Elucidating Mechanisms of Recurrent Laryngeal Nerve Injury During Thyroidectomy and Parathyroidectomy

Elucidating Mechanisms of Recurrent Laryngeal Nerve Injury During Thyroidectomy and Parathyroidectomy Samuel K Snyder, MD, FACS, Terry C Lairmore, MD, FACS, John C Hendricks, MD, FACS, John W Roberts, MD, FACS Intraoperative nerve monitoring during thyroidectomy, parathyroidectomy, or related central neck procedures can elucidate actual or potential mechanisms of recurrent laryngeal nerve (RLN) injury, especially visually intact nerves, which were previously unknown to the endocrine surgeon. STUDY DESIGN: In this prospective evaluation study, 373 patients underwent 380 consecutive thyroidectomy- or parathyroidectomy-related operations using intraoperative nerve monitoring, with 666 RLNs at risk. The success of visual and functional identification of the RLN, persistent loss of RLN function to nerve stimulation, the mechanism and location of RLN injury, and anatomy of the RLN or technical difficulties that appeared potentially risky for RLN injury were recorded. RESULTS: RLN was identified visually or functionally in 98.2% of nerves at risk. Initial intraoperative injury to the RLN occurred in 25 nerves at risk (3.75%). It was significantly more likely to be a visually intact RLN (n ⫽ 22; 3.3%) than a transected RLN (n ⫽ 3; 0.45%), p ⬍ 0.001. Paralysis persisted in 2 RLNs (0.3%). Visual misidentification accounted for only 1 RLN injury; the most common cause of injury resulted from traction to the anterior motor branch of a bifurcated RLN near the ligament of Berry (n ⫽ 7; 28%), then paratracheal lymph node dissection (n ⫽ 6; 24%), incorporating ligature (n ⫽ 4; 16%), and adherent cancer (n ⫽ 4; 16%). Fifty nerves at risk (7.5%) were identified as particularly at risk for injury, most notably those with anatomic variants (n ⫽ 26; 52%) and large or vascular thyroid lobes (n ⫽ 19; 38%). CONCLUSIONS: RLN injury during thyroidectomy or parathyroidectomy occurs intraoperatively significantly more often to a visually intact RLN than to a transected nerve. The anterior motor branch of an RLN bifurcating near the ligament of Berry is particularly at risk of traction injury. (J Am Coll Surg 2008;206:123–130. © 2008 by the American College of Surgeons) BACKGROUND:

Identification and preservation of the recurrent laryngeal nerve (RLN) is an essential aspect of thyroid and parathyroid surgery. Routine visual identification of the RLN has been shown to result in a lower incidence of RLN injury in multiple studies.1-3 But this operative standard still results in a 4% to 8% incidence of initial postoperative RLN paralysis; with continued followup, the RLN frequently recovers function, resulting in a permanent paralysis rate of 1% to 2%.4-7 A discouraging circumstance of thyroid and parathyroid surgery is to visually identify and preserve the

RLN, only to discover postoperatively a unilateral vocal cord paralysis. This calls into question the validity of the RLN identification and the mechanism of injury. The risk of injury is increased with thyroidectomy for cancer, substernal goiter, chronic thyroiditis, Graves’ disease, and reoperative neck surgery.4 These disease processes make RLN identification more difficult. Injury to the RLN can result from sharp trauma (transection), clamping, ligation, compression, traction, thermal injury, or ischemia.8 In an attempt to minimize the risk of RLN nerve injury, several different techniques of intraoperative nerve monitoring (IONM) have been developed to help identify the RLN.9-11 A common technique connects the vocalis muscle to an electromyographic monitor with implanted or surface electrodes.12-15 This is facilitated by attaching recording electrodes to the surface of the endotracheal tube,4 allowing repeated assessment of RLN function intraoperatively, so if a loss of RLN function to stimulation develops,

Competing Interests Declared: None. Received April 26, 2007; Revised June 26, 2007; Accepted July 17, 2007. From the Department of Surgery, Scott and White Memorial Hospital and Clinic; Scott, Sherwood and Brindley Foundation; and The Texas A&M University System Health Science Center College of Medicine, Temple, TX. Correspondence address: Samuel K Snyder, MD, Department of Surgery, Scott and White Memorial Hospital, 2401 South 31st St, Temple, TX 76508.

© 2008 by the American College of Surgeons Published by Elsevier Inc.

