Of Papillary Thyroid Cancer: a Nested Case–Control Study Huang Huang1, Jennifer Rusiecki2, Nan Zhao3, Yingtai Chen3,4, Shuangge Ma3, Herbert Yu3,5, Mary H

Published OnlineFirst April 4, 2017; DOI: 10.1158/1055-9965.EPI-16-0845

Research Article Cancer Epidemiology, Biomarkers Thyroid-Stimulating Hormone, Thyroid & Prevention Hormones, and Risk of Papillary Thyroid Cancer: A Nested Case–Control Study Huang Huang1, Jennifer Rusiecki2, Nan Zhao3, Yingtai Chen3,4, Shuangge Ma3, Herbert Yu3,5, Mary H. Ward6, Robert Udelsman7, and Yawei Zhang1,3

Abstract

Background: The effects of thyroid-stimulating hormone above the normal range were associated with an increased risk of (TSH) and thyroid hormones on the development of human PTC among men (OR, 1.96; 95% CI, 1.04–3.66) but not women. papillary thyroid cancer (PTC) remain poorly understood. The risk of PTC decreased with increasing TSH levels within the Methods: The study population consisted of 741 (341 wom- normal range among both men and women (Ptrend ¼ 0.0005 and en, 400 men) histologically confirmed PTC cases and 741 0.041, respectively). matched controls with prediagnostic serum samples stored in Conclusions: We found a significantly increased risk of PTC the Department of Defense Serum Repository. Concentrations associated with TSH levels below the normal range among of TSH, total T3, total T4, and free T4 were measured in serum women and with TSH levels above the normal range among samples. Conditional logistic regression models were used to men. An inverse association between PTC and TSH levels calculate ORs and 95% confidence intervals (CI). within the normal range was observed among both men and Results: The median time between blood draw and PTC diag- women. nosis was 1,454 days. Compared with the middle tertile of TSH Impact: These results could have significant clinical implica- levels within the normal range, serum TSH levels below the tions for physicians who are managing patients with abnormal normal range were associated with an elevated risk of PTC among thyroid functions and those with thyroidectomy. Cancer Epidemiol women (OR, 3.74; 95% CI, 1.53–9.19) but not men. TSH levels Biomarkers Prev; 26(8); 1–10. 2017 AACR.

Introduction cancer are poorly understood. The most well-established risk factors for thyroid cancer include increased age, female gender, Thyroid cancer has the highest prevalence of all endocrine exposure to ionizing radiation, history of benign thyroid disease, malignancies, and its incidence is increasing faster than any other and a family history of thyroid cancer (4–7). Recent studies have malignancy in both men and women (1). In the United States, identified higher body weight and height as risk factors for thyroid thyroid cancer is the ninth most common cancer, accounting for cancer (8, 9). 3.8% of all malignancies and 0.3% of all deaths from cancer (2). Thyroid-stimulating hormone (TSH) is the major growth The most common histologic type of thyroid cancer is papillary factor for thyroid cells and regulator of thyroid functions. It thyroid cancer (PTC), which accounts for more than 80% of all controls the processes that lead to increased thyroid hormone thyroid carcinomas (3). The causal factors underlying thyroid production and secretion (10). Blood concentrations of thyroid hormones [i.e., triiodothyronine (T3) and its prohormone thyroxine (T4)] inversely regulate the release of TSH through 1Department of Surgery, Yale School of Medicine, Yale Cancer Center, New a negative feedback loop at the pituitary levels. High TSH levels 2 Haven, Connecticut. Uniformed Services University of the Health Sciences, F. have been associated with PTC pathogenesis in a mouse model Edward Hebert School of Medicine, Department of Preventive Medicine & (11). Suppression of TSH is currently recommended to manage Biostatistics, Bethesda, Maryland. 3Yale School of Public Health, New Haven, Connecticut. 4Cancer Institute & Hospital, Peking Union Medical College, Chinese patients with differentiated thyroid cancer (DTC), which has Academy of Medical Sciences, Beijing, China. 5Epidemiology Program, Univer- shown benefits to patient survival (12). Thyroid hormones sity of Hawaii Cancer Center, Hawaii. 6Occupational and Environmental Epide- have also been suggested to have a tumor-promoting effect on miology Branch, Division of Cancer Epidemiology and Genetics, National Cancer several cancers, including pancreatic, breast, ovarian, and pros- 7 Institute, Rockville, Maryland. Endocrine Neoplasia Institute, Miami Cancer tate cancers (13). However, findings of epidemiologic studies Center, Miami, Florida. linking TSH and thyroid hormones to the risk of thyroid cancer Note: Supplementary data for this article are available at Cancer Epidemiology, have been inconsistent (14–40). Biomarkers & Prevention Online (http://cebp.aacrjournals.org/). The majority of early studies reported an increased risk of Corresponding Author: Yawei Zhang, Yale School of Medicine, Yale School of thyroid cancer associated with elevated TSH levels (14–30), Public Health, Yale Cancer Center, 60 College Street LEPH 440, New Haven, CT several studies found no significant association (31–39), and one 06520. Phone: 203-7856210; Fax: 203-7376023; E-mail: [email protected] reported a reduced risk (40). All studies that reported a positive doi: 10.1158/1055-9965.EPI-16-0845 association between TSH and thyroid cancer were cross-sectional 2017 American Association for Cancer Research. (14–29) or case–control studies (30). Therefore, the possibility of

