Is dysphonia permanent or temporary after anterior cervical approach?

Abstract

The rate, causes and prognosis of dysphonia after anterior cervical approach (ACA) were investigated in our clinical series. During a 10-year interval, 235 consecutive patients with cervical disc disease underwent surgical treatment using anterior approach. Retrospective chart reviews showed recurrent laryngeal nerve (RLN) injury in 3 (1.27%) patients. All three patients were men and only one patient had multilevel surgery. These patients had RLN injury after virgin surgery. Laryngoscopic examination demonstrated unilateral vocal cord paralysis in all patients who had postoperative dysphonia. No permanent dysphonia was observed in our series and patients recovered after a mean of 2 months (range 1–3 months) duration. Dysphonia after ACA was a rare complication in our clinical series. Pressure on RLN or retraction may result in temporary dysphonia.

Introduction

Anterior cervical approach (ACA) is the most common surgery of the symptomatic cervical disc disease [32]. Life-threatening complications and serious problems are waiting for the surgeons after ACA. Laryngological complications after anterior approaches to the cervical spine are most notable on vocal cord paralysis secondary to the recurrent laryngeal nerve (RLN) injury. The incidence of the RLN injury with ACA has been widely reported in the literature [2, 3, 6, 7, 9, 10, 13–15, 19, 26, 27, 30, 33]. Although dysphonia may be permanent, most cases are temporary which are lasting for weeks or months. Some proposed mechanisms of this complication include direct surgical trauma, nerve division, or ligature, pressure or stretch-induced neuropraxia, postoperative edema [3, 11, 18, 25], and endotracheal tube-related vocal fold paralysis is described. We retrospectively reviewed the cases with dysphonia after anterior cervical surgery to evaluate the causes and prognosis of this complication.

Patients and methods

Between January 1995 and September 2005, 235 consecutive patients underwent ACA at our institute. The surgery was done with a standard fashion under general anesthesia using endotracheal intubation. All surgeons routinely approached to the cervical spine using the right side. Deep retraction of the soft tissue was achieved with hazelnut sponges and the tips of blades of Cloward retractors were placed deep to the longus colli muscles on either side. During the exploration of the carotid-esophageal sulcus, we avoid making sharp dissection and keep away from the tracheoesophageal groove not to expose RLN. RLN was not visualized during the surgery in any case. Anterior discectomy and removal of osteophytes confirmed with preoperative studies were performed. Simple discectomy or fusion with allograft/autograft with or without plating, or only cage instrumentation was performed after decompression. The mean duration of surgery was 78 min (range 50–150 min).

A total of 281 levels were operated on 71 female and 164 male patients. The mean age was 46.3 years (range 30–75 years) for female patients and 39.7 years (range 20–75 years) for male patients. A right-sided approach was used in all the procedures. The most common preoperative diagnosis was radiculopathy with cervical disc hernia. Surgery was performed on one level in 197, two levels in 30, three levels in 8 patients. The levels involved in the surgery are C2-3 in 2 patients (0.7%), C3-4 in 3 patients (1%), C4-5 in 34 patients (12%), C5-6 in 86 patients (31%), C6-7 in 155 patients (55%) and C7-T1 in one patient (0.35%). The patients were hospitalized for an average of 3.2 days (range 2–7 days).

Results

On the preoperative examinations, no patient had dysphonia. Dysphonia developed in three male patients (1.27%) after ACA. For these patients, postoperative laryngoscopy was performed to evaluate possible direct mechanical trauma to the vocal folds due to endotracheal intubation. All patients showed only unilateral vocal cord paralysis consistent with RLN injury and they all had virgin surgery (Table 1). All the patients who developed dysphonia had surgery for cervical levels involving C6-7 level. No aspiration or dysphagia was observed in this series of patients.

Table 1.

Characteristics of patients with postoperative dysphonia

Level of procedure	Sex	Age (years)	Procedure	Duration of surgery (min)	Recovery time (months)
C6-7	M	39	ACD + cage	60	2
C6-7	M	42	ASD	70	3
C5-6, C6-7	M	68	ACD + fibula allograft	90	1

ACD anterior cervical discectomy, ASD anterior simple discectomy

The mean recovery period of the dysphonia was 2 months (range 1–3 months) postoperatively.

Discussion

Robinson and Smith first described the anterior cervical spinal surgery [23], and popularized by Cloward [17, 18] in the 1950s. Vocal cord paralysis is a rare neurological complication than dysphagia in cervical surgery. Temporary unilateral vocal cord paralysis after ACA was reported in the range between 0.98 and 8% and the incidence of permanent paralysis was between 0.15 and 3.5% [3, 6, 9, 14, 17, 33]. In our series, temporary unilateral vocal cord paralysis was 1.27% and no permanent vocal cord paralysis was observed.

