Complications of Thyroid Cancer Surgery in Pediatric Patients at a Tertiary Cancer Center

Abstract

Background

The incidence of complications and risk factors for hypocalcemia after pediatric thyroid cancer surgery have not been clearly defined in the literature, as most reports fail to distinguish between benign and malignant disease. The trend away from total thyroidectomy (TT) to thyroid lobectomy in low-risk disease means there is a need to clearly define the complication profile of malignant disease.

Methods

After IRB approval, a retrospective chart review was undertaken at Memorial Sloan Kettering Cancer Center for pediatric patients undergoing surgery for well-differentiated thyroid cancer from 1986-2021. Clinicopathological characteristics and complications were evaluated. Multivariable analysis was carried out to identify factors independently associated with post-operative hypocalcemia.

Results

307 pediatric patients with well-differentiated thyroid carcinoma were identified (median follow-up of 61 months). 69% underwent TT and 31% partial thyroidectomy. 40% had N0 disease, 28% N1a and 33% N1b. Post-operatively, no patients developed a neck hematoma. 1.6% had temporary unilateral vocal cord palsy (VCP) and 0.7% had permanent VCP due to recurrent laryngeal nerve (RLN) invasion. Temporary and permanent hypocalcemia occurred in 32.6% and 5.2%. After multivariable analysis, central neck dissection (CND) (OR 3.30; p<0.001) and N1 disease (OR 2.51; p=0.036) were identified as independent risk factors for temporary hypocalcemia and N-stage (OR 3.64; p=0.018) for permanent hypocalcemia.

Conclusion

Pediatric thyroid cancer surgery results in low complication rates, despite nodal metastases. Vocal cord paralysis is rare unless disease is found to be invading the RLN intraoperatively. N-stage and CND are independent risk factors for hypocalcemia, helping to identify high risk patients.

Introduction

Well-differentiated pediatric thyroid carcinoma is increasing in incidence globally and its management continues to evolve.^1-2 The 2015 American Thyroid Association (ATA) guidelines recommended total thyroidectomy (TT) for the majority of children with papillary thyroid carcinoma (PTC), the most common malignancy.³ However, TT has been shown to increase the risk of major morbidity and is a risk factor for postoperative hypocalcemia.^4-5 The significant increase in thyroid lobectomy (TL) utilization for well-differentiated pediatric thyroid cancer in recent years is likely influenced by the overall survival in pediatric thyroid cancer, coupled with presumed increased risk of complication in TT compared to lobectomy (TL).^6-7 However, there is limited data on morbidity occurring exclusively in oncologic pediatric thyroid surgery. Previous studies on complication rates include a mix of benign and malignant disease, with malignancy making up less than half of diagnoses.^8-14 Only one study has described complication rates in malignancy alone, reporting on patients exposed to radioiodine fallout after the Chernobyl accident in Belarus. All patients underwent total thyroidectomy, central and lateral neck dissection.¹⁵ The etiology of disease and extent of surgery in this study means complication rates are not representative of sporadic disease outside of nuclear fallout zones, which requires less extensive surgery.

Permanent hypocalcemia after pediatric thyroid surgery can lead to significant life-long morbidity, including nephrocalcinosis, decreased estimated glomelular filtration rate, basal ganglia calcifications and bone demineralisation.^16-17 Previous studies have had incongruent results when assessing the impact of malignancy on risk of hypocalcemia. One showed malignancy to be an independent risk factor for temporary hypocalcemia, but another did not identify any independent risk factors for permanent hypocalcemia after thyroid surgery.^5,18 It is therefore important to further investigate disease and treatment characteristics in malignant disease that increase risk of long-term morbidity. Other complications with deleterious effects after thyroid surgery include recurrent laryngeal nerve (RLN) injury (resulting in dysphonia, aspiration or airway compromise) and cervical hematoma. Due to potentially life-changing effects if complications occur, a focused analysis is necessary in order to inform the ongoing discussion on the most appropriate extent of surgery in pediatric thyroid cancer.