123

ISSN 1072-7515/08/$34.00 doi:10.1016/j.jamcollsurg.2007.07.017

124

Snyder et al

Mechanisms of Recurrent Laryngeal Nerve Injury

Abbreviations and Acronyms

IONM ITA NAR RLN

⫽ ⫽ ⫽ ⫽

intraoperative nerve monitoring inferior thyroid artery nerves at risk recurrent laryngeal nerve

resulting postoperative vocal cord paralysis can be anticipated. Frequently, a point loss of nerve conduction to stimulation can be identified with proximal nerve nonfunction and distal nerve intact function to electrical stimulation.16 To identify intraoperative mechanisms of RLN injury, a prospective evaluation of IONM during thyroidectomy, parathyroidectomy, or related central neck procedures was undertaken. The aim of the study was to determine the location and mechanisms of RLN injury and the subsequent prospect for recovery of nerve function after injury. Additionally, instances of potential nerve injury that IONM helped identify would be enumerated to better understand mechanisms of RLN injury that could occur during central neck operations.

METHODS From March 2004 to May 2006, 373 patients underwent 380 consecutive thyroidectomy- or parathroidectomyrelated operations using IONM during general anesthesia with electrode-bearing endotracheal tubes (Medtronic Xomed Nerve Integrity Monitor EMG Endotracheal Tube) connected to an electromyographic monitor (Medtronic NIM-Response). There were potentially 680 RLNs at risk (NAR). Six preoperatively paralyzed RLNs, 3 nonsalvageable cancer-invaded RLNs that were intentionally sacrificed to allow en bloc resection of a thyroid cancer, and 5 lateral neck dissections exposing only the vagus nerve were removed from analysis, leaving 666 NAR and 369 patients undergoing 375 procedures. During intubation with the electrode-bearing endotracheal tube, the anesthesiologist verified placement of the recording wires against the vocal cords. This was confirmed on the Nerve Integrity Monitor unit by demonstrating less than 1.0 K⍀ impedance for each wire. The event stimulus threshold, which determines how sensitive the Nerve Integrity Monitor will be in recording a twitch of a vocal cord, was set at 150 ␮V. The nerve stimulator wand (Medtronic Xomed Prass monopolar) used to test the RLN was set at 1.0 mA and emitted pulsed current. Direct physical touching of the motor elements of the RLN was required at these settings to produce a vocal cord contraction. Sliding the nerve stimulator over the deeper paratracheal tissues can potentially pinpoint the location of the RLN before visual exposure. The Nerve Integrity Monitor recorded a visual

J Am Coll Surg

wave spike and audible beep to each contraction above the threshold. The RLN was stimulated at its most proximally exposed portion for accurate assessment of intact nerve conduction to avoid overlooking a proximal nerve injury. A prospective registry of the utility was collected on the studied patients. Parameters recorded on the patients having IONM were the success in identifying the RLNs, the manner in which they were identified (visual, nerve stimulator, or both), and specific circumstances in which nerve stimulation particularly assisted the surgeon in identifying functional RLNs intraoperatively. Equipment or setup dysfunction that prevented assessment of the vocal cord response to RLN stimulation; altered RLN function to nerve stimulation, classified as a temporary loss of RLN function to nerve stimulation or a lower event threshold necessary to demonstrate intact RLN function; persistent loss of RLN function to nerve stimulation, classified as no recovery of RLN function intraoperatively even with a lower event threshold of 50 ␮V; nerve transection; and postoperative voice (vocal cord) function were recorded. In patients with persistent loss of RLN function to nerve stimulation intraoperatively, indicating nerve injury, repeated assessment of RLN function during the course of the operation allowed the surgeon to determine when the injury occurred and what the mechanism of injury (traction, compression, clamping, ligation, sharp trauma, or thermal injury) was. A traction injury was evident by the surgeon’s awareness of the degree and location of tension being applied to the RLN and the absence of other mechanisms of injury. When possible, the precise point at which nerve function to stimulation was lost along the RLN was recorded, while maintaining intact function to stimulation of the more distal portion of the RLN. Intraoperative anatomy or technical difficulties that the operating surgeon thought IONM identified as potentially risky for RLN injury were recorded. Flexible fiberoptic or video laryngoscopy was used selectively to evaluate instances of suspected vocal cord paralysis or instances of postoperative voice dysfunction in which vocal cord paralysis was suspected but not for instances of normal IONM without postoperative voice dysfunction. Paralysis or paresis of the vocal cord so identified was then followed with laryngoscopy every 3 months for up to 1 year for evidence of recovery, after which paralysis or paresis was defined as permanent. Analysis of these data and study design was approved by the Institutional Review Board. Statistical comparisons were made with chi-square test, Fisher’s exact test, and the test for multinomial proportions, as appropriate. A p value of 0.05 or less was defined as significant. There was no financial or professional asso-

