Outcome after long-segment tracheal resection: study of 52 cases

Abstract

OBJECTIVES

Resection of long-segment trachea is challenging, and although 50% of adult trachea can be removed, anastomotic complications arise proportionally. Different release manoeuvres have been described to gain length and reduce tension at the suture line. The aim of the study was to evaluate the outcome when different release manoeuvres have been utilized during resection and reconstruction of the trachea.

METHODS

From January 2005 to December 2015, 52 patients with long segments of trachea ≥40 mm requiring resection and reconstruction were treated at our institute. Demographic, operative and postoperative data were retrospectively analysed.

RESULTS

Fifty-two patients with long-segment tracheal disorders ≥40 mm were analysed in this stud. Transient swallowing and phonation dysfunction occurred in 17 (32.2%) patients, exclusively in patients who underwent laryngeal release. Swallowing dysfunction was Grade I in all patients, except 2 who suffered Grade II dysphagia and were relieved in the early postoperative period. Forty-five (86.5%) patients were symptom free, and 7 (13.4%) patients were symptomatic (dyspnoea on exertion and/or stridor) and required reintervention. Four (7.6%) patients responded to 1 or 2 sessions of bronchoscopic dilatation, and 3 patients were left with permanent tracheostomies. Patients with neoplastic pathology were followed up without any neoplastic recurrence.

CONCLUSIONS

Long-segment resection and reconstruction of the trachea utilizing one or more release manoeuvres can be safely done, with low complication rates. Although swallowing and phonation dysfunction after laryngeal release were commonly encountered (almost one-third of patients), they were mild, transient, self-limited and recovered within 2–3 weeks of the early postoperative period.

Long-segment tracheal resection, Tracheal stenosis, Release manoeuvres

INTRODUCTION

Tracheal disorders requiring resection and reconstruction may be iatrogenic or neoplastic. It is most commonly encountered following postintubation tracheal stenosis [1]. Radical resectioning and end-to-end anastomosis carry high success rates with low morbidity. However, several reports have addressed anastomotic complications when reconstruction is performed under tension at the suture line showing increased potential for early dehiscence or late restenosis [2]. Practically, long-segment tracheal resections have higher tensions during reconstruction unless proper tracheal mobility is achieved. Several releasing manoeuvres like blunt dissection of the anterior and posterior tracheal wall, neck to chin, infrahyoid release [3], suprahyoid release [4], inferior pulmonary ligament dissection, pericardial dissection [5] and mediastinoscopic tracheal and bilateral bronchial release (MTBBR) [6, 7] were described. These releases, used singly or in combination, can be applied to gain tracheal mobility and decrease tension at the suture line to reduce complications. This retrospective study examined the outcome of different release manoeuvres used in 52 patients who underwent tracheal resection and reconstruction (TRR) for segments ≥40 mm.

METHODS

During a 10-year period, between January 2005 and December 2015, 52 patients underwent long-segment TRR for tracheal disorders, following neoplastic and acquired postintubation stenosis. They were performed by collaborating teams of surgeons from the Otolaryngology, head and neck and Cardiothoracic Surgery departments at Kasr Alainy Hospital, Cairo University. The hospital research board approved this study.

Patient criteria and preoperative workup

All patients were subjected to history taking, clinical examination, high-resolution computed tomography neck and chest ± 3D reconstruction of the tracheobronchial tree. Prior endoscopy was conducted routinely to evaluate vocal cords, glottis, cricoid and tracheal mucosa and wall. The length of the pathological segment was measured, its proximity to the vocal cords and degree of stenosis according to Meyer and Cotton grading scale were recorded [8]. Biopsies were taken in neoplastic cases. In all cases, several previous endoscopic dilatations were done and any prior endoluminal tumours were removed to secure the airway. Patients with aspiration, bilateral vocal cord paralysis, multilevel stenosis, long segments >50% of the trachea and chronic respiratory disease requiring future intubation were excluded from this study.

All patients gave informed consent for possible postoperative complications including recurrent laryngeal nerve palsy, possible need for tracheostomy whether permanent or temporary, restenosis, swallowing or phonation dysfunction and the need for a postoperative chin stitch. However, only in co-operative and receptive patients was the danger of hyperextension during the early postoperative period explained. None of the patients within this subgroup required a chin stitch.

