1 Introduction
Ischemic change in bronchial stumps after lung cancer surgery is considered as one of several possible etiologies contributing to the development of bronchopleural fistulas (BPFs), which still represent major complications [1–4]. Impaired healing of bronchial stumps is a constant impediment to successful surgical procedures and is thought to be due to ischemia [5]. Although there have been many studies on postoperative BPFs, only limited information is available regarding changes in bronchial stumps evident on bronchofiberscopy after lung cancer resection.
More than 15 years ago, we underwent a series of BPFs after lung cancer surgery with lymph node dissection (LND) of the pulmonary hilum and mediastinum, including subcarinal nodes. Thus, we focused on localized ischemic or ulcerative changes and partial necrosis of the bronchial wall around stumps evident on bronchofiberscopy after lung cancer resection, a condition we term postoperative ischemic bronchitis (POIB) (Fig. 1). Although ischemic bronchial complications which still represent an important source of early morbidity and mortality following lung transplantation has been well analyzed, no precise report on this entity of POIB after lung cancer surgery has been published [6–8].
Bronchofiberscopic photographs of postoperative ischemic bronchitis (POIB) after lung surgery. Note localized ulcerative change in bronchial stumps (a, right intermediate trunk in a 56-year-old male; b, right upper bronchial stump in a 66-year-old male).
Bronchofiberscopic photographs of postoperative ischemic bronchitis (POIB) after lung surgery. Note localized ulcerative change in bronchial stumps (a, right intermediate trunk in a 56-year-old male; b, right upper bronchial stump in a 66-year-old male).
The aim of the present study was to analyze the frequency of POIB, the impact of development and the treatment after lung cancer surgery (Fig. 2).
Sequential POIB findings for a 59-year-old male. Note severe ischemic changes and ulceration at the left upper bronchial stump after lobectomy on the 9th postoperative day (POD). The lesion had largely resolved by the 66th POD.
Sequential POIB findings for a 59-year-old male. Note severe ischemic changes and ulceration at the left upper bronchial stump after lobectomy on the 9th postoperative day (POD). The lesion had largely resolved by the 66th POD.
2 Materials and methods
The surgical files of the Cancer Institute Hospital, Tokyo, Japan, between January 1991 and December 2002 were searched for all the cases of primary lung cancers undergoing curative pulmonary resection according to the Japanese Lung Cancer Society criteria [9]: curative resection was defined as complete resection of lung cancers with no evident cancer cell remaining based on surgical and pathologic findings. Recently based on the data of Asamura et al. [10] and our experience, subcarinal LND is not being performed for tumors of the right upper lobe and left upper segment without metastases to the hilar (station 12) and interlobar nodes (station 11), determined by intraoperative examination of frozen sections. For tumors of other lobes both superior mediastinal and subcarinal LND are performed.
During the period, surgical resection was performed for 1197 patients. Among these, there were 1015 (389 women and 626 men) who received bronchofiberscopic examination between the 5th and 15th postoperative days routinely. These examinations were performed during hospitalization. If patients had POIB, bronchofiberscopic examinations were performed again. Bronchofiberscopic files and photographs of these patients were reviewed and findings between the 8th and 14th postoperative days were evaluated retrospectively by two or more readers for localized ischemic and/or ulcerative changes or partial necrosis of the bronchial wall around stumps evident on bronchofiberscopy, termed POIB (Fig. 1). Based upon our experience, bronchial mucosal healing is well documented until the 7th postoperative day. The readers were blind to the outcome of POIB or BPF. Patients undergoing partial resection, segmentectomy, or bronchoplasty were excluded. The study was approved by our institutional review board, and each patient gave written informed consent before bronchofiberscopic examination.
The clinical variables included in the analyses were as follows: gender, age, smoking history and preoperative complications such as diabetes mellitus, pulmonary tuberculosis, pulmonary fibrosis and preoperative chemoradiotherapy. As postoperative factors, acute respiratory distress syndrome, postoperative ventilation, atrial fibrillation, other cardiovascular complications, other respiratory complications, and BPFs were analyzed.