www.aacrjournals.org OF1

Huang et al.

reverse causation or treatment effect could be of potential concern samples within 1 year of the control's matched case's midpoint of because the TSH levels were measured after diagnosis. There are their 4 samples. Controls were randomly selected with replace- only three previous prospective cohort studies. One reported a ment from the cohort of service members who were not diagnosed significantly reduced risk of thyroid cancer associated with ele- with any cancer (with the exception of non-melanoma skin vated TSH levels (40). Two smaller studies reported lower, but not cancer), as per query of the ACTUR. Controls were matched significant, TSH levels in thyroid cancer cases than in controls (38, one-to-one to cases by date of birth (1 year), gender, race/ 39). The relationship between thyroid hormones and risk of ethnicity (White, Black, Hispanic, and other), average date of the thyroid cancer has also been inconclusive (14–17, 37, 38, 40). selected 4 samples drawn (1 year), and component at diagnosis/ Two studies found that lower thyroid hormone levels were matching. Demographic and military characteristics for all cases associated with a higher risk of thyroid cancer (16, 17), whereas and controls were abstracted from DMSS. The DMSS now serves as the remaining five reported no association (14, 15, 37, 38, 40). the central repository of medical surveillance data for the U.S. In light of the inconclusive associations between TSH, thy- armed forces and contains longitudinal records which have been roid hormones, and thyroid cancer, we conducted a nested continuously updated since 1990. The system includes demo- case–control study using data from the Department of Defense graphic and military characteristics as well as military and medical (DoD) Automated Central Tumor Registry (ACTUR) and the experiences of service members throughout their military careers Defense Medical Surveillance System (DMSS), with prediag- (41). All study procedures were approved by the Uniformed nostic serum samples from the Department of Defense Serum Services University Institutional Review Board, the Walter Reed Repository (DoDSR) to investigate the associations of PTC with National Military Medical Center, the DoD Joint Pathology Cen- TSH and thyroid hormones [total T3 (TT3), total T4 (TT4), and ter, and The Human Investigation Committee of Yale University. free T4 (FT4)]. Measurement of TSH and thyroid hormones A calibrated Roche Cobas E601 analyzer was used to measure Materials and Methods the serum concentrations of TSH and thyroid hormones using Study population the manufacturer's reagents and calibrators. TSH was captured Our study population was U.S. military personnel who had between 2 monoclonal antibodies (one was biotinylated and serum samples stored in the DoDSR (41). These stored samples the other labeled with a ruthenium complex) which specific were leftover sera collected for the routine HIV test of military for sterically noninterfering epitopes of human TSH. TT3 and personnel. The DoDSR is maintained by the Armed Forces Health TT4 were dissociated from binding proteins using 8-anilino-1- Surveillance Center, U.S. Army Public Health Command. As of naphthalene sulfonic acid (ANS) and competed with the exoge- August 2013, the repository stored more than 55 million serum nous biotinylated T3 or T4 for binding to a T3- or T4-specific samples from more than 10 million individuals, most of whom antibody labeled with a ruthenium complex. FT4 directly com- were active-duty and reserve personnel. Serum samples from all peted with the exogenous biotinylated T4 for binding to a T4- military members were typically drawn at the time of service entry specific antibody labeled with a ruthenium complex. All the and, on average, every 2 years thereafter for mandatory HIV testing antibodies were captured by streptavidin-coated magnetic micro- (42, 43). particles, which were then magnetically captured by an electrode We designed an individually matched nested case–control and the application of voltage-induced emission of photons by study. Cases were identified by linkage of the ACTUR with the the ruthenium complex. The intensities of the luminescence were DoDSR database. The ACTUR was established in 1986 and is the inversely proportional to the serum concentrations of TSH and data collection and clinical tracking system for cancer cases thyroid hormones. The normal ranges for serum concentrations of diagnosed and treated at military treatment facilities among DoD TSH, TT3, TT4, and FT4 were 0.3–4.2 mU/mL, 79– 149 ng/dL, 5.0– beneficiaries, including active-duty military personnel, retired 10.6 mg/dL, and 0.80–1.80 ng/dL, respectively. All control samples military personnel, and their dependents. The registry includes were tested in the same batch as their matched case samples. On information on demographic variables, diagnostic factors, and the basis of results obtained from quality control samples (5%), tumor characteristics (44). Cases met the following criteria: (i) intrabatch coefficient of variation ranged from 3.9% to 7.7%. histologically confirmed with International Classification of Dis- eases for Oncology, third edition (ICD-O-3 for thyroid gland: Statistical analyses C739) histology codes 8021, 8050, 8052, 8130, 8260, 8290, Measurements of TSH and thyroid hormones failed in one 8330-8332, 8335, 8340-8346, 8450, 8452, and 8510; (ii) at least serum sample, leaving 741 pairs of PTC cases and matched 1.5-mL prediagnostic and 0.5-mL postdiagnostic serum samples controls included in the final analysis. The distributions of demo- stored in the DoDSR; (iii) diagnosis between 2000 and 2013; and graphic and military characteristics between cases and controls (iv) age 21 years or older at diagnosis. Cases with any prior cancers were compared by c2 tests. The correlations between TSH, TT3, (excluding non-melanoma skin cancer) recorded in the ACTUR at TT4, and FT4 were estimated using the Pearson correlation coeffi- the date of diagnosis of thyroid cancer were excluded from the cients. Given the individual-matched case–control design, con- study. A total of 800 eligible cases were identified. The histology of ditional logistic regression analyses were employed to calculate all reported thyroid cancer cases was abstracted from ACTUR. Of ORs and 95% confidence intervals (95% CI) for the association these eligible cases, 742 (92.8%) were PTC (ICD-O-3: 8050, 8260, between TSH, thyroid hormones, and PTC. Serum concentrations and 8340-8343). IDs for all cases were sent from ACTUR to Armed of TSH, TT3, TT4, and FT4 were divided into 3 categories on the Forces Health Surveillance Center via encrypted methods. Those basis of the normal range (below, within, and above the normal IDs were excluded from the eligible pool of controls. Controls' range). The normal range group was further categorized into eligibility criteria were: having at least 4 serum samples in DoDSR, tertiles on the basis of the distributions of serum concentrations and according to the matching criteria, the midpoint of those 4 among controls. Thus, there were 5 categories for each hormone:

OF2 Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 Cancer Epidemiology, Biomarkers & Prevention

