Repair of postpneumonectomy bronchopleural fistula using pedicled pericardial flap supported by fibrin glue

Abstract

Effective closure of the postpneumonectomy bronchopleural fistula (PBF) with the use of different techniques still remains a challenge for thoracic surgeons. The aim of this study was to evaluate the efficacy of modified method of PBF closure using pedicled pericardial flap (PPF) supported by fibrin glue (FG). The efficacy of the late PBF closure with the use of two surgical methods was compared. In 10 patients, the edges of the PBF were covered with FG and PPF. In the second group of nine patients, myoplasty was used to close the bronchial fistula. Postsurgical follow‐up was for 1 year. In the first group, the healing of the fistula was achieved in 100% of the cases, whereas in the second, myoplasty group, healing was achieved in only 66·67% of the cases. The number of complications was similar in both groups. Pericardial flap supported by fibrin glue can be an effective method adjunctive to the treatment of PBF in selected patients.

Introduction

Postpneumonectomy bronchopleural fistula (PBF) is a life‐threatening complication after pneumonectomy, occurring in nearly 1–12% of patients 1, 2, 3. Pathological communication between airway lumen and postpneumonectomy pleural space leads to rapid infection of pleural cavity, penetration of infectious fluid to the airways followed by pleural empyema, aspiration‐induced infection or abscesses on contralateral lung, respiratory failure and sepsis. PBF is associated with significant morbidity and a high mortality rate ranging between 18% and 50% 1, 3, 4. The treatment of PBF remains one of the most difficult challenges for thoracic surgeons and is still unsatisfactory. Most of the failures have been noted during the attempts of late PBF closure (>30 days after pneumonectomy). The surgical techniques used for PBF closure are differentiated based on the time of fistula manifestation, its size, the length and quality of bronchial wall, fixation in the mediastinum, availability of neighbouring tissues well supplied with blood and the surgeon's preferences and experience. Increasingly better treatment effect is obtained in selected patients using endoscopic intrabronchial techniques in combination with pleural space drainage 5, 6. In the remaining patients, the method of choice is an attempt to close the fistula through rethoracotomy or sternotomy access reinforced with the surrounding tissues. Different types of pedicled, well‐vascularised flaps, muscular (intercostals, greater pectoral and latissimus dorsi), pleural, pericardial fat, diaphragm, omentum or vena azygos, are used for late PBF coverage 7, 8. Having exhausted other possibilities, the treatment results in permanent deformation in the form of open‐window thoracostomy or persistent drainage. In this study, we recommend a modification of earlier known method of PBF closure with the use of pedicled pericardial flap with application of fibrin glue (PPF + FG) on the margins of the fistula. The efficacy of the method was estimated by comparing its results with the usage of traditional methods (TMs) of PBF closure in the form of myoplasty. Postoperative follow‐up comprised a period of 1 year.

Patients and methods

In the years 2000–2011, at the Department of Thoracic Surgery, General and Oncological Surgery, Medical University in Lodz, 26 patients with fistulas of main bronchi after pneumonectomy were treated. Patients with PBF came from three thoracosurgical centres in our region. In all, 241 pneumonectomies were performed. In 203 patients, non‐small cell lung cancer was an indication of pneumonectomy; in the remaining 38 patients, other lung pathologies including infectious diseases (31) such as abscesses, tuberculosis and bronchiectasia as well as benign neoplasms were detected (7). In our centre, PBF was diagnosed in 14 patients (5·81%) with the death rate of 28·57%.

In all 26 patients, bronchial stump closure was made with commercial mechanical staplers (Ethicon, Ethicon, Inc., Johnson & Johnson, Somerville, NJ and Auto Suture, Auto Suture Covidien, Mansfield, MA) during primary pneumonectomy. Stapling was performed by approximation of the membranous and the cartilaginous portion of the bronchus. No routine coverage of the bronchial stump was performed with autologous tissues to prevent against PBF.

