Insulin Resistance Associated With Differentiated Thyroid Carcinoma: Penalized Conditional Logistic Regression Analysis of a Matched Case-Control Study Data

authors:

avatar Zahra Heidari ORCID 1 , avatar Mahdi Abdani 1 , avatar Mohammad Ali Mansournia ORCID 2 , *

MD, Department of Endocrinology and Metabolism, Zahedan University of Medical Sciences, Zahedan, IR Iran
MD, MPH, PhD, Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, IR Iran

how to cite: Heidari Z, Abdani M, Mansournia M A. Insulin Resistance Associated With Differentiated Thyroid Carcinoma: Penalized Conditional Logistic Regression Analysis of a Matched Case-Control Study Data. Int J Endocrinol Metab. 2018;16(1):e14545. https://doi.org/10.5812/ijem.14545.

Abstract

Background:

In the recent years, thyroid cancer incidence has increased worldwide. It has been questioned whether factors, such as insulin resistance, are involved in this rise. The main aim of this study was to examine the association between insulin resistance and differentiated thyroid cancer.

Methods:

This case-control study was performed on 30 newly diagnosed patients with differentiated thyroid cancer and 30 healthy controls. Thirty euthyroid patients recently diagnosed with differentiated thyroid cancer, based on the fine needle aspiration cytology, were examined for insulin resistance before thyroidectomy in the euthyroid state. For each patient in the case group, one healthy euthyroid individual without thyroid nodule from general population was matched based on gender, age (± 1 year), and body mass index (BMI) (± 1). Thyroid function test, thyroid antibodies, serum glucose, serum insulin, and thyroid ultrasonography was performed for all participants. Insulin resistance was defined based on homeostasis model assessment of insulin resistance (HOMA-IR) to be more than 2.5.

Results:

In the patient group, 24 (80%) cases were female and 6 (20%) were male. Insulin resistance was more prevalent in the case group than the control group (43.3% versus 13.3%). Insulin sensitivity index was lower in the case group than in the control group (50 and 81, respectively). There was a significant positive association between serum HOMA-IR levels and differentiated thyroid cancer (OR: 2.43 for 1 unit increase in HOMA-IR, 95% CI: 1.35 - 5.51; P = 0.001). Insulin resistance was significantly associated with differentiated thyroid cancer (OR: 4, 95 % CI: 1.27 - 17.6; P = 0.016).

Conclusions:

There was a significant association between insulin resistance and differentiated thyroid carcinoma. More research with a larger sample size and prospective design are needed to determine the role of this factor in the development of differentiated thyroid cancers.

1. Background

Thyroid cancer is the most common endocrine-related malignancy worldwide. This cancer makes up 1.3% of all cancers and causes about 0.5% of cancer-related deaths each year around the world. Although the mortality of differentiated thyroid cancer is relatively low, the rate of disease relapse or persistence is high, leading to increased morbidity (1).

In the recent decades, thyroid cancer incidence has rapidly increased all over the world (2-5). It is unknown whether this rise in thyroid cancer is due to increased use of thyroid sonography and sono-guided fine needle aspiration cytology examination or whether this is due to a true rise in thyroid cancer from an undetermined etiology. Incidence of large tumors has increased as well. Furthermore, despite earlier diagnosis and better treatment for thyroid cancer, the mortality caused by this illness has not only been diminished, but has also increased (6, 7). Thus, additional factors other than early detection may contribute. Currently, risk factors for thyroid cancer include previous childhood radiation to the head and neck, family history of thyroid cancer, exposure to ionizing radiation (8), inadequate or excess iodine intake (9), which are well known but none of them can justify an increase in the incidence of thyroid cancer. Recently, other factors, such as diabetes, obesity, metabolic syndrome (10, 11), insulin resistance (12), chemical toxins (13, 14), and dietary factors (15, 16) have been reported as potential risk factors of thyroid cancer.

In a similar trend with thyroid cancer, the prevalence of obesity and as a result, insulin resistance has increased in the recent decades (17). Although the precise molecular mechanism responsible for increasing the incidence of thyroid cancer is not well defined, it may be possible to attribute some of these carcinogenic effects to insulin resistance. Insulin resistance results in hyperinsulinemia. Increased levels of insulin reduce the production of Insulin-like growth factor-1 binding proteins (IGFBPs) and this leads to increased levels of insulin-like growth factor-1 (IGF-I) in the blood. The IGF-I displays potent antiapoptotic, cell-survival, and transforming activities (18).