Vol. 206, No. 1, January 2008

Mechanisms of Recurrent Laryngeal Nerve Injury

Snyder et al

125

Table 1. Central Neck Procedures Procedure

Thyroid disease only Thyroidectomy ⫹ CLND CLND Parathyroid disease Single parathyroidectomy ⫹ thyroidectomy ⫹ CLND Subtotal parathyroidectomy ⫹ thyroidectomy ⫹ CLND Total

n

Reoperative, n

n

NAR Reoperative, n

n

RLN injury Reoperative, n

259 71 9

25 13 9

445 138 14

34 25 14

10 11 0

0 1 0

5 12 4 5 8 2 375

1 — 1 1 1 — 51

7 24 8 10 16 4 666

1 — 2 2 2 — 80

0 1 2 0 1 0 25

0 — 1 0 0 — 2

CLND, central lymph node dissection; NAR, nerves at risk; RLN, recurrent laryngeal nerve.

ciation between authors and the commercial company whose nerve monitoring product was studied.

RESULTS There were 369 patients who underwent 375 thyroidectomy- or parathyroidectomy-related central neck operations using IONM. A variety of central neck operative procedures were performed (Table 1): 339 for thyroid disease only (597 NAR); 21 for additional single gland parathyroid disease (39 NAR), with most procedures including some type of thyroidectomy or paratracheal lymph node dissection (76%); and 15 for multiglandular parathyroid disease (30 NAR), with most procedures including some type of thyroidectomy or paratracheal lymph node dissection (67%). There were 51 reoperative procedures (13.6%) in the central neck compartment, with 80 NAR (Table 1). For the 629 unilateral thyroidectomy procedures, there were 533 total lobectomies (84.7%) with 527 NAR (excludes 6 preoperatively paralyzed or nonsalvageable, cancer-invaded RLNs), 76 near-total lobectomies (12.1%)

leaving a 0.5- to 1-g posterior remnant with 76 NAR, and 20 partial lobectomies (3.2%) with 20 NAR (Table 2). There were 164 NAR exposed to possible injury with a paratracheal lymph node dissection, 141 of those in addition to a total thyroid lobectomy. RLN injury occurred in 2 of 80 NAR with an earlier central neck operation (2.5%) versus 23 of 586 NAR without an earlier central neck operation (3.9%, p ⫽ 0.76); 13 of 164 NAR with a central lymph node dissection (CLND, 7.9%) versus 12 of 502 NAR without a CLND (2.4%, p ⬍ 0.001); and 23 of 527 NAR with a total lobectomy (4.4%) versus 1 of 76 NAR with a near-total lobectomy (1.32%, p ⫽ 0.34). One RLN injury occurred during CLND after an earlier thyroid lobectomy. The recurrent laryngeal nerve was identified visually, functionally, (recordable vocal cord contraction with IONM using nerve stimulation) or both ways in 98.2% of NAR: visually only in 1.5%; RLN stimulation only in 2.4%: vagal nerve stimulation only in 1.5%; and both visually and with nerve stimulation in 92.8%. The nerve stimulator located 19.4% of NAR before visual identifica-

Table 2. Procedures Exposing the Recurrent Laryngeal Nerve to Risk of Injury Procedure

Unilateral thyroidectomy Total lobectomy (84.7%) ⫹ unilateral LND Near-total lobectomy (12.1%) Partial lobectomy (3.2%) Paratracheal lymph node dissection Parathyroid exploration Total

n

NAR

RLN injury

%

388 145 76 20 26 20 675

386 141 76 20 23 20 666*

11 12 1 0 1 0 25

2.85 8.50 1.32 0.00 4.35 0.00 3.75

*Excludes nine preoperatively paralyzed or nonsalvageable cancer-invaded RLNs. LND, lymph node dissection; NAR, nerves at risk; RLN, recurrent laryngeal nerve.

