she deteriorated rapidly and died in October, 2005. Post-mortem examination showed multiple caseating granulomas with Langhan’s-type giant cells in the lungs, liver, lymph nodes (fi gure), and spleen; tissue specimens were negative for acid fast bacilli, and tuberculosis culture and PCR. A postmortem diagnosis of disseminated miliary tuberculosis was made on the basis of histology. There is a fi ve-fold increased risk of tuberculosis associated with anti-TNFα therapy in rheumatoid arthritis. Most cases are due to reactivation of latent tuberculosis and usually occur in the early phase of treatment, although cases have been reported after 2 years. The presentation is often atypical with extra pulmonary or disseminated disease. Our patient presented nearly 2 years into her treatment—her screening tests for tuberculosis before initiation of adalmimuab were negative, although her tuberculin skin test might have been falsely negative due to steroids. Before starting adalimumab, her tuberculosis risk was assessed and it was felt that the risk of tuberculosis prophylaxis outweighed the risk of disease (she had had abnormal liver function tests on lefl unomide). There was considerable diffi culty in arriving at the diagnosis of tuberculosis in our patient because she presented with atypical symptoms and all the microbiological tests to diagnose tuberculosis were negative. Our case highlights three important points. Firstly, current screening tests can miss latent tuberculosis especially in immunosuppressed patients; newer blood tests, which detect γ-interferon, are promising. Secondly, it can be diffi cult to diagnose tuberculosis with PCR and culture; histopathology is important in diagnosing diffi cult cases and should be requested early on, the results of which are available within days. Thirdly, patients on anti-TNFα therapy who present with persistent respiratory symptoms with any suspicion of tuberculosis should be considered for empirical anti-tuberculosis treatment despite the lack of microbiological evidence and while awaiting tissue diagnosis.