PBFs occurred on the right side in 23 patients (88·46%) and on the left side in 3 patients (11·54%). Taking into account the time criterion, they were distinguished as [according to Algar et al. 8] early (up to 30 days after pneumonectomy) and late (>30 days) PBFs. Late PBF was diagnosed in 23 patients (88·64%) and early PBF in 3 patients (11·54%). A successful rethoracotomy was performed in three patients with early PBF. The bronchial stump was hand‐sutured and reinforced with the surrounding tissues (muscles and parietal pleura). In 23 patients with late PBF, the following methods were used to close the bronchial stump: PPF + FG (10), myoplasty (9), endoscopic intrabronchial closure with pleural drainage (3) and open‐window thoracostomy (1).

We focused on the comparison of the efficacy of late PBF closure via rethoracotomy access in two groups of patients: operated on by PPF + FG technique and by TMs with the use of myoplasty. Intrabronchial endoscopic technique and thoracostomy were excluded from further study. The clinical characteristics of 19 patients treated via rethoracotomy are presented in Table 1.

Table 1.

Characteristics of 19 patients with late postpneumonectomy bronchopleural fistula

Patient (sex)	Age (years)	Pneumonectomy indications	PBF manifestation (weeks)	Side	Drainage (days)	PBF diameter	Surgery method	Sepsis	Pleural packing	Complications	Hospitalisation time
1 (M)	50	Cancer	7	R	6	11 mm	PPF + FG	No	No	–	10
2 (M)	39	Multiple abscesses (TBC)	8	R	13 (amb)	TD	PPF + FG	Yes	Yes	Wound infection	17
3 (M)	65	Cancer	9	L	15 (amb)	TD	PPF + FG	No	Yes	Untightness (48 hours)	10
4 (W)		Multiple abscesses (TBC)	13	R	9	10 mm	PPF + FG	No	No	Wound infection	16
5 (M)	55	Cancer	16	R	23 (amb)	8 mm	PPF + FG	Yes	Yes	Empyema without fistula (heart failure)	13
6 (M)	47	Cancer	18	R	6	TD	PPF + FG	Yes	No	–	12
7 (M)	61	Cancer	25	P	10	12 mm	PPF + FG	No	No	Arrhythmia	15
8 (W)	53		28	R	5	6 mm	PPF + FG	No	No	–	11
9 (M)	59	Cancer	37	R	3	14 mm	PPF + FG	No	No	Empyema without fistula	9
10 (M)	54	Cancer	49	R	24 (amb)	10 mm	PPF + FG	No	No	Untightness (4 days). Empyema without fistula	11
11 (M)	67	Cancer	35	R	22 (amb)	13	MYO	No	No	PBF recurrence	32
12 (M)	51	Abscess	16	R	5	TD	MYO	Yes	Yes	Death	9
13 (M)	59	Cancer	21	L	7	7	MYO	No	No	Wound infection	14
14 (M)	49	Cancer	49	R	14 (amb)	14	MYO	No	No	Empyema without fistula	28
15 (W)	62	Bronchiectasis	8	R	12	TD	MYO	Yes	Yes	PBF recurrence	34
16 (M)	48	Cancer	27	R	4	6	MYO	No	No	–	12
17 (W)	59	Cancer	11	R	3	TD	MYO	Yes	No	PBF recurrence	24
18 (M)	64	Cancer	22	R	5	11	MYO	No	No	Subphernicus abscess wound infection	30
19 (M)	54	Cancer	14	R	8	8	MYO	No	No	–	14

PPF + FG, pedicled pericardial flap with fibrin glue; MYO, myoplasty; (M), man; (W), woman; R, right side; L, left side; TBC, tuberculosis; (amb), ambulatory drainage; TD, total dehiscence.

