Neurovascular Complications After Neck Dissection: a Prospective Analysis at a Tertiary Care Centre in South India

Abstract

Owing to high incidence of oral cancers in India, neck dissection is the most commonly performed procedure in any head and neck oncology setup. This surgery is particularly prone to various neurovascular complications due to intricate anatomy of neck, but the exact incidence has been studied less especially in Indian subcontinent. A prospective observational study was done on 105 patients of oral cancer who underwent neck dissection at Regional Cancer Centre, Trivandrum, and various short-term and long-term neurovascular complications were recorded. There was high temporary praxia of marginal mandibular and greater auricular nerve of 32.5% and 36.1%, respectively, immediate post-operative period which reduced significantly during 6 months. Eight patients had long-term shoulder abduction difficulty. There were isolated cases of IJV thrombosis and phrenic nerve paralysis. Despite our best intentions, complications arise due to complex anatomy of the neck, but if surgeon remains cognisant of the potential complications, their impact on patients’ health, pocket and hospital resources can be minimized.

Introduction

The history of neck dissection reveals a continuing quest to reduce its complications and sequelae. Since its original description by Crile in 1906 and subsequent popularization by Hays Martin in 1951, the radical neck dissection (RND) remained the standard treatment for palpable or potential cervical metastasis in head and neck cancers. Though it provided a reliable method of treating patients with head and neck cancer, it became increasingly apparent that it carried substantial morbidity, which led to development of modified procedures that attempted to reduce the adverse effects of the classical operation and yet preserve its effectiveness in oncological terms [1, 2].

As neck dissection surgical techniques have evolved, so have the type and frequency of complications. Due to intricate anatomy and presence of various nerves, vessels and lymphatic channels, neck dissections are particularly prone to various neurovascular complications. Shoulder complaints and functional impairment are common sequel of neck dissection, often attributed to injury of the spinal accessory nerve by dissection or direct trauma. Although current modifications in neck dissection aim at preserving the spinal accessory nerve, a variable degree of shoulder dysfunction still occurs in a significant number of patients. These can be attributed to spinal accessory neuropraxia and neurotmesis in addition to adhesive capsulitis and myofascial pain in upper trapezius, levator scapulae and rhomboid muscles [3, 4]. Other frequently injured nerves are greater auricular nerve (GAN) and marginal mandibular nerve (MMN); however, phrenic nerve injury is relatively low [5].

Neck dissection, mostly on the left side, carries a risk of chylous fistula, occurring in 3–5% cases. Studies have also shown internal jugular vein (IJV) thrombosis after neck dissection that ranges in 0 to 29.6% cases [6].

These complications in addition to prolonging hospital stay result in expensive and time-consuming management, thus increasing overall burden on hospital resources and patient’s pocket. Despite being most commonly performed procedure in any head and neck unit, exact incidence of various neurovascular complications and functional morbidities associated with neck dissection have been studied less especially in Indian subcontinent. With this in view, current study was undertaken and patients were studied prospectively for various short-term and long-term neurovascular complications following neck dissection.

