Intraoperative nerve monitoring during thyroidectomy: evaluation of signal loss, prognostic value and surgical strategy

Abstract

Introduction

Intraoperative neural monitoring of the recurrent laryngeal nerve has been widely used to avoid nerve injury during thyroidectomy. We discuss the results of the change in surgical strategy after unilateral signal loss surgeries using intermittent intraoperative neural monitoring in a high-volume referral centre.

Materials and methods

Details of consecutive patients who underwent thyroidectomy with intermittent intraoperative neural monitoring between January 2014 and December 2017 were prospectively recorded and retrospectively reviewed. Loss of signal was defined as recurrent laryngeal nerve amplitude level lower than 100 μV during surgery. The rate of loss of signal and change in surgical strategy during the operation were evaluated.

Results

Loss of signal was detected in 25 (5.4%) of 456 patients for whom intermittent intraoperative neural monitoring was performed. Four patients had anatomic nerve disruption and surgery was completed by an experienced endocrine surgeon making use of intraoperative neural monitoring with continuous vagal stimulation. Staged thyroidectomy was performed on 16 patients with unilateral loss of signal in whom the nerves were intact visually. Postoperative vocal cord paralysis was encountered in 18 of 21 (85.7%) patients with loss of signal, and 16 of 18 (88.8%) were improved during the follow-up period. Patients’ voices were subjectively normal to the surgeon postoperatively in 9 of 21 (42.8%) patients who were found to have loss of signal with intact nerves.

Conclusions

Intraoperative neural monitoring can be used safely in thyroid surgery to avoid recurrent laryngeal nerve injury. It enables the surgeon to diagnose recurrent laryngeal nerve injury intraoperatively to estimate the postoperative nerve function and to modify the surgical strategy to avoid bilateral vocal cord paralysis.

Introduction

Recurrent laryngeal nerve injury is the leading major complication of thyroid surgery. It decreases the quality of life of patients and exposes surgeons to medicolegal issues.^1,2 Unilateral vocal cord paralysis after unilateral nerve injury may lead to hoarseness, aspiration-related respiratory complications and dysphagia, whereas bilateral recurrent laryngeal nerve injury, although rare, may cause more severe life-threatening dyspnoea requiring tracheostomy. The rate of transient recurrent laryngeal nerve paralysis has been reported to be between 0.4% and 12%, whereas that of permanent paralysis has been reported to be less than 1% in experienced hands.^3,4

The recurrent laryngeal nerve is more vulnerable to damage in the hands of less experienced surgeons in reoperative thyroid and neck surgery, for patients with large goitres or Graves’ disease, and in patients with a history of neck radiotherapy.⁵ Intraoperative neuromonitoring of the recurrent laryngeal nerve has been widely used in the identification and evaluation to avoid injury to the nerve. This method not only assesses the function of the recurrent laryngeal nerve, but also contributes to the detection of anatomical variations during surgery and the anticipation of postoperative vocal cord functions.^6,7 Although there is a continuing debate over its effect on the reduction of recurrent laryngeal nerve injury, recent two meta-analyses supported the application of intraoperative neuromonitoring in reducing both transient and permanent injury to the recurrent laryngeal nerve during thyroidectomy, particularly in bilateral operations and reoperative cases.^5,8

The international neural monitoring study group divided nerve injuries into two categories.⁹ While there is structural damage to the recurrent laryngeal nerve in segmental injury, transmission disturbance is detected in global injury. In particular, intraoperative neuromonitoring provides for the application of two-stage thyroidectomy to prevent bilateral nerve damage leading to severe respiratory distress. The detection rate of nerve damage is reported as 94–99%.⁹ This has led to the emergence of a change of surgical strategy; hence the concept of staged thyroidectomy, postponing resection of the second side after signal loss on the first side to prevent bilateral nerve damage. Changing the surgical strategy should, however, be more than staged thyroidectomy.

We evaluated successful surgical strategy changes and the results of staged thyroidectomy procedures performed after unilateral loss of signal in patients having intraoperative neuromonitoring during the operation.

