Incidence of Thyroid Carcinoma Among Patients With Hyper and Hypothyroidism

Thyroid cancer is the most commonly encountered type of cancer in clinical practice and its incidence is continuously increasing: 1.6% of all new cancer cases are thyroid cancers.^1,2 Improvement in diagnostic methods, especially the detailed investigation of thyroid nodules, has had a major effect on this recent increase in incidence.^3,4

In endemic regions for goiter such as Turkey, the rate of cancer is reported to be lower among patients with hyperthyroidism.⁵ When thyroid gland function is inadequate, continuous overstimulation by the thyroid-stimulating hormone (TSH) results in irregular growth of the thyroid gland. Iodine deficiency, inadequate hormone synthesis by the thyroid gland, dietary or medical goitrogens increase the risk for thyroid cancer by continuously stimulating high levels of TSH.⁶ Experimental studies have shown that animals that were fed an iodine-deficient diet for a prolonged period of time, develop follicular hyperplasia, followed by nodules and adenomas at a higher rate than the control groups.^5,6 On the other hand, even though TSH is suppressed in patients with hyperthyroidism, increased TSH stimulation at the early stage results in a toxic state that stimulates tumor development through an oncogenic effect before TSH suppression occurs.⁷ Still, the relationship between increased TSH secretion and thyroid cancer is yet to be elucidated.

In the Istanbul Training and Research Hospital General Surgery Clinic, we observed that carcinoma incidence in patients with hyperthyroidism is at least as high as those with hypo- or euthyroidism. We therefore aimed to retrospectively review cases of bilateral total thyroidectomy in our hospital to investigate if hyperthyroidism has any protective role against carcinoma.

Materials and Methods

The study was approved by the local ethics committee of the Istanbul Training and Research Hospital on May 16, 2014 with Protocol Number 492.

In this study, 1200 cases undergoing bilateral total thyroidectomy in the Istanbul Training and Research Hospital General Surgery Clinic between the years 2010 and 2014 were retrospectively examined. While collecting data, TSH, unbound triiodothyronine (FT3) and unbound thyroxine (FT4) levels at presentation, demographic data such as age and sex, and postoperative pathologic diagnoses were recorded. For categorizing thyroid hormone levels, we used the 2013 Thyroid Disease Diagnosis and Treatment Guidelines⁸ of the Turkish Endocrinology and Metabolism Association as reference. Age, sex, hormone level groups, and postoperative pathologic diagnoses were independently compared. Because of the retrospective design of the study, informed consent for anonymous data collection was not taken.

The collected data was analyzed using commercial software (SPSS v 22.0, Statistical Package for Social Sciences, Inc., Chicago, Illinois). The distribution of the variables was evaluated using the Kolmogorov-Smirnov test. For statistical analysis of quantitative data, Kruskal-Wallis and Mann-Whitney U-tests were used whereas χ² test was employed for categorical variables.

Results

We examined 1200 patients in total; 210 patients (17.5%) were male and 990 (82.5%) were female. Mean age was 49.2 years (17–82); 718 of the patients (59.8%) had euthyroidism; 351 (29.3%) had hyperthyroidism; and 131 (10.9%) had hypothyroidism (Table 1).

Table 1  Classification of included patients according to age, sex, pathologic diagnoses and hormone status

View Fullscreen

The total number of patients with benign pathologies was 866 (72.2%). Among these, 576 (48.0%) had adenomatous hyperplasia; 196 (16.3%) had lymphocytic thyroiditis; 73 (6.1%) had follicular adenoma; and 21 (1.8%) had Hashimoto's thyroiditis. According to histological type, of the 334 malignant cases, papillary carcinoma made up 88%, follicular carcinoma 7%, and medullary carcinoma 5% of all thyroid malignancies.

Hormone levels and pathologic diagnoses were separately compared with the age of patients and no statistically significant difference was detected (Table 2).

Table 2  Comparison of pathologic diagnoses and hormone status according to patient age

View Fullscreen

No significant difference was found in the sex distribution of benign and malignant groups. However, analyzing the subtypes of benign and malignant groups, the ratio of male patients was higher in the medullary carcinoma group compared to follicular and papillary carcinomas and lower in the lymphocytic thyroiditis group than in other benign pathologies. Of the 196 patients with lymphocytic thyroiditis, 178 were female. The number of male patients in this group was only 18. The male-to-female ratio of patients with hypothyroidism, euthyroidism and hyperthyroidism did not differ significantly (Table 3).

Table 3  Comparison of pathologic diagnoses and hormone status according to patient sex

View Fullscreen

Comparing hormone status according to the postoperative pathologic diagnoses, 122 (36.5%) of the 334 malignant cases were hyperthyroid. On the other hand, only 26.4% (229/637) of the benign cases were hyperthyroid. Thus, hyperthyroidism was significantly more frequent (P < 0.05) among thyroid cancer patients (Table 4, Fig. 2).