126

Snyder et al

Mechanisms of Recurrent Laryngeal Nerve Injury

J Am Coll Surg

Table 3. Recurrent Laryngeal Nerve Injury Injury no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Mechanism

Location, cm (below entry point)

Vocal cord followup

Transection, paratracheal LND (misidentification) Transection, adherent cancer Transection (partial), ITA ligature Traction, bifurcated RLN Traction, bifurcated RLN Traction, bifurcated RLN Traction, bifurcated RLN Traction, bifurcated RLN Traction, bifurcated RLN Traction, bifurcated RLN Traction, paratracheal LND Unknown, paratracheal LND Unknown, paratracheal LND Traction, paratracheal LND Traction, paratracheal LND Traction, adherent cancer Traction, adherent cancer Traction, adherent cancer (anterior branch) Ligature Ligature (anterior branch) Ligature Compression Compression Traction Unknown, faulty IONM

2.5 Entry point 2.5 1.0 1.0 1.5 1.0 0.5 1.0 0.2 3.0 Entire nerve Entire nerve 1.5 3.0 0.2 2.0 2.0 1.0 1.5 0.6 1.0 1.0 Entry point Entire nerve

Paralysis Paresis Paralysis Normal Clinically normal Slight paresis Normal Normal Normal Normal Died 2 mo postop Normal Normal Normal Normal Normal Normal Paresis Normal Normal Normal Normal Normal Normal Normal

IONM, intraoperative nerve monitoring; ITA, inferior thyroid artery; LND, lymph node dissection; RLN, recurrent laryngeal nerve.

tion or in lieu of visual identification. Overall, 21% of patients had postoperative laryngoscopy. Of the 12 RLNs not identified, half of these involved just a partial thyroid lobectomy or limited parathyroidectomy. The other half that involved near-total or total thyroidectomy or superior parathyroidectomy had followup flexible laryngoscopy demonstrating normal vocal cord function. There were 10 RLNs that could not be tested intraoperatively with nerve stimulation because of faulty IONM setup that occurred when monitoring equipment dysfunction would not allow proper application or interpretation of nerve stimulation, eg, a broken wire within the endotracheal tube. Seven of these were examined with flexible laryngoscopy and one paretic vocal cord was discovered in a patient with postoperative hoarseness. This paresis resolved completely with further laryngoscopy followup (Table 3). The other three RLNs were clinically normal postoperatively, with no voice change. When intraoperative RLN visualization was difficult, intact function could be determined with RLN or vagal nerve stimulation. There were 37 RLNs (5.5%) that demonstrated altered RLN function to nerve stimulation (temporary loss of RLN function to nerve stimulation or a lower event threshold necessary to demonstrate intact RLN

function) during the operation. Twenty-five of these nerves were normal with postoperative flexible laryngoscopy, and the others were clinically normal, with no postoperative voice change. Intraoperative RLN injury was determined in 24 RLNs that had persistent loss of function to nerve stimulation, for 25 total RLN injuries (3.75%, Table 3). Injury was confirmed with postoperative laryngoscopy in 23 of 24 patients. This was significantly more common in a visually intact RLN (n ⫽ 22; 3.3%) than in a transected RLN (n ⫽ 3; 0.45%; p ⬍ 0.001). Visual misidentification accounted for only one RLN injury (transected nerve, mistaken for a branch of the inferior thyroid artery [ITA] as it passed under the ITA). One RLN was incorporated into a ligature of the ITA and partially transected. Both of these transected nerves remained paralyzed on followup, for a permanent paralysis rate of 0.3%. Three of the other injured nerves recovered partial function (0.45%), and the rest returned to normal function. One RLN superficially adherent to a cancer near the ligament of Berry was dissected bluntly to free it up, only to snap the nerve, resulting in transection at the laryngeal entry point. The RLN was sutured back to the laryngeal entry point, determined with

Vol. 206, No. 1, January 2008

Snyder et al

Mechanisms of Recurrent Laryngeal Nerve Injury

127

Table 4. Potential Recurrent Laryngeal Nerve Injury Causes of injury

Figure 1. Bifurcation of the recurrent laryngeal nerve within the ligament of Berry.