Operative procedure

All patients were approached through a collar incision, though manubriotomy was necessary for 13 (25%) patients, where the impacted region affected the mediastinal trachea. We followed the techniques described by Pearson et al. [9] and Grillo et al. [10]. In all patients, the orotracheal tube was placed, and dilation was performed using rigid bronchoscopy to allow a small calibre tube to pass below the vocal folds. After resection, the distal trachea was intubated to allow anastomosis of the posterior suture line and removed with the passage of the orotracheal tube while performing the anterior suture line. In all patients, a transverse collar incision was made; however, a manubriotomy was added if the impacted region was in the lower cervical to the mediastinum.

Through a collar incision, the superior flap was elevated in the subplatysmal plane to the level of the hyoid bone, and inferiorly flap was elevated to expose normal trachea below the level of stenosis. The cricoid and the trachea were exposed with a division of the thyroid isthmus if needed. A midline incision at the level of stenotic segment was performed to identify the upper and lower limits of the stenosis, avoiding unnecessary resection of unaffected trachea or cricoid cartilage. Injury of the recurrent laryngeal nerves was avoided by confining sharp dissection to subperichondrial plane of the lateral wall of the trachea. Sharp dissection was used to separate the membranous parts of the trachea at the level of the stenotic segment and the oesophagus.

In cases where cricoid involvement was present, the anterior arch and rim of the lateral parts of the posterior lamina of the cricoid were removed to expose healthy mucosal edges.

Upward mobilization of the distal tracheal segment was achieved through blunt dissection of the anterior surface of the cervicomediastinal trachea.

Two Vicryl^® 2/0 stitches were placed at the lateral edges of the distal trachea, the anaesthesiologist flexes the neck, bringing the proximal end of the trachea downward, while the operator uses the previous 2 stiches to bring the distal end up, approximating the anastomotic line. During this manoeuvre, the proximity and tension at suture line was evaluated, and the need for a release procedure was determined.

Before 2011, laryngeal release utilizing suprahyoid and/or infrahyoid release were the procedures that gained extra length at the suture line to provide tension-free anastomosis. However, after 2011, we added tracheal and bilateral bronchial release utilizing video-assisted mediastinoscopy [6], which reduced the need for laryngeal release in many cases during our experience.

Our ENT team followed the technique described by Montgomery [4] for suprahyoid release with the exception that they did not divide the body of the hyoid bone. Instead, the upper edge of the bone was freed from muscle attachment as far as the lingual arteries bilaterally—the same technique used during laryngectomy to avoid any shearing of bone edges on soft tissue.

When tension at the suture line persisted, our ENT team perform infrahyoid release following the technique described by Dedo and Fishman [3], and this technique was performed alone or in combination other release procedures. Care was taken not to injure the superior laryngeal nerve.

Our cardiothoracic team also utilized the technique described by Kang et al. [6], which entails tracheal and bilateral main bronchial release using video-assisted mediastinoscopy. After dissection of the anterior and posterior wall of the trachea to the carina under direct vision, the anterior and posterior walls of both main bronchi were dissected under video-assisted mediastinoscopy guidance. The bronchi were released to the proximal part of the intermediate bronchus on the right and to the origin of the upper lobe bronchus on the left. Further dissection down was performed in cases where manubriotomy was necessary.

The anastomotic line was sutured utilizing absorbable sutures (4/0 Polydioxanone) in an interrupted fashion. The knots were carefully placed outside the tracheal lumen to avoid granulation tissue formation. The suture line started midline posteriorly and advanced laterally to both sides and ended at the midline anteriorly. The posterior sutures were tied first, in the same order of placement, before the anterior sutures were taken. After all anterior sutures were placed the endotracheal tube was removed, orotracheal intubation was performed and the tube balloon passed the suture line under direct vision before tying the anterior suture line was completed.

Electrocautery and other forms of energy were not used in any case. Bleeding and oozing were controlled using sterile Egyptian cotton soaked with diluted adrenaline (1/200 000) and compression at all time. For that reason Redivac drains were used routinely to avoid haematoma formation and were usually removed within 48 h.