TSH, Thyroid Hormone, and PTC

below the normal range, lower, medium, and higher levels within All serum samples were drawn between 83 and 4,232 days the normal range, and above the normal range. The middle tertile before the cases were diagnosed with PTC. As anticipated, there within the normal range was used as reference category for all were statistically significant strong positive correlations between analyses. All the conditional logistic regression models were TT3, TT4, and FT4 (r ¼ 0.68 for TT3 and TT4, P < 0.0001; r ¼ 0.40 adjusted for body mass index (BMI; <18.5, 18.5–24.9, 25– for TT3 and FT4, P < 0.0001; and r ¼ 0.52 for TT4 and FT4, P < 29.9, and 30 kg/m2) and branch of military service (army, air 0.0001, respectively). TSH was weakly but statistically significant- force, marines and coast guard, and navy). TT3, TT4, and FT4 ly correlated with TT3, TT4, and FT4 (r ¼0.06 for TSH and TT3, P models were also adjusted for serum concentrations of TSH. ¼ 0.022; r ¼0.17 for TSH and TT4, P < 0.0001; and r ¼0.19 for However, additional adjustment for serum concentrations of TT3, TSH and FT4, P < 0.0001, respectively). Female cases had lower TT4, and FT4 in TSH models did not result in material changes in mean TSH levels than their matched controls, whereas male cases the observed associations and thus were not included in the final had higher mean TSH levels than their matched controls. None of models. Dose–response relationship was further investigated these differences were statistically significant. We also observed using Ptrend, estimated by treating serum concentrations of nonsignificantly higher mean levels of thyroid hormones among TSH and thyroid hormones as continuous variables. Stratified female cases as compared with female controls, but thyroid analyses were performed by gender, histologic subtype [classical hormone levels were similar between cases and controls among PTC (ICD-O-3: 8050, 8260, 8341–8343) and follicular variant of men. PTC (ICD-O-3: 8340)], tumor size (10 and >10 mm), and years In the overall population, serum TSH levels below the normal between serum samples drawn and PTC diagnosis (<3, 3–6, and range were associated with a significantly increased risk of PTC >6 years, based on sample size). Sensitivity analysis was also (OR, 2.65; 95% CI, 1.27–5.52; Fig. 1) compared with the middle conducted among women younger than 50 years to see whether tertile of the normal range. Paradoxically, TSH levels above the estrogen impacts the associations among premenopausal women. normal range were also associated with an increased risk of PTC All tests were 2-sided with a of 0.05. Statistical analyses were (OR, 1.58; 95% CI, 0.97–2.56) with borderline significance. conducted using SAS software, version 9.3 (SAS Institute, Inc.). Serum concentrations of TT3, TT4, and FT4 below or above the normal range were not significantly related to an elevated risk of Results PTC. Within the normal ranges, the risk of PTC decreased with increasing TSH levels (P ¼ 0.0001) and with decreasing TT3 PTC cases were more likely to have served in the Army or Air trend levels (P ¼ 0.031), but no dose–response relationships were Force at time of diagnosis, whereas controls