the help of the nerve stimulator, and it recovered partial function with followup. One patient died 2 months postoperatively and was not available for followup, and another patient with normal voice function on postoperative day 1 refused early and late laryngoscopy evaluation for visual confirmation of vocal cord function, because of an earlier experience of pain and sinusitis resulting from a laryngoscopic examination. The most common mechanism of RLN injury was a traction injury to the anterior motor branch of a bifurcated RLN in the ligament of Berry (Fig. 1). The traction injury was created in the more fragile anterior branch of the RLN just above its bifurcation. Nerve stimulation distal to this point revealed intact nerve conduction, but nerve stimulation proximal to the bifurcation did not conduct. Injury to the RLN during paratracheal lymph node dissection generally occurred more proximally secondary to traction when trying to tease the RLN away from adjacent fatty tissue or lymph nodes. Teasing an RLN away from superficial adherence to a thyroid cancer can have the same traction mechanism of injury. Three RLNs were caught in a small vessel branch ligature near the ligament of Berry that was discovered by trying to reconfirm RLN function to nerve stimulation after completion of the lobectomy. The incorporating ligature was carefully removed, indicating a visually intact nerve that had a point loss of nerve conduction at the ligature site. Compression of the RLN against the trachea while trying to obtain exposure of the ligament of Berry region produced two other RLN injuries. The most common location of RLN injury was 1.5 cm or less below the laryngeal entry point (16 RLNs, 64%). With followup, no visually intact RLN remained paralyzed. IONM assisted the surgeon in visually identifying anomalies of RLN anatomy, which potentially could have led to injury, by confirming the aberrant location of the RLN with a normal response to nerve stimulation (Table 4). These occurred in 3.9% of NAR, with the medially displaced RLN against the trachea near the inferior pole at the

Anatomic variants Medially displaced Bifurcated at ITA Medially displaced and bifurcated Medially and anteriorly displaced Anteriorly displaced Nonrecurrent Total anatomic variants Difficult dissections Large and/or vascular thyroid lobe Scarring Difficult localization Total difficult dissections

RLNs, n

%

10 7 1 1 3 4 26

1.50 1.05 0.15 0.15 0.45 0.60 3.90

19 4 1 24

2.85 0.60 0.15 3.60

ITA, inferior thyroid artery; RLN, recurrent laryngeal nerve.

thyroid being the most common anomaly. The bifurcated RLN at the ITA was suspected when nerve stimulation produced no response from the visually identified posterior sensory branch. More dissection anteriorly revealed the responsive anterior motor branch, which could have been mistaken for a vessel branch. The RLN anteriorly displaced by a thyroid nodule growing underneath it can mimic a thyroid vessel branch as well. The nonrecurrent laryngeal nerve likewise could have been mistaken for the superior thyroid artery or inferior thyroid artery. Visually identifying the RLN was extremely difficult, with very large thyroid lobes (which distort anatomy) and very vascular thyroid lobes (Table 4). This was enough of a problem in 2.85% of NAR to make RLN injury more likely. Nerve stimulation helped identify the functional RLN. Difficult scarring with reoperation placed the RLN at increased risk of potential injury in 4 NAR (0.6%). Nerve stimulation helped identify the location of the RLN in these instances. Overall, 24 RLNs (3.6%) was identified as potentially at risk of injury from a difficult dissection point of view.

DISCUSSION If one of the goals of thyroidectomy and parathyroidectomy is to minimize the incidence of RLN injury, then a detailed study of circumstances that increase the risk of injury is a natural first step in accomplishing that goal. This study is one attempt to elucidate mechanisms of injury that lead to RLN injury by using a new operative aid, IONM, which gives rapid and unique intraoperative feedback information concerning the functional status of the RLN. It is the repeatable, specific, and immediate feedback of IONM that allows the surgeon to frequently make the observations about when during the course of the opera-

