Complications of Bilateral Neck Dissection in Thyroid Cancer From a Single High-Volume Center

Key Points

Question  What is the morbidity associated with bilateral neck dissection performed for thyroid cancers?

Findings  This is a retrospective review of medical records for patients undergoing bilateral lateral neck dissections for thyroid cancers at an academic medical center. The overall rate of permanent hypoparathyroidism, defined as continued calcium supplementation at the most recent follow-up, was 37%.

Meaning  The morbidities associated with bilateral lateral neck dissections for thyroid cancers should be carefully considered against the oncologic benefit.

Importance  The morbidity of bilateral lateral neck dissection (BLND) for thyroid cancers has not been described in detail. This study delineates the specific complications arising from BLND for thyroid cancers at a single high-volume center.

Objective  To determine the morbidity associated with BLNDs for differentiated thyroid cancers at our institution.

Design, Setting, and Participants  This was a retrospective review of medical records performed to identify patients having undergone BLNDs for thyroid cancers by a single surgeon at an academic, tertiary medical center in Toronto, Ontario, Canada, from 1988 to 2015. Patients who underwent BLND for papillary, follicular, or medullary thyroid cancers were identified through operative procedure codes and review of operative and pathology reports. The indication for this procedure was suspicious bilateral lateral compartment on imaging and clinical examination. Sixty-two patients who underwent BLND for thyroid cancers, with or without total thyroidectomy and central compartment dissection, were identified.

Main Outcomes and Measures  The main outcome measures for this study were unanticipated medical or surgical complications during the operation or in the postoperative period. Secondary measures were oncologic outcomes, including regional structural or biochemical recurrence.

Results  Of the 62 patients, 24 were male (39%), and 38 (61%) were female. Their mean age was 46 years (range, 17-80 years). The overall risk of permanent hypoparathyroidism was 37%. There was 1 case of unanticipated permanent recurrent nerve paralysis and 1 case of temporary nerve paresis. Postoperative chyle fistula occurred in 6 cases (10%). There were 3 readmissions within 30 days of surgery, 1 pulmonary embolism, and 1 perioperative mortality. Fifty percent of patients had pN0 contralateral necks despite preoperative clinical suspicion. Four patients were found to have anaplastic thyroid cancers intraoperatively. Five patients (8%) developed nodal recurrence in the neck. Four patients died of their disease within available follow-up (mean, 3.2 years).

Conclusions and Relevance  Bilateral lateral neck dissection for thyroid cancers confers a significant amount of morbidity, including a significant rate of hypoparathyroidism. Knowledge of the complications of this procedure, especially in the setting of questionable survival benefit, may assist in preoperative decision-making and patient counseling.

Despite the frequency of nodal metastases in differentiated thyroid cancers, the impact on survival and disease recurrence is unclear.^1-7 There is a consensus that elective or prophylactic lateral neck dissection for papillary thyroid cancer (PTC) or follicular thyroid cancer (FTC) is not necessary.⁸ For patients with ipsilateral nodal disease, bilateral lateral compartment neck dissection (BLND) is often performed for high clinical suspicion in the contralateral aspect of the neck owing to imaging findings, elevated serum tumor markers, and/or physical examination findings. Lateral neck dissections may be indicated for patients with medullary thyroid cancer (MTC) with suspicious imaging and calcitonin levels greater than 200 pg/mL owing to the high risk of nodal metastases.⁹ Despite extensive surgery, including bilateral lateral compartments, patients with MTC frequently have persistently elevated calcitonin levels postoperatively.¹⁰

Several studies have looked at the morbidity of lateral neck dissections in thyroid cancers and yielded disparate results. Some have found differences in complication rates with the addition of a lateral neck dissection, while others have found no significant difference.^10,11 A recent large systematic review¹¹ reported conflicting results of oncologic outcomes and complications of lateral neck dissections in differentiated thyroid cancers. To our knowledge, no study has reported a comprehensive and detailed description of surgical complications for a large series of BLNDs for thyroid cancers. Knowledge of the morbidity of this procedure may inform clinical decision-making and patient counseling preoperatively.