However, information on the role of insulin resistance in differentiated thyroid cancer is limited. To the best of our knowledge, there are only 2 studies about the role of insulin resistance in DTC formation. In one study, insulin resistance was more prevalent in 27 female patients with DTC in comparison with the control group (12). In another study, Forty-one patients with DTC were compared with the control group. The difference between groups with regards to homeostasis model assessment of insulin resistance (HOMA-IR), the frequency of insulin resistance (IR) and body mass index (BMI) was not significant (19).

Due to limited information, this study aimed at further evaluating the association between insulin resistance and differentiated thyroid cancer as well as the role of this parameter as a potential novel risk factor in thyroid cancers.

2. Methods

2.1. Subjects

This case-control study was carried out on 30 euthyroid patients with differentiated thyroid carcinoma. The inclusion criteria were 1) euthyroidism (20, 21) (normal thyroid function was defined as normal thyroid stimulating hormone [TSH: 0.4 - 4.2 mIU/L], normal free thyroxine [FT4: 0.8 - 1.8 ng/dL] and normal triiodothyronine [FT3: 2.3 - 4.2 pg/mL] levels); 2) thyroid nodules ≥ 1 cm; 3) negative titers of anti-thyroid antibodies (anti-thyroid peroxidase antibody < 16 IU/mL and anti-thyoglobulin < 100 IU/mL); 4) malignant cytology according to Bethesda system classification (suspicious for malignancy or malignant).

Individuals with any of the following characteristics were excluded from the study: subjects with a history of thyroid disease, previous thyroid medication therapy at any time, overt or subclinical hyperthyroidism and hypothyroidism, history of neck irradiation or thyroid surgery, iodinated contrast material exposure in the previous 6 months, smoking, statin and antihypertensive therapy. Individuals were also excluded if they had a personal or family history of diabetes mellitus, endocrine obesity, hepatic or renal failure, schizophrenia, convulsion, and depression, which could have an impact on thyroid function tests. Pregnant and lactating females also accounted for exclusion from the study.

Thirty-four newly diagnosed patients with differentiated thyroid carcinoma based on fine needle aspiration cytology, who live in an iodine-sufficient area (22), attended the outpatient endocrinology clinic of Zahedan city, southeast of Iran. These patients were chosen as the potential study population between June 2014 and July 2016. Among them, 2 patients were excluded due to positive thyroid antibodies, and 1 patient due to subclinical hypothyroidism and 1 because of pregnancy. In all patients with positive cytology, permanent pathology confirmed the diagnosis. Finally, 30 newly diagnosed patients with differentiated thyroid carcinoma were consecutively chosen as the patient group.

For each of the 30 patients of the case group, a pair was selected from hospital staff based on matching protocol; they were individually matched according to gender, age (± 1 year) and BMI (± 1). Fasting plasma glucose was measured for ruling out diabetes mellitus in the control group. Thyroid function tests and thyroid ultrasonography were performed in all participants. Thus, 30 healthy euthyroid control participants were chosen with normal thyroid sonography. Individuals with any known acute or chronic illness were excluded.

Finally, 30 pairs were obtained considering the above-mentioned inclusion and exclusion criteria. The Zahedan University ethics committee for human studies approved the protocol (ethical code number: 92-5988). All participants provided an informed consent.

2.2. Anthropometric Measurements

Body weight (kg) was measured without shoes, using digital scales, and height (cm) was measured in a standing position without shoes. Body Mass Index was calculated based on the weight (Kg)/height (m)2.

2.3. Biochemical Evaluations

All blood samples were taken between 8:00 and 9:00 in the morning and after 12 hours of fasting. Samples were stored at -70°C until the day of the test. Serum glucose was measured using the glucose oxidase technique. Intra-assay CV was 1.0% and inter-assay CV was 4.7%. Serum insulin was evaluated using solid-phase the competitive chemiluminescent enzyme immunoassay (Diagnostic Product LIAISON, Italy) method. Inter-assay coefficient of variation (CV) was 7.39%.

Measured serum insulin and fasting plasma glucose was applied to calculate steady state beta cell function (%B), insulin sensitivity (%S), and insulin resistance (IR) for normal and pre-diabetic participants by using the homeostasis model assessment (HOMA2) calculator (University of Oxford, web site; http://www.dtu.ox.ac.uk/homacalculator/index.php). A HOMA index equal or higher than 2.5 denotes insulin resistance (23-25).

Thyroid stimulating hormone (TSH) was evaluated by using the immunoradiometric assay (Diagnostic Products LIAISON, 2011, Italy) with a reference range from 0.4 to 4.2 mIU/L. The intra-assay coefficient of variation (CV) was 1.5%. The corresponding inter-assay CV was 2.3%. Free triiodothyronine was measured using immunochemoluminescent assays by an automated analyzer (Diagnostic Products LIAISON, 2011, Italy) with a reference range from 2.3 to 4.2 pg/mL. The intra-assay CV was 5.16% and inter-assay CV was 7.9%. Free thyroxine was also measured using immunochemoluminescent assays by an automated analyzer (Diagnostic Products LIAISON, 2011, Italy) with a reference range from 0.8 to 1.8 ng/dL. The intra-assay CV was 3.4% and inter-assay CV was 3%.