128

Snyder et al

Mechanisms of Recurrent Laryngeal Nerve Injury

tion the RLN went from functional to nonfunctional to nerve stimulation; precisely where the RLN was injured, because nerve stimulation distal to a point of injury demonstrates intact function; and to deduce how the RLN was injured, by an awareness of the dissection techniques being used at the time of loss of function to nerve stimulation. Currently, the gold standard remains visual identification and preservation of the RLN.9,17 We followed that dictum and found that RLN transection was very uncommon (0.45%), but nonfunction of a visually intact RLN was significantly more often the case (3.3%). In a study reported by Chiang and colleagues5 of 704 NARs treated with total lobectomy and the standard of visual preservation of the RLN, 3 nerve transactions occurred (0.43%), and postoperative paralysis occurred as well in 37 intact RLNs (5.25%). More importantly, 35 of the 37 intact, but paralyzed, RLNs recovered completely with followup (94.6%). This experience agrees with the results of our study that nearly all intact, but injured, RLNs will eventually recover normal function. Transected RLNs do not fare nearly as well. It is unlikely that misidentification of the RLN will occur with IONM; there was only 1 in 666 NAR (0.15%) in our study. We found IONM helpful in identifying potential circumstances for RLN injury in 7.5% of NAR. It is impossible to guess what the RLN injury rate would have been in these instances without IONM, but perhaps it helped avoid RLN transection associated with misidentification. Multiple studies trying to compare RLN injury rates with and without IONM have failed to show a statistically significant difference.6,17,18 Comparative studies frequently do show a small, but nonsignificant decrease in RLN injury with IONM. With postoperative paralysis occurring intraoperatively significantly more often in visually intact RLNs, having a technique like IONM, which can predict this intraoperatively, becomes more useful. Knowledge of a unilateral nonfunctional RLN to nerve stimulation may alter the operative approach to the contralateral side.13 This did not occur in our study other than in a unilateral central lymph node dissection after earlier total thyroidectomy that was limited to the side of a preoperatively paralyzed vocal cord. We found the IONM setup we used reliable in predicting postoperative vocal cord paralysis when there was a persistent loss of RLN function (as defined in this study) to nerve stimulation after an initial normal response to nerve stimulation. A potential criticism of our study is that routine laryngoscopy was not performed preoperatively or postoperatively to verify normal vocal cord function, but only selectively, when vocal cord dysfunction was anticipated by preoperative history, intraoperative findings, or

J Am Coll Surg

postoperative voice function. We possibly could have underestimated the true incidence of preoperative or postoperative vocal cord dysfunction.4,19,20 But studies on IONM have indicated a high negative predictive value of 98% to 99.6% for postoperative vocal cord paralysis when normal RLN stimulation occurs intraoperatively.6,13,21 This study was directed more at elucidating mechanisms of RLN injury in patients with postoperative vocal cord paralysis. Understanding the vagaries of RLN anatomy is important in preventing RLN injury. Within the larynx, the RLN bifurcates into anterior and posterior branches. In as many as 30% of patients, the RLN will start to branch before reaching the laryngeal entry point beneath the cricopharyngeus muscle.4 In approximately 5% of patients, the branching occurs at the level of the RLN and ITA crossing. The anterior branch is nearly always the motor branch, but, occasionally, motor fibers are also in the posterior branch, particularly to the posterior cricoarytenoid muscle, which produces vocal cord abduction. The functional significance of this is that visual identification and preservation of only the posterior branch of the RLN will lead to vocal cord paralysis. Only functional information provided by IONM or the careful visual identification of RLN branching will avoid motor branch nerve injury. Branching of the RLN in the ligament of Berry proved to be problematic in our study. The branched nerve fibers are thinner and more fragile and are more susceptible to traction injury. The typical upward traction on the thyroid near the ligament of Berry can produce a traction injury to the closer anterior motor branch at a point just beyond the bifurcation, as demonstrated in our study. The traction pull resistance should be maximal in the anterior motor branch at the take-off from the main RLN, where the thicker main RLN and posterior branch are more relatively fixed in the tissues. Maneuvers to minimize the risk of this injury are an active awareness of the degree and duration of medial and anterior traction being applied to the RLN in the ligament of Berry, periodically easing up on the traction, altering the traction at times to a more cephalad or caudal direction, or performing a near-total thyroid lobectomy rather than a total lobectomy, which carried a lower but not statistically significant risk of RLN injury in our study. A medially displaced RLN up against the trachea near the inferior pole of the thyroid can easily be mistaken for a vessel branch entering the thyroid as the inferior pole is elevated. Usually, thyroid lobectomy is performed from a lateral to medial direction, but with a medially displaced RLN, it may necessitate a medial to lateral direction to keep the RLN in view. An anteriorly displaced RLN on the lateral surface of the thyroid or anterior surface of a poste-