Materials and methods

This study was approved by the ethical committee of Izmir Katip Celebi University Atatürk Training and Research Hospital. Details of consecutive patients who underwent thyroidectomy between January 2014 and December 2017 were prospectively recorded and retrospectively reviewed. In our clinic, surgeons have different preferences for safe thyroidectomy; some prefer to use intraoperative neuromonitoring, while others do not. Vocal cord examination of patients who are due to undergo thyroid and parathyroid surgery has been a routine approach preoperatively but postoperative vocal cord examination has been routinely performed only for patients having intraoperative neuromonitoring and on demand in other patients on the morning after surgery.

The study group included those patients with normal preoperative indirect laryngoscopy who were scheduled for total thyroidectomy with the use of intermittent intraoperative neuromonitoring from the beginning of the operation and who met the intraoperative neuromonitoring international study group guideline criteria (preoperative indirect laryngoscopy: L1; vagal nerve stimulation at the beginning of thyroidectomy: V1; recurrent laryngeal nerve stimulation at the beginning of thyroidectomy: R1; recurrent laryngeal nerve stimulation at the end of thyroidectomy: R2, vagal nerve stimulation at the end of thyroidectomy: V2; postoperative indirect laryngoscopy: L2) and patients with true signal loss (signal loss V2 in operation and vocal cord paralysis on L2).⁹ To differentiate true signal loss from technical problems we strictly adhered to the algorithm described by intraoperative neuromonitoring International study group.⁹

Patients who underwent unilateral surgery or surgery with continuous vagus stimulation initially, who had preoperative vocal cord paralysis due to previous surgery or tumour invasion, who had sternotomy for retrosternal goitre (in whom we prefer to use continuous vagus stimulation from the beginning) were excluded from the study. A final total of 456 patients were included in the study.

Expert anaesthesiologists intubated patients with an electrode tube (AVALANCHESI System, Dr. Langer Medical GmbH, Waldkirch, Germany). There were strict instructions not to use any long-acting muscle relaxant agents at any point of the procedure. The recurrent laryngeal nerve was stimulated by giving a 1–2 mA current according to international standards. Loss of signal is defined as intraoperative neuromonitoring with an amplitude in R1 greater than 100 μV and an amplitude level during surgery lower than 100 μV after the application of troubleshooting algorithm. The disappearance of R2 and V2 signals after complete dissection of the recurrent laryngeal nerve means that the recurrent laryngeal nerve may have been injured during the dissection. In such cases, the recurrent laryngeal nerve trace was investigated to discover where the nerve conduction was broken and the mechanism of injury. Postoperative indirect laryngoscopy was performed on table or postoperative day 1 for all patients.

Loss of signal, whether unilateral or bilateral, surgical strategy during the operation and follow-up results were evaluated. Reliability of the two-stage thyroidectomy was investigated. Persistent or transient vocal cord paralysis and requirement for tracheostomy were compared according to the presence or absence of intraoperative neuromonitoring.

Statistical analysis

Results were compared for variables the using chi square test or Fisher’s exact test if an observed value was less than five for categorical variables. SPSS v. 22.0 was used for all analyses. A P-value of less than 0.05 was considered statistically significant.

Results

Between January 2014 and December 2017, 1509 thyroidectomies were performed in our department. Total thyroidectomy was performed on 1118 of those patients (Table 1).

Table 1.

Types of thyroidectomy performed in our department between 2014 and 2017.

Year	Total thyroidectomy	Lobectomy	Completion thyroidectomy	Total
2014	242	70	40	352
2015	257	61	53	371
2016	317	62	27	406
2017	302	52	26	380
Total	1118	245	146	1509

Intraoperative neuromonitoring was used in 620 of 1509 patients who had thyroidectomies performed, and 456 of those patients whose surgery included intraoperative neuromonitoring who met the aforementioned criteria were included in the study. Initially, a total of 912 recurrent laryngeal nerves were accepted as being at risk. Of those patients, 350 (76.4%) were female and 106 were male and the mean age was 52.8 years (range 18–82 years). During surgery, loss of signal was obtained in 25 (5.4%) patients. Unilateral loss of signal was seen in 24 (5.2%) patients, while one (0.2%) of the patients had bilateral loss of signal.