Both groups of patients were comparable. There were eight men and two women with a mean age of 54·7 years and a median of 54·5 years in PPF + FG group. The TM group included seven men and two women with a mean age of 57·0 years and a median of 59·0 years. In the PPF + FG group, primary non‐small cell lung carcinoma was an indication of pneumonectomy in eight patients. In the remaining two patients, pneumonectomy was necessary because of complications following earlier lung resection due to multiple lung abscesses in the course of tuberculosis. In TM group, non‐small cell lung carcinoma was an indication of pneumonectomy in seven patients, of lung abscess in one patient and of bronchiectasis with recurrent bleeding in the respiratory tract in one patient. The time of PBF manifestation in PPF + FG group was from 7 to 49 weeks after pneumonectomy with a mean of 21·00 weeks and a median of 17·00 weeks. In TM group, the fistula manifested itself within the period from 8 to 49 weeks with a mean of 22·56 weeks and a median of 21·00 weeks.

The diameter of the fistulas in both groups of patients ranged from 5 mm to total dehiscence of the bronchial stump (Table 1). In three patients from PPF + FG group and in two from TM group with severe local inflammatory process in pleural cavity and coexisting sepsis confirmed in clinical picture and by high values of inflammatory markers (leucocytosis, C‐reactive protein and procalcitonin), debridement of the pleural cavity with the application of three consecutive rethoracotomies with packing according to the technique described by Schneiter et al. 9 was performed prior to the bronchial fistula closure. The aim of such management was to sanitise the pleural space and to create better conditions for PBF healing. The patients were followed up on an outpatient basis for 1 year after hospitalisation.

Surgical technique

All patients were qualified for rethoracotomy after primary pneumonectomy. During the surgery, a mechanical debridement of pleural cavity with the exposure of the edges of bronchial fistula was made. In the TM group, coverage of the bronchial stump was made with pedicled intercostal muscle bundle (six patients) and with latissimus dorsi flap (four patients). Standard surgical techniques were applied in these patients similar to those described by Hollaus et al. 10.

In the next 10 patients from the PPF + FG group, a modified technique with the use of FG was applied to close the fistula to reinforce the hand‐sutured bronchial stump and then to cover it with a pericardial flap. In detail, patients were intubated with a double‐lumen, endotracheal tube to protect the contralateral lung until a PBF was definitively ruled out. The fistula margins were sutured with four single deep mattress stitches [3‐0 or 4‐0 polydioxanone, monofilament absorbable suture (PDS)] inserted ‘en masse’ through the bronchial wall and the surrounding mediastinal tissues and opposing each other. The edges of the open bronchial fistula were injected with FG (Tissucol; Baxter, Baxter Healthcare Corporation, Deerfield, IL) and immediately the fistula was closed with the prepared stitches (Figure 1). It took about 3–5 minutes to obtain a stable fibrin clot. Then, a pedicled flap of the anterior pericardium without phrenic vessels of about 3–4 cm width was prepared. The flap was long enough to cover the fistula with no tension at the free margin of about 2 cm. The flap was attached caplike with earlier used mattress stitches. To maintain tightness, the pericardial flap was sutured with single stitches (4‐0 PDS) every 1·5–2·0 cm (Figure 2). The bronchial stump was then checked for air leakage with 30 cm H₂O sustained airway pressure. No air leaks were detected during water test. Finally, the restoration of the pericardium to prevent cardiac herniation is essential. We applied routinely propylene mesh for that purpose (Figure 3). Continuous pleural irrigation was performed postoperatively in all patients until pleural lavage was clear.

Figure 1.

Visible application of fibrin glue on internal edges of the bronchial fistula, earlier single surgical stitches were performed.

Figure 2.

Closed fistula after knotting of surgical stitches, fixation of pericardial flap attached cap‐like to the bronchial stump.

Figure 3.

Final stage – the bronchial fistula covered with pericardial flap fixed to the surrounding tissues with single stitches, pericardial sack reconstructed with polypropylene meshwork.

Result of the surgery

Healing of the fistula was obtained in all 10 patients in PPF + FG group (100%). The duration of pleural cavity drainage was from 3 to 24 days with a mean of 11·4 days and a median of 9·5 days. In four patients, in whom no clear lavage was obtained within 10 days postoperatively, the drainage was continued on the outpatient basis. The patients were discharged from the hospital with compact mobile drainage system. When the pleural space was sanitised, the drainage was removed in all four patients. The total time of hospitalisation was from 9 to 17 days with a mean of 12·4 days and a median of 11·5 days.

