Prophylactic flap coverage and the incidence of bronchopleural fistulae after pneumonectomy

Abstract

A best evidence topic in thoracic surgery was written according to a structured protocol. The question addressed was ‘In pneumonectomy patients, is buttressing the bronchial stump associated with a reduced incidence of bronchopleural fistula?’. Fifty-seven papers were found using the reported search, of which 12 represented the best evidence to answer the clinical question. The authors, journal, date and country of publication, patient group studied, study type, relevant outcomes and results of these papers are tabulated. One prospective randomized controlled trial was identified, which found significantly lower rates of bronchopleural fistula and empyema after pneumonectomy with the use of pedicled intercostal flap buttressing. Intercostal muscle flaps and pericardial flaps have been used in case series of high-risk patients, e.g. those with neoadjuvant therapy or extended resections, with low rates of subsequent bronchopleural fistulae. There is the least-reported evidence for thoracodorsal artery perforator and omental flaps. There is relatively little published evidence beyond the single randomized trial identified, with only a few comparison studies to guide clinicians. We conclude that there is evidence for flap buttressing in reducing the risk of bronchopleural fistulae after pneumonectomy in diabetic patients. Flap coverage in other high-risk situations, such as extrapleural or completion pneumonectomy, has been reported in case series with good results. Of the reported techniques, the evidence is strongest for the pedicled inter-costal flap.

INTRODUCTION

A best evidence topic was constructed according to a structured protocol. This is fully described in the ICVTS [1].

THREE-PART QUESTION

In [patients undergoing pneumonectomy] is [bronchial stump coverage] superior to [stump closure alone] in order to reduce the incidence of [bronchopleural fistula]?

CLINICAL SCENARIO

Your next patient is an elderly gentleman requiring a pneumonectomy for lung cancer. He has multiple comorbidities. You are worried about his risk of bronchopleural fistula (BPF) after surgery. Before the operation you consult the literature on types of flap coverage of the bronchial stump and the evidence that they reduce BPF risk.

SEARCH STRATEGY

Medline 1950 to May 2012 using OVID, Pubmed, Medline and Embase interface [flap.mp OR flap bronchial stump/] AND [flap pneumonectomy.mp or pneumonectomy.mp or fistula.mp].

SEARCH OUTCOME

Fifty-seven papers were found using the reported search. From these, 12 papers were identified as providing the best evidence to answer the question (Table 1).

Table 1:

Review of current literature

Author, date, journal and country Study type (level of evidence)	Patient group	Outcomes	Key results	Comments
Sfyridis et al. (2007), Ann Thorac Surg, Greece [2] Randomized controlled trial (level 1b)	70 patients from a single centre (Metaxa Anticancer Hospital, Athens) who had lung malignancy over 46 months. Patients had diabetes mellitus (DM) >5 years. All DM treatments included	Incidence of BPF postoperatively	Lower incidence of both BPF (0% vs 8.8%, respectively P = 0.02) and empyema (0% vs 7.4%, P = 0.05) in intercostal muscle flap group than control	IF in diabetic patients decrease the incidence of both BPF and empyema compared with conventional closure alone
Klepetko et al. (1999), Eur J Cardiothoracic Surg, Austria [5] Cohort study (level 2b)	Retrospective analysis of 129 consecutive patients who underwent pneumonectomy by one surgeon at a single centre during a 10-year period (University of Vienna) with 96.1% follow-up. Patients had a pneumonectomy for lung malignancy (n = 123) or infectious diseases (n = 6). Coverage techniques included pericardium, mediastinal tissue, parietal pleura, azygos vein and intercostal muscle	Incidence of BPF	BPF occurred in 1 patient (0.8%) who had pleura covering	No statistically significant result to confirm coverage of the bronchial stump reduces the incidence of BPF. The authors suggest pericardial flaps could further decrease the risk
Taghavi et al. (2005), Ann Thorac Surg, Austria [6] Case series (level 2b)	93 patients who underwent pneumonectomy for primary lung cancer, all using a pedicled pericardial flap for bronchial stump coverage. Patients were followed up for a maximum of 12 months	Incidence of clinically evident BPF. Perioperative mortality	Perioperative mortality 4.3% (n = 1), which was not believed to be flap-related. No BPF identified during follow-up	Many of these patients also reported in Klepetko et al. pedicled pericardial flap use is a safe and effective method
Lindner et al. (2010), Asian Cardiovasc Thorac Ann, Germany [7] Cohort study (level 4)	Retrospective analysis of 243 patient records from a single centre (Asklepio Fachkliniken Munchen-Gauting) over a 7-year period. These patients underwent a pneumonectomy for lung cancer. The bronchial stump was reinforced either with a PCFP or by buttressing with pleura and local tissues	Incidence of BPF. Post use of PCFP compared with coverage with pleura and surrounding tissues	No statistically significant difference in overall incidence of BPF between the use of PCFP or pleura (4.9% compared with 6%, respectively). Absolute risk difference for PCFP vs pleura is 1.1% Other risk factors for BPF included T stage 4: incidence of BPF 12% compared with T stage 1, 2, 3 and 3.5%, P = 0.017. Absolute risk difference 9%, relative risk reduction 71%	Choice of PCFP or local pleural buttressing did not affect the risk of BPF No true control group (without any bronchial buttressing) Advanced T stage and carcinomatous lymphangiosis at resection margin increases the incidence of BPF
Berna et al. (2012), Ann Thorac Surg, France [10] Case control (level 4)	10-month observational study of patients (n = 6) who had a completion pneumonectomy with a thoracodorsal TDAP bronchial stump reinforcement at a single centre	Morbidity and mortality of patients who had TDAP reinforcement postoperatively and at 30 days Radiological viability of flap on CT at 1 and 3 months	0% morbidity and mortality at 30 days postoperation Flap components (fatty tissue and skin) were all radiologically normal	Small series, showing the feasibility of this novel flap, no control group. Only completion pneumonectomies included. Short follow-up
Mineo and Ambrogi (1999), J Thorac Cardiovasc Surg, Italy [11] Case series (level 4)	10-year review of a 25-patient case series. Diaphragmatic flap used for various purposes. Only 9 used prophylactically after pneumonectomy	Perioperative mortality Diaphragmatic hernia	0% of patients 0% of patients	The diaphragmatic flap can be considered practical, safe and useful for suture line protection or treatment of BPF
Abolhoda et al. (2009), Tex Heart Inst J, USA [12] Case series (level 4)	Retrospective analysis of 10 patients during a 2-year period who had a pedicled latissiumus dorsi muscle flap to reinforce the bronchial stump. Mixed series of pneumonectomies/lobectomies for infectious, malignant or inflammatory disease. 10 (100%) patients were surgically high risk due to immunosuppression, malnourishment and severe COPD. 5 (50%) upper lobectomies, 2 (20%) pneumonectomies, 3 (30%) cavity closures with LD muscle for chronic empyema	Incidence of BPF and recurrent empyema postoperation	There were no BPFs, or empyemas postoperation	Use of the latissimus dorsi flap to seal the bronchial stump is feasible. Only 2 pneumonectomy patients included, making any further conclusion impossible