This retrospective review was approved by the Mount Sinai Hospital research ethics board. At the time of treatment, extensive discussion was conducted with the patient describing the risks and benefits of surgery. Given the retrospective nature of this review, informed consent for this study was waived. Patients were identified through medical records using operative procedure codes, specifically searching for patients who underwent operative lymph node dissections at Mount Sinai Hospital in Toronto, Ontario, Canada, with a diagnosis of thyroid cancer between 1988 and 2015. Patients with preoperative PTC, FTC, and MTC were included. A total of 898 medical records were identified. Of these, 729 medical records were available for review. Cases were screened for those performed by a single head and neck endocrine surgeon (J.L.F.). The surgeon’s dictated operative notes and pathology reports were used to identify patients who underwent BLND for thyroid cancers. Indications for surgery, complications, adjuvant therapy, and oncological outcomes were identified from the patient’s clinical records.

The indication for BLND was clinically positive nodal disease identified by physical examination and/or imaging findings. All patients underwent computed tomographic imaging of the neck preoperatively. Fine-needle aspiration was performed of the primary thyroid tumor if evident and/or the dominant nodal metastasis. For patients with equivocal nodal disease, lateral neck dissections proceeded according to patient preference after a thorough discussion of risks and benefits. A comprehensive discussion was had with the patient, disclosing the 50% risk of pN0 neck in contrast to the potential need for additional surgery in the future given the suspicious imaging findings. The surgical technique was performed as follows. A bipolar “pinch-burn” technique was used for dissection and/or hemostasis. Bilateral lateral compartment dissections were performed first, followed by total thyroidectomy (TT) (if a primary case) and finally central compartment dissection (CND). If there was a suspicion of inoperable or anaplastic disease, the thyroid would be addressed first. This population of patients was not included in the study. Lateral compartment dissections typically included levels IIa, IIb, III, IV, and Vb. Level IIb was performed when adjacent levels were grossly positive. Comprehensive bilateral dissection of the central compartment was performed between the common carotid arteries and inferiorly to the innominate artery, removing all fibrofatty tissue except parathyroids preserved on a vascular pedicle. Parathyroid glands were reimplanted into strap muscle or sternocleidomastoid and marked with hemoclips if devascularization was identified. At this academic institution, residents and fellows contribute to the surgery and care of the patient under attending physician’s supervision.

Operative times were collected from the nursing and anesthesia records and defined from “incision time” to “close time.” Hypoparathyroidism was considered permanent if the patient still required calcium and calcitriol supplementation at the most recent follow-up. Patients who were able to discontinue their supplements during the follow-up period were considered cases of temporary hypoparathyroidism. Vocal fold mobility was assessed with flexible fiberoptic laryngoscopy preoperative and postoperatively. Injury was considered temporary if the patient regained mobility within 6 months of surgery. Wound infections were defined as local erythema and tenderness requiring antibiotics. Biochemical recurrence was defined as elevated thyroglobulin or calcitonin without obvious structural disease.

Sixty-two patients who underwent BLND for thyroid cancer were identified. Patient and tumor data are reported in Table 1. Forty-eight patients underwent BLND, TT, and CND; 1 of these patients also required sternotomy for surgical access performed by thoracic surgery. The mean operative times for BLND-TT-CND, BLND-CND, and BLND alone were 303 minutes, 269 minutes, and 291 minutes, respectively. Twelve cases were for recurrence and included BLND and CND. One of these patients, who had undergone a previous TT, also underwent a Sistrunk procedure simultaneously for thyroglossal duct source of PTC. Two patients underwent only BLND because they had previously undergone CND and had no clinically evident disease in the central compartment. Four patients who were initially believed to have PTC were discovered intraoperatively to have anaplastic thyroid cancer. Limited debulking of the thyroid was performed essentially for tissue diagnosis. One patient with PTC metastases was found to have an incidental subcentimeter MTC.

