In our study, we did not find evidence that patients with T2DM and/or obesity are at higher risk of being diagnosed with thyroid cancer. However, we did find that T2DM patients are at risk of developing a goitre, probably due to the disorder of glucose metabolism itself and not thyrotropin stimulation. In general, men are diagnosed with more advanced thyroid cancer compared to women, and these males with poorer prognosis regarding thyroid cancer tend to also have diabetes.
In the period from 1998 to 2014, rapid increases in the incidence of thyroid cancer have been observed, especially for papillary thyroid cancer, which is more likely to be found in a subclinical form and therefore detected by intense scrutiny of the thyroid gland [22]. This is supported by autopsy studies that have found the presence of differentiated thyroid carcinomas (“incidentaloma”) in 8.5% of patients [23]. For this reason, we believe that it is crucial to compare benign and malignant tumours confirmed by histology. In our study, incidentalomas, especially micropapillary thyroid carcinoma with maximum diameter 7 mm, were identified in 13.8% of all thyroid cancers, and 60% of incidentalomas were detected in PDM/T2DM patients.
There is evidence of a robust association between patients with type 2 diabetes and an increased risk of common cancers, i.e. pancreatic, endometrial, breast, colorectal, and hepatocellular [24, 25], with the risk of all cancers 16% higher than the expected cancer incidence in the general population [26]. The analysis of Goto et al. 2020 provided evidence that the genetic mechanisms responsible for T2DM may not play major roles in cancer development [27]. Nevertheless, plausible biologic pathways linking diabetes to thyroid cancer risk include: 1) chronic thyrotropin stimulation; 2) hyperinsulinemia and insulin resistance; (3) antidiabetic medication; 4) concomitant increased body mass index (BMI) in the majority of diabetics, which has been in general associated with increased risk for malignancies, and (5) chronic glucose exposure [28].
Rapp et al. 2006 performed a population-based prospective cohort study investigating relationships between fasting blood glucose and the incidence of cancer. Slightly positive associations with glucose values > 5.3 mmol/l were noted for thyroid cancer (HRs of 1.34 (95% CI 0.66–2.72) in women and 1.24 (0.65–2.37) in men). Surprisingly, several cancers were also positively associated with low fasting blood glucose levels (2.2–4.1 mmol/l), including thyroid cancers in women; however, estimates were not statistically significant (29). In our study, the B, M and MB cohorts were comparable in glucose levels; respectively 5.3, 5.3 and 5.18 mmol/L (p = 0.834).
There have been many studies, mostly large-scale and employing meta-analyses, addressing the issue of diabetes and the risk of developing thyroid cancer. These studies have very conflicting results, with limited statistical significance between differentiated thyroid cancer risk and diabetes. Further, these studies have contrasting results in female and male populations. Although thyroid cancer typically occurs about three times more frequently in women than in men in the United States, rates of diabetes in men and women in the United States are similar [30–32]. The increased thyroid cancer risk varies even among patients with different kinds of diabetes mellitus, being 34%, 30%, and 51% in patients with type 1 diabetes, gestational diabetes, and T2DM, respectively [33]. In addition, these analyses are considerably limited by the reliance on self-reported data. In our previous study, 33% of patients with T2DM and especially PDM were newly diagnosed by our screening [34]. Data from the US has clearly shown that diabetes mellitus may be frequently undiagnosed; 35% of adults > 20 years of age and 50% of those > 65 years of age may have prediabetes [35]. The study by Fang et al 2018 examined cancer risks for a total of 51,324 registered cancer-free individuals that had been newly diagnosed with T2DM. Within 1 year following the onset of diabetes onset, women with T2DM had a higher risk of thyroid cancer. The results suggest that T2DM patients are at a higher risk of certain cancers; this risk particularly increases shortly after diabetes diagnosis, which is likely due to detection bias caused by increased visits to medical institutions [36]. Our study results concur with these conclusions: PDM/T2DM were present in 25% of all patients, which is higher than in the general population.
We did not observe an increased risk of thyroid cancer in our diabetic patients. Further, the cohort of prediabetic and diabetic patients are those with the most expressed insulin resistance and hyperinsulinemia, and thus have larger thyroid glands and tendency to have larger thyroid nodules, which could lead to earlier thyroid cancer manifestation and detection. In contrast, the cohort with proven thyroid cancer had smaller thyroid nodules compared to diabetic patients with typical goitre.
We also observed larger thyroid glands and a tendency to have larger thyroid nodules in men compared to women. Furthermore, men were tended to be diagnosed with more advanced thyroid cancer, often characterised by unfavourable mutations such as in TERT. These males with high-risk thyroid carcinoma had more often diabetes. We have to emphasise that the sample size was very small, making definitive conclusions impossible.
Insulin resistance with hyperinsulinemia likely leads to thyroid growth, and we suppose that the thyroid growth trigger is general to thyroid tissue and not just to single thyroid nodules. We do not suggest, however, that chronic thyrotropin stimulation is the crucial cause of goitrogenesis. This is despite the fact that our B cohort had significantly lower TSH levels (p < 0.001) than the M cohort, as TSH values were on average in the normal range in all cohorts and the B cohort had even larger thyroid nodules (p = 0.026).
