OBJECTIVE Despite the advances in surgical technology, bronchopleural fistulas (BPFs) still occur and are often fatal. We studied the risk factors for BPF formation after lung cancer operation to clarify the indication of preventive bronchial stump coverage. In addition, the reliability of our methods of bronchial closure was evaluated. METHODS We reviewed 557 consecutive bronchial stumps, corresponding to 547 patients without any coverage in pulmonary resection for lung cancer between 1989 and 1998. According to nine variables, stumps that made dehiscence were compared with uneventful ones using contingency table analysis. The incidence of BPFs according to each method of bronchial closure was calculated. RESULTS BPFs developed in ten patients (1.8%). Compared with the lobar bronchus (LB), the main bronchus (MB; P<0.01; odds ratio, 23.0) and the intermediate bronchus (IB; P=0.03; odds ratio, 10.7) carried a high risk. Previous ipsilateral thoracotomy (P<0.01; odds ratio, 37.9) and preoperative chemotherapy and/or radiotherapy (P=0.02; odds ratio, 13.2) increased the risk. The incidence of BPFs with manual suture, stapling devices only, reinforcement suture at the distal side of staplers, or reinforcement suture at the proximal side of staplers was 1.8, 5.0, 1.9 and 1.0%, respectively. CONCLUSIONS The main and intermediate bronchial stumps, and the stumps in cases with previous ipsilateral thoracotomy or receiving induction therapy are prone to BPFs. Preventive coverage should be considered for these stumps. Our methods for reinforcement of stapled stumps are thought to be reliable.