Closure of Large Intrathoracic Airway DefectsUsing Extrathoracic Muscle FlapsAntoine J. H. Meyer, MD, Thorsten Krueger, MD, Domenico Lepori, MD,Michael Dusmet, MD, John-David Aubert, MD, Philippe Pasche, MD, andHans-Beat Ris, MDThoracic Surgery Unit, Department of Radiology, Division of Pulmonary Medicine, Department of Head and Neck Surgery,CHUV, University of Lausanne, Lausanne, Switzerland

Background. Prospective assessment of pedicled ex-trathoracic muscle flaps for the closure of large intratho-racic airway defects after noncircumferential resection insituations where an end-to-end reconstruction seemedrisky (defects of > 4-cm length, desmoplastic reactionsafter previous infection or radiochemotherapy).

Methods. From 1996 to 2001, 13 intrathoracic muscletranspositions (6 latissimus dorsi and 7 serratus anteriormuscle flaps) were performed to close defects of theintrathoracic airways after noncircumferential resectionfor tumor (n � 5), large tracheoesophageal fistula (n � 2),delayed tracheal injury (n � 1) and bronchopleuralfistula (n � 5). In 2 patients, the extent of the trachealdefect required reinforcement of the reconstruction byuse of a rib segment embedded into the muscle flapfollowed by temporary tracheal stenting. Patient fol-low-up was by clinical examination bronchoscopy andbiopsy, pulmonary function tests, and dynamic virtualbronchoscopy by computed tomographic (CT) scan dur-ing inspiration and expiration.

Results. The airway defects ranged from 2�1 cm to 8�4cm and involved up to 50% of the airway circumference.They were all successfully closed using muscle flaps withno mortality and all patients were extubated within 24hours. Bronchoscopy revealed epithelialization of thereconstructions without dehiscence, stenosis, or recur-rence of fistulas. The flow-volume loop was preserved inall patients and dynamic virtual bronchoscopy revealedno significant difference in the endoluminal cross surfaceareas of the airway between inspiration and expirationabove (45 � 21 mm2), at the site (76 � 23 mm2) and belowthe reconstruction (65 � 40 mm2).

Conclusions. Intrathoracic airway defects of up to 50%of the circumference may be repaired using extrathoracicmuscle flaps when an end-to-end reconstruction is notfeasible.

(Ann Thorac Surg 2004;77:397–405)© 2004 by The Society of Thoracic Surgeons

The increasing use of complex reconstructions of theintrathoracic airways and neoadjuvant therapy fol-

lowed by resection for thoracic malignancies have re-newed an interest in extrathoracic muscle flaps for me-diastinal reinforcement after surgery [1]. Theintrathoracic transposition of the serratus anterior or thelatissimus dorsi muscle have been described in thetreatment of infected residual spaces [2–7], postpneumo-nectomy empyema with or without bronchopleural fis-tula [8], prosthesis infection after superior vena cavareplacement [9], tracheal erosion caused by posttrau-matic aortic aneurysm [10], nonmalignant tracheoesoph-ageal fistula [11], prophylactic mediastinal reinforcementafter complex tracheobronchial reconstructions [12], andafter radiotherapy [13]. The latissimus dorsi and serratusanterior muscle can be harvested easily through a pos-terolateral thoracotomy incision and their intrathoracictransposition is not associated with a significant addi-tional morbidity [12].

Several reports have described the use of intrathoracic

transposition of chest wall muscles for the closure ofpostpneumonectomy bronchial stump insufficiencywhere a primary suture of the debrided stump seemedunwise [8, 14]. The muscle was sutured into the bronchialdefect without attempting a primary closure of thestump. We have extended this technique to close in-trathoracic airway defects of variable localization andextent after noncircumferential resections in situationswhere an end-to-end reconstruction seemed risky. Thisstudy focuses on the functional and morphologic resultsusing extrathoracic muscle flaps for the closure of largeintrathoracic airway defects after noncircumferentialresection.

Patients and Methods

All patients who underwent closure of intrathoracic air-way defects after noncircumferential resection using anextrathoracic muscle flap (latissimus dorsi or serrratusanterior muscle) between 1996 and 2001 were the subjectof this study. Patient follow-up was prospectively byclinical and radiologic examination, repeated bronchos-copy and biopsy, pulmonary function tests including

Accepted for publication July 21, 2003.

Address reprint requests to Dr Ris, Thoracic Surgery Unit, CHUV,CH-1011 Lausanne, Switzerland; e-mail: hans-beat.ris@chuv.hospvd.ch.

© 2004 by The Society of Thoracic Surgeons 0003-4975/04/$30.00Published by Elsevier Inc doi:10.1016/S0003-4975(03)01462-0

GEN

ERA

LT

HO

RA

CIC

flow-volume loop measurements, and virtual dynamiccomputed tomographic (CT) bronchoscopy during inspi-ration and expiration.

