Bronchoscopic assessment of the evolution of endobronchial tuberculosis.

BACKGROUND

We previously classified forms of endobronchial tuberculosis (EBTB) into seven subtypes by bronchoscopic finding: actively caseating, edematous-hyperemic, fibrostenotic, tumorous, granular, ulcerative, and nonspecific bronchitic.

STUDY OBJECTIVE

To evaluate the value of this classification in predicting the therapeutic outcome of EBTB.

DESIGN

A prospective study with serial bronchoscopy performed from the diagnosis of EBTB to the completion of antituberculosis chemotherapy.

PARTICIPANTS

Eighty-one patients with biopsy-proven EBTB.

INTERVENTIONS

Fiberoptic bronchoscopy was done every month until there was no subsequent change in the endobronchial lesions, every 3 months thereafter, and at the end of treatment.

RESULTS

Twenty-two of the 34 cases of actively caseating EBTB changed into the fibrostenotic type, and the other 12 healed without sequelae. Seven of the 11 cases of edematous-hyperemic EBTB changed into the fibrostenotic type, and the other 4 healed. Nine of the 11 cases of granular EBTB, 6 cases of nonspecific bronchitic EBTB, and 2 cases of ulcerative EBTB resolved completely. However, the other two cases of granular EBTB changed into the fibrostenotic type. Seven cases of fibrostenotic EBTB did not improve despite antituberculosis chemotherapy. These various changes in bronchoscopic findings occurred within 3 months of treatment. In 10 cases of tumorous EBTB, 7 progressed to the fibrostenotic type. In addition, new lesions appeared in two cases, and the size of the initial lesions increased in another two cases, even at 6 months after treatment.

CONCLUSIONS

The therapeutic outcome of each subtype of EBTB can be predicted by follow-up bronchoscopy during the initial 3 months of treatment, with the exception of the tumorous type. In tumorous EBTB, close and long-term follow-up is advisable because the evolution of the lesions during treatment is very complicated and bronchial stenosis may develop at a later time.