Among 25 patients (including one with bilateral palsy), 21 (84%) patients had visually intact nerve throughout the course, 2 (8%) patients had inadvertent nerve disruption during surgery, and 2 (8%) underwent unilateral nerve resection because of tumoral invasion. Among patients having visually intact nerves, one (4%) had surprisingly bilateral loss of signal at the end of surgery (fig 1).

Figure 1.

The study summary.

Management of loss of signal was decided during the operation according to the type of injury (Table 2). Loss of signal due to inadvertent recurrent laryngeal nerve disruption was diagnosed at the first side of thyroidectomy in those two patients. Those patients were referred to an endocrine surgeon and repair of recurrent laryngeal nerve was performed to restore anatomical integrity using end-to-end 7/0 Prolene sutures. The indications for surgery were revisited and, due to bulky thyroid papillary cancers, the consultant surgeon decided to continue the contralateral site by using continuous vagal stimulation and total thyroidectomies were completed without any complication.

Table 2.

Recurrent laryngeal nerve injury type, surgical strategy and outcome.

Injury type	First side (n)	Second side (n)	Bilateral (n)	Strategy	Follow- up results
Inadvertent nerve disruption	2			Per-operative consultation + repair + proceed to surgery with continuous vagal nerve stimulation	uVCP (n = 2)
Nerve sacrificed due to bulky tumoral disease	2			Proceed to surgery with continuous vagal nerve stimulation	uVCP (n = 2)
Loss of signal + visually intact nerves	18			Two-stage thyroidectomy planned	uVCP (n = 1)
		2		No additional process required	uVCP (n = 1)
			1	No additional process required

uVCP, unilateral vocal cord paralysis (permanent).

In two other patients, nerves were sacrificed during the first site of dissection and an experienced endocrine surgeon decided to complete the surgery with continuous vagus stimulation, again due to papillary thyroid cancer without any complication. Vocal cord paralysis on the first side was revealed during the follow-up postoperative indirect laryngoscopy in all four patients. Nerve resection-related loss of signal was 100% predictive for vocal cord paralysis.

The patient with bilateral loss of signal had no loss at the end of the first side of resection, but while losses were being evaluated during the resection of second side, loss of signal was detected on both recurrent laryngeal nerves although the whole course of recurrent laryngeal nerve’s anatomically were intact. After following the troubleshooting algorithm, true loss of signal was confirmed. The patient was carefully extubated and the vocal cords were examined. Bilateral vocal cord palsy detected but the rima opening was almost 3 mm and vocal cords were not fixed in the median position. Then we decided to follow-up with steroid administration without tracheostomy. Interestingly, postoperative bilateral vocal cord paralysis was not reflected to the voice of the patient. The patient’s voice was normal, and recovery from the right vocal cord paralysis was observed on the second postoperative day and he was discharged from the hospital. The left vocal cord recovered during the third postoperative week.

Loss of signal with the visually intact nerve was detected in remaining 20 patients, on the first side of resection in 18 patients and on the second side in two patients. Staged thyroidectomy was planned for 18 patients after the application of the troubleshooting algorithm, and the operative indication was revised. No additional process was required for patients with loss of signal on the second side (Table 2).

Staged thyroidectomy was planned for 18 patients after recovery but performed in 16 patients (64%). Completion thyroidectomy was not essential for one (4%) patient with a final diagnosis of benign thyroidal lesion, and the other patient (4%) refused a second surgical process due to the risk of potential complications (Table 3). No complications occurred in the staged completion thyroidectomies.

Table 3.

Staged thyroidectomy procedures in 25 patients with loss of signal.