Inclusion criteria consisted of patients requiring clo-sure of an intrathoracic airway defect after noncircum-ferential resection in situations where an end-to endreconstruction seemed risky or not feasible. Indicationsfor this technique in our series were the following:redo-operations for chronic bronchopleural fistula (BPF)with a short bronchial stump and empyema; repair of theairway for nonmalignant tracheoesophageal fistula (TEF)and delayed tracheal injury; and closure of airway defectsafter noncircumferential resection for tumors involvingthe intrathoracic trachea or the carina. In all patients thedecision to refrain from end-to-end reconstruction wastaken before the operation and was based on the lengthof required resection and the estimated risk of excessiveanastomotic tension.

TechniqueThe latissimus dorsi or serratus anterior muscle weredissected through a posterolateral thoracotomy incisionwhile preserving their proximal vascular blood supply[12]. Preference was given to the latissimus dorsi muscle,although the serratus anterior muscle was used if thelatissimus dorsi muscle had been divided by a previousintervention. Noncircumferential resections of the cen-tral intrathoracic airways were performed through astandard posterolateral thoracotomy. The extrathoracicpedicled muscle flap was transposed into the chest cavityvia an accessory thoracotomy through the bed of aresected segment of the second rib. It was sutured intothe airway defect with resorbable interrupted suturesunder slight tension and with bronchoscopic control inorder to maintain stability of the airways and to preventprotrusion of the muscle into the lumen. Thirty percent ofthe airway circumference was the maximal circumferen-tial extent of resection accepted for reconstruction by useof a muscle flap alone.

In patients where there was a defect of more than 30%of the airway circumference, mechanical reinforcementof the reconstruction was obtained by embedding a ribsegment into the muscle flap (Fig 1). In this situation,thoracotomy was performed through the bed of thefourth rib, which was resected after deperiostation. A15-cm long segment of the deperiosted rib was dividedalong its long axis while preserving its width using alueur (Fig 1A). A pouch was then created within thelatissimus dorsi flap after intrathoracic transposition. Forthis purpose, the muscle was carefully split along itsmuscle fibers with scissors and by blunt dissection and apouch of 4-cm width and 15-cm depth was created alongthe long axis of the muscle. The tailored rib was thenembedded in this muscle pouch (Fig 1B). The latissimusdorsi muscle was sutured to the edges of the airwaydefect in a way that allowed the embedded rib segmentto bridge the airway defect in its long axis (Figs 1C and 2).This was followed by deployment of an endotrachealsilicon Y stent (Hood Laboratory, Pembroke, MA) at theend of the intervention just after extubation. The stentwas kept in place for 4 weeks and then removed under

bronchoscopic control (Fig 3). Fifty percent of the airwaycircumference was the maximal extent of resection ac-cepted for coverage by use of a reinforced muscle flabusing an embedded rib segment.

Fig 1. Closure of a tracheal airway defect after resection of morethan one third, but less than 50%, of the airway circumference usinga latissimus dorsi muscle reinforced by an embedded rib segment. (a)A 15-cm long segment of the fourth rib is harvested during thoracot-omy and divided along its long axis while preserving one corticalis.(b) A pouch is created within the muscle by blunt dissection alongits fibers, which allows embedding of the rib segment containing onecorticalis. (c) The muscle is sutured into the airway defect by inter-rupted resorbable sutures in a way that the embedded rib segmentbridges the defect in its long axis.

Fig 2. Intraoperative documentation demonstrating the tracheal de-fect with the endotracheal tube in place (*). The latissimus dorsimuscle flap (LD) with the embedded rib (RIB) is sutured into thecranial part of the airway defect.

398 MEYER ET AL Ann Thorac SurgMUSCLE FLAPS FOR AIRWAY RECONSTRUCTION 2004;77:397–405

GEN

ERA

LT

HO

RA

CIC

Follow-UpMorbidity and 30-day mortality were recorded. At 3 and6 months after the operation, the reconstructions wereassessed by use of pulmonary function tests, measure-ment of the flow-volume loop and bronchoscopy includ-ing biopsy at the level of the reconstruction with histo-logic assessment of the specimen. A dynamic assessment

of the reconstruction during respiration was performed 6months after the operation by use of virtual dynamic CTbronchoscopy with three-dimensional reconstructionsabove, at the level of and below the reconstruction duringinspiration and expiration (Fig 4). The endoluminal crosssurface areas of the airway were measured above, at thelevel of and below the reconstruction during inspirationand expiration and the difference between inspirationand expiration was recorded for each level assessed[15–17]. This examination was performed in all patientsexcept for those requiring closure of airway defects forBPF since no airway instability had been reported inprevious series after airway repair for BPF using muscleflaps.

The patients were followed on clinical and radiologicgrounds until December 2002, or until death with respectto recurrent infection, fistula, or malignancy. The pairedStudent’s t test and a bidirectional hypothesis were usedfor statistical analysis where appropriate and significancewas accepted at p less than 0.05.

