NEGATIVE AND POSITIVE PREDICTIVE VALUES OF NERVE MONITORING IN THYROIDECTOMY

Transcription

1 ORIGINAL ARTICLE NEGATIVE AND POSITIVE PREDICTIVE VALUES OF NERVE MONITORING IN THYROIDECTOMY Claudio R. Cernea, MD, 1,2 Lenine G. Brandão, MD, 1,2 Flavio C. Hojaij, MD, 2 Dorival De Carlucci, Jr, MD, 1,2 José Brandão, MD, 1,2 Beatriz Cavalheiro, MD, 1,2 Adriana Sondermann, MD 1,2 1 University of São Paulo Medical School, São Paulo, Brazil. cerneamd@uol.com.br 2 Department of Head and Neck Surgery and Albert Einstein Jewish Hospital, São Paulo, Brazil Accepted 22 October 2010 Published online 16 March 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI: /hed Abstract: Background. Recurrent nerve injury is 1 of the most important complications of thyroidectomy. During the last decade, nerve monitoring has gained increasing acceptance in several centers as a method to predict and to document nerve function at the end of the operation. We evaluated the efficacy of a nerve monitoring system in a series of patients who underwent thyroidectomy and critically analyzed the negative predictive value (NPV) and positive predictive value (PPV) of the method. Methods. NIM System efficacy was prospectively analyzed in 447 patients who underwent thyroidectomy between 2001 and 2008 (366 female/81 male; 420 white/47 nonwhite; 11 to 82 years of age; median, 43 years old). There were 421 total thyroidectomies and 26 partial thyroidectomies, leading to 868 nerves at risk. The gold standard to evaluate inferior laryngeal nerve function was early postoperative videolaryngoscopy, which was repeated after 4 to 6 months in all patients with abnormal endoscopic findings. Results. At the early evaluation, 858 nerves (98.8%) presented normal videolaryngoscopic features after surgery. Ten paretic/paralyzed nerves (1.2%) were detected (2 unexpected unilateral paresis, 2 unexpected bilateral paresis, 1 unexpected unilateral paralysis, 1 unexpected bilateral paralyses, and 1 expected unilateral paralysis). At the late videolaryngoscopy, only 2 permanent nerve paralyses were noted (0.2%), with an ultimate result of 99.8% functioning nerves. Nerve monitoring showed absent or markedly reduced electrical activity at the end of the operations in 25/868 nerves (2.9%), including all 10 endoscopically compromised nerves, with 15 falsepositive results. There were no false-negative results. Therefore, the PPV was 40.0%, and the NPV was 100%. Conclusions. In the present series, nerve monitoring had a very high PPV but a low NPV for the detection of recurrent nerve injury. VC 2011 Wiley Periodicals, Inc. Head Neck 34: , 2012 Keywords: thyroidectomy; nerve monitoring; laryngeal nerves; recurrent nerve; laryngeal paralysis Correspondence to: C. R. Cernea Oral Presentation at the 4th World Congress of the International Federation of Head and Neck Oncological Societies, Seoul, Korea, June 15 19, VC 2011 Wiley Periodicals, Inc. Since the landmark work of Emil Theodor Kocher, morbidity of thyroidectomy has been reduced to a very low level. 1 However, to keep this safety, the head and neck surgeon must be very familiar with structures adjacent to the thyroid gland, including the laryngeal nerves and the parathyroid glands. Nerve monitoring (NM) with the NIM system (Medtronic, Jacksonville, FL) was introduced by Lamade et al 2,3 in 1996 and 1997 and has gained increasing popularity among surgeons during the last decade through the remarkable publications by Randolph et al. 4,5 In brief, the patient is intubated with an endotracheal tube containing special surface electrodes that are positioned at the level of the vocal folds. Thus, when the laryngeal nerves are stimulated with a probe during surgery, electromyography results are recorded, enabling a quantitative evaluation of the neurophysiological status of the nerve. Horne et al 6 sent a questionnaire to 1685 surgeons in the United States; about 44.9% of the respondents who performed thyroidectomies used NM during the operation, at least in selected cases. Despite the utility of this method, controversial points of view have been reported. Some authors did not find any utility in the use of NM. 7 9 Others believe that nerve monitoring should be used only in a reoperative setting or when additional difficulty during the operation is anticipated (large retrosternal goiters or significant central nerve compartment metastatic disease, among others) Moreover, few articles had critically evaluated the positive predictive value (PPV) and negative predictive value (PPV) of recurrent nerve paralysis in thyroidectomy with NM, to clarify the true value of NM in predicting a laryngeal nerve with abnormal function at the end of the operation. Regarding PPV, results oscillated from 11.6% to 92.1%, with most articles describing low percentages, whereas NPV was much more efficient, varying from 96.1% to 99.7%. 10,12 17 This is a study of a cohort of consecutive patients who underwent thyroid operations in which NM was Predictive Value of Nerve Monitoring in Thyroidectomy HEAD & NECK DOI /hed February

