Efficacy of Intraoperative Neuro-Monitoring to Localize the External Branch of the Superior Laryngeal Nerve

Introduction

Injury of the external branch of the superior laryngeal nerve (EBSLN) alters the motor innervation of the inferior pharyngeal constrictor muscle, as well as the cricothyroid muscle whose function is to displace the thyroid cartilage down and forward to facilitate vocal cord tensing (1,2). This is manifested clinically after thyroidectomy. The symptoms include a monotonous voice that lacks acute tones and an inability to elevate the tone of voice suddenly (3,4). This implies a clear alteration in quality of life, particularly in people whose professional activity is related to their voice. It is difficult to emit a simple alarming shout if the EBSLN is affected. Hence, injury affects all patients (5 –8).

Prevalence of EBSLN injury varies widely and ranges from 0% to 58% (1,5,9). This wide range depends on the postoperative assessment used to detect it. This can be achieved clinically or through laryngoscopy, stroboscope, and/or electromyography. Taking postoperative electromyography as a standard parameter to detect EBSLN injury, the latter has been reported by Jansson et al. in 58% (10), by Aluffi et al. in 14% (9), and by the Thyroid Clinic of the General Hospital of Mexico in 14% of patients (11) when the EBSLN is not intentionally visually localized. This occurs in 8% of patients when it is intentionally located. Therefore, intraoperative localization of the EBSLN is undoubtedly important for the well-being of every patient subjected to a thyroidectomy.

The methods and techniques designed to avoid injury to the EBSLN vary greatly and range from intraoperative technical recommendations (12 –15) to the use of IONM. In fact, the use of IONM has been reported to be able to achieve a successful localization of EBSLN in 100% of cases (16 –18). There are multiple controlled clinical studies that compare the visual localization of the EBSLN against the use of IONM. This demonstrates that the latter method is much more efficient in localizing the nerve (19 –22).

Nevertheless, no proof yet exists that visual localization definitively confirms that the identified structure is the EBSLN. In fact, a possible explanation for the increased prevalence of ESBNL injury assessed through postoperative electromyography could be because the structure identified by the surgeon as EBSLN is actually not this structure.

Thus, in order to determine whether the anatomical localization—intensely proposed by the authors' research group and others—is indeed as effective as assumed, the objective of this study was to determine whether visual localization of EBSNL coincides with its localization through IONM during thyroidectomy and if the latter method influences the frequency of injuries.

Materials and Methods

A prospective, cross-sectional, comparative, observational study was performed in 240 superior thyroid poles from 148 patients who underwent first-time thyroid surgery (92 total thyroidectomies and 56 lobectomies) at the Thyroid Clinic of the General Hospital of Mexico from December 2013 to December 2014.

The patients provided written informed consent for the surgical procedure, as well as the study and publication of the results according to the institutional review board. The study protocol was approved by the hospital's Ethics and Research Committee.

All surgeries were performed under general anesthesia. No paralytic agents were used except for an initial dose of succinylcholine at 1–2 mg/kg for intubation. Neither lidocaine jelly nor any lubricants on the tube were used (23).

For the IONM technique, the NIM-Response^® 3.0 System (Medtronic, Jacksonville, FL) was used in all operations with surface electrodes integrated with an endotracheal tube (NIM Contact^® reinforced EMG Endotracheal Tube). This tube was inserted between the vocal folds by an anesthetist under direct visualization during intubation. Adequate tube position was confirmed by direct visualization of the surface electrodes and the vocal folds in the final surgical patient's position. A standardized technique of neuro-monitoring of the recurrent laryngeal nerve (RLN) was used, including indirect vagal response evaluation at the beginning and at the end of the operation according to recommendations formulated by the International Intraoperative Monitoring Study Group (23). The nerves were stimulated using a monopolar electrode and an interrupted stimulation technique at 1 mA, 100 ms impulse duration, and 4 Hz frequency. A positive signal was determined by observing contraction of the cricothyroid muscle (“cricothyroid twitch”) and/or a formal glottic response acoustic alarm; a clear EMG-evoked action potential waveform >100 μV was obtained (17,24,25). The response amplitudes were recorded.

The variables of interest corresponded to two maneuvers. First, the superior thyroid pedicle was exposed through a partial section of the internal border of the sternothyroid muscle. It was then bluntly resected from the sternothyroid–laryngeal triangle (Jolles space). The EBSLN was identified visually (no magnifying loupe to identify the EBSLN visually was used). The EBSLN was classified according to the classification established by Cernea et al. (visual identification) (26).