Results

From 1996 to 2001, 13 patients underwent closure of largeintrathoracic airway defects using an extrathoracic mus-cle flap (Table 1). There were 11 men and 2 women witha mean age of 56 years old (range 14–70 years old). Themean follow-up time after the operation was 24 months,ranging from 6 to 72 months. The type of reconstructionperformed in the 13 patients is illustrated in Figure 5.

Redo-operations for chronic BPF with a short bronchialstump and empyema were performed in 5 patients. Inthese 5 patients, heavy desmoplastic reaction was fore-seen at reoperation and the inability to close the airway

Fig 3. Closure of a tracheal defect of 8�4 cm by a latissimus dorsiflap reinforced by an embedded rib segment after noncircumferentialresection of more than one-third, but less than 50%, of the trachealcircumference. Postoperative result assessed by bronchoscopy8-weeks after the repair and 4-weeks after stent removal (same pa-tient as in Fig 2).

Fig 4. Assessment of tracheocarinal recon-structions using muscle flaps by virtual bron-choscopy using three-dimensional reconstruc-tion of a spiral computed tomographic scanduring respiration: (a) measurements of theendoluminal cross-surface areas of the airwayperformed above, at the level, and below thereconstruction; (b) during inspiration; (c) dur-ing expiration.

399Ann Thorac Surg MEYER ET AL2004;77:397–405 MUSCLE FLAPS FOR AIRWAY RECONSTRUCTION

GEN

ERA

LT

HO

RA

CIC

defect by direct suture. One patient presented with a BPFand empyema following right upper lobectomy whowould not tolerate a completion pneumonectomy due tocongestive heart failure. He underwent decortication,debridement of the airway, and closure of the 2 � 1 cmdefect of the intermediate and main stem bronchus usingan extrathoracic muscle flap (Fig 5A). Four patientsunderwent closure of a carinal airway defect rangingfrom 2 � 2 to 3 � 2 cm after debridement of a large BPFand a short bronchial stump following right pneumonec-tomy (Fig 5B). All 4 patients had previous mediastinalirradiation ranging from 48 to 64 Gy. After double lumenintubation, rethoracotomy was performed and the serra-tus anterior muscle was dissected. The chest cavity wascleaned and the bronchial stump was debrided, whichrequired division of the azygos vein and dissection of thecarina. The muscle was sutured into the airway defectafter intrathoracic transposition. The cavity was packedwith Betadine-soaked towels that were changed everysecond day through a transitory open window thoracos-tomy until the chest cavity was covered by granulationtissue. Obliteration of the chest cavity was then per-formed by filling the cavity with Clagett solution and the

skin was closed. Closure of the airway was successful inall patients without recurrent fistula or empyema.

Closure of central airway defects for nonmalignant TEFand delayed tracheal injury were performed in 2 and 1patient, respectively (Fig 5D). One TEF occurred 4 yearsafter Ivor-Levis operation and mediastinal irradiationbetween the distal intrathoracic membranous part of thetrachea and the gastric pull-up and measured 3 � 1 cmafter debridement without evidence of tumor recurrence.A 14-year-old patient presented with a destroyed rightlung due to recurrent bronchoaspiration related to alarge congenital TEF at the level of the distal intrathoracictrachea which measured 2 � 4 cm after debridement. Theairway defect was closed using an intrathoracic trans-posed extrathoracic muscle flap, which was sutured intothe defect and interposed between the primarily repairedesophagus and gastric pull-up, respectively. The secondrequired a pneumonectomy for destroyed lung. Onepatient presented with a delayed postintubation injury ofthe intrathoracic trachea leaving a defect of 8 � 2 cm thatwas closed using an extrathoracic muscle flap. In all 3patients a closure of the airway using a muscle flap wasanticipated due to the extent of the defect and thecircumstances rendering an end-to-end reconstructionhazardous. No recurrence of fistula was observed in bothpatients with TEF during follow-up. Postoperative heal-ing after delayed tracheal injury was uneventful despitethe fact that intubation had been required for almost 2weeks prior to the repair.

