• 
• 

Three Part Question

In [patients with a bronchopleural fistula] is a [minimal access or bronchoscopic approach superior to conventional re-thoracotomy] for BPF closure in terms of [survival].

Clinical Scenario

You perform a right pneumonectomy for a T2 N1 M0 non-small-cell lung cancer, in a 77-year old diabetic with a history of ischaemic heart disease.
On the sixth postoperative day he deteriorates, with a copious productive cough, worsening gas exchange, pyrexia and rising inflammatory markers. Chest X-ray shows a fall in fluid level, and fibreoptic bronchoscopy confirms a bronchopleural fistula(BPF), with dehiscence of one third of your staple line. The intercostal drain you site recovers 700ml of purulent fluid. You feel that he would not survive a repeat thoracotomy and wonder if a bronchoscopic approach might work?

Search Strategy

Medline 1950-2007 database, and EMBASE database 1980-2007,using the Ovid interface .
[exp Bronchial fistula/ OR bronchopleural fistula.mp OR bronchial dehiscence.mp OR exp respiratory tract fistula/] AND [exp thoracoscopy/OR thoracoscopy.mp OR exp thoracic surgery, video-assisted/OR exp surgical procedures, minimally invasive/OR minimally invasive.mp OR minimal access.mp] AND [outcome.mp OR outcome$.mp OR exp intraoperative complications/OR exp empyema OR empyema.mp OR complication$.mp OR exp treatment outcome/]. The Cochrane Register of Systematic Reviews was also searched for relevant studies. Results were restricted to studies on human subjects published in the english language.

Search Outcome

The search identified 1052 abstracts in Medline and Embase. Abstracts were hand-searched, and several identified for full manuscript assessment. No randomised trials, 2 narrative review articles, several case series and many case reports were identified for manuscript review. Reports of two patients or less were excluded from further study. Reports containing both post-lobectomy and post-pneumonectomy were assessed on their post-pneumonectomy content alone. The papers finally considered to be the best evidence are summarized in the table.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Hollaus et al, 1998, Austria	45 post-resection BPFs, 40 (89%) postpneumonectomy, treated initially by bronchoscopic therapy. Fistulas >8mm diameter excluded. All male, median age 60.29 years. No stump recurrences of malignancy.	Case Series Level 4	Mortality, progression to surgical treatment, resolution of fistula and empyema.	9 patients (including 4 lobectomies) cured by endoscopic therapy alone (20% of total, 12.8% of pneumonectomy BPFs). 7 (15.6%) BPF resolved but chronic empyema remained 16 (35.6%) went on to surgery, in-patient mortality 8 (20%). Further 17 post-discharge deaths in mean follow-up 779 days. Overall mortality 57%	Large retrospective series. Treated with fibrin glue alone (fistulae <3mm) or fibrin and spongy calf bone (>3mm). Mixed series including 5 (11%) lobectomies. Rates of cure (resolution of both BPF and empyema) low in postpneumonectomy group, conversion to open surgery and mortality relatively high.
Varoli et al, 1998, Italy	35 patients with post-resection BPF, 19 postpneumonectomy, the remainder postlobectomy. 33 (94%) with associated empyema.	Case Series Level 4	Clinical resolution, time to clinical resolution, mortality.	11 (58%) postpneumonectomy fistulas "completely resolved" within 2-8 weeks (mean 5.5). Healed fistulas ranged in size from 2-10mm. 8 patients with "total" dehiscence did not heal, all "intermediate" or "late" fistulas. 5 patients (26%) died within 2 weeks of fistula appearance.	Retrospective series. Treatment with intercostal tube drainage for empyema, followed by repeated bronchoscopic submucosal injection of polidocanol (weekly). Definition of clinical resolution unclear. Some reporting of results does not differentiate between the post-pneumonectomy and post-lobectomy groups.
Han et al, 2006, China	6 patients with post-resection BPF, 5 post-pneumonectomy, 1 post-lobectomy. Pleural drains sited pre-procedure. Stump length varied from 18-35mm.	Case Series Level 4	Correct stent placement on bronchoscopy and helical CT, clinical resolution of BPF and empyema.	Stents successfully placed in all patients. 4 of 5 post-pneumonectomy fistulas closed and empyema resolved (67%). In 1 other patient, fistula closed but empyema remained. No mortality at mean follow up of 316 days (range 150-636).	Retrospective review. Radiological placement of covered metal stents, specifically designed with a blind-ended side arm to obstruct the stump. Device placed under topical anaesthesia. Complete medium-term follow-up with good results and complete survival in a small series. Size of fistulas treated not stated but presumably large.
Scappaticci et al, 2000, Italy	12 post-pneumonectomy BPFs within a series of 20 post-resection BPFs, 6 early-onset and 6 late.	Case Series Level 4	Clinical resolution of BPF and empyema, survival	Of 6 early BPFs, 3 <5mm closed successfully with resolution of associated empyema in one affected patient. In 3 BPFs >5mm no resolution of BPF or associated empyemas. 2 died, 1 later successfully treated with surgery In 6 late "small" fistulas, 5 successfully treated endoscopically, 1 progressed to surgery. Overall BPF cure rate 66.7%, empyema cure rate 10% (of 10), survival 83%.	Retrospective series. Bronchoscopic application of methyl-2-cyanoacrylate glue. Criteria for resolution of BPF and empyema not stated. Duration of follow-up not stated. Low cure rate for empyema- 0% in 6 late presentation BPFs.
Kiriyama et al, 2002, Japan	6 post-pneumonectomy BPFs within series of 8 post-resection BPFs. 2 with stump recurrence of malignancy, 1 with stump aspergilloma. All fistulae <2mm diameter. Onset from 1-9 months post-operatively. 5 of 6 (83%) had intercostal drain prior to bronchoscopy.	Case Series Level 4	Clinical resolution of BPF, survival, presence of ongoing clinical pleural space infection.	All initially closed with laser therapy, but only 2 without recurrence (33%). Recurrences occurred 3-12 days later. Follow-up 1.2-108 months. 4 deaths (66%), 3 from recurrent malignancy, 1 from invasive aspergilloma. No pleural space infection in 5 of 6 (83%).	Retrospective case series. Nd:YAG applied to peri-fistula endothelium via the flexible bronchoscope under local anaesthesia and sedation. Authors note that all successes were at higher energy doses (15-20W). Relatively low overall success rate.
Sabanathan and Richardson, 1994, UK	3 postpneumonectomy BPFs, 6 weeks (2 patients) to 32 weeks (1 patient) post-operatively. All with associated empyema, fistula size not stated.	Case Series Level 4	Resolutionof BPF and empyema, progress to surgery, mortality.	Treatment by immediate chest drain, followed by bronchoscopic closure with n-butyl-2-cyanoacrylate glue. Space then irrigated. Follow-up 3months-3 years. No mortality, 2 progressed to omentopexy, all resolved.	No mortality and resolution in all cases. However, 2 (66%) went on to open surgery by omentopexy, so the contribution of endoscopic gluing is unclear.