Antithyroid peroxidase (normal range < 16 U/mL) and antithyroglobulin (normal range < 100) was measured by immunochemoluminescent assays employing commercial kits (Diagnostic Products LIAISON).

2.4. Thyroid Ultrasound

Thyroid sonography was performed for all participants by a sonologist using a 7.5-MHz linear probe (Aloka Co. Ltd., Tokyo, Japan). All patients with thyroid nodules larger than 1 cm were evaluated by fine needle aspiration biopsy (FNAB). In the control group, there was no nodule with any size.

2.5. Cytology

All FNAB samples were evaluated by an experienced pathologist. The cytology was reported as malignant nodules, according to the Bethesda system in all cases. Patients with diagnosis of thyroid cancer or suspicious of having it were included. Permanent pathology confirmed the diagnosis.

2.6. Statistical Analysis

By studying 30 case and 30 control participants, an unmatched study had a 95 % power to detect a mean difference in HOMA-IR value of 1.6 between case and control groups at the 5% significance level, assuming standard deviation of HOMA-IR value to be 1.7 (14) (so the standardized mean difference would be about 1 SD, which is clearly considered as a clinically important effect size) (26). As the study is matched on strong confounders, such as age and gender, the sample size required for a matched study is indeed expected to be lower than that required for the unmatched study. Thus, the power of the current matched study should be greater than 95% (27-29). Continuous variables were presented as mean (standard deviation) and categorical variables as counts (percentage).

The researchers used conditional logistic regression to estimate the odds ratios (with 95% confidence interval) between HOMA-IR and differentiated thyroid cancer adjusted for matched confounders including age, gender, and BMI. Fractional polynomials and locally weighted scatter plot smoothing (LOWESS) algorithm were used to determine the appropriate scale of continuous variables in models.

Pair-matched analysis only relies on discordant pairs (exposed case-unexposed control and unexposed case-exposed control) and these involve small counts in our study as the sample size is not large, and more importantly the prevalence of exposure is very high or very low. Therefore, the standard conditional logistic regression analysis is susceptible to sparse-data bias. Following Greenland et al. the researchers performed penalized conditional logistic regression using a F(2,2) prior for the matched odds ratio (29-31). The F(2,2) prior encodes the 95% prior odds ratio interval of 1/39 to 39. To apply F(2,2) prior, the researchers augmented the original dataset with 2 sets of discordant pairs: one containing an exposed case and an unexposed control and the other with an unexposed case and an exposed control. This is the case for binary variable: insulin resistance (1: HOMA-IR ≥ 2.5, 0: HOMA-IR < 2.5). For the continuous variable HOMA-IR, the researchers centered the variable (subtracted the mean) so that 0 and 1 are meaningfully different values for the covariate. Based on the suggestions of Greenland et al. this study presented 95% profile-likelihood confidence intervals for odds ratios along with P values based on likelihood-ratio tests (30, 31). All analyses were done by Stata version 12 (Stata Corporation, College Station, TX).

3. Results

The study included 30 patients and 30 controls matched to the cases based on age, gender, and BMI, in pairs. In the patient group, 24 (80%) cases were female and 6 (20%) were male. The mean (SD) age was 34.40 (12.73) years old in the patient group and 33.79 (12.93) in the control group. The mean BMI was 26.1 in both groups. Serum concentrations of FT4, FT3, TSH, Anti-TPO, and Anti-Tg were within the normal range and there was no significant difference between case and control groups. The mean FT4 level was higher in the case group compared to the control group (Table 1).

Table 1.

Clinical and Laboratory Characteristics of Study Subjectsa

VariablesCase Group (n = 30)Control Group (n = 30)
Gender (female), n (%)24 (80)24 (80)
Age, y34.4 (12.7)34.1 (12.8)
BMI, kg/m226.1 (5.4)26.1 (5.3)
TSH, mIU/L2.0 (1.0)2.3 (1.1)
FT4, ng/dL1.2 (0.32)0.95 (0.12)
FT3, pg/mL2.5 (0.73)2.7 (0.35)
TPOAb, IU/mL5.6 (4.0)4.6 (4.3)
TgAb, IU/mL21.2 (20.4)18.4 (28.3)

As depicted in Table 2, the mean fasting plasma glucose, fasting plasma insulin, and HOMA-IR were higher in the case than in the control group. Insulin resistance was more prevalent in the case group than the control group (43.3% versus 13.3%). Insulin sensitivity index was lower in the case group than in the control group (50 and 81, respectively).