Closure of the airway after noncircumferential resec-tion for tumors involving the intrathoracic trachea or thecarina was performed in 5 patients. Three patients pre-sented with a noncircumferential squamous cell carci-noma of the intrathoracic trachea with an extension ofgreater than 4 cm in length rendering primary resectionand end-to-end anastomosis hazardous. The first patientwas treated with three cycles of cisplatin (100 mg/m2) anddoxetacel (75 mg/m2) followed by a block of concomitantboost accelerated radiotherapy to a total dose of 64 Gy.He exhibited a near complete response as judged onendoscopic and CT evaluation and the residual tumorseemed completely resectable by noncircumferential re-section. The airway defect measured 8 � 4 cm aftercomplete resection and was closed using a latissimusdorsi flap reinforced by an embedded rib segment (Figs1–3, 5D). The second patient revealed a 4-cm long,noncircumferential tumor of the membranous part of theintrathoracic trachea and was treated by noncircumfer-ential resection in healthy tissue leaving an airway defectof 5 � 3 cm, which was also closed by use of a latissimusdorsi flap reinforced by an embedded rib segment (Fig5D). The third patient revealed local tumor recurrenceafter resection and end-to-end reconstruction of theintrathoracic trachea for squamous cell carcinoma. Therecurrent tumor was completely resected by noncircum-ferential resection and the resulting airway defect of5 � 2 cm was closed by a muscle flap without reinforce-ment (Fig 5D). Two patients had a centrally located T4nonsmall cell lung cancer (NSCLC) involving the mainstem bronchus and the tracheobronchial angle withoutevidence of contralateral lymph node disease. They were

Table 1. Characteristics of Patients UndergoingNoncircumferential Tracheocarinal Resections andReconstruction Using Extrathoracic Muscle Flaps

Age(years) Indication

Size of AirwayDefect (cm) Flap

69 BPF after RUL for NSCLC 2 � 1 SA59 BPF after right P and RT for

NSCLC3 � 2 SA

47 BPF after right EPP and RT formesothelioma

2 � 2 SA

70 BPF after right P and RT forNSCLC

2 � 2 SA

58 BPF after right EPP and RT formesothelioma

2 � 2 SA

53 T4 NSCLC RUL after RT-chemotherapy and P

4 � 2 SA

60 T4 NSCLC RUL after RT-chemotherapy and P

4 � 2 LD

70 Recurrent squamous carcinomaof trachea after resection withend-to-end anastomosisfollowed by RT

5 � 2 LD

69 Squamous carcinoma of tracheaafter RT-chemotherapy

8 � 4 LD � R

47 Squamous carcinoma of thetrachea

5 � 3 LD � R

14 congenital TEF and destroyedlung requiring P

2 � 4 LD

64 acquired TEF after Ivor-Levisesophagectomy and RT foresophageal cancer

3 � 1 SA

45 Delayed postintubation trachealinjury

8 � 2 LD

BPF � bronchopleural fistula; EPP � extrapleural pneumo-nectomy; LD � latissimus dorsi muscle; NSCLC � nonsmall celllung cancer; P � pneumonectomy; R � muscle flap reinforced byembedded rib segment; RT � radiotherapy; RUL � right upperlobectomy; SA � serratus anterior muscle; TEF � tracheoesoph-ageal fistula.

400 MEYER ET AL Ann Thorac SurgMUSCLE FLAPS FOR AIRWAY RECONSTRUCTION 2004;77:397–405

GEN

ERA

LT

HO

RA

CIC

treated with three cycles of cisplatin (100 mg/m2) anddoxetacel (75 mg/m2) followed by a block of concomitantboost accelerated radiotherapy to a total dose of 44 Gy.Partial response occurred and resection requiring pneu-monectomy in both patients was performed. A carinalwedge resection was performed in tumor-free tissueincluding a 2-cm long resection of the lateral tracheo-bronchial angle and the airway defect of 4 � 2 cm wasclosed by a muscle flap (Fig 5C).

All 5 patients had a complete resection with microscop-ically tumor-free resection margins. Three patients dieddue to metastatic disease during follow-up. One patientdeveloped a local recurrence at the site of trachealresection 12-months after resection.

Postoperative MorbidityThere was no 30-day mortality. Postoperative extubationwithin the first 24 hours was achieved in every patientwithout the need for reintubation, tracheostomy, or pro-longed mechanical ventilation. No stent dislocation wasobserved in the 2 patients requiring temporary trachealstenting for 4 weeks. Two patients required coverage of awound dehiscence arising over a winged scapula after aserratus anterior transposition by use of a contralaterallatissimus dorsi flap.

Endobronchial AssessmentEvery patient had repeated postoperative endoscopiccontrols during hospitalization, as well as at 3 and 6months postoperatively. No airway complications wererecorded, no stenosis, muscle protrusion, airway instabil-ity, suture dehiscence, or recurrence of bronchopleural

fistula were observed. This also holds true for the 2patients with resection of the airway of up to 50% of thecircumference closed by a latissimus dorsi flap reinforcedby an embedded rib segment and stented for 4 weeks. At3 months, bronchoscopy indicated that the muscle flapwas covered by an intact epithelial layer in all patients.Bronchoscopic biopsies harvested at the site of repairdemonstrated islets of squamous cell epithelium inte-grated into granulation tissue at 3 months, and apseudostratified ciliated epithelium of respiratory typecovering the muscle flap at 6 months after the operation(Fig 6).