Comment(s)

Bronchopleural fistula is a disastrous complication following pneumonectomy, with high morbidity. Incidence after pneumonectomy in a recent series was 12%, and mortality with fistula was 67%[Hollaus, 1997]. Impaired respiratory mechanics, contralateral lung contamination and chronic pleural sepsis all contribute to its poor outcome. Conventional surgical strategies such as thoracoplasty or chest wall fenestration carry significant mortality in themselves, and long-term morbidity with pain and deformity. Reducing the morbidity of therapy with a minimal access approach is therefore appealing. The fistula site is accessible by bronchoscopy and thoracoscopy which may avoid major surgery in these high risk patients may reduce risk. Our search strategy produced reports of several different procedures. They can be broadly categorised as (1) tracheobronchial stenting, (2) occlusion with glue or other materials, and (3) procedures to induce scar tissue formation at the fistula. We decided to exclude from our table series of two cases or less. Case reports of other treatments, including thoracoscopic closure, were excluded on these grounds. The search also identified two recent topic reviews[Ng 1995, Lois]. Endobronchial stenting has been used to isolate fistulas in several case reports, but our search identified only 1 series of more than 2 patients[Han 2004]. The authors used specially-designed covered metal stents (with a blind-ending arm to fit the bronchial stump) in 6 patients. The mortality rate at a mean follow-up of 316 days was 0%, with 67% of patients cured of both fistula and empyema. Patients in this series underwent follow-up CT scan and bronchoscopy to confirm adequate stent placement. This initial study is encouraging, although it is not population-based, making the effect of case-selection hard to quantify. It does however establish the feasibility of the technique in selected patients. Fibrin glue occlusion was used in a large series of 45 post-resection fistulas (40 post-pneumonectomy) <8mm in diameter[Hollaus 1998, Hollaus 1997]. Only 20% of cases were cured with this technique, with resolution of both fistula and empyema. However, the majority had an adverse outcome, with 20% dying as in-patients, 35.6% progressing to surgery and 15.6% left with a chronic empyema. Another series using synthetic glue in 12 postpneumonectomy fistulas had a better fistula resolution rate (66.7%) and survival of 83%, but disappointingly only 10% of empyemas resolved[5]. The authors' impression was that small and late-presentation fistulas responded better to glue, although the numbers were too small for statistical analysis. Sabanathan also used cyanoacrylate glue in 3 patients, with no mortality, but resolution of BPF and empyema was only achieved in 1 patient (33%). The other 2 required open omentopexy. Lastly, reported scar-forming techniques have included the Nd:YAG laser[Kiriyama] and sub-mucosal injection of the vein sclerosant polidocanol [Varoli]. Resolution of BPF ranged from 33%-58%, and mortality from 26%-66%. Both of these techniques involved repeated bronchoscopic interventions in most patients. Other techniques included bronchoscopic submucosal alcohol injection[Takaoka] and thoracoscopic gluing or clipping[Torre]. The pooled mortality in amongst 90 patients reported in these 5 series (including 5 BPFs following lobectomy) is 39.6%. Of 85 postpneumonectomy BPFs reported in these series, 30.1% were cured of both BPF and associated empyema (when present) by bronchoscopic treatment alone. These outcomes could clearly still be improved. Fistula size, time of onset from surgery, bronchial stump length and aetiology of stump failure (e.g. tumour recurrence) vary. It may be that some strategies are useful for specific types of BPF. For example, several authors have suggested that glue and scarring techniques seem better suited to smaller fistulas[Hollaus 1998, Varoli, Scappaticci]. Concentrating on fistula subsets with specific techniques may improve results.