2 used. The main objective was to compare the results of this NM, that is, intraoperative electromyography, with postoperative laryngoscopy, considered the gold standard method to detect vocal fold paresis or paralysis, focusing specifically on PPV and NPV of NM. MATERIALS AND METHODS Patients. A series of consecutive patients who underwent thyroidectomy with intraoperative NM was evaluated. Between 2001 and 2008, 447 patients underwent operation. There were 366 females and 81 males, with age ranging from 11 to 82 years (median, 43 years). Four-hundred twenty individuals were white, and 27 were nonwhite. Most of the thyroid glands in this study were normally sized or slightly enlarged (89.0%). A small percentage of the patients had toxic goiters (7.1%) or substernal goiters (5.8%). No patient had a preoperative abnormal videolaryngoscopy result. There were 421 total thyroidectomies and 26 partial thyroidectomies leading to 868 nerves at risk. All patients were intubated with the endotracheal tube containing surface electrodes (NIM System), which were positioned at the glottic level, in contact with both true vocal folds. The main reason for use of this system is its availability in Brazil; to our knowledge, it is the only system that can be routinely imported in our professional environment. We have no financial conflict of interest. The position of the tube was carefully checked before the incision to ensure proper orientation concerning the vocal folds, and the tube was securely fixed to the patient by the anesthesiologist to avoid rotation or vertical displacement, which could lead to false-positive results. During the operation, the patient was not paralyzed. The access to the thyroid gland was obtained through a very small (approximately 3 cm) incision. Therefore, at least in our experience, we had not gained access to the Xth nerve to enable its stimulation. The nerve stimulator was set at 0.5 ma, and the inferior laryngeal nerve was stimulated to confirm identification. If no initial response was observed, the intensity was raised to 1.0 ma. A typical electromyographic recording would show a peak after stimulation with amplitude above 100 lv. All inferior laryngeal nerves were identified, and their neurophysiological activity was recorded before the corresponding thyroid lobe was resected. After the resection, the functional integrity of the nerve was again checked and compared with the previous result. Especially during the last 8 years, with the addition of the harmonic scalpel, we have been performing conventional thyroidectomy through a very small (approximately 3 cm) incision. Therefore, because of this limited access, no vagal stimulation was attempted. Marked decrease (less than 20% of the amplitude obtained before the resection) or absence of motion after electrical stimulation was considered as positive for nerve paresis or paralysis. Immediate postoperative recovery was uneventful in most patients, who were usually discharged from the hospital on the next day, except for one patient who experienced bilateral vocal fold paralysis. He was not tracheostomized and left the hospital on the fourth postoperative day, receiving steroids and with an adequate airway. The gold standard to evaluate inferior laryngeal nerve function in this study was early postoperative videolaryngoscopy, performed from the second to the fifth postoperative day. Those patients with abnormal endoscopic findings underwent another endoscopic evaluation 4 to 6 months after the operation. Therefore patients with abnormal intraoperative NM but normal postoperative videolaryngoscopy results were considered to have false-positive results, whereas normal intraoperative NM with an abnormal postoperative videolaryngoscopy result was considered falsenegative. Accordingly, the PPV of NM was calculated dividing the number of true-positive results (positive NM and abnormal videolaryngoscopy result) by the total of positive NM results; the NPV was calculated by dividing the number of true-negative results (negative NM and normal videolaryngoscopy result) by the total negative NM. The possible causes for the falsepositive cases were evaluated. This study was approved by the institutional review board of our institution. RESULTS There were 868 nerves at risk. At the early postoperative laryngoscopic evaluation, 858 nerves (98.8%) had normal mobility. Ten paretic or paralyzed nerves (1.2%) were detected. Among these, 2 patients presented with unexpected unilateral paresis, 2 patients had unexpected bilateral paresis, 1 patient had unexpected unilateral paralysis, 1 patient had unexpected bilateral paralyses, and 1 patient had expected unilateral paralysis. This last patient had a papillary carcinoma with large metastasis on level VI with gross invasion of the recurrent nerve, which was excised. The patient who experienced unexpected bilateral recurrent nerve paralysis had a papillary carcinoma associated with subacute thyroiditis, which was successfully treated with steroids 6 months earlier. Despite the absence of pain or any other inflammatory symptoms at the time of the thyroidectomy, the dissection of the thyroid gland and of the 2 recurrent nerves was remarkably difficult because of extensive inflammation. At the late videolaryngoscopy, only 2 permanent unilateral nerve paralyses were noted (0.2%), corresponding to the 2 cases with unilateral paralysis diagnosed at the initial videolaryngoscopy, with an ultimate result of 99.8% functioning nerves. These 2 patients were the only ones who presented true absence of signal at the end of the operation, probably 176 Predictive Value of Nerve Monitoring in Thyroidectomy HEAD & NECK DOI /hed February 2012

3 reflecting complete neuropraxy, despite the apparent anatomic integrity of the nerve. The patient with bilateral vocal fold paralysis had a complete recovery 4 months after the operation. NM showed absent (2 nerves) or markedly reduced electrical activity (23 nerves) at the end of the operations in 25/868 nerves (2.9%), including all 10 endoscopically compromised nerves, with 15 falsepositive nerves. There were no false-negative nerves. Therefore the PPV was 40.0%, and the NPV was 100%. The causes of the false-positive nerves are depicted on Table 1. In 4 cases, no cause for this discrepancy was identified. DISCUSSION Laryngeal nerve injury is 1 of the most feared complications of thyroid surgery. Certainly, the thorough knowledge of laryngeal nerve anatomy and its variations, as well as of the danger points concerning a thyroidectomy, such as Berry s ligament, is of paramount importance to decrease the risk of injury. 18 In recent years, some authors have advocated the use of some form of nerve monitoring during thyroidectomy to increase even further the safety level. In 1984, Woltering et al 19 proposed a method for recurrent nerve stimulation with mean amperage of 1.3 ma, monitored by a double-cuffed endotracheal tube. In their series of 12 patients, the nerve was correctly identified and stimulated in all cases, with normal postoperative vocal fold function checked with indirect laryngoscopy in all cases. Some years later, Mermelstein et al 20 described the use of electrodes attached to the endotracheal tube and in close contact with the vocal folds in a series of 28 patients who underwent thyroid and parathyroid operations. The acceptance of NM has increased lately. Horne et al 6 sent a questionnaire to 1685 surgeons in the United States. A total of 685 (40.7%) of the questionnaires were returned, and 81% (555) of the respondents reported performing thyroidectomy. About 44.9% of the respondents who performed thyroidectomies used NM during the operation, but only 28.6% (159) reported using intraoperative monitoring for all cases. Respondents were 3.14 times more likely to currently use intraoperative monitoring if they used it during their training. Surgeons currently using intraoperative recurrent laryngeal nerve (RLN) monitoring during thyroidectomy were 41% less likely to report a history of permanent RLN injury. NM with the NIM system was popularized by Randolph 4,5 and has become the most frequently used system for nerve monitoring. This system has the advantage of not only identifying the laryngeal nerves but also being able to produce a recordable documentation that serves as a reliable