Pulmonary Function TestsThe flow-volume loop was preserved in every patient atassessment 3 and 6 months after the operation. The meanpostoperative forced expiratory volume at 50% expiration(FEF50) was 2.18 � 0.64 L/s and the mean postoperativeforced inspiratory volume at 50% inspiration (FIF50) 3.6 �0.92 L/s. The mean differences between the predicted andpostoperative measured FEF50 and FIF50 values were8.4% and 1.7%, respectively (p � 0.05).

Virtual bronchoscopy during inspiration and expira-tion was performed in all but 3 patients (patients 3, 5, and11 of Table 1) 6 months after the operation and revealedno significant difference of the endoluminal cross surfacearea of the airway between inspiration and expirationabove, at the level of and below the reconstruction,respectively (p � 0.05; Table 2). The mean cross surfacearea above the level of the resection was 364 � 96 mm2

and decreased during expiration by 13% � 6%. At themiddle of the reconstruction these values were respec-

Fig 5. Pattern of intrathoracic airway defects closed by muscle flaps: (a) defect in the main stem and intermediate bronchus resulting from BPFafter right upper lobectomy in a patient who would not tolerate completion pneumonecomy; (b) defect at the level of the carina after debride-ment of BPF following right pneumonectomy and radiotherapy; (c) tracheocarinal defect after pneumonectomy and carinal wedge resection fol-lowing neoadjuvant radiochemotherapy for centrally located NSCLC; (d) tracheocarinal defect after noncircumferential resection for trachealmalignancies, nonmalignant tracheoesophageal fistula, or delayed tracheal injury (see text). (BPF � bronchopleural fistula; NSCLC � non-small cell lung cancer.)

401Ann Thorac Surg MEYER ET AL2004;77:397–405 MUSCLE FLAPS FOR AIRWAY RECONSTRUCTION

GEN

ERA

LT

HO

RA

CIC

tively 356 � 156 mm2 and 21% � 6%. Below the recon-struction they were 348 � 134 mm2 and 19% � 11%. In thetwo reconstructions where the latissimus dorsi flap wasreinforced by an embedded rib segment, the CT scan at6 months revealed an intact rib segment without signs ofresorbtion, sequestration or heterotopic calcification ofthe muscle.

Comment

End-to-end reconstructions of the intrathoracic tracheo-carinal airways have demonstrated excellent results inexperienced hands. However, a postoperative mortalityrate of 12.7% has been reported in 135 patients undergo-ing carinal resections with a 30% mortality after left

carinal pneumonectomy [18–21]. The major predictors ofmortality and morbidity have been the length of theresection and the development of anastomotic complica-tions. Excessive anastomotic tension has been reported tobe the single greatest cause of failure after trachealoperations. However, this cannot always be avoideddespite careful patient selection and meticulous surgicaltechnique, especially if long segments have to be re-sected, or if a resection has to be performed in thepresence of desmoplastic reactions of the airway afterprevious surgery, infections or mediastinal irradiation[22]. Because there is actually no clinically establishedsubstitute available for complete tracheal or carinal re-placement [23], partial noncircumferential airway resec-tion followed by autologous reconstructions may be avalid alternative procedure in situations where completeresection and end to end anastomosis seems risky [24].Intrathoracic transposition of chest wall muscles, namelythe serratus anterior or latissimus dorsi muscle havebeen used for many years for mediastinal reinforcementas well as for the obliteration of residual infected spaces.They have demonstrated their usefulness as well as lowmorbidity in relation to their harvesting. The mechanicalstrength and versatility of these flaps have inspiredseveral authors to use them for closure of bronchopleuralfistula associated with a postpneumonectomy empyemain order to avoid a hazardous primary closure of theairways [8, 14, 25]. The muscle is sutured to the edges ofthe debrided bronchial stump without any attempt toclose the stump primarily.

We have extended this concept by using these muscleflaps as substitutes for closing large intrathoracic airwaydefects, in situations where circumferential resection andend-to-end anastomosis seemed not feasible or risky.Indications for this technique in our series were redo-operations for chronic BPF with a short bronchial stumpand empyema; repair of the airway for nonmalignantTEF and delayed tracheal injury; and closure of theairway after noncircumferential resection for tumors in-volving the intrathoracic trachea (tracheal tumors) or thecarina (centrally located NSCLC) after neoadjuvantradiochemotherapy.

Redo-Operations for Chronic BPFA chronic postpneumonectomy bronchopleural fistula isusually managed by debridement and closure of thebronchial stump with interrupted sutures, reinforcementof the mediastinum by omentum or a muscle flap andpacking of the cavity through a transitory open windowthoracotomy [5, 7, 8]. Secondary obliteration of the cavityis performed using the Clagett procedure once the cavity

Fig 6. Histologic assessment of a bronchoscopic biopsy performed atthe site of airway closure by a latissimus dorsi flap 6-months afterthe operation, revealing a pseudostratified ciliated epithelium of re-spiratory type covering the muscle flap (magnification �10).