Material and Methods

In this study, 105 patients of oral cavity squamous cell carcinoma who underwent neck dissection at the Regional Cancer Centre, Trivandrum, between 1st January 2015 and 31st December 2015, were studied prospectively. Patients who were post-chemotherapy/radiotherapy or had history of previous head and neck surgery were excluded. Informed consent was obtained from all patients and data was recorded on predesigned pro forma. It included demographic profile, clinical details with stage of the disease, type of neck dissection and post-operative complications. All patients had minimum 6-month follow-up. Various post-operative complications assessed included marginal mandibular nerve (MMN) injury, greater auricular nerve (GAN) injury, phrenic nerve injury, shoulder disability, internal jugular vein (IJV) thrombosis and chyle leak. For assessing MMN injury, angle of mouth deviation was noted on post-operative day (POD) 1 and at 6th month follow-up. GAN injury was detected by assessing numbness of ear lobule using monofilament nylon fiber at POD 1 and at 6th month follow-up. Pre-operatively, all patients underwent chest X-ray and Doppler neck to record the status of phrenic nerve and internal jugular vein (IJV), respectively. Postoperatively, all patients underwent chest X-ray to detect abnormal elevation of ipsilateral diaphragm to detect phrenic nerve injury. Also, patients underwent colour Doppler sonography of ipsilateral neck on POD 3–5 to detect IJV thrombosis. All those patients who had evidence of phrenic nerve injury or IJV thrombosis in immediate post-operative period underwent repeat chest X-ray or Doppler USG neck, respectively, at 6 months. All patients had shoulder abduction range recorded pre-operatively as grade 0 < 90°, 1 = 90°–120°, 2 = 120°–140°, 3 = 140°–160°, 4 = 160°–180° and 5 > 180°. Shoulder dysfunction was again recorded at 6 months post-operatively by assessing the following parameters: presence of shoulder droop, scapular flaring, shoulder abduction strength and shoulder abduction range. Criteria for detection of chyle leak included presence of white turbid fluid in drain. If chyle leak detected, then whether it was managed conservatively or required surgical re-exploration was recorded in pro forma.

All patients underwent neck dissection along with excision of primary tumour. Type of neck dissection performed has been classified into “radical” neck dissection (RND), “modified radical” neck dissection (MRND) or “selective” neck dissection (SND) procedures. RND defined as removal of all lymph nodes from level I to level V and all the three nonlymphatic structures (sternocleidomastoid muscle, the internal jugular vein [IJV] and the spinal accessory nerve [SAN]), while MRND removed the lymph nodes of the same levels but preserved the latter three structures. MRND being further divided into type 1–3 based on number of structures preserved. The SND includes removal of levels I–III (supra-omohyoid neck dissection, SOHND) and levels I–IV (extended supra-omohyoid neck dissection, ESOHND) [7]. All neck dissections were performed via transverse skin crease incision or reverse hockey stick incision.

Statistical Analysis

The results were presented as means and standard deviation for continuous variables and were compared using Student’s t test or Mann-Whitney test. Categorical variables were described by relative frequency and were compared by Chi-square or Fisher’s exact test. P value < 0.05 was considered significant.

Results

A total of 105 cases of oral cavity cancers who underwent neck dissection between 1 January and 31 December 2015 at Surgical Oncology Division, Regional Cancer Centre, Trivandrum, were included in the study. Of these, four could not complete 6-month follow-up. Three of these expired within 6 months of surgery, two because of disease advancement and one due to unrelated cause. One patient was lost to follow-up for unknown reasons. So, 105 cases were analysed for immediate post-operative complications while 101 were analysed for long-term complications.

Demographic Data

The mean age of presentation was 56.5 years with male to female ratio of 2.5:1. Most common site of carcinoma in oral cavity was found to be tongue (73%) followed by buccal mucosa (12%), alveolus, floor of mouth, and hard palate. Stage at presentation was almost equally divided among stage I, II, III, IV with carcinoma tongue more in early stage and buccal mucosa, alveolus presenting more in stage IV (Table 1).

Table 1.

Demographic and clinical characteristics

Parameters	Number
Case distribution
Number of patients	105
Male/female	2.5:1
Age range	24–85 years
Mean	56.5 years
Comorbidities
HTN	50
DM	37
Obesity (BMI ≥ 30)	4
Habits
Tobacco chewing	38
Pan chewing	53
Smoking	38
Alcohol	26
Primary site
Tongue	77
Buccal mucosa	13
Alveolus	11
Others	4
Stage
I	28
II	21
III	31
IV	25
Type of neck dissection
SOHND	8
ESOHND	70
MRND	29
Side of neck dissection
Right	57
Left	46
Bilateral	2
Neck incision
Transverse	84
Reverse hockey stick	21
Closure/reconstruction of primary tumour site
Primary closure	81
PMMC	16
Nasolabial	1
Others	7
Total No of events at 6 months	54
Marginal mandibular nerve injury	11
Greater auricular nerve injury	14
Phrenic nerve injury	3
Internal jugular vein thrombosis	1
Scapular flaring	7
Shoulder droop	10
Arm abduction < 140°	8
Chyle leak	0

