Complications of Central Neck Dissection in Patients with Papillary Thyroid Carcinoma: Results of a Study on 1087 Patients and Review of the Literature

Introduction

T he tendency of papillary thyroid carcinoma (PTC) to spread to central and lateral neck lymph nodes is well known. Therapeutic central and lateral neck dissections are performed with total thyroidectomy in patients with suspected or clinically overt lymph node metastases.

Nevertheless, the incidence of neck lymph node micrometastasis is high, with an estimated range from 38% to 61% of cases (1 –5).

Neck ultrasonography, whose accuracy is increased by fine-needle aspiration cytology and fine-needle aspiration thyroglobulin measurements, reliably evaluates the lateral compartment of the neck before thyroid surgery. This procedure can detect nonpalpable lymph node metastases (6,7), thereby helping to plan therapeutic lateral dissections with total thyroidectomy. However, no completely reliable tool currently exists to preoperatively detect lymph nodes metastases in the central compartment of the neck. For instance, the accuracy of neck ultrasonography is limited by the fact that paratracheal lymph nodes are minute and are located beneath the thyroid gland. Moreover, the air-filled trachea represents an additional element of disturbance (8,9). Instead, neck ultrasonography in a previously dissected compartment shows great accuracy in detecting paratracheal recurrence or persistence of the disease (8). For this reason, some authors recommend that prophylactic central neck dissection (CND) can be routinely performed with total thyroidectomy for the treatment of patients affected by PTC with clinically negative neck lymph nodes (2,4,10,11). However, this routine use of prophylactic CND is the subject of an ongoing debate regarding this treatment approach. On the one hand, the procedure allows accurate pathologic staging of lymph nodes of the central compartment and, at the same time, treatment of the micrometastases that may be responsible for the recurrence or the persistence of the disease (12). Although some authors report that the lymph node metastasis from PTC is related to an increased risk of locoregional recurrence, without any significant adverse effect on long-term prognosis (13,14), recent studies suggest a possible role of lymph node metastasis from PTC in reducing survival (15).

On the other hand, the morbidity that the routine use of prophylactic CND adds to the total thyroidectomy must be taken into account; a number of authors have reported an increased rate of injury to the recurrent laryngeal nerve (RLN) and hypoparathyroidism (1,16).

The aims of this study were to characterize the morbidity that CND adds to total thyroidectomy, to review current practices on this subject reported in the English language literature, and to discuss our treatment protocol in patients affected by PTC with clinically negative neck lymph nodes.

Methods

This retrospective controlled observational study was approved by the institutional review committee of the Arcispedale Santa Maria Nuova (ASMN), Center for Clinical and Basic Research (IRCCS), Reggio Emilia, Italy.

All patients seen at the Otolaryngology Unit of the Azienda Ospedaliera ASMN, Istituto di Ricovero e Cura a Carattere Scientifico, Reggio Emilia, Italy, with a histopathological diagnosis of PTC between 1980 and 1996 were treated with total thyroidectomy. Concomitant therapeutic CND was performed only in patients with clinically overt neck lymph node metastasis in the central or lateral compartment. Since 1997, patients cytopathologically diagnosed with PTC with clinically negative neck lymph nodes have undergone total thyroidectomy with concomitant prophylactic ipsilateral CND. Concomitant completion contralateral paratracheal lymph node neck dissection is performed at our institution when lymph node metastases are present on intraoperative frozen-section pathology. By definition, prophylactic CND consists of removal of lymph node basins of the central compartment, and it implies that lymph node metastases had not been detected, either clinically or by imaging. Unilateral CND consists of removal of prelaryngeal, pretracheal, and one paratracheal nodal basin, whereas bilateral CND consists of removal of prelaryngeal, pretracheal, and both the right and the left paratracheal nodal basins (17).

In the present study, we retrospectively reviewed the clinical records of patients with histopathological diagnosis of PTC and clinically negative neck lymph nodes surgically treated at the Otolaryngology Unit of the ASMN-IRCCS between January 1, 1997, and July 31, 2010. Patients were excluded if they had undergone (i) previous thyroid or parathyroid surgery; (ii) previous neck surgery; (iii) previous neck irradiation; (iv) concomitant surgery for hyperparathyroidism; (v) surgery for locoregional recurrence; and (vi) completion thyroidectomy. The study included 1087 patients, who were grouped on the basis of the treatment received.