During the earlier procedures, chin to manubrium braided silk sutures were placed to prevent neck hyperextension, which were removed 1 week after surgery. Later procedures were accompanied by preoperative educational instruction, and we relied on patients to prevent neck hyperextension when possible.

Our anaesthesia team provided extra care to ensure a smooth recovery for all patients before extubation by avoiding unnecessary neck movement during extubation, particularly in those without chin stitches. The neck was supported by several pillows to provide comfortable neck flexion. Cord mobility and laryngeal oedema are meticulously excluded before sending patients to the intensive care unit.

Postoperative and follow-up

Patients remained in the intensive care unit for a 24-h observation period and were kept on antibiotics, antiemetics and proton pump inhibitors for 1 week after surgery. Before discharge, symptoms of airway obstruction, phonation or swallowing dysfunction were addressed, and a fibre-optic bronchoscopy was conducted to assess the suture line. Our department’s postoperative policy included weekly follow-ups for all patients during the first month followed by monthly check-ups for 6 months. Patients with phonation or swallowing dysfunction were referred to speech therapy clinic at our institute for further management.

Anastomotic success was assessed clinically by the recurrence of symptoms of airway obstruction, which was confirmed by bronchoscopic demonstration of restenosis. If necessary, decisions were then made on the need and type of reintervention that was necessary. We considered the procedure a success if the following conditions were met: (i) asymptomatic patients with or without routine bronchoscopies demonstrating restenosis and (ii) symptomatic patients who underwent one or more endoscopic dilatation sessions and remained free of symptoms during their follow-up visits. On the other hand, the procedure was considered a failure if symptomatic patients were not improving despite repeated endoscopic dilatations and whose airway obstruction needed a permanent airway appliance (stent or tracheostomy).

RESULTS

Between January 2005 and December 2015, we performed TRR in 179 patients, the 52 of which were long tracheal segments ≥40 mm were analysed in this study. There were 37 males, with overall mean age of 34.5 years. The most common cause of tracheal resection was acquired due to postintubation tracheal stenosis. The main cause for intubation was head injury following road traffic accident (21 patients). Associated cricoid (cricotracheal) stenosis was present in 23 patients. Endoscopic assessment of all patients revealed Grade II stenosis in 30 patients and Grade III stenosis in 22 patients according to the Meyer–Cotton grading scale [8]. The estimated segment length ranged from 40 to 52 mm long (endoscopic) with a mean of 43.78 mm. Table 1 details this preoperative patient data.

Table 1:

Preoperative patient data

Variables
Age (years)
Range	18–56
Mean ± SD	34.5 ± 10.8
Sex
Men	37
Women	15
Ratio (men/women)	2.5:1
Causes of stenosis, n (%)
Acquired PITS	45 (86.5)
Neoplastic	7 (13.5)
Adenoid cystic carcinoma	5 (9.6)
Carcinoid	2 (3.8)
Duration of intubation (days)
Range	3–39
Mean ± SD	16.2 ± 7.96
Reason of intubation, n (%)
RTA	21 (40.3)
Major surgery	6 (11.5)
Suicide	6 (11.5)
Chest infection	5 (9.6)
ARDS	5 (9.6)
Abortion	2 (3.8)
Duration of development of stridor (days)
Range	14–120
Mean ± SD	47.7 ± 28.6
Comorbidity, n (%)
Hypertension	4 (7.6)
Neurological deficit	3 (5.7)
Anticoagulation	3 (5.7)
Diabetes mellitus	2 (3.8)
Endoscopic assessment, n (%)
Anatomical site of stenosis
Cervical trachea	35 (67.3)
Mediastinal trachea	13 (25)
Cricotracheal	4 (7.6)
Degree of stenosis
Grade II	30 (57.6)
Grade III	22 (42.3)
Length of stenotic segment
Range (mm)	40–52
Mean	43.78

ARDS: acute respiratory distress syndrome; PITS: postintubation tracheal stenosis; RTA: road traffic accident; SD: standard deviation.