Table 2.

Homeostasis Model Assessment Parameters of Study Subjectsa

VariablesCase Group (n = 30)Control Group (n = 30)
Fasting plasma glucose, mg/dL101.5 (9.2)97.4 (7.9)
Fasting insulin, µu/mL18.7 (10.8)11.2 (5.9)
HOMA-IR2.4 (1.4)1.5 (0.76)
Insulin resistance, n (%)13 (43.3)4 (13.3)
Beta cell function, %138.7 (47.0)110.4 (37.7)
Insulin sensitivity index, %50.0 (22.2)81.0 (45.2)

Correlation of PTC size with clinical and laboratory parameters is shown in Table 3.

Table 3.

Correlation of DTC Size With Clinical and Laboratory Parameters in the Patient Group

VariableDTC Size
Correlation R-ValueP Value
BMI0.120.52
Age0.110.54
HOMA-IR0.250.17
Beta cell function -0.100.57
Insulin sensitivity index-0.150.41
TSH0.160.39
FT4-0.100.58
FT3-0.030.84
TPOAb0.200.26
TgAb0.080.64

There were 11 discordant pairs with exposed (HOMA-IR ≥ 2.5) cases and 2 discordant pairs with unexposed (HOMA-IR < 2.5) cases, so the ordinary matched odds ratio is 11/2 = 5.5.

A positive correlation was observed between serum HOMA-IR levels and differentiated thyroid cancer (OR: 2.43 for 1 unit increase in HOMA-IR, 95 % CI: 1.35 - 5.51; P = 0.001). Insulin resistance was significantly associated with differentiated thyroid cancer (penalized OR: 4, 95 % CI: 1.27 - 17.56; P = 0.016). Note that penalization using F(2,2) prior shrinks the ordinary matched OR to 1 by adding one to each of the discordant pairs, i.e. penalized OR= (11 + 1)/(2 + 1) = 4.

Figures 1 shows the Lowess regression curves for the associations between HOMA-IR and DTC. Figure 1 suggests that including HOMA-IR as a linear term in the logistic regression model provides a descent fit.

The Lowess Regression Curve for the Association Between Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) and Differentiated Thyroid Carcinoma (DTC)
The Lowess Regression Curve for the Association Between Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) and Differentiated Thyroid Carcinoma (DTC)

4. Discussion

In this study, the researchers investigated the association between HOMA parameters and differentiated thyroid cancer in an Iranian population. The study suggested that insulin resistance is more prevalent in patients with DTC. As revealed earlier, the researchers classified participants to 2 groups based on HOMA-IR value in the present study. The odds of DTC in the subjects with insulin resistance (HOMA-IR ≥ 2.5) was significantly (4-fold) higher than that of the subjects with HOMA-IR < 2.5. These results showed an association between IR and differentiated thyroid cancer.

Recently a controversial topic in the field of thyroid disease is the association of IR with thyroid nodule and goiter. These studies have reported the association of IR and thyroid morphological changes (32-37). However, there is limited data on the association between IR and thyroid cancer (12, 19). The current study results are consistent with the results of Rezzonico’s study. In this study, insulin resistance was observed in 50% of patients with DTC and only in 10% of the control group (12). In contrast, in Balkan’s study, no association was found between IR and thyroid carcinoma (19).

In the recent decades, the incidence of thyroid cancer has increased throughout the world (38-40). In a similar manner, prevalence of insulin resistance in parallel with obesity has increased in the recent years. Insulin resistance is a pathognomonic characteristic of subjects with metabolic syndrome, obesity, type 2 diabetes mellitus, polycystic ovarian disease, and pre-diabetes. Insulin resistance is characterized by an inadequate physiological response of peripheral tissues to insulin, which leads to hyperinsulinemia (41). Increased insulin levels, decrease the production of insulin-like growth factor-1 (IGF-1) binding proteins and increase levels of free IGF-1. The IGF-I displays potent antiapoptotic, cell-survival, and transforming activities. It has been observed that IGF-1 and its receptor are overexpressed in differentiated thyroid cancers suggesting that overexpression of these factors may contribute to thyroid tumorigenesis (18, 42). Insulin/IGF-1 signaling pathway is involved in modulation of thyroid gene expression, thyrocyte proliferation, differentiation, and malignant transformation (43, 44). IGF-1, IGF-2, IGF-1 receptor, and insulin receptor isoforms are over-expressed in thyroid cell precursors and obviously decreased in differentiating thyroid cells. Increased levels of insulin and IGF-I in the bloodstream increase the growth of the progenitor cells in the thyroid carcinoma (45). Genetic events, such as point mutation, generate an activation of the MAPK pathway, which usually induces cell proliferation and dedifferentiation (46, 47). It is probable that increased insulin levels due to hyperinsulinemia might also stimulate this MAPK pathway, thyroid proliferation and, finally, development of thyroid cancer.