Group A was the control group, containing 394 patients (36.2%) who had total thyroidectomy alone. This group included all patients surgically treated for benign thyroid disease with an incidental diagnosis of PTC on final histopathology, patients with cytopathological diagnosis of follicular lesion and diagnosis of PTC follicular variant on final histopathology, and patients in whom concomitant prophylactic ipsilateral CND was not performed because anatomosurgical conditions put the RLN at a great risk of injury.

Group B consisted of 385 patients (35.4%) who had total thyroidectomy and concomitant prophylactic ipsilateral CND. This group included all patients with PTC without evidence of ipsilateral pretracheal and paratracheal lymph node metastasis on intraoperative frozen-section pathology.

Group C consisted of 308 patients (28.4%) who had total thyroidectomy and concomitant prophylactic bilateral CND. This group included all patients with PTC and evidence of lymph node metastasis on intraoperative frozen-section pathology and patients with cytopathological or intraoperative diagnosis of PTC in both lobes or in the isthmus of the thyroid gland.

Vocal cord motility was assessed preoperatively and postoperatively by means of indirect laryngoscopy and flexible fiber optic laryngoscopy. RLN injury was defined as a postoperative impairment of the motility of one or both vocal cords, compared to the preoperative finding, ranging from mild reduction of motility to a picture of a fixed vocal cord. The cases of RLN injury after an intentional section of a nerve invaded by gross tumor were not considered in the analysis of data. RLN injury was considered transient in patients who regained normal vocal cord motility within 6 months after surgery; otherwise, it was considered permanent.

Albumin-adjusted total serum calcium level was measured preoperatively and once daily on postoperative days 1 to 3, and then at follow-up visits. Transient hypoparathyroidism was defined as postoperative hypocalcemia with an albumin-adjusted total serum calcium level lower than 2.0 mM (8.0 mg/dL; normal range, 8.5–10.2 mg/dL). Permanent hypoparathyroidism was defined as persistent hypocalcemia 6 months after surgery requiring calcium and vitamin D supplements.

Transient or permanent RLN injury and transient or permanent hypoparathyroidism were taken as primary endpoints for the statistical analysis and were assessed at follow-up visits. Based on these endpoints, a follow-up of at least 9 months was considered appropriate. Primary endpoints of the study were evaluated by comparing the rates of various complications in the three study groups. Statistical analysis of the differences in clinical variables between groups was conducted by the Chi-square test. Statistically significant levels are presented as p-values, with observed differences considered statistically significant when p≤0.05. The categorical variables considered in the univariate analysis were transient RLN injury (yes vs. no), permanent RLN injury (yes vs. no), transient hypoparathyroidism (yes vs. no), and permanent hypoparathyroidism (yes vs. no). All results are presented as odds ratios (ORs) with their associated 95% confidence intervals (95% CIs). SPSS software (version 18.0; SPSS, Inc., Chicago, IL) was used for data analysis.

The pertinent articles we reviewed were retrieved from PubMed or as cited in relevant articles. Literature searches included the guidelines provided by the most authoritative national and international associations in the field of thyroid disease and oncology. Guest Editorials and Letters to the Editor were also considered. Only articles published in English were reviewed.

Results

Transient RLN injury was found in 14 out of 394 patients (3.6%) in Group A, in 15 out of 385 patients (3.9%) in Group B, and in 17 out of 308 patients (5.5%) in Group C (Table 1). Permanent RLN injury was found in 4 patients (1%) in Group A, in 2 patients (0.5%) in Group B, and in 7 patients (2.3%) in Group C (Table 2). There were no cases of bilateral RLN injury. Analysis of data showed no significant differences in the rates of transient (p=0.404) and permanent (p=0.099) RLN injury between the three study groups.

Transient hypoparathyroidism was documented in 109 patients (27.7%) in Group A, in 139 patients (36.1%) in Group B, and in 160 patients (51.9%) in Group C. Statistical analysis showed a significantly higher rate of transient hypoparathyroidism among the patients in Group B (Table 3) and in Group C (Table 4) than in patients in Group A (A vs. B: p=0.014; OR: 1.477; 95% CI: 1.091–2.001; A vs. C: p<0.001; OR: 2.827; 95% CI: 2.065–3.870).