Surgical Details

Most common type of neck dissection done in our department was extended supraomohyoid neck dissection, ESOHND (65%), followed by modified radical neck dissection (27%) and supraomohyoid neck dissection, SOHND (8%). Incision used was either transverse or reverse hockey stick in all cases. Among the patients who underwent MRND, internal jugular vein was ligated in one case due to tumour infiltration. Sternocleidomastoid was removed in 16 patients who underwent pectoralis major myocutaneous flap (PMMC) as reconstructive procedure to accommodate for the bulk of flap. No patient in the study population underwent radical neck dissection.

Neurovascular Complications

Total number of events recorded at 6th month follow-up was 54, with marginal mandibular nerve injury present in 11, greater auricular nerve injury in 14, phrenic nerve paralysis in 3, IJV thrombosis in 1, scapular flaring in 7, shoulder droop in 10 and arm abduction disability in 8 patients (Table 1). A detailed discussion on above is as follows:

• A.Greater auricular nerve injury

Greater auricular nerve praxia/injury as measured by ear lobe numbness by monofilament test was present in 36.1% (n = 38) of the cases on POD1. On follow-up at 6 months, some numbness was still present in 13.8% (n = 14) cases.

• B.Marginal mandibular nerve injury

Marginal mandibular nerve weakness as recorded by angle of mouth deviation was analysed in 86 patients after excluding 19 patients who had lip splitting incision. On POD1, 32.5% (n = 28) patients had some angle of deviation which reduced to 13.4% (n = 11) at 6 months. Majority of these patients with persistent angle of mouth deviation had only mild deviation evident only on smiling (Fig. 1). However, two patients (2.5%) had visible angle of mouth deviation with drooling of saliva.

• C.Phrenic nerve injury

Fig. 1.

Angle of mouth deviation persistent at 6 months

For detecting injury to phrenic nerve, all study patients underwent a pre-operative and a post-operative chest X-ray. All patients had normal pre-operative chest X-ray. Five patients were found to have raised ipsilateral raised hemidiaphragm suggesting phrenic nerve palsy in post-operative chest X-ray. However, none of the patient developed respiratory complication in post-operative period. These five patients were again evaluated with chest X-ray at 6 months and three of them had persistent raised hemidiaphragm while it resolved in two patients (Fig. 2).

• D.Shoulder dysfunction

Fig. 2.

Raised right hemi diaphragm due to phrenic nerve paralysis

Shoulder dysfunction was recorded in terms of arm abduction, scapular flaring, shoulder drooping and muscle strength pre-operatively and at 6 months from surgery. Spinal nerve is most prone to injury during level II and level V dissection. All patients in this study had clearance of level II A and B irrespective of type of neck dissection. Level V dissection was done only during MRND.

Spinal accessory nerve was preserved in all the patients intraoperatively. One patient had congenital shoulder anomaly, hence was excluded from the analysis along with four others who expired and lost to follow-up. Of 100 remaining patients, scapular flaring was present in 7 patients, shoulder droop in 10 patients (Fig. 3) with decreased range of arm abduction < 140° in 8 patients.

• E.Internal jugular vein (IJV) thrombosis

Fig. 3.

Mild shoulder droop on right side

Internal jugular vein was analysed for thrombosis pre-operatively and post-operatively between post-operative day 3–5 using ultrasound Doppler study. Findings were defined as abnormal if there was evidence of a flow defect indicating partial or complete thrombosis. All patients had normal vein calibre with normal flow on pre-operative scan. One patient underwent ligation of internal jugular vein intra-operatively as mentioned previously and hence was excluded from this analysis. Only 1 of 104 patients was found to have ipsilateral IJV thrombosis on post-operative scan and that too resolved on scan done after 6 months.