were more likely to trend observed for TT4 and FT4. have served in the Navy, Marines, or Coast Guard (Table 1). Cases TSH levels below the normal range were associated with had a slightly larger BMI than controls, but the difference was not increased risk of PTC among women (OR, 3.74; 95% CI, 1.53– statistically significant. Since the cases were individually matched 9.19) but not among men (OR, 1.07; 95% CI, 0.25–4.62; Fig. 2) to controls on the basis of age, gender, and race/ethnicity, the compared with the middle tertile of the normal range. In addition, distributions of these variables were similar between cases and an increased risk of PTC in relation to TSH levels above the normal controls. range was observed only among men (OR, 1.96; 95% CI, 1.04– 3.66) but not among women (OR, 1.09; 95% CI, 0.49–2.46). The risk of PTC decreased with increasing TSH levels within the Table 1. Distributions of selected characteristics among PTC cases and matched P ¼ controls normal range among both men and women ( trend 0.0005 Cases Controls and 0.041, respectively). However, lower TSH levels within the (N ¼ 741) (N ¼ 741) normal range were associated with an increased risk of PTC and n (%) n (%) P the association was stronger in women (OR, 1.53; 95% CI, 1.03– Age at diagnosis, y 2.28) than in men (OR, 1.17; 95% CI, 0.80–1.71). In contrast, <30 210 (28.3) 203 (27.4) higher TSH levels within the normal range were associated with a 30–39 309 (41.7) 322 (43.5) reduced risk of PTC among men (OR, 0.65; 95% CI, 0.44–0.95). 40–49 185 (25.0) 179 (24.2) An inverse association between TT3 levels above the normal range 50 37 (5.0) 37 (5.0) 0.92 Gender and risk of PTC was observed only among men (OR, 0.59; 95% CI, Male 400 (54.0) 400 (54.0) 0.36–0.98); while the risk of PTC increased with increasing serum Female 341 (46.0) 341 (46.0) 1.00 concentrations of TT3 among women (overall Ptrend ¼ 0.019), no Race significant associations with TT4 and FT4 were observed. White 467 (63.0) 466 (62.9) When the analyses were stratified by histologic subtype among Black 131 (17.7) 132 (17.8) men (Fig. 3), TSH levels above the normal range were borderline Hispanic 68 (9.2) 68 (9.2) fi Other 55 (7.4) 55 (7.4) signi cantly associated with an increased risk of classical PTC Unknown 20 (2.7) 20 (2.7) 1.00 (OR, 1.94; 95% CI, 1.00–3.77). A significantly inverse dose– BMI, kg/m2 response relationship was observed between TSH levels within <25 256 (34.6) 285 (38.5) the normal range and risk of classical PTC (Ptrend ¼ 0.0010) but 25–29.9 148 (20.0) 129 (17.4) not follicular variant PTC (Ptrend ¼ 0.16). TT3 levels above the 30 16 (2.2) 10 (1.4) normal range were significantly associated with a decreased risk of Missing 321 (43.3) 317 (42.8) 0.23 – Service classical PTC (OR, 0.53; 95% CI, 0.30 0.94). When the analyses Army 299 (40.4) 253 (34.1) were stratified by histologic subtype among women (Fig. 4), TSH Air Force 193 (26.1) 150 (20.2) levels below the normal range were associated with a significantly Marines and Coast Guard combined 64 (8.6) 91 (12.3) increased risk of classical PTC (OR, 2.72; 95% CI, 1.09–6.78). The Navy 185 (25.0) 247 (33.3) <0.0001 lower TSH levels within the normal range were associated with an