Vol. 206, No. 1, January 2008

Snyder et al

rior thyroid nodule can also be easily mistaken for a vessel branch to the thyroid. IONM helped identify this anomaly in 4 NAR (0.6%) in our study. Visual identification alone requires finding it more proximally before much mobilization of the thyroid. The nonrecurrent laryngeal nerve requires a high index of suspicion when mobilizing the superior pole of the right thyroid or when the RLN cannot be identified in its usual location. Toniato and associates22 reported on their experience with 31 non-RLNs and reported a 12.9% incidence of paralysis. More commonly, the non-RLN followed a transverse course to pass above or below the ITA, but the less common direct course near the superior thyroid artery was more likely to be injured. Thyroid cancer surgery and reoperative central neck surgery have been cited in other studies as risk factors for RLN injury.5,6,17 In our study using IONM, reoperations did not carry an increased risk of RLN injury. In four patients, IONM clearly helped identify an RLN within a scarred operative field. Thyroid cancer leads to an increased risk of RLN injury, either from the nerve being adherent to a thyroid cancer or cancer-embedded lymph node or from the need to dissect along the course of the RLN with a paratracheal lymph node dissection. Our study supported a significantly increased risk of injury to the RLN with a central lymph node dissection for thyroid cancer. Traction injury in our study was observed to occur when a blood vessel, nerve branch, lymph node, or tumor was being firmly teased away from the RLN, briefly causing excessive localized traction that was confirmed with IONM to be the point of injury, with loss of function to nerve stimulation proximal to the point of injury. In one patient, excessive blunt mobilization near the RLN resulted in nerve transection. Small tracheal or esophageal nerve branches from the RLN or adherence to metastatically involved lymph nodes make it difficult to separate the RLN from the surrounding paratracheal lymphatic tissue. Careful sharp dissection close to the nerve to free it from superficial points of adherence, particularly under loop magnification, is helpful in avoiding excessive localized traction near the RLN. Once the RLN is visually identified, it is important to mobilize the thyroid by dividing vessels, while keeping the RLN in view. Excessive small vessel bleeding complicates this task, particularly near the ligament of Berry, where there is a rich plexus of vascularity. In our study, four ligated vessels incorporated the RLN into the ligature, three with small vessels in the ligament of Berry. It was believed that as the ligature was passed down over the clamped vessel, it was secured a little further beneath the clamp than planned to also catch the tented-up, nearby RLN. This ligature, including the RLN, was not clearly realized until after lobectomy, when a nonfunctional nerve was traced to the point

Mechanisms of Recurrent Laryngeal Nerve Injury

129

of ligature and documented to have normal nerve stimulation distal to the ligature. Without IONM identifying this RLN injury, permanent paralysis of the RLN would have occurred. The three visually intact nerves later recovered normal function with followup. Careful ligation of vessels near the RLN under direct vision is necessary to avoid this problem. Loop magnification can help accomplish this. Continuous blood aspiration and patience until the exact bleeding site is identified are required to achieve safe and precise hemostasis with troublesome bleeding near the RLN. Overall, the region of the ligament of Berry is the most difficult to dissect and the most likely to have an injury to the RLN, as demonstrated by our study and others.4,23 A near-total thyroid lobectomy, when feasible, may be a safe option to minimize the risk of RLN injury near the ligament of Berry. In conclusion, recurrent laryngeal nerve injury during thyroidectomy, parathyroidectomy, or related central neck procedures in the setting of intraoperative neuromonitoring and visual identification occurs significantly more often to a visually intact nerve than a transected nerve. Thyroid cancer procedures have significant risk of RLN injury related to adherent cancer or paratracheal lymph node dissection. The anterior motor branch of a recurrent laryngeal nerve bifurcating near the ligament of Berry is particularly at risk of traction injury. Anatomic variants of the RLN, such as medial displacement, anterior displacement, bifurcation at the ITA, and nonrecurrence, are potentially at increased risk of injury from misidentification. Author Contributions Study conception and design: Snyder Acquisition of data: Snyder, Lairmore, Hendricks, Roberts Analysis and interpretation of data: Snyder, Lairmore, Hendricks Drafting of manuscript: Snyder, Lairmore, Hendricks Critical revision: Snyder, Lairmore, Hendricks, Roberts

REFERENCES 1. Jatzko GR, Lisborg PH, Muller MG, et al. Recurrent nerve palsy after thyroid operations—principal nerve identification and a literature review. Surgery 1994;115:139–144. 2. Wagner HE, Seiler C. Recurrent laryngeal nerve palsy after thyroid gland surgery. Br J Surg 1994;81:226–228. 3. Hermann M, Alk G, Roka R, et al. Laryngeal recurrent nerve injury in surgery for benign thyroid disease: effect of the nerve dissection and impact of individual surgeon in more than 27,000 nerves at risk. Ann Surg 2002;235:261–268. 4. Randolph GW. Surgical anatomy of the recurrent laryngeal nerve. In: Randolph GW, ed. Surgery of the thyroid and parathyroid glands. Philadelphia, PA: Saunders; 2003:300–342.