Some pathology reports, particularly prior to the implementation of synoptic reporting from early in the search period, did not specify the exact primary tumor size, extrathyroidal extension, or nodal counts. Of 59 patients with available node counts, the average nodal yield was 77 lymph nodes. The average number of positive nodes was 15 (range, 0-95 nodes). Interestingly, 47% of patients did not have pathologic disease in the contralateral lateral aspect of the neck despite preoperative clinical suspicion. There were 9 patients with no nodal metastases whatsoever found on final pathologic results. Five of these patients had MTC, and 4 had PTC. The indications for surgery in these cases was suspicious imaging findings with calcitonin levels greater than 200 pg/mL in MTC, or elevated thyroglobulin level with suspicious imaging in recurrent PTC cases. One patient with PTC had a biopsy-proven lateral node that was not identified in the pathology specimen. This patient had clear postoperative imaging, did not receive postoperative radioactive iodine, and had no recurrences in 4 years of follow-up. Another patient had preoperatively known distant metastases in the bone and lung but interestingly did not have any nodes identified on final pathologic results. The indications for postoperative external beam radiation included gross invasion of major structures, including the pharynx and/or esophagus and trachea, or palliation for anaplastic pathology. The patients with unknown adjuvant therapy were lost to follow-up.

Surgical complications, including those specifically within the population with anaplastic cancers, are described in Table 2. Intraoperative chyle leaks were repaired immediately; none of these patients had postoperative chyle leaks. All postoperative chyle leaks resolved with conservative measures and did not require reoperation. Patients are tested intraoperatively for chyle leak using Valsalva and/or abdominal compression techniques.

Other medical complications were as follows. One perioperative mortality occurred. This patient had MTC and died suddenly at a group home 11 days postoperatively, and an autopsy was not performed. One patient with anaplastic disease developed unilateral acute vision loss 10 days postoperatively owing to tumor embolus. Postoperative urinary retention, an anesthetic complication that can prolong hospital stay, occurred in 2 patients.

Three patients remained intubated overnight at the discretion of anesthesia requiring intensive care unit (ICU) admission. One additional patient, who was found to have anaplastic disease, required ICU admission for bilateral cord weakness, known preoperatively, and subsequently required a tracheotomy during the hospital admission. Two patients had pharyngeal and/or esophageal invasion requiring local resection and temporary nasogastric tube feeding.

The observed neurovascular complications are as follows. There were no cases of deep vein thrombosis, but there was 1 case of postoperative pulmonary embolism. Three patients had postoperative spinal accessory dysfunction. There were 2 cases of unexpected recurrent laryngeal nerve (RLN) paresis, 1 temporary and 1 permanent. Cases in which the nerve was nonfunctioning preoperatively or sacrificed owing to tumor involvement were not considered complications. However, for thoroughness, there were 9 patients with direct tumor involvement of the RLN requiring sacrifice; 4 of these had preoperative paresis. Owing to direct passage through tumor, 1 patient required vagus nerve sacrifice, and 1 patient required spinal accessory sacrifice.

The rates of hypoparathyroidism for each specific procedure type are described in Table 3. Permanent hypoparathyroidism occurred in 19 cases (37%). Because this was defined as patients still taking calcium and calcitriol supplements at the time of last follow-up, this value included 4 patients with follow-up of less than 6 months. Neither patient who underwent BLND alone developed hypoparathyroidism. Twelve patients (19%) were able to discontinue supplementation during their follow-up periods. Interestingly, 2 patients receiving standard postoperative supplementation required readmission for hypercalcemia. One of these patients was prescribed a thiazide diuretic. An additional patient was readmitted for subcutaneous air, without obvious source on imaging, which resolved with conservative management.

To our knowledge, this study represents the first detailed review of complications for a large population of patients undergoing BLND for thyroid cancers. The few series of BLND for thyroid cancers that exist report oncologic outcomes and do not consistently report surgical complications.^10-13 Most reviews of lateral neck dissections for thyroid cancers include only a handful of BLNDs.^6,14-16 Ito et al¹⁷ reported a number of elective lateral neck dissections, which resulted in increased complications, without a difference in oncologic outcomes. Noguchi et al¹⁸ described 98 patients with BLND and reported 4 cases of RLN palsy and 1 case of Horner syndrome. The same group reported 135 patients who underwent BLND for thyroid cancer and reported 1 case of Horner symptoms and 5 RLN palsies.¹³ Oshima et al¹⁹ described outcomes for 86 patients who underwent BLND with TT and reported only 1 complication, a transient chyle leak. The reporting of complications for BLNDs performed for thyroid cancers is frequently incomplete and secondary to the reporting of oncologic outcomes. This study details complications from a large group of BLND performed for thyroid cancers at a high-volume center.