According to the site and extent of stenosis, 26 (50%) patients had tracheal resection with cricotracheal anastomosis and 22 (42%) patients had tracheotracheal anastomosis. Thyrocricotracheal anastomosis was performed in 4 patients, who underwent additional cricoid resection. Cricoid resection ranged from resection of the arch in 2 patients and a rim of the lateral wall and lamina in the other 2 patients. The resected segments ranged from 40 to 54 mm with a mean of 44.23 ± 3.8 mm.

A variety of release procedures were used to achieve tension-free anastomosis, as described in Table 2. In all patients, the anterior wall of the trachea was bluntly dissected to the level of the carina, and no additional manoeuvre(s) were necessary in 8 (15.3%) patients. Additional suprahyoid release was performed in 8 (15.3%) patients, infrahyoid release in 1 (1.9%) patient and both manoeuvres were necessary in another 8 (15.3%) patients.

Table 2:

Details of the surgery

Details of the surgery	n = 52
Collar incision plus manubriotomy	13 (25)
Site of reconstruction anastomosis
Tracheotracheal	22 (42)
Cricotracheal	26 (50)
Thyrocricotracheal	4 (8)
Types of release incisions
None	8 (15.3)
Suprahyoid	8 (15.3)
Suprahyoid + infrahyoid	8 (15.3)
Infrahyoid	1 (1.9)
Infrahyoid + MTBBR	7 (13.4)
MTBBR	19 (36.5)
Suprahyoid + infrahyoid + MTBBR	1 (1.9)
Size of resected segment, mean ± SD	40–54 mm (44.23 ± 3.8 mm)

Values are represented as n (%) unless otherwise specified.

MTBBR: mediastinoscopic tracheal and bilateral bronchial release; SD: standard deviation.

We began using the MTBBR manoeuvre in 2011 following Kang et al. [6]. This release manoeuvre was used in 19 (36.5%). With the exception of 7 patients, whom required additional infrahyoid release to be performed. In 1 patient both supra and infrahyoid release were coupled with mediastinoscopic bilateral bronchial release.

All patients were extubated in the operating room and sent for observation for 24 h in the intensive care unit. In this study, 2 patients were re-explored due to complications. One for bleeding and the other for haematoma evacuation. One patient developed mild and non-progressive surgical emphysema, which was treated accordingly. One patient had bilateral cord paralysis after resectioning of the adenoid cystic carcinoma, which extended but did not infiltrate the surroundings, a tracheostomy was immediately placed. During follow-ups, improved cord movement was noted, and the patient regained speech. However, the patient refused any further reintervention after stenosis developed at the suture line.

Swallowing and phonation dysfunction were present in 17 (32.2%) patients and were exclusive to those patients who underwent laryngeal release. Swallowing dysfunction was Grade I in all patients, except 2 who suffered Grade II dysphagia [12]. It is worth mentioning that most patients improved within 10–20 days after reassurance, followed by visits to a speech and swallowing therapy clinic. Phonation dysfunction responded to voice rest and speech therapy.

In this study, restenosis was detected in 7 (13.4%) patients and occurred between 2 and 12 weeks postoperation. They were symptomatic and ranged from Grade II stenosis to Grade III stenosis. All but 3 patients responded to 1 or 2 sessions of endoscopic dilatation, who were tracheotomized. We did not encounter any patient with dehiscence in our study. Table 3 shows a summary of postoperative complications.

Table 3:

Postoperative course and follow-up

Postoperative course and follow-up	Number of patients (%) and outcome
Exploration for bleeding	2 (3.8)
Surgical emphysema	1 (1.9)
Swallowing and phonation dysfunction
Number of patients	17 (32.2)
Type of release
Laryngeal	17/25 (68)
Tracheal	0/27 (0)
Grade of dysphagia
Grade I (dysphagia to solid foods)	15 (28.8)
Grade II (dysphagia to soft foods)	2 (3.8)
Management	Reassurance and speech clinic
Course	Transient, regressive and self-limited to early postoperative period (10–20 days)
Restenosis
Number of patients	7 (13.4)
Appearance of symptoms	2–12 weeks postoperatively (1 after 2 weeks, 2 after 5 weeks, 1 after 7 weeks, 2 after 10 weeks and 1 after 12 weeks)
Management	4 patients responded to endoscopic dilatation
	3 patients tracheostomized
Tracheostomy
For bilateral cord paralysis and restenosis	1 (1.9)
For restenosis and failed dilatation	2 (3.8)
Freedom of symptoms (5 months–10 years)	45 (86.3)