Clinical Bottom Line

In 85 post-pneumonectomy BPFs there was a 30% cure rate using a range of bronchoscopic techniques. The mortality was 40% reflecting the very high mortality with this complication. Bronchoscopic treatment has so far only been reported in small case series but may in the future offer further treatment options in patients too unwell to undergo re-thoracotomy.

References

1. Hollaus PH, Lax F, Janakiev D, Lucciarini P, Katz E, Kreuzer A, Pridun NS. Endoscopic treatment of postoperative bronchopleural fistula: experience with 45 cases. Ann Thorac Surg 1998; 66:923-7.
2. Varoli F, Roviaro G, Grignani F, Vergani C, Maciocco M, Rebuffat C. Endoscopic treatment of bronchopleural fistulas. Ann Thorac Surg 1998;65:807-9
3. Han X, Wu G, Li Y, Li M. A novel approach: treatment of bronchial stump fistula with a plugged, bullet-shaped, angled stent. Ann Thorac Surg 2006;81:1867-71.
4. Scappaticci E, Ardissone F, Ruffini E, Baldi S, Mancuso M. Postoperative bronchopleural fistula: endoscopic closure in 12 patients. Ann Thorac Surg 2000; 69:1629-30.
5. Kiriyama M, Fujii Y, Yamakawa Y, Fukai I, Yano M, Kaji M, Sasaki H. Endobronchial neodymium:yttrium-aluminum garnet laser for non-invasive closure of small proximal bronchopleural fistula after lung resection. Ann Thorac Surg 2002;73:945-9.
6. Sabanathan S, Richardson J. Management of postpneumonectomy bronchopleural fistulae. A review. J Cardiovasc Surg 1994;35:449-57
7. Hollaus PH, Lax F, El-Nashef BB, Hauck HH, Lucciarini P, Pridun NS. Natural history of bronchopleural fistula after pneumonectomy: a review of 96 cases. Ann Thorac Surg 1997;63:1391-7.
8. Ng CSH, Wan S, Lee TW, Wan IYP, Arifi AA, Yim APC. Post- pneumonectomy empyema: current management strategies. ANZ J Surg 2005;75:597-602.
9. Lois M, Noppen M. Bronchopleural fistulas: an overview of the problem with special focus on endoscopic management. Chest 2005;128(6):3955-3965.
10. Takaoka K, Inoue S, Ohira S. Central bronchopleural fistulas closed by bronchoscopic injection of absolute ethanol. Chest 2002:122(1):374-8.
11. Torre M, Chiesa G, Ravini M, Vercelloni M, Belloni PA. 1987: Endoscopic gluing of bronchopleural fistula. Updated in 1994. Ann Thorac Surg 1994;58(3):901-2.