Table 2. Assessment of Endoluminal Cross-Surface Areas of Airway (Mean � SD) After Closure of Intrathoracic AirwayDefects With Muscle Flaps by Dynamic Virtual CT Bronchoscopy During Inspiration and Expiration

Endoluminal Cross-Surface Areas (mm2) Inspiration Expiration Difference p Value

Above the reconstruction 364 � 96 318 � 99 45 � 21 p � 0.05At the level of the reconstruction 356 � 156 280 � 148 76 � 23 p � 0.05Below the reconstruction 348 � 134 283 � 111 65 � 40 p � 0.05

CT � computer tomography; SD � standard deviation.

402 MEYER ET AL Ann Thorac SurgMUSCLE FLAPS FOR AIRWAY RECONSTRUCTION 2004;77:397–405

GEN

ERA

LT

HO

RA

CIC

and the mediastinum is covered by granulation tissue [6].However, primary closure of an insufficient short right-sided bronchial stump after mediastinal irradiation maynot be feasible due to the desmoplastic reaction of thetissues and recurrent fistulisation may occur. Recentreports have demonstrated that extrathoracic pedicledmuscle flaps can be sutured under slight tension into thebronchial defect without attempting primary closure ofthe stump and that healing of BPF and empyema can beobtained [8, 12, 14]. Our results confirm these findings; allof the postpneumonectomy bronchopleural fistula in ourseries healed after closure of the airway by use ofextrathoracic muscle flaps, debridement and packing ofthe cavity followed by its obliteration according toClagett and Geraci [6]. Moreover, 4 of 5 patients treatedin this way had previous mediastinal irradiation up to 64Gy. In addition, this technique was also applied withsuccess to a patient with chronic empyema after upperright lobectomy and a large bronchopleural fistula withnecrotic edges rendering a sleeve resection dangerousand who would not have tolerated a completionpneumonectomy.

Airway Repair for Nonmalignant TEF and DelayedTracheal InjuryIn 2 patients, intrathoracic muscle transfer was success-fully used for closure of a large airway defect followingdebridement of nonmalignant TEF at the level of thedistal trachea, and in one of an 8-cm long, delayed injuryof the intrathoracic membranous part of the trachea.Small TEF can be managed by division of the fistula anddirect repair of the trachea and the esophagus, but largertracheal defects may require tracheal resection and re-construction [26]. In case of extensive defects of themembranous wall exceeding the length of the tracheathat can be safely resected, it has been suggested toprimarily close longitudinally the posterior wall of thetrachea, or if the reconstruction seems impossible, tocover the defect by a strap muscle and to maintain airwaypatency with a T-tube [27]. Our results indicate thatclosure of large defects of the membranous wall of theintrathoracic trachea using extrathoracic muscle flapsmay be a valid alternative to a hazardous primary repairof the airway, especially after previous mediastinal irra-diation. A tension-free reconstruction of the airways wasachieved by suturing the muscle flap into the airwaydefect. At the same time, this allowed the interposition ofthe muscle flap between the repaired esophagus andairway [19, 20, 23, 28]. Both patients were extubatedwithout requiring stents and did not have symptoms ofstridor or dysphagia during follow-up.

The same holds true for the closure of lengthy intratho-racic defects of the membranous trachea after delayedinjury. These lesions usually result from endotrachealintubation with longitudinal tears of the membranouspart of the intrathoracic trachea and can involve a longsegment of the trachea. Although fresh lesions can berepaired by direct suture, distal and delayed lesions mayrequire formal debridement and reconstruction of theairway [29]. Our results suggest that closing such anairway defect after debridement may safely be done by

suturing an extrathoracic muscle flap into the debridedairway defect in situations where an exceedingly longinjury precludes intrathoracic end to end reconstructionof the trachea.

Closure of Airway Defects After NoncircumferentialResection for Tumors Involving Intrathoracic Tracheaor CarinaExtrathoracic muscle flaps were also used with success inour series for the repair of trachea or carinal defects afterresection of tumors, especially after neoadjuvant radio-chemotherapy. Two patients with centrally locatedNSCLC involving the tracheobronchial angle receivedneoadjuvant radiochemotheray and underwent pneumo-nectomy and carinal wedge resection, the resulting air-way defect of 4 � 2 cm was reconstructed by use of apedicled latissimus dorsi muscle flap. Several reportshave demonstrated that long-term survival can be ob-tained after resection of T4 tumors involving the carinaprovided these tumors have undergone neoadjuvantchemotherapy or radiochemotherapy with clinical re-sponse [30, 31]. However, carinal resection and carinalpneumonectomy bear a high morbidity and mortalityrate after neoadjuvant therapy. Our results suggest thatcomplete resection may be obtained with a carinal wedgepneumonectomy instead of a sleeve pneumonectomyand that the airway defect may be successfully closedusing extrathoracic muscle flaps which obviates an end toend reconstruction of the airway after induction therapy.