Permanent hypoparathyroidism was documented in 25 patients (6.3%) in Group A, in 27 patients (7%) in Group B, and in 50 patients (16.2%) in Group C. Statistical analysis showed no difference (p=0.818) in the rate of permanent hypoparathyroidism between the patients in Group A and Group B (Table 5). Instead, a significantly higher rate of permanent hypoparathyroidism was documented in the patients in Group C (Table 6) than in the patients in Group A (p<0.001; OR: 2.860; 95% CI: 1.725–4.743).

Discussion

The aims of this study were to characterize the morbidity that routine use of CND adds to total thyroidectomy, to review current practices on this topic in the literature in English, and to discuss our treatment protocol in patients affected by PTC with clinically negative neck lymph nodes. We reported the rates of RLN injury and hypoparathyroidism in a series of 1087 consecutive patients operated on at the Otolaryngology Unit of the ASMN-IRCCS over a 14-year period.

The animated debate on CND is due to the lack of high-quality comparative prospective randomized trials; while the procedure seems to offer diagnostic and therapeutic advantages, some authors have reported a significant increase in the morbidity it adds to total thyroidectomy. Thus, the central focus of the debate is not the therapeutic use of the procedure, which is a mandatory completion of total thyroidectomy in patients with PTC and suspected or clinically overt lymph nodes metastases, but it's the routine use as prophylaxis. Current recommendations in the literature regarding the use of prophylactic CND in patients affected by PTC are therefore discordant (18 –22), as the summary in Table 7 illustrates.

Doherty (16) affirms that there is not enough evidence to support or refute the hypothesis that CND associated with the total thyroidectomy as initial treatment of PTC with clinically negative neck lymph nodes is beneficial; however, in experienced hands, this procedure should not increase the patient risk. This may benefit disease-free survival and possibly overall survival as well, and may limit the extent of adjuvant therapy. Instead, other authors report that lymph node metastases are related to a higher risk of recurrence and to a significant negative impact on survival (15,24,25). Steward (16) reports no evidence of benefit in terms of lower recurrence or mortality rates for thyroidectomy and prophylactic CND compared to thyroidectomy alone in the management of PTC. A number of authors suggest that lymph node metastases do not influence long-term prognosis (13,14,26 –29). It has been reported that 38%–61% of patients are affected by PTC host lymph node micrometastases in the central compartment, yet only 7%–15% of them will develop clinically overt regional metastasis (13,30). In Steward's opinion, these data suggest that it is preferable that experienced surgeons perform therapeutic reoperation for central nodal metastasis on a small number of patients than for less-experienced surgeons to perform prophylactic CND in all patients. Nevertheless, the rates of permanent hypoparathyroidism and RLN injury have been reported to increase significantly during deferred therapeutic CND (31,32). Generally, PTC metastasis first involves lymph nodes in the central compartment and then those in the lateral compartment. However, skip metastases leaping the central compartment have been reported in 11.2% to 19.7% of cases (12,33). This increases the doubt as to whether prophylactic CND can completely eradicate the neoplastic disease.

Most authors agreed with the contention that CND associated with total thyroidectomy was more harmful than total thyroidectomy alone, particularly with regard to hypoparathyroidism rather than RLN injury (1,4,27,29,34 –38). However, other authors have reported that the rate of RLN injury and permanent hypoparathyroidism in patients treated with total thyroidectomy and CND may be overestimated (2,4,11,39,40). In the present study, we evaluated the morbidity of CND by comparing the rates of these two complications in three groups of patients affected by PTC with clinically negative neck lymph nodes and treated with total thyroidectomy, total thyroidectomy with ipsilateral CND, or total thyroidectomy with bilateral CND, respectively. Analysis of the data showed that the rates of transient RLN injury and hypoparathyroidism were 4.2% and 37.5%, respectively. Our data were in line with those in the literature, which report a rate of transient RLN injury ranging from 0% to 7.3% and transient hypoparathyroidism ranging from 14% to 60% (1,2,4,11,19,34). Concerning permanent complications, we found a rate of RLN injury and hypoparathyroidism of 1.2% and 9.4%, respectively. These data were also in line with the literature, which reports a rate of permanent RLN injury ranging from 0% to 5% and of permanent hypoparathyroidism ranging from 4% to 11% (1,3,9).