Thyroid Stimulating Hormone is the most important factor regulating the growth and differentiation of thyroid cells and is a known mitogen in the presence of insulin in cell cultures. Furthermore, TSH could also reduce the apoptosis process in response to various factors (48). In the present study, the serum TSH level was not significantly different between the 2 groups and was within the normal range. Therefore, the results of this study are not explained with TSH.

Carcinogenesis is a complex process and some other pathologic mechanisms, such as vitamin D deficiency, variation in deodinase expression, and chronic inflammation, may play a role during thyroid carcinogenesis (49-51).

The present study had several limitations. First was the relative small number of participants. To reduce the sparse-data bias, the researchers performed a penalized conditional logistic regression analysis with a weekly informative prior (30, 32, 52). Second, selecting controls from the hospital staff may introduce selection bias. However, the exact source population of the cases in this study is difficult to specify, and any choice of controls only approximates the actual population. Third, this study was a cross-sectional observational study that cannot naturally establish the causal relationship between insulin resistance and thyroid cancer. Fourth, the glucose clamp method, the gold standard test, was not used, which is invasive and costly; instead, HOMA-IR was used as a surrogate measure of insulin resistance. In this regard, there is an acceptable correlation between results of insulin resistance based on HOMA and the euglycemic clamp. Thus, HOMA-IR is widely used in clinical and epidemiological studies to measure insulin resistance (53, 54).

Having a control group with no thyroid nodules, assessment of cytological and histopathological outcome of each thyroid nodule, anti-thyroid antibodies measurement, and matching of age, gender and body mass index were potential advantages of this study.

In conclusion, insulin resistance is more prevalent in patients with differentiated thyroid cancer and shows a significant association. Conducting studies with larger sample sizes and prospective design could enhance the power of future studies in this area. Further studies are needed to determine the contribution of this factor in thyroid cancer incidence.

Acknowledgements

References

  • 1.

    Davies L, Morris LG, Haymart M, Chen AY, Goldenberg D, Morris J, et al. American Association of Clinical Endocrinologists and American College of Endocrinology Disease State Clinical Review: The Increasing Incidence of Thyroid Cancer. Endocr Pract. 2015;21(6):686-96. [PubMed ID: 26135963]. https://doi.org/10.4158/EP14466.DSCR.

  • 2.

    Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61(2):69-90. [PubMed ID: 21296855]. https://doi.org/10.3322/caac.20107.

  • 3.

    Amphlett B, Lawson Z, Abdulrahman GJ, White C, Bailey R, Premawardhana LD, et al. Recent trends in the incidence, geographical distribution, and survival from thyroid cancer in Wales, 1985-2010. Thyroid. 2013;23(11):1470-8. [PubMed ID: 23488941]. https://doi.org/10.1089/thy.2012.0573.

  • 4.

    Colonna M, Uhry Z, Guizard AV, Delafosse P, Schvartz C, Belot A, et al. Recent trends in incidence, geographical distribution, and survival of papillary thyroid cancer in France. Cancer Epidemiol. 2015;39(4):511-8. [PubMed ID: 26003877]. https://doi.org/10.1016/j.canep.2015.04.015.

  • 5.

    Jung KW, Won YJ, Oh CM, Kong HJ, Lee DH, Lee KH, et al. Cancer Statistics in Korea: Incidence, Mortality, Survival, and Prevalence in 2014. Cancer Res Treat. 2017;49(2):292-305. [PubMed ID: 28279062]. https://doi.org/10.4143/crt.2017.118.

  • 6.

    Kent WD, Hall SF, Isotalo PA, Houlden RL, George RL, Groome PA. Increased incidence of differentiated thyroid carcinoma and detection of subclinical disease. CMAJ. 2007;177(11):1357-61. [PubMed ID: 18025426]. https://doi.org/10.1503/cmaj.061730.

  • 7.

    Enewold L, Zhu K, Ron E, Marrogi AJ, Stojadinovic A, Peoples GE, et al. Rising thyroid cancer incidence in the United States by demographic and tumor characteristics, 1980-2005. Cancer Epidemiol Biomarkers Prev. 2009;18(3):784-91. [PubMed ID: 19240234]. https://doi.org/10.1158/1055-9965.EPI-08-0960.