Procedure	Patients		Complications
	(n)	(%)
Loss of signal/patients	25/456	5.4
Staged thyroidectomy:
Planned	18
Performed	16	3.5	0
Not performed	2	0.4
Reason for not performed:
No requirement	1	0.2
Patient refusal	1	0.2

Postoperative indirect laryngoscopy revealed vocal cord paralysis in 18 of 21 patients (19 from 22 nerves at risk) with signal loss and unimpaired nerve integrity (86.4%). Postoperative vocal cord paralysis was not detected in any of the 431 patients without signal loss. The positive predictive value and the negative predictive value of intraoperative neuromonitoring was 86.4% and 100%, respectively, for the visually intact nerves with loss of signal.

During the follow-up period for 19 patients with signal loss who had intact nerves, vocal cord paralysis improved in 17 (89.4%) nerves and persisted in 2 (10.6%). Finally, 6 of 25 patients had persistent vocal cord paralysis during the follow-up period. The median time for improvement of nerve function was one month (range two days to three months). Patients’ voices were subjectively normal to the surgeon postoperatively in 9 of 21 (42.8%) patients who were found to have loss of signal with intact nerves.

Among 912 recurrent laryngeal nerves at risk in 456 patients, we experienced loss of signal in 25 (5.4%) patients and recurrent laryngeal nerve injuries in 22 patients (21 unilateral and 1 bilateral injury) in total. The overall recurrent laryngeal nerve injury rate in the group was 4.8%; 6 (1.3%) patients had persistent injuries, while 16 (3.5%) had transient injuries (Table 4).

Table 4.

Rates of recurrent laryngeal nerve injury types.

	Patients (n = 456)
	(n)	(%)
Initial recurrent laryngeal nerves at risk	912	100
Recurrent laryngeal nerve injury	22	4.8
Persistent	6	1.3
Transient	16	3.5

A total of 1118 patients who had bilateral total thyroidectomy were divided into an intraoperative neuromonitoring group (n = 456) and a non-intraoperative neuromonitoring group (n = 662) to compare results regarding vocal cord paralysis. In the intraoperative neuromonitoring group, only one (0.2%) patient had bilateral vocal cord paralysis without requirement for a tracheostomy. In the non-intraoperative neuromonitoring group, two patients had bilateral vocal cord paralysis (0.3%) and only one required a tracheostomy during the immediate postoperative period (Table 5).

Table 5.

Comparison of the complications in bilateral total thyroidectomy patients with or without intraoperative neuromonitoring application.

Intraoperative neuromonitoring	Patients (n)	Bilateral vocal cord paralysis		Tracheostomy (n)
		(n)	(%)
Applied	456	1	0.2	0
Not applied	662	2	0.3	1
Total	1118	3	0.2	1

Discussion

Intraoperative neuromonitoring of recurrent laryngeal nerve during thyroid surgery has been widely used to prevent the permanent injury of the nerve. However, the role and the benefits of intraoperative neuromonitoring are still controversial.

The literature is replete with conflicting results about the use of intraoperative neuromonitoring and its effect on recurrent laryngeal nerve injury. The initial prospective and retrospective studies and meta-analyses failed to show a decreased number of persistent or transient paralysis of the recurrent laryngeal nerve.^3,10 However a 2014 meta-analysis supported the use of intraoperative neuromonitoring in thyroid surgery with significant decrease in injury rates.¹¹ Rulli et al demonstrated that intraoperative neuromonitoring had a significant impact on preventing transient injuries; however, they failed to display such an effect on permanent injuries. Sun et al demonstrated that the use of intraoperative neuromonitoring reduced both overall and permanent recurrent laryngeal nerve palsy in thyroid reoperations.⁸ Finally, Bai et al found that intraoperative neuromonitoring played a preventive role of total, transient and permanent injury of recurrent laryngeal nerve in patients undergoing bilateral thyroidectomy.⁵ Moreover, the meta-analysis found that surgery with intraoperative neuromonitoring was associated with fewer total and transient recurrent laryngeal nerve injuries in hospitals where operative volume was less than 300 nerves at risk per year and fewer total and permanent recurrent laryngeal nerve injuries in the hospitals where the operating volume was equal or greater than 300 nerves at risk per year.

