Rifampicin induced pneumonitis or bronchogenic spread of tuberculous empyema through a bronchopleural fistula?

We read with great interest the article by Kunichika et al which describes a probable case of pneumonitis induced by rifampicin. However, we think that the case needs further clarification. We note that on admission the chest radiograph did not show a simple left pleural effusion but an air-fluid level, suggestive of a bronchopleural fistula. A bronchopleural fistula associated with tuberculosis usually follows trauma or a surgical procedure, but it can also occur spontaneously in patients with longstanding tuberculous empyema because the tuberculous process establishes an open pathway between the bronchus and pleura. Tuberculous empyema may be present for a long time with few clinical symptoms, and patients may come to clinical attention only when undergoing a routine chest radiograph or after a bronchopleural fistula or empyema develops. The chest radiograph on admission also showed a thickened pleura and loss of volume of the left lung, suggesting longstanding pleural disease. The authors do not specify the characteristics of the pleural fluid of the patient. In the chest radiograph taken on day 9 we think that the shadows in the right lung follow an alveolar rather than an interstitial pattern (an air bronchogram may be present). A possible alternative explanation for these findings might be the bronchogenic spread of a tuberculous empyema through a bronchopleural fistula. The drainage of the pleural fluid should theoretically have prevented this, but we note that there was still a small amount of pleural fluid on day 9 and therefore it is possible that some fluid could have passed through a bronchopleural fistula. A steroid induced reduction in the inflammatory response associated with this spread might explain the clinical and radiographic improvement. BAL lymphocytosis would be consistent with this hypothesis. The CD4/CD8 ratio in the BAL fluid cannot reliably differentiate between this possibility and lung induced pneumonitis. Drug induced disease can be associated with either high or low CD4/CD8 ratios in BAL fluid. The CD4/CD8 ratio in pulmonary tuberculosis is variable, although in most patients it is within the normal range. The results of the drug lymphocyte stimulation test are highly suggestive of rifampicin induced pneumonitis, but it cannot be considered 100% specific. Furthermore, we think that the antituberculous chemotherapy regimen needs to be clarified. As we understand it, the patient was treated only with streptomycin for about 2 months. Single drug treatment in tuberculosis is not appropriate because it is associated with a high risk of drug resistance. Also, isoniazid and ethambutol (without streptomycin?) were later administered for another 6 months. This seems a rather short course of treatment for a case of smear positive tuberculosis, unless an initial (2 months) course with at least three drugs was implemented. In fact, chronic tuberculous empyema with bronchopleural fistula can result in treatment failure, probably because the thick pleural walls can limit penetration of drugs into the pleural space. The long term course of the patient needs to be described in more detail. The authors only state that within 6 months there was no recurrence of the abnormal shadows on the chest radiographs, which seems a somewhat brief description of this interesting case.