• F.Chyle leak

There were two intra-operative thoracic duct ligations for suspected intra-operative chyle leak. However, none of the patient developed clinical chyle leak in post-operative period.

Effect of Type of Neck Dissection on Complications Associated with Neck Dissection

Greater auricular nerve injury was found to be significantly associated with modified radical neck dissection as compared with selective neck dissection (ESOHND and SOHND) while there was no significant difference noted with other complications (Table 2). Increased rate of GAN injury in MRND group can be due to removal of sternocleidomastoid in majority of these patients for PMMC flap reconstruction and hence increasing probability of nerve injury. Though there was slightly higher shoulder dysfunction noted in MRND group compared with selective neck dissection, difference could not reach statistical significance. These were no significant association found between complications and obesity, hypertension, diabetes or habits.

Table 2.

Effect of type of neck dissection on neurovascular complications associated with neck dissection

		MRND	Selective ND	P value
Marginal mandibular nerve injury	Present	2	9	0.981
	Absent	9	62
Greater auricular nerve injury	Present	9	5	0.003
	Absent	19	68
Phrenic nerve injury	Present	0	3	0.558
	Absent	28	70
IJV thrombosis	Present	0	1	1.000
	Absent	28	72
Scapular flaring	Present	4	3	0.091
	Absent	24	70
Shoulder droop	Present	5	5	0.199
	Absent	23	68
Arm abduction < 140°	Present	3	5	0.682
	Absent	25	68

P < 0.05 is significant

Discussion

Complications of neck dissection affect every surgeon regardless of experience and technical skill. Owing to high incidence of oral cavity cancer in India, neck dissection is the most commonly performed procedure in any head and neck oncology setup. Therefore, it is imperative that complication of such surgery and their attendant morbidity are minimized.

Marginal mandibular nerve is a branch of facial nerve which supplies the depressor anguli oris, depressor labii inferioris, mentalis and orbicularis oris. Thus, injury to the nerve can result in facial asymmetry, deviation of angle of mouth, drooling of saliva and difficulty of speech and chewing. Marginal mandibular nerve is most prone to injury during level IB lymph nodal dissection which forms integral part of both selective and modified neck dissection. According to literature, dissection around nerve for identifying it leads to temporary neuropraxia in 16–23% cases, which usually lasts for few weeks [5, 8, 9]. Despite preserving marginal mandibular nerve in all patients, present study reported a slightly higher rate of neuropraxia of marginal mandibular nerve in 32.5% patients in immediate post-operative period; it decreased to 13.4% at 6 months. Majority of these patients had only slight deviation of mouth, with drooling of saliva present in only two (2.5%) patients.

In our study, ear lobe numbness suggesting possible GAN trauma was present in 36.1% patients in immediate post-operative period which later reduced to 13.8% during 6-month duration. This implies that both marginal mandibular and greater auricular nerve praxia after neck dissection are far more common than we think but often are neglected, as during post-operative period, both surgeon and patient are concerned about other major post-operative issues related with complex surgery like feeding, pain, wound management etc. One other possible reason for relatively higher incidence of neuropraxia in our study could be that most of the other studies are retrospective analysis and mild symptoms like numbness of ear lobule or slight angle of mouth deviation are often not recorded in case sheets. This emphasizes on the fact that just preserving the nerve anatomically is not enough and every attempt should be made to minimize handing of these nerves during dissection and any trauma due to traction or electrocautery should be avoided.