Next, the second maneuver was performed. This consisted of corroborating via IONM whether the anatomical element indeed corresponded to the functional EBSLN whose response was considered positive when a twitch of the cricothyroid muscle or a formal glottic response was obtained by stimulating the identified structure with 1 mV (24,25). If this was not the case, the entire region was mapped to locate the EBSLN and classified according to the Cernea classification. However, this example used IONM. The same surgeon performed the whole procedure. Once the EBSLN was localized through IONM, the vessels of the superior pedicle were sealed, and cut via a Focus harmonic scalpel (Ethicon Endo-Surgery, Inc., Cincinnati, OH). After completion of the superior thyroid pole resection, the functional integrity of the nerve was documented through a new electrical stimulation of the EBSLN and a positive cricothyroid twitch response or a formal glottic response (24,25).

An independent medical phonetician did voice and laryngoscopy evaluation. Preoperative evaluation of all patients included direct fiber-optic laryngoscopy to check vocal cord mobility performed 24 hours before surgery. Patients with preoperative vocal cord lesions were excluded from the study. Postoperative follow-up consisted of a direct fiber-optic laryngoscopy 24 hours and three months after the surgical procedure to check vocal cord mobility again. This was particularly applied for deviation of the petiole of the epiglottis to the side of cricothyroid muscle weakness during extreme high-pitch voice to establish acute and chronic unilateral ESLN denervation (4,8). Further failure of the glottis to elongate was recorded when the patient attempts to glide from modal to high register phonation to determine the presence of bilateral cricothyroid paralysis (24). At this time, all patients were asked about any voice change with emphasis on deeper voice or an inability to produce high-pitched sounds (24).

The statistical analysis assessed the concordance between visualization and IONM using the kappa test with linear weighting. The difference between visual and IONM identification was assessed via chi-square test to establish a significance level at p < 0.05. The odds ratio was 95% of the confidence interval. The pre- and post-resection EBSLN response was analyzed by Student's t test at a significance level of p < 0.05.

Results

Of the 148 patients, 142 were women and six were men. The average age was 41 years (range 18–78 years) with a mode of 35 years and a median of 39 years (SD = 14.4 years).

Of the 148 cases, 56 (37.8%) corresponded to thyroid lobectomies and 92 (62.2%) to total thyroidectomies with diagnoses of 14 (9.45%) follicular adenomas, 42 (28.35%) nodular colloid goiters, 8 (5.45%) multinodular goiters, and 84 (56.75%) papillary carcinomas of the thyroid. All RLN at risk were preserved.

Of the 240 superior thyroid poles studied, IONM identified the EBSLN in 234 (97.5%), and 190 (79.1%) EBSLN were identified visually: OR = 10.35 [CI 4.37–24.65]. Comparison of the localization capacity between both methods revealed a difference through chi-square with a p-value of <0.0001.

Of the 190 EBSLN visually identified, 150 (78.9%) were corroborated through IONM. The identified structure was indeed the EBSLN. Analysis via kappa with linear weighting revealed a value of 0.362 with a standard error of 0.0467 [CI 0.2686–0.4554]. This indicates fair agreement between the visual and IONM identifications. Comparison between IONM and visual localization according to the Cernea classification was as follows:

Cernea type 1. Of 114 nerves defined as type 1 via IONM, only 58 coincided with the 70 that were visually identified; the remaining 12 were identified through the IONM mapping as Cernea 2a with an OR = 2.8 [CI 1.92–4.19] in favor of IONM localization (p < 0.0001).

Cernea type 2a. Ninety-four nerves were identified as type 2a through IONM, coinciding with 70/92 identified visually; the remaining included 18 type 1 and four type 2b. The OR = 1.56 [CI 1.06–2.28] is in favor of IONM localization (p < 0.02).

Cernea type 2b. IONM identified 26 type 2b nerves confirming 22/28 that were identified visually. The remaining six were identified as type 2a through IONM. The OR = 1.20 [CI9 0.66–2.18] is in favor of IONM localization (p = 0.54).

Cernea Ni. It was not possible to identify the nerve visually in 50 cases. Of these, IONM could not offer identification in six cases. Of the 44 remaining cases identified through IONM, 38 corresponded to type 1 and six to type 2a. The OR = 10.26 [CI 4.30–24.45] is in favor of the IONM localization (p < 0.0001). The outcome of classifications is shown in Table 1.

All 234 nerves identified by IONM presented cricothyroid twitch (pre- and post-resection). Of these, 168 (71.7%) had glottic response and an EMG-evoked action potential waveform >100 mV. The pre-resection average EBSLN response was 271.99 mV (SD = 75.45 mV), and the post-resection average response was 264.73 mV (SD = 74.98 mV). Comparison between the pre- and post-resection response through a two-tailed Student's t-test yielded a p-value of 0.3769 [CI −8.88 to 23.41].

The postoperative assessment with direct laryngoscopy and subjective evaluation of the voice did not find any evidence of alteration at 24 hours and three months of follow-up in the 234 nerves identified by IONM as well in the six cases in whom the nerve was not located. All patients completed the entire follow-up.