In TM group, closure of the fistula was obtained in six patients (66·67%). The duration of pleural cavity drainage was from 3 to 22 days with a mean of 8·89 days and a median of 7·00 days. The compact drainage was continued in two patients after hospital discharge. It was finally removed, respectively, on days 14 and 22 postoperatively. The time of hospitalisation was from 9 to 34 days with a mean of 21·89 days and a median of 24·00 days. Comparison of the treatment results in both groups is presented in Figure 4.

Figure 4.

Comparison of the results of the therapy in patients with postpneumonectomy bronchopleural fistula (PBF) treated with the use of two surgical techniques: TM (traditional methods: myoplasty and omentoplasty) and PPF + FG (pedicled pericardial flap with fibrin glue administration).

In TM group, one patient died during hospitalisation because of circulatory or respiratory failure. The following local complications were observed in this group of patients: three recurrencies of bronchial fistula, two cases of wound infection, one recurrence of empyema without fistula and one subphrenic abscess. In two patients with PBF recurrence, persistent drainage was applied as further treatment, and in one case, endobronchial stent was inserted to control the fistula. The patient with subphrenic abscess required transcutaneous drainage.

In PPF + FG group, no hospital death was noted. In two patients, initially a small air leak was observed to the passive drainage system and it stopped, respectively, after 2 and 4 days. Within 6 postoperative months, recurrence of pleural empyema was seen in three patients but without bronchial fistula. In two patients, we managed to sterilise the pleural cavity with continuous pleural lavage. In the last of these patients, the empyema maintained despite the drainage and 4 weeks after the surgery it resulted in heart failure caused by tight reconstruction of the pericardium, which required reoperation. Two weeks later, the patient died. Among less severe complications, infection of the thoracotomy wound in two patients and arrhythmia in one patient were observed.

Discussion

The treatment of late PBF with pleural empyema is a difficult and unrewarding task, reluctantly undertaken by thoracic surgeons. Surgeries via thoracotomy access with translocation of autologous tissue flaps or via sternotomy are usually complicated and dangerous and the results of treatment are unsatisfactory. Thus, an attempt of PBF closure by endoscopic methods is an attractive option. Endoscopic treatment of PBFs is based on the delivery of biological glue, coils, covered stents and sealants to the PBF site 5. If appropriate selection of patients is done, the healing of fistulas can reach 100% 5. However, this technique can be applied only in selected patients and most of them require another surgery. According to Hollaus et al., only selected patients, with fistulas <8 mm and no life‐threatening sepsis originating from empyema, can be qualified for primary bronchoscopic treatment 11. The overall rate of fistula closure was 35·6%, unfortunately 35·6% of patients required surgical intervention 11.

The success rate of PBFs surgical closure has been reported in the range from 75% to 85% with the mortality rate from 19·5% to 71% 12, 13, 14. At present, the base for a successful therapy of an open bronchial stump is its closure and reinforcement with well‐vascularised tissue 7. Because of ischaemia at the bronchial edge and infected environment, the choice of the autologous tissue for coverage of PBF appears to be of crucial importance for optimal results. Various pedicled flaps have been used to cover the bronchial stump depending on the local conditions, their availability as well as the preferences and experience of the surgeon. They can originate from thoracic muscles, diaphragm, pericardial fat or pericardiophrenic graft, pleura, omentum or azygos vein 7. The idea of the use of pericardial flap for reinforcement of bronchial stump during pneumonectomy is not a new one and it was described for the first time by Brewer et al. as an alternative method to the pericardial fat graft 15.

Each of these methods has its limitations. Pleural flaps, although being the most frequently used structure, usually have a disadvantage that they are extremely thin and do not have adequate blood supply 7, 16. In the study by Lindner et al. in 100 patients subjected to pneumonectomy treated by pleural coverage of bronchial stump, PBF occurred in 6% of the cases 17.