One of the more striking findings within this population is the rate of permanent postoperative hypoparathyroidism. The largest reviews of BLNDs for papillary and follicular thyroid cancers do not report the incidence of postoperative hypoparathyroidism in their patients.^12,13,19 A number of authors have reported an increase in postoperative temporary and/or permanent hypoparathyroidism with the addition of central and lateral neck dissection.^6,14,20,21 Roh et al²² reported a temporary and permanent rate of postoperative hypocalcemia as 20% and 4.9%, respectively, with nodal dissection of any type. One prospective review reported a transient hypoparathyroidism rate of 80% with the addition of BLND.²⁰ Scollo et al¹⁰ reported a permanent hypoparathyroidism rate of only 4% in 101 patients with MTC undergoing BLND. Scollo et al¹⁰ identified all parathyroid glands in each case, marking them with methylene blue and then autotransplanting any parathyroid gland tissue confirmed by frozen section.

One issue with the disparate results of postoperative hypoparathyroidism is its variable definition. Some authors define permanent as more than 6 months, some report it as more than 1 year; some authors use biochemical markers, and others use the clinical indicator of oral supplementations.^21,22 In the present series, the rate of postoperative permanent hypoparathyroidism may be overestimated because a number of patients had limited follow-up and could have been maintained on calcium and vitamin D supplementation for other reasons. It is the practice at our institution for patients with thyroid cancers to be referred to an endocrinologist postoperatively who takes over management of hypoparathyroidism.

Surgical mechanisms of the high rate of hypoparathyroidism with bilateral neck dissection could be disruption of the inferior thyroid arteries at their thyrocervical trunk origin during comprehensive lateral and central neck dissection. Thorough central compartment dissection consisting of complete removal of fibrofatty and/or nodal tissue between the carotids to the level of innominate is performed routinely. Unfortunately, this also can contribute to parathyroid devascularization. The population in this series also had a high burden of disease: multiple patients with direct esophageal and neural invasion, high average nodal yield, and a significant proportion of patients requiring postoperative external beam radiation. Three of 4 patients with anaplastic disease developed hypoparathyroidism. It is possible that some of the glands were previously devascularized by the disease process, and the lateral neck dissections and limited central compartment work destabilized the remaining glands. The average nodal yield and proportion of positive nodes were greater in this series than other reports.¹²

While concern for postoperative hypoparathyroidism and hypocalcemia is of great concern in this population, there is a risk of oversupplementation. Patients with temporary hypoparathyroidism may recover function prior to the discontinuation of supplements. There were 2 readmissions for iatrogenic hypercalcemia in this series. Physicians should be wary of the effect of thiazide diuretics in patients requiring postoperative supplementation. Other than anecdotal reports, there is no literature detailing rates of iatrogenic hypercalcemia and contributing factors following total thyroidectomy.

The rate of spinal accessory injury in this report is significantly lower than rates in prior studies of lateral neck dissections.¹⁴ This finding illuminates one of the flaws of this retrospective review for complications. Without specific evaluation and subsequent documentation, spinal accessory palsy and recovery may be underestimated. The incidence of chyle leaks in this report is similar to the incidence in other studies of lateral neck dissections in thyroid cancers.^14,22-24 Our rate of unintended RLN paresis is similar to those of prior reports. There was only 1 case of temporary nerve paresis, and 1 case of unexpected permanent paresis. As shown in Table 1, patients with anaplastic cancers contributed significantly to the number of complications. The disease in these cases can be infiltrative into major structures, resulting in more technically challenging surgery.

The other medical and surgical complications listed in Table 2 highlight the additional potential morbidity associated with this procedure. Infections, such as pneumonia and local cellulitis, require treatment with antibiotics, close monitoring, and often prolong hospital stay. Urinary retention, while not specific to this procedure, is a distressing concern to patients and requires invasive instrumentation with Foley catheter replacement and creates a source of potential infection.