  • 8.

    Dom G, Tarabichi M, Unger K, Thomas G, Oczko-Wojciechowska M, Bogdanova T, et al. A gene expression signature distinguishes normal tissues of sporadic and radiation-induced papillary thyroid carcinomas. Br J Cancer. 2012;107(6):994-1000. [PubMed ID: 22828612]. https://doi.org/10.1038/bjc.2012.302.

  • 9.

    Meinhold CL, Ron E, Schonfeld SJ, Alexander BH, Freedman DM, Linet MS, et al. Nonradiation risk factors for thyroid cancer in the US Radiologic Technologists Study. Am J Epidemiol. 2010;171(2):242-52. [PubMed ID: 19951937]. https://doi.org/10.1093/aje/kwp354.

  • 10.

    Borena W, Stocks T, Jonsson H, Strohmaier S, Nagel G, Bjorge T, et al. Serum triglycerides and cancer risk in the metabolic syndrome and cancer (Me-Can) collaborative study. Cancer Causes Control. 2011;22(2):291-9. [PubMed ID: 21140204]. https://doi.org/10.1007/s10552-010-9697-0.

  • 11.

    Zhao ZG, Guo XG, Ba CX, Wang W, Yang YY, Wang J, et al. Overweight, obesity and thyroid cancer risk: a meta-analysis of cohort studies. J Int Med Res. 2012;40(6):2041-50. [PubMed ID: 23321160]. https://doi.org/10.1177/030006051204000601.

  • 12.

    Rezzonico JN, Rezzonico M, Pusiol E, Pitoia F, Niepomniszcze H. Increased prevalence of insulin resistance in patients with differentiated thyroid carcinoma. Metab Syndr Relat Disord. 2009;7(4):375-80. [PubMed ID: 19320560]. https://doi.org/10.1089/met.2008.0062.

  • 13.

    Biondi B, Arpaia D, Montuori P, Ciancia G, Ippolito S, Pettinato G, et al. Under the shadow of vesuvius: a risk for thyroid cancer? Thyroid. 2012;22(12):1296-7. [PubMed ID: 23083444]. https://doi.org/10.1089/thy.2012.0002.

  • 14.

    Pellegriti G, De Vathaire F, Scollo C, Attard M, Giordano C, Arena S, et al. Papillary thyroid cancer incidence in the volcanic area of Sicily. J Natl Cancer Inst. 2009;101(22):1575-83. [PubMed ID: 19893009]. https://doi.org/10.1093/jnci/djp354.

  • 15.

    Jung SK, Kim K, Tae K, Kong G, Kim MK. The effect of raw vegetable and fruit intake on thyroid cancer risk among women: a case-control study in South Korea. Br J Nutr. 2013;109(1):118-28. [PubMed ID: 22455656]. https://doi.org/10.1017/S0007114512000591.

  • 16.

    Clero E, Doyon F, Chungue V, Rachedi F, Boissin JL, Sebbag J, et al. Dietary patterns, goitrogenic food, and thyroid cancer: a case-control study in French Polynesia. Nutr Cancer. 2012;64(7):929-36. [PubMed ID: 23061901]. https://doi.org/10.1080/01635581.2012.713538.

  • 17.

    Bessesen DH. Update on obesity. J Clin Endocrinol Metab. 2008;93(6):2027-34. [PubMed ID: 18539769]. https://doi.org/10.1210/jc.2008-0520.

  • 18.

    Vella V, Sciacca L, Pandini G, Mineo R, Squatrito S, Vigneri R, et al. The IGF system in thyroid cancer: new concepts. Mol Pathol. 2001;54(3):121-4. [PubMed ID: 11376121].

  • 19.

    Balkan F, Onal ED, Usluogullari A, Tuzun D, Ozdemir D, Inancli SS, et al. "Is there any association between insulin resistance and thyroid cancer? : A case control study". Endocrine. 2014;45(1):55-60. [PubMed ID: 23564559]. https://doi.org/10.1007/s12020-013-9942-x.

  • 20.

    Spencer CA, Takeuchi M, Kazarosyan M. Current status and performance goals for serum thyrotropin (TSH) assays. Clin Chem. 1996;42(1):140-5. [PubMed ID: 8565217].

  • 21.

    Kratzsch J, Fiedler GM, Leichtle A, Brugel M, Buchbinder S, Otto L, et al. New reference intervals for thyrotropin and thyroid hormones based on National Academy of Clinical Biochemistry criteria and regular ultrasonography of the thyroid. Clin Chem. 2005;51(8):1480-6. [PubMed ID: 15961550]. https://doi.org/10.1373/clinchem.2004.047399.