Phrenic nerve paralysis is another nerve complication of neck dissection and leads to ipsilateral hemidiaphragm elevation with or without mediastinal shift on chest radiograph and may lead to pulmonary complication in post-operative period. Preservation of the fascial layer over the nerve and anterior scalene muscle is the primary method for prevention of injury. In a retrospective study, unilateral phrenic nerve paralysis was observed in 14 (8%) of 176 consecutive neck dissections. None of the patients with post-operative phrenic nerve paralysis displayed severe symptoms [10, 11]. Present study recorded unilateral phrenic nerve paralysis in 4.7% cases which further decreased to 2.8% at re-evaluation at 6 months and none of the patient developed any pulmonary complication neither in immediate post-op, nor later during follow-up period.

This implies that mostly such an injury while performing neck dissection will often go unnoticed; however, in few cases with already compromised pulmonary function or underlying lung disease, it can lead to substantial morbidity and even mortality post-operatively. Thus, importance of preserving the fascia and careful dissection cannot be ignored.

Chylous leakage is an uncommon complication after neck dissection for which several treatment modalities have been described. The risk for developing chylous fistula following radical neck dissection averages 1–2.5% with majority (75–92%) being on left side [12]. In a study by Teymoortash et al., 2 of 98 patients who underwent selective neck dissection developed chylous leak post-operatively and were managed conservatively [13]. In current study, there were two cases with intra-operative detection of chyle leak both of which were controlled using figure of 8 stitch, and forced Valsalva manoeuvre was done by anaesthetist to check for leak by increasing intrathoracic pressure. None of the patient developed chyle leak in post-operative period highlighting the importance of intra-operative identification.

Apart from the intra-operative injury or ligation of internal jugular vein, post-operative thrombosis represents a possible complication of neck dissection.

In a study by Quraishi et al., incidence of IJV thrombosis has been reported to be 25% which decreased to 6% on long-term follow-up [14]. Other studies have reported the incidence of IJV thrombosis ranging from 0 to 29.6%. Harada et al. and Teymoortash have reported 100% patency rate after neck dissection [6, 13].

In our study also, only one patient developed IJV thrombosis in immediate post-operative period which also resolved in 6-month Doppler scan. This low rate of IJV thrombosis can be attributed to use of scalpel knife instead of cautery while dissecting tissue over IJV.

With regards to functional results after neck dissection, impairment of the spinal accessory nerve may cause complex clinical syndrome in terms of “shoulder syndrome”. It is described as pain, atrophy and weakness of shoulder girdle. The spinal accessory nerve (SAN) is a motor nerve, innervating the sternocleidomastoid and trapezius muscle. It is believed that sacrificing the accessory nerve during neck dissection results in paralysis of descending and transverse part of trapezius muscle resulting in shoulder syndrome. Theoretically, neck dissection sparing SAN should result in no or slight shoulder dysfunction, although this is not always the case. Various studies have reported shoulder complaints in 31–60% of patients who underwent SAN sparing neck dissection [15–18]. According to Van Wilgen et al., almost 14% patients have temporary shoulder complaints in the first post-operative year [19]. The present study reported a shoulder dysfunction of 13% at 6 months after surgery in the form of limited arm abduction (8%), scapular flaring (7%) and shoulder droop (10%). Although it has been suggested that patients who underwent SND have less damage to spinal accessory nerve and least shoulder dysfunction as compared with MRND, the current study did not find any significant difference present in terms of shoulder dysfunction among the two procedures. Similar results have been reported by Koybasioglu et al. [20] Also, shoulder exercises in post-operative period may contribute towards minimizing the impact.

Conclusion

Over the past century, neck dissection has become the most commonly performed procedure in head and neck oncology, be it performed in isolation or as integral part of major resection and reconstruction. Though it is performed with such regularity, associated morbidities should not be underestimated and every effort should be made to minimize them by gaining familiarity with the complex anatomy of the region. A meticulous, almost protocol-driven approach should be employed. Despite the best planning, complications can still occur, but their impact can be minimized by a vigilant and proactive approach in perioperative period.