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 OF3

Huang et al.

Cases Controls OR* (95% CI)

TSH (μU/mL) <0.30 28 11 2.65 (1.27−5.52) 0.30−1.19 280 230 1.37 (1.04−1.79) 1.20−1.93 215 230 1.00 1.94−4.20 166 236 0.75 (0.56−1.00) >4.20 51 34 1.58 (0.97−2.56) P trend**(within the normal range) 0.0001 P trend**(overall) 0.90 Total T3***(ng/dL) <79 4 2 2.59 (0.45−14.80) 79−117 155 177 0.83 (0.61−1.12) 118−132 208 192 1.00 133−149 197 187 1.00 (0.75−1.34) >149 177 183 0.88 (0.64−1.21) P trend**(within the normal range) 0.031 P trend**(overall) 0.18 Total T4***(μg/dL) <5 4 3 1.12 (0.23−5.50) 5−7.7 209 196 1.21 (0.92−1.59) 7.8−8.8 203 232 1.00 8.9−10.6 214 224 1.06 (0.80−1.41) >10.6 110 86 1.39 (0.95−2.04) P trend**(within the normal range) 0.37 P trend**(overall) 0.46 Free T4***(ng/dL) <0.80 6 2 3.00 (0.56−16.20) 0.80−1.16 140 243 1.14 (0.86−1.50) 1.17−1.29 210 233 1.00 1.30−1.80 274 258 1.15 (0.88−1.52) >1.80 11 4 2.18 (0.59−8.05) P trend**(within the normal range) 0.88 P trend**(overall) 0.35

0.00 1.00 2.00 3.00 4.00 5.00 6.00

Figure 1. Risk of PTC associated with serum concentrations of TSH and thyroid hormones. , Conditional logistic regression, adjusted for BMI and branch of military service. , Estimated by continuous variables. , Additionally adjusted for serum concentration of TSH.