In the case of a diaphragm and omentum flap, in the late postsurgical period, the diaphragm is rigid and shortened; it is difficult to transfer the meshwork from abdominal cavity 18. Usually, the patient is in severe condition, and the surgery takes an extra 50–70 minutes and the other body cavity is opened and exposed to contamination. In such conditions, omentoplasty can be unsuccessful in about 50% of patients 19.

Intercostal muscle flaps have been used preferentially in some institutions 20. Their vascularisation can be sometimes poor at the end of the operation, and they are frequently crushed because of the application of rib retractor during surgery. In the later period, there is a possibility of the development of heterotopic ossification based on the periosteum of the rib, which can be a serious problem 21. In our experience, intercostal muscle flap is easy for preparation; however, it is too narrow and there are problems with fixing it around the bronchial stump. It does not adhere well to the bronchial stump and it has a tendency to contract and to necrosis. Other muscle flaps, latissimus dorsi and serratus anterior flaps, have larger surface but they contract easily, have poor circulation and necrosis is often observed, and thus they are not effective in bronchial stump repair. In our experience, the results of PBF closure with the use of myoplasty turned out to be worse than the coverage of the fistula using pericardial flap supported by fibrin glue.

As far as pericardium is concerned, in the current literature, the use of pericardial fat or pericardium itself is described for bronchial stump coverage during pneumonectomy to prevent the development of fistula and very seldom as the separate method of late PBF management 22. Frequent insufficiencies as regards the length, width, consistency, vascularisation and the risk for pericarditis are mentioned among the drawbacks of pericardial flaps 19, 23. However, these remarks are related to flaps such as pericardial fat and pleural but not to pericardial flaps. In our experience, we have observed that the pericardial flap demonstrates useful properties. Owing to the location of pericardium in the neighbourhood of PBF and the diameter of the pericardium it is easily accessible and thus, there is no problem with adjusting its size to bronchial fistula coverage. None of our patients treated by PPF+FG method manifested ischaemia or necrosis. As far as the consistency is concerned, the pericardial sack wall becomes thicker (to 5–10 mm) because of earlier inflammatory process and thus, it becomes more resistant. After fixing with single stitches, the pericardial flap demonstrates effective adhesiveness to the bronchial stump. It should be remembered that at the end of the surgery the pericardium must be reconstructed to avoid heart dislocation. In this study, no purulent pericarditis was observed in the early postoperative period despite the potential risk of this complication. The use of biological glue is an important element of the described PBF closure technique. The use of FG for reinforcement of the bronchial stump after pulmonary resection is well known 24. In our opinion, the biological glue improves the tightness obtained after knotting the ends at the fistula edges. This technique is used relatively frequently during endoscopic procedure 5, 6. It is applied less frequently as the adjunctive method in PBF closure during rethoracotomy.

Comparing the results of PBF treatment in our patients, obtained with the use of two methods, it can be stated that the described PPF + FG technique is more effective as regards the number of healed fistulas and shorter hospitalisation time. However, the mean time of pleural cavity drainage turned out to be longer than that in patients operated on by a TM method. The absence of perioperative death should be emphasised as a benefit of this technique. The number of complications in general and death rate within the 1‐year follow‐up were similar. Obviously, the efficacy of the PPF+FG method should be confirmed in the observation of larger patient population. The results of such a study will be the subject of consecutive publication.

Conclusion

Pericardial flap can be an effective method adjunctive to the treatment of PBF in selected patients. Good properties of pedicled pericardial flap increase the chances of healing of the fistula closed with surgical stitches and fibrin glue. It seems that during pneumonectomy, the neighbouring autologous tissues should be used routinely more frequently for bronchial stump coverage to prevent bronchial fistula development.

Author contribution

Sławomir Jabłoński was involved in study design, data collection, data interpretation, manuscript preparation, literature search and funds collection; Marian Brocki was involved in data collection and literature search; Piotr Klejszmit and Leszek Kutwin were involved in data collection; Marcin Wawrzycki was involved in data collection, data interpretation and literature search and Jacek Śmigielski was involved in data collection and literature search.