Even in large series, permanent recurrent laryngeal nerve paralysis is seen to be less than 1%, whereas transient recurrent laryngeal nerve paralysis is seen in 2–3% of patients.10 In our series, permanent recurrent laryngeal nerve paralysis rate in surgeries using intraoperative neuromonitoring was 1.3% and the transient recurrent laryngeal nerve paralysis rate was 3.5%. We cannot give the comparison of recurrent laryngeal nerve injury and vocal cord paralysis between the intraoperative neuromonitoring used and non-used groups because the exact injury rates in the latter group is unknown since they did not have routine vocal cord examination postoperatively.

The increasing use of intraoperative neuromonitoring in thyroid surgery has revealed the need to develop new strategies in case of a loss of signal occurs on the first side of resection in patients initially intended for total thyroidectomy. The most common strategy is the staged thyroidectomy when loss of signal occurs during resection of the first site using intraoperative neuromonitoring, to avoid bilateral recurrent laryngeal nerve injury, hence tracheostomy.¹²

Although some studies have claimed that the negative predictive value of intraoperative neuromonitoring is above 90% in patients with signal loss, the positive predictive value is low, below 40%, so no change in surgical strategy is necessary.^3,13 We and others agree that loss of signal during thyroidectomy is a good predictor of postoperative vocal cord paralysis if the troubleshooting algorithm is properly applied. Caló et al reported that the positive predictive value of intraoperative neuromonitoring was found to be 78.4%, the negative predictive value was 99.8%, specificity was 99.4% and sensitivity was 93.4%.¹⁴ In two other studies, the positive predictive value was estimated to be 76.7–77% and the negative predictive value was found to be 99.4–95% performed with intermittent intraoperative neuromonitoring,^15,16 whereas the positive predictive value was 88.2% and the negative predictive value was found to be 99.8% with continuous monitoring.¹⁵

In our study, the positive and negative predictive values of loss of signal were 86.4% and 100%, respectively, which indicated that intraoperative neuromonitoring was a reliable prognostic tool for postoperative vocal cord function. According to those results, signal acquisition with intraoperative neuromonitoring detects intactness of recurrent laryngeal nerve with 100% accuracy, while intraoperative neuromonitoring can detect recurrent laryngeal nerve injury with 86.4% accuracy.

The reason why the positive predictive value of loss of signal during intraoperative neuromonitoring is not 100% has not been fully understood. Vocal cord paralysis was more frequent in laryngoscopy performed in the early postoperative period, but subsequent laryngoscopy showed improvement of the condition. This finding suggests that neuropraxia is the main factor, and recovers rapidly. In some studies, it has been shown that the signal improves in 40–90% of patients during surgery. No standardisation has been identified on the duration of regeneration.^12,17

The most common and recommended surgical strategy in case of loss of signal during the planned total thyroidectomy with intraoperative neuromonitoring is to terminate the operation and leave the resection of the contralateral thyroid lobe to the second session (ie staged thyroidectomy). The main purpose is to prevent bilateral vocal cord paralysis. It is logical that the second operation would be safer and the risk of bilateral vocal cord paralysis would be avoided after the recovery of recurrent laryngeal nerve function in patients with visually intact nerves. Staged thyroidectomy can be performed in cases of thyroid cancer and benign thyroid disease.¹⁸ According to a survey of German surgeons’ associations, in the case of loss of signal, 90% of surgeons changed their surgical strategy and left the contralateral lobe resection into the second session.^19,20 The unnecessary second surgery in case of false negative loss of signal is the major potentially deleterious consequence of intraoperative neuromonitoring.

Sometimes there is no need to change surgical strategy in loss of signal. For example, if loss occurs during the second side of resection with the visually intact nerve, the resection can be completed safely and the patient can be reassured with expectant management.