elevated risk of follicular variant of PTC (OR, 11.31; 95% CI, carcinoma (OR, 3.62; 95% CI, 0.85–15.48). The lower TSH 3.10–41.31). There was an increasing trend in risk of classical PTC levels within the normal range were associated with an elevated with increasing TT3 levels (Ptrend ¼ 0.021), but no statistically risk of PTC microcarcinoma (OR, 2.47; 95% CI, 1.10–5.55) but significant association between TT3 levels and risk of follicular not PTC >10 mm (OR, 1.30; 95% CI, 0.79–2.14). TSH levels variant of PTC was observed. within the normal range were inversely associated with PTC When the analyses were stratified by tumor size among men >10 mm (Ptrend ¼ 0.027). No significant associations were (Fig. 5), the higher TSH levels within the normal range were found for TT3, TT4, and FT4 for both PTC microcarcinoma associated with a reduced risk of PTC with tumor size greater and PTC > 10 mm among men and women. than 10 mm (OR, 0.57; 95% CI, 0.35–0.92) but not PTC Further stratified analyses were conducted by the years between microcarcinoma (10 mm; OR, 0.99; 95% CI, 0.47–2.12). serum samples drawn and PTC diagnosis (Supplementary Table TSH levels within the normal range were inversely associated S1). No clear pattern was observed with timing of the serum with PTC >10 mm (Ptrend ¼ 0.0003), whereas, there was no samples drawn, although the numbers of cases was small after trend for PTC microcarcinoma (Ptrend ¼ 0.21). When the stratification. analyses were stratified by tumor size among women (Fig. Sensitivity analyses showed that the associations between risk 6), TSH levels below the normal range were associated with of PTC and serum concentrations of TSH, TT3, TT4, and FT4 did asignificantly increased risk of PTC >10 mm (OR, 4.98; 95% CI, not change materially after restricting the analyses to women 1.30–19.06) and a nonsignificantly elevated risk of PTC micro- younger than 50 years.

OF4 Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 Cancer Epidemiology, Biomarkers & Prevention

TSH, Thyroid Hormone, and PTC

Male (n = 800) Female (n = 682)

Cases Controls OR* (95% CI) Cases Controls OR* (95% CI)

TSH (μU/mL) <0.30 4 4 1.07 (0.25−4.62) 24 7 3.74 (1.53−9.19) 0.30−1.19 131 109 1.17 (0.80−1.71) 149 121 1.53 (1.03−2.28) 1.20−1.93 125 123 1.00 90 107 1.00 1.94−4.20 103 145 0.65 (0.44−0.95) 63 91 0.81 (0.51−1.28) >4.20 37 19 1.96 (1.04−3.66) 14 15 1.09 (0.49−2.46) P trend**(within the normal range) 0.0005 0.041 P trend**(overall) 0.28 0.27 Total T3***(ng/dL) <79 3 2 2.00 (0.32−12.61) 1 0 ⎯ 79−117 96 101 0.84 (0.57−1.25) 59 76 0.80 (0.49−1.31) 118−132 133 119 1.00 75 73 1.00 133−149 116 101 1.10 (0.75−1.62) 81 86 0.93 (0.58−1.51) >149 52 77 0.59 (0.36−0.98) 125 106 1.15 (0.74−1.78) P trend**(within the normal range) 0.12 0.045 P 0.93 0.019 trend**(overall) Total T4***(μg/dL) <5 3 2 1.14 (0.15−8.87) 1 1 1.07 (0.06−19.05) 5−7.7 149 126 1.31 (0.92−1.84) 60 70 0.95 (0.58−1.54) 7.8−8.8 121 141 1.00 82 91 1.00 8.9−10.6 110 111 1.13 (0.77−1.68) 104 113 0.99 (0.65−1.52) >10.6 17 20 1.05 (0.48−2.28) 93 66 1.51 (0.93−2.43) P trend**(within the normal range) 1.00 0.71 P 0.55 0.053 trend**(overall) Free T4***(ng/dL) <0.80 30 ⎯ 3 2 2.08 (0.31−13.97) 0.80−1.16 110 99 1.19 (0.78−1.82) 130 144 1.02 (0.70−1.49) 1.17−1.29 107 121 1.00 103 112 1.00 1.30−1.80 176 177 1.04 (0.72−1.51) 98 81 1.33 (0.87−2.05) >1.80 4 3 2.26 (0.38−13.59) 71 3.58 (0.36−36.01) P 0.47 0.22 trend**(within the normal range) P 0.51 0.11 trend**(overall)

0.00 1.00 2.00 3.00 4.00 5.00 6.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00

Figure 2. Risk of PTC associated with serum concentrations of TSH and thyroid hormones, stratiﬁed by gender. , Conditional logistic regression, adjusted for BMI and branch of military service. , Estimated by continuous variables. , Additionally adjusted for serum concentration of TSH.