While not the primary objective of this study, it is important to address the pathologic and oncologic findings. The indications for BLND are high clinical suspicion of nodal metastases from imaging, physical examination, and/or laboratory findings, including biochemical and cytologic results. Despite high suspicion preoperatively of bilateral nodal metastases, nearly half of patients did not have contralateral nodes on final pathology results. This is consistent with results in prior reports.¹⁸ Authors have suggested that patients with certain risk factors, such as male sex, large primary tumors, cancer in both lobes, isthmus primary, and/or gross extrathyroidal extension may undergo elective BLND because they are at high risk of contralateral metastases in the future.^17,18,25 Ohshima et al²⁵ reported that of patients undergoing ipsilateral neck dissection for PTC, 1.8% later developed contralateral nodal disease over an average follow-up of 12 years. While this is a low rate, this group was more likely to develop distant metastases (P < .001) and had a lower 10-year survival rate (83.7% vs 99.3%; P < .001) than patients without nodal recurrence. The same authors¹⁹ also reported that upfront BND for clinical positive bilateral nodal disease, rather than salvage BND for regional failure, improves distant control and survival. Most reports, however, contradict the notion that lateral neck dissection improves survival outcomes. According to a recent large systematic review,¹¹ the development of lateral compartment regional disease rarely, if ever, affects life expectancy, and the impact of prophylactic lateral neck dissections is questionable. At our institution, if preoperative imaging is equivocal on the contralateral side, a discussion is held with the patient regarding the risks and benefits of the addition of a contralateral neck dissection. Often, patients choose to undergo BLND rather than expectant management of the contralateral side.

Interestingly, many of the patients in this series had lateral nodal metastases with microcarcinomas. Other authors have found that tumor size does not affect the development of nodal disease; rather, the development of regional disease reflects the biologic nature of that tumor.^22,26

This study has a number of limitations. Primarily, the study is limited by its format as a retrospective review of available medical records. The endocrine pathologic characteristics analyzed are not uniform, as MTC, FTC, PTC, and anaplastic tumors are all included. A number of patients had limited follow-up (<6 months) limiting complete assessment of certain complications including hypoparathyroidism.

This series details the morbidity associated with therapeutic BLNDs for thyroid cancers. Despite comprehensive surgical management, a small percentage of patients recurred in the neck. The impact of these recurrences, however, is questionable, and management can usually be untaken without significant morbidity. The findings may inform preoperative counseling of patients and promote careful consideration of the risk-benefit ratio when performing BLND for thyroid cancers, and perhaps a less aggressive approach to surgical management may be appropriate.

The risk of hypoparathyroidism and other complications are detailed in this series of BLNDs for thyroid cancers from a single high-volume institution. Despite high preoperative clinical suspicion, in particular on imaging, contralateral sides of the neck often are pathologically node-negative. Furthermore, comprehensive neck dissections for thyroid cancers results in notable morbidity, while the survival impact of this procedure is debated. Knowledge of the risk-benefit ratio of the procedure may inform preoperative patient counseling and decision-making.

Corresponding Author: Caitlin McMullen, MD, Department of Otolaryngology–Head & Neck Surgery, Mount Sinai Hospital, University of Toronto, 600 University Ave, Room 401, Toronto ON M5G 1X5, Canada (caitlin.mcmullen@uhn.ca).

Accepted for Publication: October 8, 2016.

Published Online: January 12, 2017. doi:10.1001/jamaoto.2016.3670

Author Contributions: Dr McMullen had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Rocke, Freeman.

Acquisition, analysis, or interpretation of data: McMullen, Rocke.

Drafting of the manuscript: McMullen.

Critical revision of the manuscript for important intellectual content: Rocke, Freeman.

Obtained funding: Freeman.

Administrative, technical, or material support: Rocke, Freeman.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and none were reported.

Previous Presentation: This study was presented at the American Head & Neck Society Ninth International Conference on Head and Neck Cancer; July 19, 2016; Seattle, Washington.