Pulmonary Mycobacterium szulgai Infection and Treatment in a Patient Receiving Anti-Tumor Necrosis Factor Therapy

Jakko van Ingen; Martin Boeree; Matthijs Janssen; Erik Ullmann; Wiel de Lange; Petra de Haas; Richard Dekhuijzen and Dick van Soolingen


Nat Clin Pract Rheumatol. 2007;3(7):414-419. 

In This Article

Discussion of Diagnosis

The patient developed a nontuberculous mycobacterial infection while receiving anti-tumor necrosis factor (TNF) therapy for the treatment of RA. The similar mycobacterial genotypes, which were obtained using random amplified polymorphic DNA analysis, indicated that the same strain of M. szulgai was isolated on all three occasions (Figure 4). Pulmonary M. szulgai isolates from other patients within the same geographical region and the same period, which were mostly cultured in the same hospital laboratory, seemed to be unrelated to those obtained from this patient; therefore, contamination of this patient's cultures, or reinfection, is unlikely to have occurred. Theoretically, repeated exposure to a single environmental source might have caused monoclonal reinfections; however, this possibility is considered improbable. It is assumed, therefore, that the primary M. szulgai culture, which was cultured 3 years prior to treatment, already represented true infection. The increases in erythrocyte sedimentation rate and serum C-reactive protein concentration in a period when RA was stable also supported this assumption.

The patient's early symptoms, radiological features and M. szulgai isolates probably represented subclinical or latent nontuberculous mycobacterial infection; however, the American Thoracic Society diagnostic criteria for nontuberculous mycobacterial infection (see "Sidebar: Summary of the American Thoracic Society Diagnostic Criteria for Nontuberculous Mycobacterial Infection")[1] were not met, and the patient was not treated for nontuberculous mycobacterial infection at this time. Histological samples were obtained from the patient at the initial presentation, and the results of the samples were interpreted as sarcoidosis. The differential diagnosis also included tuberculosis, M. szulgai infection, fungal infection, and occupational pulmonary disease such as asbestosis, berylliosis and silicosis. The absence of asbestos, beryllium or silica exposure, and absence of cultured M. tuberculosis complex bacteria or fungi, however, narrowed the differential diagnosis to sarcoidosis and M. szulgai infection.

The possibility that the patient's histological findings resulted from the nontuberculous mycobacterial infection, influenced by anti-TNF therapy, is supported by the finding that bronchiolitis and epithelioid cell granuloma, with or without central necrosis, have been recorded in histological studies of patients with nontuberculous mycobacterial infection.[3] Also, the absence of radiological abnormalities 2 years prior to the initial presentation indicated that there was an association between the patient's radiological features and the M. szulgai culture. Mycobacteria might have a role in the etiology of sarcoidosis;[4] therefore, M. szulgai could theoretically have caused sarcoidosis in this patient. This possibility is unlikely, however, because mycobacterial DNA fragments, not viable mycobacteria, have been demonstrated in tissue samples of patients with sarcoidosis, and the bacteriological course of this patient's disease is strongly indicative of mycobacterial infection rather than sarcoidosis.[4]

Treatment with etanercept was discontinued in this patient, and treatment with adalimumab was initiated. Adalimumab is closely related to the anti-TNF agent infliximab, which is effective in the treatment of sarcoidosis.[5] Paradoxically, anti-TNF agents (mainly etanercept) have been described as a possible causative agent of sarcoidosis;[6] this finding strengthened the diagnosis of sarcoidosis in this patient during the early stages, as he was being treated with etanercept at this time. Adalimumab inhibits the action of TNF and the production of interferon-γ (INFγ);[7] these properties increase patients' susceptibility to nontuberculous mycobacterial infection. This patient's infection, which was cultured at the initial presentation, was clinically stable during treatment with etanercept but slowly deteriorated after initiation of adalimumab therapy. This observation might demonstrate the high extent to which TNF and INFγ inhibition is achieved with adalimumab.

TNF has an important role in granuloma formation and maintenance, which is essential in host defense against mycobacteria;[8] therefore, mycobacterial infection or reactivation of latent infection can be expected as a complication of anti-TNF treatment. Reports have documented the emergence of nontuberculous mycobacterial infection in patients who receive anti-TNF treatment. In 2004, Wallis et al.[9] reported 29 unspecified nontuberculous mycobacterial infections in patients treated with anti-TNF agents from the FDA adverse events reporting system. These reports are not necessarily 29 true nontuberculous mycobacterial infections, as the reporting procedure might not be sufficient to fulfill the American Thoracic Society diagnostic criteria. Two true nontuberculous mycobacterial infections are described in separate case reports;[10,11] one concerned a psoas muscle abscess caused by M. avium induced by etanercept use, and the other a M. abscessus skin infection induced by infliximab use. These infections occurred after 12 and 13 months of anti-TNF treatment, respectively, as in this patient, whose symptoms deteriorated after 12 months of adalimumab therapy. These reports highlight that the length of time between the start of anti-TNF treatment and the onset of nontuberculous mycobacterial disease is longer than for the onset of tuberculosis, where a median interval of 12 weeks is observed.[8,9]

Among patients who receive anti-TNF therapy, tuberculosis infection or reactivation is more common than nontuberculous mycobacterial infection or reactivation.[8,9,10,11] Several possibilities exist for the cause of this finding, including a higher prevalence of latent tuberculosis compared with latent nontuberculous mycobacterial disease. Another explanation could be the ability of tuberculosis to infect previously healthy individuals, whereas nontuberculous mycobacteria, being opportunistic pathogens, generally affect patients with local (e.g. COPD, as in this patient) or systemic impaired immunity. The finding could also arise from the possibility that TNF has a less prominent role in the pathogenesis of nontuberculous mycobacterial infection than it has in tuberculosis infection. The concept of 'latent' nontuberculous mycobacterial infection is controversial, and missed diagnoses of active nontuberculous mycobacterial infection might also contribute to a higher recorded prevalence of tuberculosis than nontuberculous mycobacterial infections in patients who receive anti-TNF therapy.

Various authors have produced recommendations for the screening of latent tuberculosis in candidates eligible for anti-TNF therapy, which usually involves tuberculin skin testing and chest X-ray.[8] For nontuberculous mycobacteria, these tests might not suffice: tuberculin skin test results are mostly negative in patients with nontuberculous mycobacterial infections, and chest X-rays show diverse, partly species-specific patterns.[1] In candidates for anti-TNF treatment with known lung disease, specifically COPD and bronchiectasis, their sputum should be repeatedly cultured for mycobacteria.

In this patient, it should be considered that COPD itself predisposed him to nontuberculous mycobacterial infection, independently of anti-TNF therapy. The localized pulmonary presentation in this patient, as opposed to the extrapulmonary or disseminated mycobacterial infections seen in patients who receive anti-TNF treatment, suggests that the host response to infection was relatively intact.[8] It is not possible, therefore, to establish whether or not anti-TNF treatment was the sole, definite cause of his M. szulgai infection. As the number of indications for anti-TNF treatment increases, more patients with additional risk factors for nontuberculous mycobacterial infection might receive treatment with this class of agents; therefore, the case patient might represent an emerging issue of nontuberculous mycobacterial infection in anti-TNF recipients.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.