Direct surgical trauma to the RLN seems unlikely and observation of RLNs is so rare, the most likely mechanisms of injury would seem to be from indirect stretch or focal pressure on the nerve. But the studies have suggested that the anatomic course and relative resistance to stretch of the right RLN place this side at greater risk to injury [8, 21, 28]. RLN on the right side originates from the vagus nerve. It loops around the innominate artery and ascends in the tracheoesophageal groove, entering the larynx from behind the cricothyroid joint. Tew and Mayfield reported, if anterior approach occurs on the left side [26]; RLN loops around the aorta, and then ascends in a similar fashion and left RLN has a longer loop and lies better protected within the tracheoesophageal groove. Netterville et al. [21] suggested that anatomical differences in the RLN would lead to additional nerve strain during retraction and higher rates of right-sided paralysis. On the other hand; in the cadaver study, Menck et al. found no significant differences in the relative position of the right-side and left-side RLNs relative to the tracheoesophageal groove and concluded that either side was appropriate for ACAs [20]. Heeneman actually found a higher rate of permanent paralysis for left-sided approaches: a 16.6% rate for the left versus a 2.5% rate for the right [14]. Since all of our cases underwent surgery using the right side by different surgeons and our RLN complication rate is comparable with the literature, we suggest that the anatomic course of RLN may not be the major reason for causing of dysphonia. Whilst there are conflicting suggestions in the literature on this topic, we prefer the right side approach as a classical fashion of our institute which provides better surgical ability for the right-handed surgeons.

Although Flynn found severe RLN palsies were uncommon, they comprised the largest number of neurological complications (16.7%) of ACDF [9]. In perhaps the most comprehensive review of the literature on the otolaryngological complications of the anterior approach to the cervical spine, Winslow and Meyers [29] found that the incidence of hoarseness was 0.06–11%, with persistence occurring in 0–3.5%. Frempong-Boadu et al. [11] examined 23 patients undergoing ACDF preoperatively and postoperatively. Vocal cord paresis occurred in two patients postoperatively, one resolved by 1 month, the other was persistent. In our study, all three patients had temporary paralysis of the RLN and they recovered in a mean of 2 months (range 1–3 months) (Table 1). We suggest that if the surgeon causes no obvious harmful manipulation to the RLN, the possibility of occurring permanent paralysis is rare according to our results. It is an other possibility that patients may have subclinical RLN palsy that may have been missed. A study design with a routine postoperative observation of vocal cords may show the incidence of actual subclinical RLN palsy after ACA.

Bazaz et al. [1] reported that postoperative vocal cord paralyses were more common in female patients. In our series all the patients who had dsyhponia were male. Dysphonia was more common in patients with possible non-union in at least one level and singing difficulties in those in whom with more surgeries to the anterior neck region [1]. In our series, there is no non-union problem. Yue et al. [31] reported that the presence of dysphonia in the early postoperative period was not related to age, sex, smoking status, duration of surgery, number of levels operated, number of anterior neck surgeries or the plating system used. In our series, three patients had temporary RLN paralysis (1, 27%) and we could not find relation between sex and age of patients, and also levels of cervical disc hernia and reoperation, or implantation.

Endotracheal intubation alone may cause vocal cord paresis and paralysis. Increased cuff pressures may impinge upon the nerve, pushing it against the thyroid lamina. Bulger et al. [3] have suggested that vocal fold paralysis during ACDF may be due to endotracheal intubation. Intubation trauma can cause both permanent and temporary vocal fold paralysis. Cavo’s review of 30 cases of vocal fold paralysis secondary to endotracheal intubation after procedures not involving the head, neck, or thorax noted 21 complete and 2 partial recoveries with 2 permanent paralyses (5 patients were lost to follow-up) [4]. The results are strikingly similar to Kriskovich’ s rates of temporary and permanent paralysis [17]. Jensen et al. used the intraoperative laryngeal electromyographic and endotracheal tube cuff pressure monitoring and found that intraoperative increases in cuff pressure and diminished electromyographic activity occurred in patients with higher rates of postoperative hoarseness. These data further support the role of retractor/endotracheal tube interactions in vocal fold paralysis after ACDF [16]. Since the duration of surgery is not related to RLN injury in our series, we suggest that the most important factor was excessive retractor pressure than duration of pressure.

We suggest that the possible etiologies for this complication are sharp dissection, pinching of the RLN by retractors, stretching of the nerve with retraction, postoperative edema, and nerve involvement in suture, direct trauma to the cricoarytenoid joint and reoperation in the same level. Properly endotracheal intubation, careful blunt dissection and surgical technique, correct retractor placement beneath the bodies of longus colli muscles away from the tracheoesophageal groove, are critical to preventing direct surgical trauma to the nerve.

This study represents the very low rates of temporary vocal cord paralysis after cervical disc surgery and fusion. During the surgery, well-known anatomy always helps the neurosurgeon. The surgeon may safely approach the cervical spine from the side of personal preference and experience. Surgical technique including sharp dissection and excessive retraction seems to be the most important pitfall. RLN injury may not be related to the surgical side, the anatomic course of the RLN, patient’s age or sex, duration of surgery and/or reoperation.