The follow-up period ranged from 5 months to 10 years. A total of 45 (86.5%) patients (39 with postintubation stenosis and 6 patients with neoplastic pathology) were symptom free during this time. Seven (13.4%) patients (6 patients with postintubation stenosis and 1 patient with neoplastic pathology) were symptomatic (shortness of breath during exertion and/or stridor) and required reintervention. The 7 patients with neoplastic pathology had no tumour recurrence. Only 1 patient had bilateral vocal cord paralysis and restenosis and underwent tracheostomy. In total, 3 patients (one of neoplastic pathology and two of postintubation stenosis) were left with permanent tracheostomies.

DISCUSSION

Several studies have attempted to find a solution for a reliable, tension-free and safe anastomosis after long-segment tracheal resection. Mulliken and Grillo [5] showed that a combination of cervical and mediastinal release allowed for the removal of ∼5 cm (about 8 rings) with cervical flexion and that right hilar bronchial dissection permitted the removal of an additional 1.4 cm. In a later study, Grillo found that transthoracic division of the pulmonary ligament allowed the removal of almost 3 cm. Intrapericardial dissection permitted the removal of an additional 1 cm, and dividing the left main bronchus with reimplantation to the bronchus intermedius, an additional 2.7 cm [10].

Inferior mobilization of the upper trachea can be achieved by the downward decent of the larynx utilizing laryngeal release either suprahyoid [4], infrahyoid [3] or combined release [8]. There has been variability in extra length achieved by each procedure reported in the literature, ranging from 1 to 2.5 cm for each manoeuvre [13–15]. In our experience, the length removed after each different laryngeal release procedure varied by patient depending on tissue elasticity, muscle bulk, age and length of the neck. This intraoperative length achievement determines whether combined manoeuvres are needed to for a tension-free anastomosis.

In our study, 85% of patients needed some form of a release incision and mobilization to have a safe tension-free anastomosis. The most commonly used single release procedure was the tracheal and bilateral bronchial release (19/52 patients, 36.5%) via mediastinoscopy. Following the report by Kang et al. [6] in 2011, we frequently used this type of release because the superior mobilization of the distal trachea is far more important, and less demanding, than the inferior mobilization of the upper trachea for a tension-free anastomosis. More than half the patients in our study (52%) had blunt or MTBBR.

Mediastinoscopic dissection allowed the distal tracheal mobilization under direct vision, which avoided troublesome bleeding, which occasionally followed routine blind blunt dissection of the anterior wall of the trachea. Moreover, this method achieved more length than blunt tracheal dissection, ranging from 3 to 3.5 cm in some patients. Utilization of the MTBBR manoeuvre has reduced the need for laryngeal release in almost 33% of our patients. Similar to our observations, Kirschbaum and Teymoortash [7] used the same technique and reported achieving 4 cm in extra length while performing a long tracheal resection of a neoplasm.

Auchincloss and Wright [2] reviewed the incidence, risk factors, prevention and treatment of complications of TRR. They stated that complications, particularly those related to airway anastomosis, are infrequent but can be devastating. Several factors including length of resected trachea, pre-existing tracheal appliance, prior tracheal resection and medical comorbidities such as diabetes, have been shown to increase the risk of anastomotic complications. The largest series of TRR was published by Wright et al. [16], and it included 901 patients who underwent TRR and laryngo-TRR and the incidence of anastomotic complications in that series was 81 of 901 (9%) patients. This is similar to the 7 out of 52 (13.4%) patients had symptomatic postoperative anastomotic complications; all in the form of restenosis. We did not encounter anastomotic separation or dehiscence in our study.

Many authors have identified the length of the resected segment as the most important prognostic factor leading to anastomotic complications. Wright et al. [16] reported that the highest rate of anastomotic complications occurred in patients whose resected segment lengths were 4–6.5 cm, at a rate of 14%. Similarly, the average length of resected segments in our study was 4.4 cm ± 0.38 cm with a range of 4–5.4 cm, and the anastomotic complication rate was 13.4%. Out of the 7 patients in our series with anastomotic restenosis, 4 patients responded well to repeated tracheal dilatation, while 3 required permanent tracheostomy.