As surgical parameters, degree of LND, the method of bronchial stump closure, extent of pulmonary resection, histologic types of resected materials, and pathologic stages were investigated. Staging definitions for T (primary tumor), N (regional lymph nodes), and M (distant metastasis) components were according to the International Staging System for Lung Cancers [11,12]. Histopathologic studies were conducted according to the World Health Organization criteria [13].
Statistical analyses were made with the χ2 or Fisher's exact tests, p < 0.05 being considered significant.
There were four surgeons involved in this study. All pulmonary resections and LNDs were performed in the same way.
3 Results
There were no complications due to bronchofiberscopic examination. In the total of 1015 patients, POIB was documented in 29 (2.5%; males = 26, females = 3) (Table 1). The average age was 63 years (range 46–79 years) and there was no difference in prevalence below and above 65 years of age (p = 0.53, Table 1). POIB was significantly more frequent in males (p = 0.005, Table 1).
Patient demographics
Patient demographics
The most common site for POIB was the right intermediate trunk stump (n = 4; 7.1%), followed by left upper (n = 8; 3.4%), right lower (n = 5; 3.4%), right middle (n = 2; 3.3%), and left lower (n = 4; 3.2%) lobar bronchial stumps. However, the differences among locations were not statistically significant (p = 0.66) (Tables 1 and 2). Among the pneumonectomy group only one POIB was documented.
Clinical and surgical profiles for the patients with postoperative ischemic bronchitis (POIB) (n = 29)
Clinical and surgical profiles for the patients with postoperative ischemic bronchitis (POIB) (n = 29)
As to the method of bronchial stump closure in the POIB group, 14 underwent stapling and the other 15 received interrupted simple sutures. The difference between the types of bronchial closure was not statistically significant (p = 0.49, Table 1). All 29 POIB patients underwent LND of the pulmonary hilum and mediastinum, combined with lobectomy, bilobectomy, or pneumonectomy. POIB was found in 27 patients undergoing subcarinal LND, whereas it occurred in only two without this procedure (p = 0.0001, Table 1).
Among 29 POIB patients, 6 were stage IA, 5 were IB, 3 were IIA, 5 were IIB, 4 were IIIA, and 6 were IIIB (p = 0.6; Table 1). The histologic types were distributed as follows: adenocarcinoma was most common (14 cases, 50%), followed by six squamous cell carcinomas (19%), three large cell carcinomas (12%), three small cell carcinomas (7.7%), one adenosquamous carcinoma (3.8%), and two cases of other types (7.7%) (p = 0.57; Table 1).
In six (21%) of the POIB group, infectious symptoms such as fever, leukocytosis and elevation of serum C-reactive protein were observed and additional supportive therapy with antibiotics was necessary. In these patients the POIB was followed bronchofiberscopically, and it satisfactorily improved and healed (Fig. 2).
In the series of all 1015 patients, no pneumonectomy patient had BPFs, and BPFs eventually resulted in 3 patients (10%) out of 29 with POIB and in 4 (0.4%) out of 986 without this, being statistically significant (p < 0.0001). In the former three (all males), POIB was seen in one each in the left upper bronchus, right intermediate trunk and right lower bronchus, and endoscopic findings such as ulceration with evident partial necrosis could be identified in all three. One patient aged 61 years died of multiple organ failure due to a bronchopleural fistula at the right intermediate bronchial stump. Another patient aged 62 years had prolonged air leakage and symptoms of infection. In spite of additional therapy with antibiotics, BPF at the left upper bronchial stump and pyothorax occurred two months after the operation. As a result of the emergency surgery including coverage of the stump fistula with a flap of muscle, thoracoplasty followed by chest tube drainage of the infected space and appropriate antibiotic therapy, the BPF was cured. The remaining patient, aged 65 years, demonstrated BPF of the right lower bronchus. With chest tube drainage of the pleural space, spontaneous closure of the stump occurred. In the remaining 20 patients with no infection and BPFs, the POIB was followed bronchofiberscopically, and it satisfactorily improved and healed without any treatment.