comparison with the baseline results and also offers a legal certificate of the nerve status at the end of the procedure. Table 1. Causes of nerve monitoring false-positive nerves. Cause of false-positive nerves No. of cases Equipment malfunction 5 Equipment operator s fault 6 Undetermined 4 However, in spite of the obvious advantages of being able to get a real-time monitoring of laryngeal nerve function during thyroidectomy, some authors did not consider that the method adds real safety benefit, whereas others have questioned the use of nerve monitoring in all operations. Robertson et al 7 reported a retrospective cohort study of 116 recurrent nerves at risk monitored with NIM system with 120 nonmonitored nerves. Temporary recurrent nerve paresis occurred in 4.24% of the nerves in the control group and 3.45% in the monitored group (p ¼.89), showing no advantage in the study group. Similarly, Witt 8 found no difference in the frequency of recurrent nerve paresis or paralysis comparing a group of 107 unmonitored nerves with 83 monitored nerves. Attalah at al 9 analyzed the use of nerve monitoring in a series of patients undergoing what was considered high-risk thyroidectomy, with nonmonitored nerves used as controls. They found relatively elevated rates of transient and permanent nerve dysfunction in both groups (8.8% of temporary paralysis in the monitored group, in comparison with 9.1% in the unmonitored group; 3.9% of nerves at risk in the monitored group, in comparison with 3.8% in the unmonitored group). The conclusion of this study was that nerve monitoring offered no additional benefit for patients undergoing high-risk thyroidectomy. In contrast, some authors recommend NM in select patients or in specific situations, including revision thyroidectomy, large retrosternal goiters, or significant documented central compartment metastases. Hermann et al 11 published a prospective study involving 502 RLNs. NM was performed with the Neurosign 100 device (Carmarthenshire, United Kingdom) by transligamental placement of the recording electrode into the vocalis muscles. They concluded that NM was for the identification of the RLN, especially in large tumors, aberrant anatomy, or the reoperative setting. However, the method was ineffective for the prediction of postoperative outcome. Chan et al 11 used the same method to evaluate 271 RLNs at risk during thyroidectomy, subdivided in 2 groups: low risk (primary surgery for benign disease) and high risk (malignancy and recurrent disease). The rates of transient and permanent RLN palsy on the basis of nerves at risk were 4.8% (n ¼ 13) and 0.7% (n ¼ 2), respectively. There were 241 true-negative (positive signal and no cord palsy), 15 false-positive (negative signal but no cord palsy), 8 true-positive (negative or reduced signal and cord palsy), and 7 false-negative (positive signal but cord Predictive Value of Nerve Monitoring in Thyroidectomy HEAD & NECK DOI /hed February

4 palsy) results. The sensitivity, specificity, and positive and negative predictive values were 53%, 94%, 35%, and 97%, respectively. For the high-risk group, the sensitivity and positive predictive value increased to 86% and 60%, respectively. They concluded that NM should be used only in patients undergoing high-risk thyroidectomy. On the other hand, some authors believe that NM is indicated in all thyroidectomies, offering precious information to the surgeon, in addition to the gold standard, which is the combination of visual identification and gentle handling of the RLN. In fact, many experienced surgeons believe that some high-risk situations, such as nonrecurrent inferior laryngeal nerves, may be faced unexpectedly. After an extensive systematic review of the literature, Dralle et al 15 made the following statement: Apart from navigating the surgeon through challenging anatomies, NM may lend itself as a routine adjunct to the gold standard of visual nerve identification. Barczyński et al 17 conducted a prospective randomized trial evaluating 1000 RLN at risk during thyroidectomy divided in 2 equal groups: 1 undergoing NM with the Neurosign 100 system, and 1 control group, with only visual identification. In the first group, the prevalence of RLN injury, transient RLN paresis, and permanent RLN palsy was, respectively, 2.3% (p ¼.007), 1.9% (p ¼.011), and 0.4% (p ¼.368) lower than with visualization alone. A drawback of this study was the absence of any kind of recording, because the