Three patients underwent a noncircumferential resec-tion of the intrathoracic trachea for squamous carcinoma.One of these patients had a recurrent tumor after previ-ous tracheal resection. The defects ranged from 5 to 8 cmin their long axis, and from 30% to 50% of the airwaycircumference. In order to maintain the mechanical sta-bility of the airways after a resection of greater than 5-cmlength and greater than 30% of their circumference, thereconstruction was reinforced by embedding a rib seg-ment into the muscle which bridged the defect in its longaxis, followed by the placement of an endotracheal sili-con Y-stent for 4 weeks. The stent was well tolerated forup to 4 weeks by the patient and prevented the airwayscollapse in the early postoperative phase. The Y-shapedstent allowed its maintenance in the airways and did notinterfere with the epithelialisation of the muscle flap.Complete epithelialization of the muscle flap wasachieved in both patients with a rib segment embeddedin the flap.

Circumferential tracheal or carinal resections with endto end anastomosis remains the standard for the treat-ment of tracheal malignancies [32]. However, primarytracheal tumors often present with locally advanceddisease and tumor length is the most important determi-nant of resectability. The safe limits of tracheal resectionare highly individual but an intrathoracic tracheocarinallesion of greater than 4-cm length cannot be bridgedwithout excessive tension [32]. Downstaging of the tumorsize by means of neoadjuvant treatment in order toincrease the resectability rate may be an attractive ap-proach, however, preoperative radiotherapy can havedesastrous effects on anastomotic healing. Our results

403Ann Thorac Surg MEYER ET AL2004;77:397–405 MUSCLE FLAPS FOR AIRWAY RECONSTRUCTION

GEN

ERA

LT

HO

RA

CIC

suggest that noncircumferential resection after neoadju-vant chemoradiotherapy may offer a complete resectionand that airway defects of less than one third of circum-ference ranging up to 6 cm may be closed using extrath-roacic muscle flaps. Defects of up to 8 cm in length andup to one-half the circumference may be bridged usingmuscle flaps reinforced by an embedded rib segment.However, these observations need confirmation in alarger series of patients.

Bronchoscopic evaluation and pulmonary functiontesting of the reconstructions revealed patent airways inall patients without stenosis or muscle flap protrusion. Inaddition, the dynamics of the reconstructions were as-sessed by use of virtual dynamic CT bronchoscopy dur-ing respiration in order to detect stenotic or malacicsegments of the intrathoracic airway [33, 34]. The endolu-minal airway surfaces above, at the level, and below thereconstruction were measured at maximal inspirationand expiration and the differences in endoluminal airwaysurfaces during respiration were not significantly differ-ent at each level assessed.

Our results emerging from a limited number of pa-tients suggest that closure of large intrathoracic airwaydefects is feasible by use of chest wall muscle flapsfollowing non-circumferential resection of up to 50% ofthe airway circumference and result in a functional andmorphologic integrity of the repaired airways. However,this is only indicated when an end-to-end reconstructionis not feasible, which remains the standard procedure,especially in patients with tracheal malignancy.

References1. Anderson TM, Miller JI Jr. Use of pleura, azygos vein,

pericardium, and muscle flaps in tracheobronchial surgery.Ann Thorac Surg 1995;60:729–33.

2. Abrashanoff X. Plastische Methode des Schliessung vonFistelgangen, welche von inneren Organen kommen. Zen-tralbl Chir 1911;38:186–91.

3. Robinson S. The treatment of chronic non-tuberculous em-pyema. Collected Papers Mayo Clin 1915;7:618–44.

4. Arnold PG, Pairolero PC. Intra-thoracic muscle flaps. Anaccount of their use in the management of 100 consecutivepatients. Ann Surg 1990;211:656–62.

5. Pairolero PC, Arnold PG, Trastek VF, Meland NB, Kay PP.Post-pneumonectomy empyema: the role of intrathoracicmuscle transposition. J Thorac Cardiovasc Surg 1990;99:958–68.

6. Clagett OT, Geraci JE. A procedure for the management ofpostpneumonectomy empyema. J Thorac Cardiovasc Surg1963;45:141–5.

7. Deschamps C, Pairolero PC, Allen MS, Trastek VF. Manage-ment of postpneumonectomy empyema and bronchopleuralfistula. Chest Surg Clin N Am 1996;6:519–27.

8. Allen MS. Bronchopleural fistula. Chest Surg Clin N Am1992;2:823–37.

9. Spaggiari L, Solli P, Veronesi G, Pastorino U. Intrathoracicmyoplasty for prosthesis infection after superior vena cavareplacement for lung cancer. Ann Thorac Surg 2002;74:1231–3.

10. Russo P, Orszulak TA, Arnold PG, Abel MD, Stanson AW.Concomitant repair of a chronic traumatic aortic aneurysmwith tracheal erosion. Ann Thorac Surg 1987;43:559–60.

11. Oliaro A, Rena O, Papalia E, et al. Surgical management ofacquired non-malignant tracheo-oesophageal fistulas. J Car-diovasc Surg 2001;42:257–60.