  • 22.

    Delshad H, Mehran L, Azizi F. Appropriate iodine nutrition in Iran: 20 years of success. Acta Med Iran. 2010;48(6):361-6. [PubMed ID: 21287473].

  • 23.

    Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412-9. [PubMed ID: 3899825].

  • 24.

    Ascaso JF, Pardo S, Real JT, Lorente RI, Priego A, Carmena R. Diagnosing insulin resistance by simple quantitative methods in subjects with normal glucose metabolism. Diabetes Care. 2003;26(12):3320-5. [PubMed ID: 14633821].

  • 25.

    Yokoyama H, Emoto M, Fujiwara S, Motoyama K, Morioka T, Komatsu M, et al. Quantitative insulin sensitivity check index and the reciprocal index of homeostasis model assessment in normal range weight and moderately obese type 2 diabetic patients. Diabetes Care. 2003;26(8):2426-32. [PubMed ID: 12882874].

  • 26.

    Mansournia MA, Altman DG. Invited commentary: methodological issues in the design and analysis of randomised trials. Br J Sports Med. 2017. [PubMed ID: 28756393]. https://doi.org/10.1136/bjsports-2017-098245.

  • 27.

    Greenland S, Mansournia MA. Limitations of individual causal models, causal graphs, and ignorability assumptions, as illustrated by random confounding and design unfaithfulness. Eur J Epidemiol. 2015;30(10):1101-10. [PubMed ID: 25687168]. https://doi.org/10.1007/s10654-015-9995-7.

  • 28.

    Mansournia MA, Hernan MA, Greenland S. Matched designs and causal diagrams. Int J Epidemiol. 2013;42(3):860-9. [PubMed ID: 23918854]. https://doi.org/10.1093/ije/dyt083.

  • 29.

    Mansournia MA, Jewell NP, Greenland S. Case-control matching: effects, misconceptions, and recommendations. Eur J Epidemiol. 2017:[Epub ahead of print]. https://doi.org/10.1007/s10654-017-0325-0.

  • 30.

    Greenland S, Mansournia MA, Altman DG. Sparse data bias: a problem hiding in plain sight. BMJ. 2016;352. i1981. [PubMed ID: 27121591]. https://doi.org/10.1136/bmj.i1981.

  • 31.

    Mansournia MA, Geroldinger A, Greenland S, Heinze G. Separation in Logistic Regression - Causes, Consequences, and Control. Am J Epidemiol. 2017. [PubMed ID: 29020135]. https://doi.org/10.1093/aje/kwx299.

  • 32.

    Rezzonico J, Rezzonico M, Pusiol E, Pitoia F, Niepomniszcze H. Introducing the thyroid gland as another victim of the insulin resistance syndrome. Thyroid. 2008;18(4):461-4. [PubMed ID: 18346005]. https://doi.org/10.1089/thy.2007.0223.

  • 33.

    Junik R, Kozinski M, Debska-Kozinska K. Thyroid ultrasound in diabetic patients without overt thyroid disease. Acta Radiol. 2006;47(7):687-91. [PubMed ID: 16950706]. https://doi.org/10.1080/02841850600806308.

  • 34.

    Rezzonico J, Rezzonico M, Pusiol E, Pitoia F, Niepomniszcze H. Metformin treatment for small benign thyroid nodules in patients with insulin resistance. Metab Syndr Relat Disord. 2011;9(1):69-75. [PubMed ID: 21128816]. https://doi.org/10.1089/met.2010.0026.

  • 35.

    Rezzonico J, Rezzonico M, Pusiol E, Pitoia F, Niepomniszcze H. High prevalence of thyroid nodules in patients with achrocordons (skin tags). Possible role of insulin-resistance. Medicina (B Aires). 2009;69(3):302-4. [PubMed ID: 19622476].

  • 36.

    Ayturk S, Gursoy A, Kut A, Anil C, Nar A, Tutuncu NB. Metabolic syndrome and its components are associated with increased thyroid volume and nodule prevalence in a mild-to-moderate iodine-deficient area. Eur J Endocrinol. 2009;161(4):599-605. [PubMed ID: 19633072]. https://doi.org/10.1530/EJE-09-0410.

  • 37.

    Heidari Z, Mashhadi MA, Nosratzehi S. Insulin Resistance in Patients with Benign Thyroid Nodules. Arch Iran Med. 2015;18(9):572-6. [PubMed ID: 26317597].

  • 38.

    Haggar FA, Preen DB, Pereira G, Holman CD, Einarsdottir K. Cancer incidence and mortality trends in Australian adolescents and young adults, 1982-2007. BMC Cancer. 2012;12:151. [PubMed ID: 22520938]. https://doi.org/10.1186/1471-2407-12-151.