130

Snyder et al

Mechanisms of Recurrent Laryngeal Nerve Injury

5. Chiang FY, Wang LF, Huang YF, et al. Recurrent laryngeal nerve palsy after thyroidectomy with routine identification of the recurrent laryngeal nerve. Surgery 2005;137:342–347. 6. Chan W, Lang BH, Lo C. The role of intraoperative neuromonitoring of the recurrent laryngeal nerve during thyroidectomy: a comparative study on 1000 nerves at risk. Surgery 2006;140: 866–873. 7. Aytac B, Karamercan A. Recurrent laryngeal nerve injury and preservation in thyroidectomy. Saudi Med J 2005;26:1746– 1749. 8. Myssiorek D. Recurrent laryngeal nerve paralysis: anatomy and etiology. Otolaryngol Clin North Am 2004;37:25–44. 9. Hermann M. Main topics: neuromonitoring in thyroid surgery—who benefits? Eur Surg 2003;35:225–226. 10. Tschopp KP, Gottardo C. Comparison of various methods of electromyographic monitoring of the recurrent laryngeal nerve in thyroid surgery. Ann Otol Rhinol Laryngol 2002;111:811– 816. 11. Marcus B, Edwards B, Yoo S, et al. Recurrent laryngeal nerve monitoring in thyroid and parathyroid surgery: the University of Michigan experience. Laryngoscope 2003;113:356–361. 12. Horn D, Rotzscher VM. Intraoperative electromyogram monitoring of the recurrent laryngeal nerve: experience with an intralaryngeal surface electrode. Langenbecks Arch Surg 1999; 384:392–395. 13. Thomusch O, Sekulla C, Machens A, et al. Validity of intraoperative neuromonitoring signals in thyroid surgery. Langenbecks Arch Surg 2004;389:499–503. 14. Pearlman RC, Isley MR, Ruben GD, et al. Intraoperative monitoring of the recurrent laryngeal nerve using acoustic, free-run,

15. 16. 17. 18.

19.

20. 21. 22. 23.

J Am Coll Surg

and evoked electromyography. J Clin Neurophysiol 2005;22: 148–152. Petro ML, Schweinfurth JM, Petro AB. Transcricothyroid, intraoperative monitoring of the vagus nerve. Arch Otolaryngol Head Neck Surg 2006;132:624–628. Snyder SK, Hendricks JC. Intraoperative neurophysiology testing of the recurrent laryngeal nerve: plaudits and pitfalls. Surgery 2005;138:1183–1192. Dralle H, Sekulla C, Haerting J, et al. Risk factors of paralysis and functional outcome after recurrent laryngeal nerve monitoring in thyroid surgery. Surgery 2004;136:1310–1322. Robertson ML, Steward DL, Gluckman JL, et al. Continuous laryngeal nerve integrity monitoring during thyroidectomy: does it reduce risk of injury? Otolaryngol Head Neck Surg 2004; 131:596–600. Randolph GW, Kamani D. The importance of preoperative laryngoscopy in patients undergoing thyroidectomy: voice, vocal cord function, and the preoperative detection of invasive thyroid malignancy. Surgery 2006;139:357–362. Sinagra DL, Montesinos MR, Tacchi VA, et al. Voice changes after thyroidectomy without recurrent laryngeal nerve injury. J Am Coll Surg 2004;199:556–560. Beldi G, Kinsbergen T, Schlumpf R. Evaluation of intraoperative recurrent nerve monitoring in thyroid surgery. World J Surg 2004;28:589–591. Toniato A, Mazzarotto R, Piotto A, et al. Identification of the nonrecurrent laryngeal nerve during thyroid surgery: 20-year experience. World J Surg 2004;28:659–661. Yalcxin B. Anatomic configurations of the recurrent laryngeal nerve and inferior thyroid artery. Surgery 2006;139:181–187.