Table 4  Comparison of pathologic diagnoses according to hormone status

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Conversely, analyzing the rate of malignancy according to hormone status, 122 (34.8%) of the 351 patients with hyperthyroidism had malignant and 229 (65.2%) had benign diseases. Patients with hypo- and euthyroidism totaled 849, of which 212 (25%) had malignant and 637 (75%) had benign diseases. According to this data, risk of malignancy was significantly higher among patients with hyperthyroidism compared to the other 2 groups combined.

Discussion

Numerous hypotheses were proposed for the mechanism of thyroid carcinoma. According to Hancock et al,⁹ sustained TSH stimulation as a result of goitrogens causes hypertrophy of the thyroid gland, which is followed by neoplasia. However, the mechanism of thyroid cancer development as a result of excess TSH secretion has not yet been clarified.

In patients with Graves' disease, carcinogenic effect is assumed to be caused by the interaction of high levels of thyroid stimulating immunoglobulin (TSI)—instead of TSH—with the TSH receptors on the thyroid tissue. In these patients, because of the constant stimulation by TSI, cancer tends to follow a more aggressive course.³

The study by Rieger et al,¹⁰ which included 1848 patients, aimed to determine the prevalence of cancer among patients with hyperthyroidism in an endemic goiter region; the rate of cancer was found to be 0.76%. The authors concluded that in regions endemic for goiter, the rate of cancer is lower among patients with hyperthyroidism.¹⁰

Toward the end of the 90s, Ruggieri et al¹¹ reported a carcinoma incidence of 5% in hyperthyroidism and 14.3% in euthyroidism in their study with 82 patients and declared that this rate of cancer among patients with hyperthyroidism is the highest in the literature. Gabriele et al¹² have retrospectively evaluated 425 patients with hyperthyroidism and detected cancer in only 7 (1.65%), thus excluding patients with hyperthyroidism from the risk group.

A large single-center study from an iodine-sufficient region in Asia¹³ and a meta-analysis of 56 studies from 2016¹⁴ have demonstrated that high TSH levels was not only an independent risk factor for thyroid malignancy, but was also a predictor for advanced stage and aggressive disease characteristics.

On the contrary, some recent studies have remarked the increasing rates of malignancies among patients with hyperthyroidism and reported malignancy rates close to that of cold nodules in patients with Graves' disease, autonomous nodules and multiple nodules.^15,16

In our study conducted in an iodine-deficient country, we retrospectively analyzed data from 1200 patients who had undergone bilateral total thyroidectomy and compared the incidence of cancer among patients with hyperthyroidism and those with hypothyroidism. The rate of malignancy among hyperthyroid patients was significantly higher than those with hypothyroidism and euthyroidism. This finding contradicts the classical knowledge and most of the literature that suggest a strong negative correlation between hyperthyroidism and thyroid malignancy.

The fact that Turkey is an iodine-deficient region endemic for goiter might be the key point underlying this deviation from the literature. In the cross-sectional study by Du et al,¹⁷ more than 2000 subjects from three regions with different iodine nutrition levels were included and the effects of the iodine status were analyzed. The frequency of both thyroid nodules and hyperthyroidism were independently higher in the iodine-deficient region. It is possible that in our study, the relationship between hormone status and pathologic results was not causative, but that both were related to the iodine level of the patients. Unfortunately, no data was collected in our study regarding the iodine status of the subjects.

In the present study, benign cases consisted of 866 patients in total, of which 576 (67%) had adenomatous hyperplasia; 196 (23%) had lymphocytic thyroiditis; 73 (8%) had follicular adenoma; and 21 (2%) had Hashimoto's thyroiditis. These ratios are consistent with the literature.

The large study population of 1200 patients was one of the strong points of the study. It allowed the analyses of multiple variables and drawing conclusions by increasing the power of the study. Having been conducted in an endemic region for goiter is another aspect of the study that makes it noteworthy.

The major weakness of our study was its retrospective design, resulting in selection bias. Since the study population consisted of patients undergoing bilateral total thyroidectomy, it included all cases with malignant fine needle aspiration biopsy results but not those who had small nodules with benign pathology that are conservatively followed up or those for whom unilateral thyroidectomy was preferred. Many euthyroid patients with small, asymptomatic nodules are likely not diagnosed at all. This situation might have altered our results so that the rate of malignancies appeared higher than it actually is. Similarly, undiagnosed euthyroid patients were not included, causing further inaccuracy.

Larger, multicentric studies are needed to confirm our results. Prospective randomized studies with a screening protocol for subject selection might eliminate the inaccuracies that were inevitable in our study because of its very design.

Conclusions

In this study with a large patient population, the rate of malignancy among patients with hyperthyroidism was significantly higher than those with either hypothyroidism or euthyroidism. Unlike what is previously accepted, hyperthyroid patients also carry a risk for thyroid carcinoma. Thus, these patients should be followed up for cancer at least as closely as those with hypothyroidism and euthyroidism.