Materials and Methods

Following Institutional Review Board approval, a retrospective review of a prospectively maintained database of pediatric patients undergoing surgery for thyroid malignancy between 1986-2021 at Memorial Sloan Kettering Cancer Center was undertaken. Pediatric patients were defined as being 21 years-of-age or younger at the time of surgery, as per the American Academy of Pediatrics definition.¹⁹ Data was collected on patient demographics, diagnosis, tumor, node and metastasis (TNM) stage (8^th Edition American Joint Committee on Cancer), gross extrathyroidal extension (ETE), vascular invasion, margins, microscopic extranodal extension (ENE), extent of thyroid surgery, use of recurrent laryngeal nerve monitor, parathyroid gland reimplantation, neck dissection and adjuvant radioactive iodine (RAI).²⁰ Pre- and post-operative serum biochemistry and serum parathyroid hormone results were reviewed for each patient. Incidence of dysphonia, laryngoscopy findings (mirror or flexible laryngoscopy), hematoma and wound infection incidence were collected. Temporary and permanent hypocalcemia were defined by the American Association of Clinical Endocrinologists and American College of Endocrinology definition: temporary being up to 12 months in duration following surgery, and permanent lasting longer than 12 months.²¹ Temporary and permanent recurrent laryngeal nerve palsy were also defined as up to and over 12 months. Statistical analysis was carried out using R version 4.2.0 (R Core Team, 2022). A subgroup analysis was carried out for patients aged under 18 years. The complication rates were compared between < 18 year olds and 18 – 21 year olds. Univariable and multivariable logistic regression analysis was undertaken to identify factors associated with permanent hypocalcemia.

Results

Patient and Tumor Characteristics

A total of 307 pediatric patients with well-differentiated thyroid carcinoma were identified, with a median follow-up of 61 months. Table 1 demonstrates patient and tumor characteristics. Median age of surgery was 18.61 years (Interquartile range [IQR] 4.5 years; range 4-21 years) and 78% (238/307) were female. The distribution of ages is demonstrated in Figure 1. Of the PTC subtypes, classical was the most common (52%; 160/307), followed by follicular subtype (16%; 50/307), tall cell (7.8%; 24/307) and diffuse sclerosing subtypes (6.2%; 19/307). Follicular carcinoma and oncocytic (Hurthle) cell carcinoma accounted for 2.6% (8/307) and 1% (3/307) respectively. There were 3 cases of solid/trabecular and 2 cases of cribriform-morula subtypes. 12% (38/307) of PTCs did not have a subtype assigned on histopathology. The majority of patients had T1 or T2 disease (253/307; 82%). N0 disease occurred in 40% (123/307); 27% (84/307) were N1a and 33% (100/307) were N1b. Distant metastases were identified in 4.5% (14/307) of patients at presentation. Gross extrathyroidal extension was identified intra-operatively in 14% (43/307). Vascular invasion was identified in 26% (71/274) and positive margins in 17% (47/274). Extranodal extension occurred in 23% (51/223).

Table 1.

Patient and tumor characteristics

Characteristic	N = 3071
Age	18.61 (16.0, 20.5)
Gender
Female	238 (78%)
Male	69 (22%)
T Stage
1	175 (58%)
2	78 (26%)
3	32 (11%)
4	19 (6.2%)
N Stage
N0	123 (40%)
N1a	84 (27%)
N1b	100 (33%)
M Stage
0	293 (95%)
1	14 (4.5%)
Histology
Classical PTC	160 (52%)
FV-PTC	50 (16%)
PTC Unknown	38 (12%)
TCV-PTC	24 (7.8%)
DSV-PTC	19 (6.2%)
Follicular Carcinoma	8 (2.6%)
HCC	3 (1.0%)
Solid/Trabecular PTC	3 (1.0%)
Cribriform-Morular	2 (0.7%)
Any Gross ETE
No	263 (86%)
Yes	43 (14%)
Mitotic Activity
No	226 (86%)
Yes	38 (14%)
Vascular Invasion
No	203 (74%)
Yes	71 (26%)
Extra Nodal Extension
No	172 (77%)
Yes	51 (23%)
Margins
Negative	227 (83%)
Positive	47 (17%)
Multifocality
No	162 (56%)
Yes	128 (44%)

¹Median (IQR); n (%)

Missing observations excluded from counts.

Figure 1:

Histogram of patient age at surgery

Treatment

Treatment characteristics are shown in Table 2. 69% (211/307) underwent TT and 31% (96/307) underwent less-than TT (88 lobectomies, 4 isthmusectomies, 2 subtotal thyroidectomies and 2 Sistrunks procedures). Neck dissection was undertaken in 51% (156/307) of patients. Central (CND) and lateral neck dissection (LND) was undertaken in 26% (79/307), CND alone in 20% (61/307) and LND alone in 5% (16/307). Operative management was undertaken by the Head and Neck Service in 77% (237/307), Pediatric Surgery Service in 21% (64/307) and General Surgery Service in 2% (6/307). Adjuvant radioactive iodine was given in 45% (139/307) of patients. No patients were operated on as a day-case or as outpatients. In our institution, no patients are routinely started on pre-operative calcium or vitamin D supplementation. Post-operatively all patients follow a standardized protocol for hypocalcemia which is dependent on the postoperative calcium tests at 6- and 24-hours following surgery.