Concerning preoperative history, a significant risk factor was diabetes mellitus (n = 4, POIB patients, p = 0.03), whereas pulmonary tuberculosis (n = 3, POIB patients, p = 0.45), fibrosis (n = 2, POIB patients, p = 0.23) and chemoradiotherapy (n = 3, POIB patients, p = 0.28) did not demonstrate any association with POIB (Table 1). Furthermore, no links were evident with acute respiratory distress syndrome (n = 1, POIB patient, p = 0.15), postoperative ventilation (n = 0, POIB patient, p = 0.81) and atrial fibrillation (n = 0, POIB patient, p = 0.68), although postoperative respiratory complications including pneumonia and pyothorax did demonstrate significant associations (n = 5, POIB patients, p < 0.0001; Table 1). All but two female patients were smokers with an average smoking habit of 54.6 pack-years (range 11–96 pack-years) (p < 0.0001; Table 1).
No differences were evident in POIB incidence among the four surgeons involved in this study.
4 Discussion
Our data show that POIB is closely related to BPF occurrence. Moreover, it is considered that POIB presenting with ulceration and partial necrosis can lead to BPFs, defective healing of stumps being due to ischemia [1,2]. However, based on our large study population, BPFs resulted in only 3 (10%) out of 29 POIB patients. Therefore, although the BPF is a serious and potentially life-threatening complication, most POIB are not fatal [4].
The importance of peribronchial tissue in wound healing is well established and LND plays a role in BPF formation [14–16]. Among our 29 POIB patients, more than 90% underwent subcarinal LND, suggesting this to be a major impact factor for POIB. Therefore, we consider that postoperative ischemia of the bronchial wall around stumps might be the cause of POIB. As far as we know, an association between POIB and subcarinal LND has not been indicated in prior publications. The subcarinal area is traversed by feeding bronchial arteries, and dissection here must be performed as carefully as possible to preserve some of these nutrient vessels [17]. Recently, Asamura et al. [10] clearly demonstrated that subcarinal LND was not always necessary for resection of tumors of the right upper lobe and left upper segment. In this study, POIB frequently occurred in patients undergoing lobectomy with subcarinal LND. Therefore, in two particular groups with tumors located in the right upper lobe or left upper segment, lobectomy accompanied with limited LND, omitting subcarinal LND may allow avoidance of POIB.
Both manual suture and stapled closure have been found to be safe and reliable methods of bronchial closure, and differences among the types of closure show no relation to POIB, as confirmed here [18,19].
According to the present study, such clinical variables as gender, smoking habits, diabetes mellitus and postoperative respiratory infection can contribute significantly to POIB. Although all these factors are well known as important risk factors for the lung cancer operations, there is particularly a strongly significant association between smoking and occurrence of POIB [4]. Perioperative factors such as pulmonary tuberculosis, pulmonary fibrosis, preoperative chemoradiotherapy, acute respiratory distress syndrome, postoperative ventilation, atrial fibrillation and other cardiovascular complications, on the other hand do not appear to be related to POIB occurrence.
In conclusion, although the rate of POIB is low, several definite background influences were identified in this study, which should be taken into consideration by surgeons [4,20]. For patients who are male smokers with diabetes mellitus and suffering from postoperative respiratory infection, particular attention and anti-infection therapy are needed to prevent POIB. In patients with POIB presenting with ulceration and evident partial necrosis, additional supportive therapy with antibiotics may be necessary. Moreover, surgeons must bear in mind the possibility of POIB occurrence, especially in cases undergoing right middle and lower, left upper, right lower or right middle lobectomy accompanied with subcarinal LND. Furthermore, in appropriate groups with right upper lobectomy and left upper lobectomy for tumors of the upper segment, limited mediastinal LND may help prevent POIB.
This study was supported by grants-in-aid from the Ministry of Education, Sports, Culture, Science and Technology, as well as grants from the Ministry of Health, Labour and Welfare, the Smoking Research Foundation, and the Vehicle Racing Commemorative Foundation. We thank Dr. Shigehiro Tsuchiya (1956–2005) for his clinical advice.