We also focused attention on assessing any glucose disorders during euthyroidism, because both hypo and hyperthyroidism have detrimental effects on optimal glucose tolerance. A recent study by Roh et al. 2022 showed that patients with thyroid cancer who underwent thyroidectomy were more likely to develop T2DM than the matched controls. There was a U-shaped dose-dependent relationship between the levothyroxine dosage and T2DM risk. TSH levels were not accessible in Roh et al. 2022 study, but it can be hypothesised that these patients could be over or underdosed [16].
Patients with insulin resistance have a higher prevalence of thyroid nodules. A study by Rezzónico et al 2009 showed that patients with thyroid cancer are more insulin resistant (p < 0.001) with a positive association between insulin resistance and BMI. Those results could have been influenced by TSH suppressive medication, and only 40 probands were involved in the study [37]. We were able to calculate insulin sensitivity by the HOMA-IR index just in 54 patients, with glucose metabolism testing done under euthyroidism. We did not observe any significant differences between HOMA-IR and HOMA S% parameters between our B, M and MB cohorts, except for lower HOMA B% in the MB cohort (p = 0.039). HOMA B% expresses the beta cell secretory capacity, and the higher the value, the more insulin the beta cells have to secrete to respond to blood glucose levels. Our MB cohort was younger and more insulin sensitive than the B and M cohorts, which explains the lower HOMA B% levels in the MB cohort. We must also emphasise that the number of patients in our MB cohort was very small.
More than half of diabetic subjects are middle aged, and the incidence rises with increasing age in both sexes [38]. In contrast, new cases of thyroid cancer represent just 2.2% of all new cancer cases, with a median age at diagnosis of 51 years, peaking at age of 45–54 (21.6%)(https://seer.cancer.gov/statfacts/html/thyro.html). From this point of view, we suggest that trends of T2DM compared to thyroid cancer are not related. Also, in our study the age of patients diagnosed with thyroid cancer peaked at 53 years in comparison to T2DM at 63.5 years old.
An extensive review published a few years ago estimated that 20% of all cancers might be caused by obesity [39]. The prevalence patterns of obesity resemble those of T2DM. In addition to cancer risk, obesity is associated with more aggressive pathological tumour features and worse outcomes in patients with breast, ovarian, prostate, and colorectal cancers [40]. The relationship between obesity and thyroid cancer is uncertain, however. Although a positive association between obesity and thyroid cancer risk has been observed in women in several case–control and prospective studies [41, 42], not all studies have agreed and the results are even less consistent in men [43]. In a study by Dieringer et al. 2015 in an age-related subgroup analysis, a higher BMI was correlated with more lymph node involvement (p = 0.004), lymphatic invasion (p = 0.003) and tumour multiplicity (p = 0.008) in patients ≥ 45 years of age. Interestingly, when the patients were stratified according to age, those 45 years old or older showed a significant relationship between BMI and clinicopathological features, but those younger than 45 years did not [17]. Another meta-analysis on 35,237 patients supported those results, with more pronounced associations as BMI increased [44]. However, in the case of thyroid cancer, great heterogeneity exists between studies with regard to potential associations between obesity and clinicopathological characteristics. In our study we did not observe any significant differences in BMI between the B, M and MB cohorts (p = 0.452). As we already mentioned, diabetic males seem to present with more advanced thyroid cancer, and these males are on average obese.
To the best of our knowledge, there are no similar studies comparing benign, malignant and low-risk thyroid carcinomas in PDM/T2DM with known insulin resistance. We can not compete with the large-scale studies in sample size, but there are often biases in the way information is obtained and recorded in comparison to small trials [45, 46]. In our retrospective cohort study, we devoted great attention to collecting all available data and at the same time point from medical, pharmacological and laboratory reports. Therefore, the benefits of our study are detailed medical histories combined with laboratory testing including molecular genetics and screening of diabetes, along with the gold standard of histological confirmation. Furthermore, glucose metabolism was determined under euthyroidism, so we could exclude the undesirable effects of hypo/hyperthyroidism. We did not use any questionnaires and/or self-reported data due to their lower reliability. Considering the limitations of our study, the retrospective design could have introduced some biases in data collection, and there was just a small sample of patients for evaluating HOMA-IR due to the low availability of C-peptide levels. Similarly, molecular testing was not performed in all patients according to Bethesda cytological groups, because FNA samples for molecular testing were not accessible in all cases. We were also not able to evaluate initial thyroid function in our prediabetic and diabetic patients.
In conclusion, while ageing increases the chance of becoming diabetic and/or obese, our data generally do not support the hypothesis that patients with prediabetes/type 2 diabetes and obesity are at a higher risk of thyroid cancer. However, these patients are at higher risk to have a goitre. In contrast, it seems that a subgroup of males with type 2 diabetes are at risk of being diagnosed with more progressive thyroid cancer. Insulin resistance with hyperinsulinemia could be a more potent candidate for thyroid growth stimulation compared to stimulation by thyrotropin or autoimmune thyroid disease in these patients. We suppose that particular attention should be given to males with metabolic disturbances. Further studies are needed to confirm our results, but we believe that our study can partly explain the heterogeneity found in previous studies.