Interesting findings emerged when we compared the three study groups in our study. Data analysis showed that both unilateral CND and bilateral CND were not associated with an increase in the rate of transient or permanent RLN injury when these procedures were associated with total thyroidectomy. The rate of transient hypoparathyroidism, instead, rose significantly when either unilateral or bilateral CND was associated with the total thyroidectomy. Conversely, the rate of permanent hypoparathyroidism significantly increased when bilateral, but not ipsilateral, CND was associated with the total thyroidectomy. These findings were consistent with those of other authors (1,2,4,9,27,29,34 –38,41).

The increased rate of permanent hypoparathyroidism can be explained by the fact that bilateral CND may increase the risk of traumatic or ischemic damage to parathyroid glands caused by surgical manipulation on both sides (42). Every effort must thus be made to preserve vascular supply to the superior parathyroid gland. However, it is often difficult to do so for the inferior parathyroid gland, as it is more prone to ischemic injury during CND. In cases of ischemia of the inferior parathyroid gland during CND, the latter is removed, sliced into a few 1–2-mm pieces, confirmed on intraoperative frozen-section pathology to avoid autograft of a lymph node metastasis, and autotransplanted into an individual pocket created within the sternal belly of the sternocleidomastoid muscle. Based on our experience, we believe that total thyroidectomy associated with prophylactic CND ipsilateral to the tumor, intraoperative frozen-section pathology, and completion contralateral paratracheal lymph node neck dissection in presence of lymph node metastasis may represent an effective strategy in the treatment of PTC with clinically negative neck lymph nodes. In a recent study (32), Roh stated that subclinical metastases were highly prevalent in the neck lymph nodes of the central compartment ipsilateral to the tumor of patients with PTC>1 cm in diameter. Contralateral central metastases, although uncommon, are associated with ipsilateral central metastases. These findings may guide the need for and extent of prophylactic unilateral or bilateral CND. As reported by others (3,8,43,44), we think that intraoperative inspection of paratracheal lymph nodes cannot reliably identify unapparent lymph node metastases. Therefore, unlike other authors (1,3,8,9,36,38,45,46), we advocate the use of intraoperative frozen-section pathology as a rapid way to achieve a more reliable staging of lymph nodes of the central compartment of the neck and thus to determine the extent of surgery. This would limit the use of bilateral CND, more likely to cause permanent hypoparathyroidism, in patients with intraoperative pathological finding of lymph node metastasis.

The treatment protocol adopted at our institution is dictated by the fact that it is not currently possible to characterize preoperatively the prognostic factors that may be related to the presence of lymph node micrometastasis in patients affected by PTC with clinically negative neck lymph nodes. The rationale of our policy is to achieve acceptably accurate staging of the lymph nodes of the central compartment of the neck, to treat such lymph nodes, and to limit the morbidity that the routine use of prophylactic CND can add to total thyroidectomy. Thus, future studies on this topic will bring substantial new elements to the debate to better stratify patients when deciding who should undergo CND.

The strength of this study lies in the fact that the outcome analysis was conducted on a large series. To the best of our knowledge, this represents the largest series on this topic in the English language literature, with all patients of the study population treated according to the same protocol at a single institution (Otolaryngology Unit of the ASMN-IRCCS). The rates of various complications were evaluated in relation to the extent of surgery on the lymph nodes of the central compartment of the neck, and were compared to those seen in the control group.

The main limitation of the study is represented by the fact that, despite the size of the study population, this is a retrospective observational study. Thus, its level of evidence is weaker than that of high-quality randomized controlled prospective studies.

Conflicting data in the literature make it currently impossible to advocate the routine use of prophylactic CND in the treatment of patients with PTC with clinically negative lymph nodes. The increased rates of transient and permanent hypoparathyroidism in our series, even without a higher rate of RLN injury, suggest that a critical review of indications for the routine use of prophylactic CND in association with total thyroidectomy is warranted. Our data seem to suggest that this procedure should not be performed routinely, and that its use should be evaluated and discussed with the patient, given the morbidity it can add to the total thyroidectomy.