Deschamps et al. (2001), Ann Thorac Surg, USA [4] Case series (level 4)	Retrospective analysis of 713 pnemonectomies over a 13-year period at a single centre (Mayo Clinic, Rochester, MN, USA). Pneumonectomies were performed for malignancy or benign disease Separate statistical analyses of risk factors for both empyema and BPF are presented The bronchial stump was reinforced in 175 (15.6%), 111 of which used a muscle flap. These comprised 93 serratus anterior flaps, 14 latissimus dorsi flaps and four combination flaps	Development of empyema postoperatively BPF development postoperatively	Overall, 7.5% suffered empyema and 4.5% BPF. Benign disease, FEV1, diffusion capacity of lung to carbon monoxide (DLCO), low preoperative haemoglobin, right pneumonectomy, bronchial stump reinforcement (P = 0.007), completion pneumonectomy, chest tube removal timing, blood transfusion all increased the incidence of empyema. All factors above were statistically significant P ≤ 0.05−P < 0.0001 on univariate analysis Risk factors for BPF include benign disease, low FEV1, DLCO, right pneumonectomy, suture closure of the bronchial stump, bronchial stump reinforcement (P = 0.03) timing of chest tube removal, increased IV fluids in first 12 h, blood transfusion. All factors above were statistically significant P ≤ 0.05−P < 0.0001 on univariate analysis The authors briefly mention multivariable modelling, which did not demonstrate any independent risk factors for BPF or empyema	In this large series, reinforcement of the stump was associated with a higher unadjusted risk of both BPF and empyema. This relationship was not confirmed with multivariable modelling. With a relatively small proportion (15.6%) receiving reinforcement in a retrospective series, it is possible that the technique was reserved for high-risk patients, and the result is due to confounding by patient selection Suture of the bronchial stump alone did increase the risk of BPF compared with stapled closure A subgroup analysis of the different coverage techniques used for reinforcement did not occur
D'Andrilli et al. (2009), Ann Thorac Surg, Italy [13] Case series (level 4)	45 patients underwent omental flap transposition for bronchial stump protection at time of pneumonectomy (43, 95%) or early BPF post-pneumonectomy (2, 5%)	Mortality Major complication rate Non-life threatening complication rate	2.1% (1 of 45 patients) 9% (4 patients) 11.1% (5 of 45 patients)	No postomental flap BPFs reported. Authors state omental flap transposition is effective at prevention and treatment of post-pneumonectomy BPF. Low number of patients despite 11-year experience
Hamad et al. (2008), Eur J Cardiothorac Surg, Italy [8] Case series (level 4)	9-year experience with a pericardial flap to protect the bronchial stump following extrapleural pneumonectomy for malignant mesothelioma (50 patients)	Incidence of fistula Perioperative mortality	0% (0 of 50) patients 4% (2/50 patients)	The authors conclude that pericardial flaps are feasible, safe and effective without increasing operative time No control group was available for comparison
Lardinois et al. (2002), Eur J Cardiothorac Surg, Switzerland [9] Case series (level 4)	Prospective non-randomized comparison of full-thickness pedicled DF (13 patients) against IF (13 patients) for bronchial stump coverage. All patients underwent preoperative induction therapy Chemo-radiotherapy given in stage IIIa bulky multilevel N2 disease and stage IIIb in 5 (38%) patients in the intercostal flap group and in 7 (54%) patients with the diaphragmatic flap. Chemotherapy alone given in the remainder	30-day mortality Complications Gastro-oesophageal reflux Postoperative PFTs	Zero IF: 8% (1 in 13) vs DF: 38% (5 in 13). Zero reported symptoms I PFT occurred in pedicled diaphragm flap and 0 PFT in intercostal muscle flap	Trend towards a lower morbidity with IF than DF, but results not statistically significant. 3 visceral herniations were seen in DF patients vs none in IF
Algar et al. (2001), Ann Thorac Surg, Spain [3] Case series (level 4)	242 consecutive patients from one centre who had a pneumonectomy for lung cancer (Reina Sofia Hospital in Cordoba) 32 (13%) underwent neoadjuvant chemotherapy or radiotherapy 178 (74%) had autologous tissue coverage of the bronchial stump Bronchial stump coverage included intercostal muscle in 80 (44.9%) patients, mediastinal fat pad in 64 (36.1%), parietal pleura in15 (8.4%), phrenic pedicle in 10 (5.6%, azygos vein in 6 (3.3%) and other tissues in 3 (1.7%)	Incidence of postoperative BPF Perioperative mortality and morbidity were also reported	Overall mortality 5.4%, morbidity rate 59%. arrhythmia respiratory insufficiency and re-operation were the most commonly reported complications 13 patients (5.4%) developed BPF, of which 4 (30.8%) died Absence of bronchial stump coverage was an independent predictor of BPF in the final multivariable model (RR 1.65 uncovered vs covered, P = 0.039) Long bronchial stumps increased BPF risk (0.29 per additional mm, p = 0.0004), side (RR 1.61 right vs left, p = 0.059), postoperative mechanical ventilation (RR 2.74, P = 0.02) and COPD (RR 2.07, P = 0.028) were also independent predictors of BPF. Increased predicted postoperative FEV1 showed a trend towards some protective effect, P = 0.09)	This multivariate analysis identified independent risk factors for developing BPF including the absence of bronchial stump coverage The independent effect of stump coverage was significant, in this large single centre experience Several techniques were used, of which intercostal muscle and mediastinal fat pads were most common The great majority did not undergo neoadjuvant therapy

