Closure of BPF following pulmonary resection is manda­tory not only because of symptoms, such as continuous cough and expectoration, but also because it is very likely that the pleural cavity will be continuously reinfected as long as the fistula is open. In patients with this condition, who often are debilitated, a second thoracotomy represents a high-risk procedure. A conservative approach with pro­longed pleural drainage and antibiotics first was tried in our patient. Four unsuccessful applications of fibrin glue with a fiberoptic bronchoscope also were attempted. Unfortu­nately, only a temporary closure of the fistula was seen on two of these occasions. This failure of endobronchial fibrin glue application was not due to induced necrosis by cauter­ization by silver nitrate since three attempts of broncho­scope sealing were performed using fibrin glue alone. As previously suggested, it may result from a rapid dissolution of the fibrin plug. To our knowledge, such a failure has not been previously described; however, there are few reports of this new method. A thoracoplasty was chosen rather than a more direct surgical approach because of the patient s poor condition. Procedures such as suturing the bronchial stump or covering it with various flaps of intercostal muscle or omentum might have been too aggressive for the patient. Finally, as the thoracoplasty also failed, we performed a thoracoscopy and could visualize the space with the remain­ing air-fluid level and the recessus to the fistula. Talc was injected into the orifice of the BPF and the adjacent pleura. This approach of direct access to the pleural cavity may have been more successful because of the well known pleural sclerosing effect of talc. The thoracoscope route rather than bronchoscopy also may have exposed a greater pleural surface to the sclerosing agent.

French authors recently have reported a similar case with successful closure of such a fistula using a thoracoscope with a YAG laser. That is a more expensive procedure not available to every hospital. However, it also shows that a chronic bronchopleural fistula can be closed using a pleural route.

In conclusion, we suggest that pleural talc under thora­coscopy could be used in persistent bronchopleural fistula to avoid morbidity and even death associated with such a debilitating condition. It may be that the reduction of space by the thoracoplasty procedure allows this technique to be successful. However, it remains to be shown whether such a procedure should be performed before any second surgical approach in order to avoid such a major aggression in patients with poor general condition.