Table 2.

Treatment Characteristics

Characteristic	N = 3071
Thyroidectomy
Total thyroidectomy	211 (69%)
Less than total	96 (31%)
Neck Dissection
No	151 (49%)
Central and Lateral	79 (26%)
Central	61 (20%)
Lateral	16 (5.2%)
Central Neck Dissection
No CND	165 (54%)
Unilateral CND	51 (17%)
Bilateral CND	91 (30%)
Radioactive iodine
No	168 (55%)
Yes	139 (45%)

¹n (%)

Missing observations excluded from counts. Percents rounded

Complications

Complications are shown in Table 3. No patients developed a neck hematoma and 2 patients (0.7%) had post-operative wound infection. Five patients (1.6%) had temporary unilateral vocal cord palsy that returned to full function as confirmed on flexible laryngoscopy or mirror examination of the larynx. There were no cases of permanent vocal cord palsy from inadvertent injury. However, 2 patients did have permanent vocal cord palsy due to planned resection of RLN invaded by tumor. Across the time period of the study, recurrent laryngeal nerve monitoring was utilized in 55.0% (169/307) of cases. Of the 5 patients that had post-operative temporary recurrent laryngeal nerve palsy, recurrent laryngeal nerve monitoring was utilized in one case (1/5; 20%). Recurrent laryngeal nerve monitoring was not in routine use at the time of the other 4 cases.

Table 3.

Post-Operative Complications

Characteristic	N = 3071
Hematoma
No	307 (100%)
Yes	0 (0%)
Wound infection
No	305 (99.3%)
Yes	2 (0.7%)
Vocal cord palsy
No	300 (97.7%)
Temporary	5 (1.6%)
Permanent *	2 (0.7%)
Hypocalcemia
No	191 (62.2%)
Temporary	100 (32.6%)
Permanent	16 (5.2%)

¹n (%)

^*planned resection due to tumor invasion

Temporary post-operative hypocalcemia occurred in 32.6% (100/307) of patients. Permanent post-operative hypocalcemia occurred in 5.2% (16/307). All patients that had permanent hypocalcemia had total thyroidectomy and 81% (13/16) had a neck dissection (10 central and lateral, 3 central only). No patients who developed permanent hypocalcemia had abnormal pre-operative serum calcium levels. A sub-group analysis of complications was undertaken for patients under 18 years of age at time of surgery (136 patients, Supplementary Table 1). There was no significant difference in rate of transient vocal cord palsy (2/136; 1.5% in < 18 year versus 3/171; 1.8% in 18-21 years p = >0.99), permanent vocal cord palsy (1/136; 0.7% versus 1/171; 0.6%; p = >0.99) or permanent hypocalcemia (10/136; 7.4% versus 6/171; 3.5%; p = 0.13). The transient hypocalcemia rate was higher in patients aged < 18 years than those aged 18 – 21 (57/136; 42% versus 43/171; 25%; p = 0.002).

Table 4 shows results of univariable and multivariable logistic regression analysis to identify risk factors for temporary hypocalcemia. Age (Odds Ratio [OR] 0.93; p = 0.05), T3 (OR 3; p = 0.005), T4 (OR 2.94; p = 0.028), central neck dissection (OR 4.69; p < 0.001), N1a (OR 3.91; p < 0.001) and N1b disease (OR 4.66; p < 0.001) were identified as risk factors for temporary hypocalcemia after univariable analysis. Central neck dissection (OR 3.30; p < 0.001) and N1a/b disease (N1a OR 2.51; N1b OR 0.96; p = 0.036) were identified as independent risk factors for temporary hypocalcemia. Table 5 shows results of univariable and multivariable logistic regression analysis to identify risk factors for permanent hypocalcemia. Central neck dissection (OR 3.72; p = 0.026), T4 disease (OR 6.4; p = 0.007) and N1b disease (OR 5.45; p = 0.010) were identified as risk factors for permanent hypocalcemia after univariable analysis. Increased N-Stage (N1b OR 3.64; p = 0.018) was an independent risk factor for permanent hypocalcemia after multivariabe analysis. Younger age was not associated with greater risk of permanent hypocalcemia (OR 0.94; p = 0.417).

Table 4.