RESULTS

Sfyridis et al. [2] performed a randomized controlled trial on 70 diabetic patients undergoing pneumonectomy. They randomized patients 1:1 to either intercostal muscle flap reinforcement or no flap coverage. There was a significant reduction in BPF risk (absolute risk reduction 8.8%, relative risk reduction 100%, P = 0.02) and empyema (absolute reduction 7.4%, relative 100%, P = 0.05) in those undergoing flap coverage.

Several large series have also been reported.

Algar et al. [3] reported 242 pneumonectomies for malignancy at a single centre. The majority underwent stump coverage. Although several techniques were used, intercostal flaps (44.9%) and pericardial fat pads (PCFPs) (36.1%) predominated. A multivariable analysis of risk factors for BPF was undertaken. Failure to cover the stump was an independent risk factor for BPF, with a relative risk of 1.65, P = 0.039. Other risk factors included long bronchial stumps and right-sided resections. There were no statistically significant differences between the types of coverage used, but the authors preferred the use of intercostal muscle.

Deschamps et al. [4] analysed 713 pneumonectomy patients. In contrast to Algar, univariate analysis showed that the bronchial stump reinforcement was associated with an increased incidence of BPF (7.5% with flap vs 3.6% without flap, P = 0.037). Multiple techniques were used including muscle, parietal pleura and pericardium, serratus anterior, latissimus dorsi and a combination of muscle flaps. However, only 15.6% of patients underwent stump coverage, and the relationship was not confirmed on multivariable modelling. It is possible that the observed crude association was due to confounding by comorbidity.

Kleptko et al. [5] reported a cohort study of 129 patients. Multiple coverage techniques were used, with the pericardial flap preferentially chosen by the surgeon in high-risk patients. Coverage of the bronchial stump decreased the incidence of BPF. In the high-risk group, where pericardial flaps were used exclusively, there were no BPFs.

Taghavi et al. [6] studied 93 patients undergoing pericardial flaps. These were drawn from the same institution as Klepetko et al. during a mostly overlapping period. No BPFs were seen.

Lindner et al. [7] analysed 243 patients who underwent flap coverage using either a PCFP or pleura and other local tissues. They did not find a statistically significant difference in the incidence of BPF between using PCFP or pleural flap stump coverage. BPF were seen in both groups (4.9 and 6.0%, respectively). However, the authors favoured the PCFP due to its ease of use and minimal complications.

Hamad et al. [8] report a series of 50 patients undergoing bronchial stump coverage with pericardium following extra-pleural pneumonectomy for malignant mesothelioma. There were 2 deaths (4% perioperative mortality) due to cardiac complications, but no BPFs.

Other flap techniques have been reported in smaller series.

Lardinois et al. [9] present a non-randomized consecutive comparison study of 26 patients (1 undergoing either intercostal muscle or diaphragm flaps DF). Groups were matched by age, gender, side of pneumonectomy and tumour stage. 30-day mortality was zero. Complications of atelectasis, pneumonia, visceral herniation and BPF were seen in 8% after intercostal muscle flaps (IF) vs 38% in the DF group (P value not stated). No symptoms of gastro-oesophageal reflux (GORD) were reported in either group. The authors concluded that both methods were effective.

Berna et al. [10] reported six cases using the thoracodorsal artery perforator flap (TDAP) for bronchial reinforcement. No BPFs were seen.

Mineo and Ambrogi [11] reviewed 25 patients using diaphragmatic flaps for various indications, including 9 patients following pneumonectomy. Complications included myocardial infarction (1/9 patients), pulmonary oedema (1/9) and late fistula caused by recurrence at the bronchial stump (1/9). There were no perioperative deaths and no diaphragmatic herniae (follow-up 7–54 months). This was in contrast to Lardinois et al. [9], who noted three visceral diaphragmatic herniations in 13 patients who had diaphragmatic flaps placed.

Abolhoda et al. [12] presented a small case-series of high-risk patients undergoing a lobectomy or pneumonectomy (1 patient), using a pedicled Latissimus muscle flap coverage of the bronchial stump. No BPFs were seen.

D'Andrilli et al. [13] showed a small case-series using omental flap for coverage of the bronchial stump. There were no early BPFs, but no long-term follow-up.

In summary, the only randomized controlled trial by Sfyridis et al. [2] shows a reduced incidence of BPF and empyema following intercostal flap reinforcement.

There is the evidence from a separate large multivariable analysis that failure to place a flap is an independent predictor of subsequent BPF.

The optimal flap is unclear. Pedicled intercostal muscle is the only flap tested in a randomized trial, but pericardium, PCFP and pleura have also been reported in large series. Smaller series have shown the feasibility of other flaps.

CLINICAL BOTTOM LINE

There is the randomized trial evidence (level 1b) that IF reduce the BPF risk in diabetics. Other level 2 evidence supports flap coverage as a protective measure against BPF.

Coverage of the bronchial stump, probably by intercostal muscle pedicled flap, should be considered in patients considered to be at high risk of BPF.

Conflict of interest: none declared.