A change in surgical strategy in case of loss of signal should not solely include staged thyroidectomy. It is necessary to separate patients in two groups in terms of loss of signal during resection of the first side and to determine the surgical strategy accordingly. The first group includes those patients with disrupted anatomic nerve integrity. We agree that the first step should be the re-evaluation of the indications for the operation to proceed when faced with loss of signal during the first site of resection. If the surgical indication is suspicious or the surgeon is hesitant to continue for any reason, surgery should be terminated and the decision for staged thyroidectomy should be taken. If the anatomical integrity of the nerve is disrupted and there is no expectant nerve recovery, another strategy should be applied if total thyroidectomy is warranted for bulky thyroid cancers. It would be very difficult for the operating surgeon to proceed with the other side knowing that the first site nerve has suffered disruption, in respect to their experience or emotional status, self-confidence, stress and medicolegal concerns. In such situations, we recommend terminating the surgery and referring the patient for staged thyroidectomy to an experienced endocrine surgical centre, if an experienced endocrine surgeon is not available in the hospital. But if available, we recommend consulting the endocrine surgeon intraoperatively, to re-evaluate the indications for surgery together. If a decision to proceed is made, the other site can be resected together with the strict use of continuous vagal stimulation.

In our series, surgery was continued with the help of an experienced endocrine surgeon in all four cases with the careful use of continuous vagus stimulation. In those four cases, no injury was occurred on the second side. Continuous intraoperative neuromonitoring alerted the surgeon to an undesired traction injury at the beginning and enabled completion of surgery at the same session without further nerve injury.

The second group includes patients with visually intact nerves. It is logical and safe to leave the completion of thyroidectomy to another stage after the expectant recovery of the injured nerve. The majority of signal losses are due to traction damage. Recovery of nerve function will occur in most patients in a short time. For that reason, staged thyroidectomy in these cases is a wise choice. In our series, staged thyroidectomy was planned in 18 patients with loss of signal on the first side, and 16 thyroidectomies were performed without complication. In one case, the operation indications disappeared (a follicular neoplasm was reported as benign adenoma), while one patient did not accept the surgery.

Bilateral vocal cord paralysis is a serious complication that may result in tracheostomy, and intraoperative neuromonitoring should prevent its occurrence. In our study, a tracheostomy was needed for patients operated without intraoperative neuromonitoring, but no tracheostomy was needed in intraoperative neuromonitoring group. However, we have no explanation for the bilateral vocal cord paralysis using intraoperative neuromonitoring for our patient. After completion of the first site of resection, recurrent laryngeal nerve was visually and anatomically intact with the intact R2 and V2 signals. But the lack of V2 signal on the second site alerted us and we checked and confirmed the anatomical integrity of that site. Then we checked the first site and saw the lack of both R2 and V2 signals. We thought there was a latent diffuse injury on the first site and segmental injury on the second site and found the point where the signal transmission ceased. Presumably, it was caused by traction or gradual thermal injury. We administered intravenous steroids, extubated the patient and examined the vocal cords. The cords were paralysed with an acceptable opening in the rima (3 mm). We observed the patient in the hospital and vocal cord paralysis on the second side dissolved two days postoperatively and the patient was discharged. The vocal cord paralysis on the other side recovered at three weeks postoperatively.

Conclusion

Evaluation of signal loss during thyroidectomy with intraoperative neuromonitoring is challenging. The first reaction should be the correct application of the troubleshooting algorithm to establish true signal loss when faced with it. If the recurrent laryngeal nerve is visually and anatomically intact, staged thyroidectomy is a logical option, whatever the indication for the operation. If the recurrent laryngeal nerve is disrupted, the surgeon should re-evaluate the indications for surgery. If total thyroidectomy is warranted for cancer, it is logical to terminate the surgery and to refer the patient for completion of thyroidectomy to a more experienced endocrine surgery centre if an experienced endocrine surgeon is not available in the hospital. But if available, one strategy could be to consult intraoperatively, re-evaluate the indications together and resect the other site with the strict use of continuous vagal stimulation at the same session without any nerve damage.