Logistic regression analysis of post-operative temporary hypocalcemia

	Univariable			Multivariable
Characteristic	OR1	95% CI1	p-value	OR1	95% CI1	p-value
Age	0.93	0.86, 1.00	0.050
Gender
Male	—	—
Female	1.04	0.59, 1.87	0.890
T Stage
1	—	—
2	1.18	0.65, 2.10	0.587
3	3.00	1.39, 6.55	0.005
4	2.94	1.12, 7.83	0.028
CND						<0.001
No	—	—		—	—
Yes	4.69	2.82, 7.97	<0.001	3.30	1.77, 6.25
N Stage						0.036
N0	—	—		—	—
N1a	3.91	2.06, 7.64	<0.001	2.51	1.24, 5.15
N1b	4.66	2.52, 8.90	<0.001	2.05	0.96, 4.44

¹OR = Odds Ratio, CI = Confidence Interval

Table 5.

Logistic regression analysis of post-operative permanent hypocalcemia

	Univariable			Multivariable
Characteristic	OR1	95% CI1	p-value	OR1	95% CI1	p-value
Gender
Male	—	—
Female	1.27	0.40, 5.66	0.714
Age	0.94	0.83, 1.10	0.417
T Stage
1	—	—
2	0.63	0.09, 2.65	0.572
3	2.48	0.51, 9.50	0.206
4	6.40	1.45, 22.2	0.007
CND
No	—	—		—	—
Yes	3.72	1.26, 13.5	0.026	1.77	0.43, 8.93	0.4
N Stage
N0	—	—		—	—	0.018
N1a	0.48	0.02, 3.84	0.530	0.38	0.02, 3.32
N1b	5.45	1.67, 24.5	0.010	3.64	0.74, 21.8

¹OR = Odds Ratio, CI = Confidence Interval

Parathyroid reimplantation was carried out in 6.2% (19/307) of patients. There was no significant difference in transient hypocalcemia between patients that did and did not undergo parathyroid autotransplantation (42.1%; 8/19 versus 31.9%; 92/288; p = 0.36) and there was no significant difference in permanent hypocalcemia (10.5%; 2/19 versus 4.9% 14/288).

Permanent hypocalcemia and vocal cord palsy rates were compared across four ten-year time periods (Supplementary Table 2). There was a significant reduction in permanent hypocalcemia rate (p = 0.012) over this period. There was no significant difference in vocal cord palsy rate across the four time periods

Discussion

There is limited data on complication rates of thyroidectomy for pediatric thyroid cancer. All single center studies are from pediatric hospitals that feature a mix of benign and malignant disease, with malignancy being the indication for surgery in 25% to 46.9% of cases.^8-14 The permanent hypocalcemia rate in these studies ranged from 0.6% to 21.7%, although variation between studies occurred in extent of thyroid surgery and length of follow-up.^8,11 Of the studies that specified follow-up, the longest average follow-up was 32 months.¹⁰ A systematic review of hypocalcemia after thyroidectomy in children gave a pooled incidence of 35.5% transient and 4.2% permanent hypocalcemia. However, the study concluded that there was inconsistent outcome reporting.²² The heterogeneity in diagnoses, inclusion criteria and need for neck dissection in these studies means an accurate picture of complications in pediatric thyroid malignancy is lacking.

A nationwide study of 361 pediatric patients readmitted to hospital post thyroid surgery showed an increased risk of hypocalcemia in patients diagnosed with thyroid cancer (20.9% of cases).²³ However, this only included patients who were re-admitted and lacked prolonged follow-up for permanent hypocalcemia. Hanba et al (2017) also identified a postoperative hypocalcemia rate of 20% in their analysis of partial and total thyroidectomies via the Kids Inpatient Database.²⁴ Again, the study design meant permanent hypocalcemia rate could not be identified. In our study, permanent post-operative hypocalcemia occurred in 5.2%, which is comparable to systematic review of pediatric thyroid surgery for benign and malignant disease.²² Studies assessing clinical risk factors for hypocalcemia after thyroid surgery in pediatric patients have also produced inconsistent results and systematic review highlighted the urgent need for further research to address this.²⁵ Chen et al identified total thyroidectomy, bilateral lateral neck dissection, Graves disease and malignancy as risk factors for postoperative hypocalcemia after multivariable analysis.⁵ Maksimoski et al did not find that malignancy increased risk of hypocalcemia and identified central neck dissection as a risk factor in their retrospective study of tertiary pediatric hospital outcomes in North America.²⁶ A Scandinavian national registry review did not find any risk factors for permanent hypoparathyroidism after multivariate logistic regression. All of these studies included a mixture of patients with benign and malignant disease. Our univariable and multivariable regression analysis is the first to assess risk factors for permanent hypocalcemia in malignant disease alone and identified N-stage as an independent risk factor for permanent hypocalcemia. All patients who had permanent hypocalcemia had undergone total thyroidectomy. Patients with cervical metastases should therefore be identified for pre-operative counselling on the potential need for long-term calcium replacement and require close post-operative monitoring. In considering extent of surgery, the risk of permanent hypocalcemia should be considered when deciding upon total thyroidectomy over thyroid lobectomy.