Discussion

A usual behavior is to minimize the effect of injuring the EBSLN and to circumscribe its importance to people having a voice-related professional activity (4,9,10,27 –29). However, safe thyroid surgery undoubtedly also implies preserving the physical and functional integrity of the EBSLN in all patients (28,29).

The intentional and systematic visual search for the EBSLN and its classification according to Cernea et al. (26) revealed a difference in the frequency of the different subtypes as a function of geography (1). This frequency does not correspond to the original frequencies described by Cernea et al. (26). Injuries to the EBSLN at the Thyroid Clinic of the General Hospital of Mexico demonstrated by postoperative electromyography occur in 8%, even if the nerve is searched intentionally (11). These results lead to two conclusions: the first is that using IONM to identify the EBSLN is undoubtedly a more adequate method because it allowed for the identification of 97.5% compared with only 79.1% identified visually. This reveals a significant statistical difference. The OR value implies that it is 10 times more feasible to identify this nerve with IONM each time that it is achieved visually.

Although there are already large randomized studies in the literature comparing the identification of the EBSLN between visualization and the addition of IONM (19 –22), most compare two separate groups—one via visual identification and the other via IONM identification. Both conclude that the use of IONM is superior. However, the results raise the question of whether the structure identified as EBSLN by visualization really is the EBSLN. This is because none of these studies objectively proved that the structure identified visually was indeed the EBSLN. This study answers this question.

A recent study (30) compared identification of the EBSLN by visualization alone versus visualization + IONM to minimize EBSLN injury. This group concluded that using IONM and performing a meticulous dissection of the upper thyroid pole improves EBSLN identification. This study has some biases. It informs only the percentage of incidence and identification of Cernea type 1, 2a, and 2b while comparing nerves identified by dissection and IONM confirmation with cricothyroid twitch. Hence, the report is incomplete. Consequently, the report does not offer the final Cernea type of the 41 nerves that were not visually identified (Cernea Ni). The report does not report 40% of the nerves initially identified visually and not corroborated by IONM. The report does not use an appropriate interobserver test such as the kappa test with linear weighting. In fact, many of the EBSLN that were not visually identified and not corroborated by IONM are at risk of being injured. This is the essential part of EBSLN monitoring. The present study clarifies this situation. Undoubtedly, all previously performed studies are relevant because they constitute the basis on which knowledge is improved. This allows us to pursue new techniques and surgical approaches to provide a safe surgical practice.

Furthermore, IONM use drastically changed the Cernea classification. Although it was originally argued that it was not feasible to reproduce the frequency of the different types originally reported (26), which implied the existence of some ethnic variation (1), it is now clear that there is a similar frequency. This discrepancy is likely due to the fact that the initial classification was performed using the dissection of 30 cadaveric superior thyroid poles (26). This seems to be corrected via studies in actual thyroidectomies (1,31). Hence, it was impossible to obtain an exact test to certify the structure identified as EBSLN. Through the present study, we now know that the frequency of the different types of nerves is similar to the original description of Cernea et al. (26). However, this was only achieved by using IONM. Thus, the latter eliminates the error factor associated with visual identification.

The second conclusion is that the use of IONM diminishes the possibility of injuries markedly. In fact, it was 0% in this work based on 100% pre- and post-resection twitch and recorded glottic responses (with no statistical difference) through EMG. In addition, there were no alterations in the postoperative laryngoscopy and subjective voice assessment—this was mainly due to adequate identification of the EBSLN. Furthermore, positions 2a and 2b are the ones most exposed to risk of injury. The present results reveal that when using IONM, 18 (7.5%) cases were identified that had not been identified visually with an injury risk, yielding a kappa value of fair agreement. This is very similar to the 8% found in a previous study (11), implying that this frequency of injuries is directly related to the inadequate identification of the nerve and not with the manipulation per se as initially believed (11).

The lack of visual and IONM identification in six cases is difficult to explain. All maneuvers for visual identification were performed, as well as IONM mapping of the whole inferior constrictor muscle of the pharynx and the cricothyroid muscle. Neither was successful. Furthermore, long-lasting muscle relaxants also failed when used according to the criteria of the “Electrophysiologic recurrent laryngeal nerve monitoring during thyroid and parathyroid surgery: international standards guideline statement” (23). However, the postoperative evaluation—according to the mentioned parameters—revealed the six patients to be asymptomatic. The EBSLN was likely already inside the pharynx. This may be the cause of the lack of stimulation via IONM.

Finally, there should be no doubt about the role played by IONM in localizing the EBSLN during thyroidectomy as well as its importance. It allows for EBSLN identification in 97.5% of cases. It does not concur with visual identification in all evaluated nerves and, as shown in this study, is clearly superior. In this study, EBSLN morbidity was 0%.