NM was based only on the acoustic signal. One of the weaknesses of our study is the absence of vagal stimulation. In fact, according to Thormusch et al, 10 prevagal and postvagal stimulation increase the accuracy on nerve monitoring in thyroid surgery. The accuracy of indirect stimulation by vagal simulation is higher than by only direct RLN stimulation. However, because of the very small incision that we have been using, we considered that exposure of the vagus nerve was rather difficult. Despite the controversy, many previous studies have suggested that perhaps the main role of the NM was to demonstrate the functional status of the RLN at the end of the operation. We tried to evaluate the PPV and NPV in our own series of consecutive patients, with 868 RLNs at risk, and to compare our results with those available in the literature. In the present series, the PPV was 40%, and the NPV was 100%, and these results were comparable with the published series using nerve monitoring, as depicted in Table 2. Thus in our experience, as well as in other previous publications, the recording of a normally functioning RLN at the end of a thyroid lobectomy was a reliable predictor of its normal function. Moreover, in a recent publication including 1333 patients, Goretzky et al 21 stated that failed nerve stimulation is specific enough to reevaluate the surgical strategy concerning the contralateral lobectomy, eventually considering delay of the surgical approach. Table 2. Positive predictive value and negative predictive value of nerve monitoring in the literature. Author No. of nerves PPV (%) NPV (%) Thomusch et al 10 15, Chan and Lo Beldi et al Barczyński et al Present series Abbreviations: PPV, positive predictive value; NPV, negative predictive value. Nevertheless, it is important to keep in mind that the use of NM does not replace the careful and gentle handling of the RLN during a thyroidectomy, both from the legal, as well as from the ethical point of view. Angelos 22 pointed out very clearly: Whether RLN monitoring is used or not during thyroidectomy, the determining factor that brings a lawsuit is whether the nerve is injured. Because the literature shows no difference in rates of RLN injury whether neuromonitoring is used or not, the most prudent approach for the surgeon is to use the techniques that have been shown to minimize RLN injury, that is, careful, meticulous dissection with the attempt to visualize the RLN in every case possible. We also tried to examine the 15 false-positive nerves in this series. Equipment malfunction was the cause in 5 cases, precluding the effectiveness of the nerve stimulation of the recording of the nerve function, and all these cases occurred in the beginning of our experience. According to Dralle et al, 15 several technical pitfalls may occur, including inadequate use of paralyzing agents, that may impair the correct interpretation of the NM. In 6 cases, again during the initial period of our experience, inadequate operation of the monitoring system by inexperienced technicians, who did not operate the system properly (4 cases) or because of inadvertent endotracheal tube mobilization (2 cases) was considered the cause of the false-positive nerves. In 4 cases, no reasonable explanation was found. It is possible that, in these instances, a very transient injury of the RLN could have occurred. However, another possible reason could be the medial traction of the trachea. After the completion of the lobectomy, the RLN returns to its original anatomic situation, on the tracheoesophageal groove. Sometimes, to adequately expose it, it may be necessary to dislocate the trachea medially, impairing the proper motion of the correspondent vocal fold and, consequently, reducing the amplitude of the electromiography record. This situation occurred in some cases in the present series, and when no pressure was exerted on the trachea, a normal electromiography result was obtained. As a matter of fact, no unexplained false-positive nerve was found during the last 2 years of this study, when the final electrophysiological check of the RLN was performed with no tracheal traction. Therefore, we recommend that the final 178 Predictive Value of Nerve Monitoring in Thyroidectomy HEAD & NECK DOI /hed February 2012