This study had no cases of post-resection anastomotic separation/dehiscence during the 10-year follow-up period. Although we had a relatively long-segment resection range of 40–54 mm, the key to a healthy and safe anastomosis performing resections tension and ischemia free. This emphasizes the importance of the different release manoeuvres and adequate mobilization of the proximal and distal tracheal segments. However, this should not be done at the expense of devitalizing the trachea and causing ischemia. The blood supply of the cervical trachea enters almost exclusively through the lateral walls. Therefore, care should be taken not to skeletonize more than 1 cm above and below the anastomotic line to safeguard against postoperative ischemia of the anastomosis with consequent dehiscence and/or restenosis.

A suprahyoid release, first described by Montgomery [4] in 1974, gives an extra 1–2 cm of length following upper tracheal resection. In our study, a suprahyoid release was used on 17 (32%) patients whether singly or combined with other release incisions. The suprahyoid release manoeuvre impairs normal motion of the larynx during swallowing. Furthermore, extended tracheal resection may serve the same function by tethering the larynx and preventing normal elevation [16]. In addition, some component of recurrent laryngeal nerve palsy may cause some degree of dysphagia and phonation dysfunction that usually resolves itself in 2–3 weeks [11]. Grillo [17] has emphasized this possibility in his review; however, the degree of dysphagia varies among patients, from mild transient to moderate or even severe and permanent. Peskind et al. [13] who performed total supralaryngeal releases (combined suprahyoid and infrahyoid release) noted transient dysphagia in most of their patients. However, in Peskind’s series, there were a few patients with persistent dysphagia that recovered after speech and swallowing therapy.

Negm et al. [11] performed a suprahyoid release on 24 patients with laryngeotracheal stenosis, and 12 of them (50%) complained of postoperative dysphagia. In our study, swallowing and phonation dysfunction were present in 17 (32.2%) patients and were exclusive for those patients who underwent laryngeal release (7 patients who underwent suprahyoid release, 6 patients who underwent combined release and 4 patients who underwent infrahyoid release). Swallowing dysfunction was Grade I (difficult eating solid foods) [12] in all patients except 2 patients who suffered Grade II dysphagia (difficult eating soft foods) [12].

Our ENT team, unlike Montgomery [4], dissected the hyoid bone as far as the lingual arteries bilaterally and avoided transection of the hyoid body. A point of argument with this technique would be the high incidence of swallowing and phonation dysfunction in our series. Yet, the symptoms were mild, transient and regressive during the 2–3-week early postoperative period. Review of the various published series shows that difficulty swallowing and phonation dysfunction consistently follows laryngeal release procedures whether supra, infrahyoid or combined. The grade of dysphagia varies from mild to severe, but in most reports, it was transient and patients recovered [11, 13, 14, 16, 17].

Within our study group, mild transient dysphagia occurred with all forms of laryngeal release, and combining the 2 procedures did not increase risk of morbidity in patients. Nonetheless, our management approach, combing patient reassurance, diligent postoperative check-ups with speech and swallowing therapy at speech clinic, helped facilitate satisfactory recovery in all patients. In our study, due the transient occurrence and mild form of swallowing and phonation dysfunction, we did not deem it necessary to conduct further investigations.

A chin stitch has been proposed to protect against disruption of the suture line due to hyperextension during the early postoperative period [16]. In practising this step, we observed devastating psychological impacts on many patients. While we agree that in children and patients with a less than average intelligence the chin stitch is mandatory, we have been practising the alternative of preoperative patient education with patients who are co-operative and intelligent. In these selected patients, we avoid a chin stitch, using extreme care during anesthetic recovery and early postoperative period instead. We found the use of several pillows sufficient combined with the patient’s self-control sufficient to achieve and maintain comfortable head flexion and avoid hyperextension during the early postoperative period.

CONCLUSION

In conclusion, long-segment resection and reconstruction of the trachea utilizing different release manoeuvres can be safely done, with low complication rates. Although swallowing and phonation dysfunction after laryngeal release were commonly encountered in almost one-third of our patients, they were mild, transient, self-limited and patients recovered during early postoperative period.

Conflict of interest: none declared.