In our study there was a significantly increased rate of transient hypocalcemia in patients <18 years in comparison to those 18-21 years of age (57/136; 42% versus 43/171; 25%; p = 0.002). However, age was not an independent prognostic risk factor for hypocalcemia after univariable and multivariable analysis. This suggests that younger age alone is not the reason for increased risk of transient hypocalcemia, but other clinicopathologic factors such as nodal disease in younger patients is contributory. The increased rate of transient hypocalcemia in <18 years did not carry over into permanent hypocalcemia, which is more important when considering long-term morbidity.

Parathyroid gland autotransplantation is associated with increased risk of postoperative and protracted hypoparathyroidism in adult neoplastic thyroid surgery, with a meta-analysis of 10,531 patients reporting a relative risk of 1.75 in patients undergong autotransplantation compares to those that did not (p < 0.001)²⁶. In our study, an increased proportion of patients that underwent autotransplantation had transient and permanent postoperative hypocalcemia. However, this was not a statistically significant difference.

Permanent RLN palsy rate in pediatric thyroid surgery ranges from 0.4% to 2.5% in single center studies that include benign and malignant disease.^8,10 One study that looked at papillary thyroid cancer outcomes alone gave a high permanent surgical morbidity rate of 14%, including 5% permanent RLN palsy.¹⁵ However, this study was from a unique cohort of patients after radioiodine fallout in Belarus with all patients undergoing total thyroidectomy, central and lateral neck dissections. Therefore, the results are not directly applicable to sporadic malignancy, treated with more conservative thyroid surgery and therapeutic rather than elective neck dissection. Our data more accurately represents the complications associated with thyroid surgery for sporadic malignancy, in which 0.7% of patients had permanent RLN palsy. Both cases were due to invasion of the nerve by tumor with deliberate planned RLN resection. This is comparable to the most favourable rate in single center studies, that also include benign disease, and the rate of 0.8% identified in across multiple tertiary pediatric units in North America.^8,27 International consensus guidelines exist on intraoperative nerve monitoring (IONM) in pediatric patients, and it is generally accepted that IONM enhances safety and RLN identification.²⁸ In our study, one patient with temporary vocal cord palsy had IONM, which emphasizes that IONM should be used as an adjunct and not a replacement for direct nerve visualization.

Post-operative hematoma can be a serious life-threatening complication in thyroid surgery and previous studies have indicated the risk is higher in patients who undergo surgery due to Graves Disease.⁸ Our hematoma rate of 0% indicates that the risk is extremely low in pediatric malignancy in the hands of experienced thyroid surgeons.

High volume surgeons have been shown to have lower 90-day complication rates in pediatric thyroidectomy surgery than low volume surgeons.²⁹ The majority of surgeons at our institution undertake both adult and pediatric thyroidectomies for malignant disease, resulting in a high individual volume of thyroid surgery. Previous studies have suggested a link between speciality and complication rates, with pediatric general surgeons having superior outcomes to pediatric otolaryngologists.^29,30 However, there was no significant difference in complication rates in our study between specialities.

However, it is important to discuss the limitations of our study. The retrospective nature of the study means that it is susceptible to the inherent limitations associated with retrospective studies. This includes the lack of standardized prospective data collection. For example, not all patients had routine post-operative laryngoscopy, although all dysphonic patients did undergo laryngeal examination by mirror or flexible laryngoscopy. Documentation of symptoms of clinical hypoparathyroidism was also variable, particularly in more historic reports.

Conclusion

There is a lack of data regarding operative complication rates for pediatric thyroid malignancy. Our study is the first to report applicable results for malignancy alone and shows pediatric thyroid surgery at a tertiary cancer center results in low complication rates, despite the majority of patients having nodal metastases. Vocal cord paralysis is rare unless disease is found to be invading the RLN intraoperatively. Advanced N-stage and central neck dissection are independent risk factors for post-operative hypocalcemia and these patients should be appropriately counselled and monitored.

Synopsis.

Central neck dissection and N1 disease are risk factors for hypocalcemia after pediatric thyroid carcinoma surgery. Post-operative vocal cord paralysis is rare unless disease is found to be invading the recurrent laryngeal nerve intraoperatively.