12. Widmer MK, Krueger T, Lardinois D, Banic A, Ris HB. Acomparative evaluation of intrathoracic latissimus dorsi andserratus anterior muscle transposition. Eur J CardiothoracSurg 2000;18:435–9.

13. Regnard JF, Icard P, Deneuville M, et al. Lung resection afterhigh doses of mediastinal radiotherapy (sixty grays or more).Reinforcement of bronchial healing with thoracic muscleflaps in nine cases. J Thorac Cardiovasc Surg 1994;107:607–10.

14. Smolle-Juettner FM, Pierer G, Schwarzl F, Pinter H, Ratzen-hofer B, Prause G, Friehs G. Life-saving muscle flaps intracheobronchial dehiscence following resection or trauma.Eur J Cardiothorac Surg 1997;12:351–5.

15. Ferreti GR, Knoplioch J, Bricault I, Brambilla C, Coulomb M.Central airway stenoses: preliminary results of spiral–CT-generated virtual bronchoscopy simulations in 29 patients.Eur Radiol 1997;7:854–9.

16. Rodenwaldt J, Schorn C, Grabbe E. Virtuelle Endoskopie desoberen Luftwege mit der Spiral-CT. Radiologe 2000;40:233–9.

17. Burke AJ, Vining DJ, McGuirt WF Jr, Postma G, Browne JD.Evaluation of airway obstruction using virtual endoscopy.Laryngoscope 2000;110:23–9.

18. Mansour KA, Lee RB, Miller JI Jr. Tracheal resections:lessons learned. Ann Thorac Surg 1994;57:1120–5.

19. Pearson FG, Todd TRJ, Cooper JD. Experience with primaryneoplasms of trachea and carina. J Thorac Cardiovasc Surg1984;88:511–8.

20. Grillo HC, Mathisen DJ. Primary tracheal tumors: treatmentand results. Ann Thorac Surg 1990;49:69–77.

21. Mitchell JD, Mathisen DJ, Wright CD, et al. Clinical experi-ence with carinal resection. J Thorac Cardiovasc Surg 1999;117:39–52.

22. Mulliken JB, Grillo HC. The limits of tracheal resection withprimary anastomosis: further anatomical studies in man.J Thorac Cardiovasc Surg 1968;55:418–21.

23. Stamatis G. Extensive tracheal resection and reconstruction.In: Friedman M, ed. Operative techniques in otolaryngol-ogy–head and neck surgery. Philadelphia: Saunders, 1997:142.

24. Sharpe DA, Moghissi K. Tracheal resection and reconstruc-tion: a review of 82 patients. Eur J Cardiothorac Surg1996;10:1040–5.

25. Zimmermann T, Muhrer KH, Padberg W, Schwemmle K.Closure of acute bronchial stump insufficiency with a mus-cular latissimus dorsi flap. J Thorac Cardiovasc Surg 1993;41:196–8.

26. Reed MF, Mathisen DJ. Tracheoesophageal fistula. ChestSurg Clin N Am 2003;13:271–89.

27. Mathisen DJ, Grillo HC, Wain JC, Hilgenberg AD. Manage-ment of acquired nonmalignant tracheoesophageal fistula.Ann Thorac Surg 1991;52:759–65.

28. Saito H, Tatsuzawa T, Kikkawa H, et al. Transpericardialbronchial closure with omentopexy for postpneumonectomybronchopleural fistula. Ann Thorac Surg 1989;47:312–3.

29. Shrager JB. Tracheal trauma. Chest Surg Clin N Am 2003;13:291–304.

30. Grunenwald DH, Andre F, Le Pechoux C, et al. Benefit ofsurgery after chemoradiotherapy in stage IIIB (T4 and/orN3) non-small cell lung cancer. J Thorac Cardiovasc Surg2001;122:796–802.

31. Rendina EA, Venuta F, De Giacomo T, et al. Inductionchemotherapy for T4 centrally located non-small cell lungcancer. J Thorac Cardiovasc Surg 1999;117:225–33.

32. Gaissert HA. Primary tracheal tumors. Chest Surg Clin NAm 2003;13:247–56.

33. Chinn RJ, Yang GZ, Congleton J, Mellor J, Geddes DM,Hansell DM. Three-dimensional computed tomographybronchoscopy using clinical datasets: a comparison withfiberoptic bronchoscopy. Clin Radiol 1997;52:830–6.

34. Mc Adams HP, Palmer SM, Erasmus JJ, et al. Bronchialanastomotic complications in lung transplant recipients:virtual bronchoscopy for non-invasive assessment. Radiol-ogy 1998;209:689–95.

404 MEYER ET AL Ann Thorac SurgMUSCLE FLAPS FOR AIRWAY RECONSTRUCTION 2004;77:397–405

GEN

ERA

LT

HO

RA

CIC