5 record of the RLN after the lobectomy should be obtained with no medial traction of the trachea. We try to obtain the final recording with a 0.5-mA intensity of the stimulus, which is highly predictable of normal inferior laryngeal nerve function, according to the important publication by Donellan et al. 23 In conclusion, in this study comprising a large series of 868 RLNs undergoing NM during thyroidectomy, PPV and NPV were, respectively, 40.0% and 100.0%, suggesting that this method is highly effective for the correct prediction of a normally functioning nerve at the end of the operation. In addition, it is recommended not to dislocate the trachea medially when obtaining the final record to avoid false-positive results because of vocal fold compression. REFERENCES 1. Cernea CR, Brandão LG, Hojaij FC, et al. How to minimize complications in thyroid surgery? Auris Nasus Larynx 2010;37: Lamade W, Fogel W, Rieke K, Senninger N, Herfarth C. Intraoperative monitoring of the recurrent laryngeal nerve. A new method. Chirurg 1996;67: Lamadé W, Meyding-Lamadé U, Hund E, Senninger N, Herfarth C. Transtracheal monitoring of the recurrent laryngeal nerve. Prototype of a new tube. Chirurg 1997;68: Randolph GW. Surgical anatomy of the recurrent laryngeal nerve. In: Randolph GW (editor). Surgery of thyroid and parathyroid glands. Philadelphia: Elsevier, p Randolph GW, Kobler JB, Wilkins J. Recurrent laryngeal nerve identification and assessment during thyroid surgery: laryngeal palpation. World J Surg 2004;28: Horne SK, Gal TJ, Brennan JA. Prevalence and patterns of intraoperative nerve monitoring for thyroidectomy. Otolaryngol Head Neck Surg 2007;136: Robertson ML, Steward DL, Gluckman JL, Welge. Continuous laryngeal nerve integrity monitoring during thyroidectomy: does it reduce risk of injury? Otolaryngol Head Neck Surg 2004;131: Witt RL. Recurrent laryngeal nerve electrophysiologic monitoring in thyroid surgery: the standard of care? J Voice 2005;19: Atallah I, Dupret A, Carpentier AS, Weingertner AS, Volkmar PP, Rodier JF. Role of intraoperative neuromonitoring of the recurrent laryngeal nerve in high-risk thyroid surgery. J Otolaryngol Head Neck Surg 2009; 38: Thomusch O, Sekulla C, Machens A, Neumann HJ, Timmermann W, Dralle H. Validity of intra-operative neuromonitoring signals in thyroid surgery. Langenbecks Arch Surg 2004;389: Hermann M, Hellebart C, Freissmuth M. Neuromonitoring in thyroid surgery: prospective evaluation of intraoperative electrophysiological responses for the prediction of recurrent laryngeal nerve injury. Ann Surg 2004;240: Chan WF, Lo CY. Pitfalls of intraoperative neuromonitoring for predicting postoperative recurrent laryngeal nerve function during thyroidectomy. World J Surg 2006;30: Beldi G, Kinsbergen T, Schlumpf R. Evaluation of intraoperative recurrent nerve monitoring in thyroid surgery. World J Surg 2004;28: Tomoda C, Hirokawa Y, Uruno T, et al. Sensitivity and specificity of intraoperative recurrent laryngeal nerve stimulation test for predicting vocal cord palsy after thyroid surgery. World J Surg 2006;30: Dralle H, Sekulla C, Lorenz K, Brauckhoff M, Machens A; German IONM Study Group. Intraoperative monitoring of the recurrent laryngeal nerve in thyroid surgery. World J Surg 2008;32: Cavicchi O, Caliceti U, Fernandez IJ, et al. The value of neurostimulation and intraoperative nerve monitoring of inferior laryngeal nerve in thyroid surgery. Otolaryngol Head Neck Surg 2009;140: Barczyński M, Konturek A, Cichoń S. Randomized clinical trial of visualization versus neuromonitoring of recurrent laryngeal nerves during thyroidectomy. Br J Surg 2009;96: Cernea CR, Hojaij FC, De Carlucci Jr., et al. Recurrent laryngeal nerve: a plexus rather than a nerve? Arch Otolaryngol Head Neck Surg 2009;135: Woltering EA, Dumond D, Ferrara J, Farrar WB, James AG. A method for intraoperative identification of the recurrent laryngeal nerve. Am J Surg 1984;148: Mermelstein M, Nonweiler R, Rubinstein EH. Intraoperative identification of laryngeal nerves with laryngeal electromyography. Laryngoscope 1996;106: Goretzki PE, Schwarz K, Brinkmann J, Wirowski D, Lammers BJ. The impact of intraoperative neuromonitoring (IONM) on surgical strategy in bilateral thyroid diseases: is it worth the effort? World J Surg 2010; 34: Angelos P. Recurrent laryngeal nerve monitoring: state of the art, ethical and legal issues. Surg Clin N Am 2009;89: Donnellan KA, Pitman KT, Cannon CR, Replogle WH, Simmons JD. Intraoperative laryngeal nerve monitoring during thyroidectomy. Arch Otolaryngol Head Neck Surg 2009;135: Predictive Value of Nerve Monitoring in Thyroidectomy HEAD & NECK DOI /hed February