  • 39.

    Ellison LF, Wilkins K. Canadian trends in cancer prevalence. Health Rep. 2012;23(1):7-16. [PubMed ID: 22590801].

  • 40.

    Dal Maso L, Lise M, Zambon P, Falcini F, Crocetti E, Serraino D, et al. Incidence of thyroid cancer in Italy, 1991-2005: time trends and age-period-cohort effects. Ann Oncol. 2011;22(4):957-63. [PubMed ID: 20952599]. https://doi.org/10.1093/annonc/mdq467.

  • 41.

    Li C, Ford ES, McGuire LC, Mokdad AH, Little RR, Reaven GM. Trends in hyperinsulinemia among nondiabetic adults in the U.S. Diabetes Care. 2006;29(11):2396-402. [PubMed ID: 17065674]. https://doi.org/10.2337/dc06-0289.

  • 42.

    Gursoy A. Rising thyroid cancer incidence in the world might be related to insulin resistance. Med Hypotheses. 2010;74(1):35-6. [PubMed ID: 19720470]. https://doi.org/10.1016/j.mehy.2009.08.021.

  • 43.

    Kimura T, Van Keymeulen A, Golstein J, Fusco A, Dumont JE, Roger PP. Regulation of thyroid cell proliferation by TSH and other factors: a critical evaluation of in vitro models. Endocr Rev. 2001;22(5):631-56. [PubMed ID: 11588145]. https://doi.org/10.1210/edrv.22.5.0444.

  • 44.

    Riedemann J, Macaulay VM. IGF1R signalling and its inhibition. Endocr Relat Cancer. 2006;13 Suppl 1:S33-43. [PubMed ID: 17259557]. https://doi.org/10.1677/erc.1.01280.

  • 45.

    Malaguarnera R, Frasca F, Garozzo A, Giani F, Pandini G, Vella V, et al. Insulin receptor isoforms and insulin-like growth factor receptor in human follicular cell precursors from papillary thyroid cancer and normal thyroid. J Clin Endocrinol Metab. 2011;96(3):766-74. [PubMed ID: 21123448]. https://doi.org/10.1210/jc.2010-1255.

  • 46.

    Namba H, Nakashima M, Hayashi T, Hayashida N, Maeda S, Rogounovitch TI, et al. Clinical implication of hot spot BRAF mutation, V599E, in papillary thyroid cancers. J Clin Endocrinol Metab. 2003;88(9):4393-7. [PubMed ID: 12970315]. https://doi.org/10.1210/jc.2003-030305.

  • 47.

    Kimura ET, Nikiforova MN, Zhu Z, Knauf JA, Nikiforov YE, Fagin JA. High prevalence of BRAF mutations in thyroid cancer: genetic evidence for constitutive activation of the RET/PTC-RAS-BRAF signaling pathway in papillary thyroid carcinoma. Cancer Res. 2003;63(7):1454-7. [PubMed ID: 12670889].

  • 48.

    Rapoport B, Chazenbalk GD, Jaume JC, McLachlan SM. The thyrotropin (TSH) receptor: interaction with TSH and autoantibodies. Endocr Rev. 1998;19(6):673-716. [PubMed ID: 9861544]. https://doi.org/10.1210/edrv.19.6.0352.

  • 49.

    Freedman DM, Looker AC, Abnet CC, Linet MS, Graubard BI. Serum 25-hydroxyvitamin D and cancer mortality in the NHANES III study (1988-2006). Cancer Res. 2010;70(21):8587-97. [PubMed ID: 20847342]. https://doi.org/10.1158/0008-5472.CAN-10-1420.

  • 50.

    Macejova D, Ondkova S, Brtko J. Vitamin D3 affects expression of thyroid hormone receptor alpha and deiodinase activity in liver of MNU-treated Sprague-Dawley rats. Gen Physiol Biophys. 2009;28(4):363-70. [PubMed ID: 20097959].

  • 51.

    Sfanos KS, De Marzo AM. Prostate cancer and inflammation: the evidence. Histopathology. 2012;60(1):199-215. [PubMed ID: 22212087]. https://doi.org/10.1111/j.1365-2559.2011.04033.x.

  • 52.

    Greenland S, Mansournia MA. Penalization, bias reduction, and default priors in logistic and related categorical and survival regressions. Stat Med. 2015;34(23):3133-43. [PubMed ID: 26011599]. https://doi.org/10.1002/sim.6537.

  • 53.

    Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care. 2004;27(6):1487-95. [PubMed ID: 15161807].

  • 54.

    Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care. 1998;21(12):2191-2. [PubMed ID: 9839117].