Discussion increased risk of PTC among men. There was an inverse associ- In this large case–control study based on prediagnostic serum ation between PTC and TSH levels within the normal range measures and with sufﬁcient power to stratify by gender, we found among both men and women. The observed associations varied that serum TSH levels below the normal range were associated somewhat by histologic subtypes (classical vs. follicular variant with an elevated risk of PTC among women but not men. TSH PTCs) and by tumor size (10 vs. >10 mm) among men and levels above the normal range were only associated with an women. The gender effect on the association between TSH and

Classical papillary (n = 324) Follicular variant papillary (n = 76)

Cases Controls OR* (95% CI) Cases Controls OR* (95% CI)

TSH (μU/mL) <0.30 2 4 0.77 (0.12−5.10) 2 4 1.78 (0.15−21.02) 0.30−1.19 105 109 1.07 (0.69−1.66) 26 109 1.55 (0.68−3.51) 1.20−1.93 95 123 1.00 30 123 1.00 1.94−4.20 87 145 0.60 (0.39−0.92) 16 145 0.84 (0.35−1.99) >4.20 35 19 1.94 (1.04−3.77) 2 19 1.58 (0.18−13.86) P trend**(within the normal range) 0.001 0.16 P trend**(overall) 0.28 0.24 Total T3***(ng/dL) <79 2 2 2.59 (0.21−31.25) 1 2 1.16 (0.05−26.46) 79−117 76 101 0.83 (0.53−1.29) 20 101 0.97 (0.39−2.41) 118−132 107 119 1.00 26 119 1.00 133−149 98 101 1.20 (0.78−1.83) 18 101 0.88 (0.34−2.28) >149 41 77 0.53 (0.30−0.94) 11 77 0.90 (0.27−3.03) P **(within the normal range) trend 0.067 1.00 P 0.8 0.93 trend**(overall) Total T4***(μg/dL) <5 3 2 1.76 (0.14−22.23) 0 2 - 5−7.7 120 126 1.32 (0.90−1.94) 29 126 1.15 (0.46−2.86) 7.8−8.8 98 141 1.00 23 141 1.00 8.9−10.6 89 111 1.07 (0.70−1.65) 21 111 1.61 (0.54−4.48) >10.6 14 20 1.18 (0.49−2.82) 3 20 0.42 (0.06−3.13) P trend**(within the normal range) 0.30 0.72 P 0.66 trend**(overall) 0.76 Free T4***(ng/dL) <0.80 30 ⎯ 0 0 ⎯ 0.80−1.16 88 99 1.07 (0.66−1.73) 22 99 1.74 (0.65−4.66) 1.17−1.29 89 121 1.00 18 121 1.00 1.30−1.80 141 177 0.99 (0.65−1.52) 35 177 1.23 (0.51−2.97) >1.80 3 3 1.71 (0.26−11.48) 13 ⎯ P 0.61 0.55 trend**(within the normal range) P 0.45 0.93 trend**(overall)

0.00 1.00 2.00 3.00 4.00 5.00 6.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00

Figure 3. Risk of PTC associated with serum concentrations of TSH and thyroid hormones among males, stratiﬁed by histological subtypes. , Conditional logistic regression, adjusted for BMI and branch of military service. , Estimated by continuous variables. , Additionally adjusted for serum concentration of TSH.

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 OF5

Huang et al.

Classical papillary (n = 275) Follicular variant papillary (n = 66)

Cases Controls OR* (95% CI) Cases Controls OR* (95% CI)

TSH (μU/mL) <0.30 21 7 2.72 (1.09−6.78) 3 7 ⎯ 0.30−1.19 109 121 1.09 (0.70−1.72) 40 121 11.31 (3.10−41.31) 1.20−1.93 80 107 1.00 10 107 1.00 1.94−4.20 52 91 0.69 (0.42−1.14) 11 91 2.31 (0.53−10.06) >4.20 12 15 0.93 (0.39−2.23) 2 15 4.92 (0.33−73.13) P trend**(within the normal range) 0.11 0.11 P trend**(overall) 0.37 0.21 Total T3***(ng/dL) <79 1 0 ⎯ 0 0 ⎯ 79−117 48 76 0.83 (0.48−1.42) 11 76 0.35 (0.07−1.80) 118−132 59 73 1.00 16 73 1.00 133−149 62 86 0.94 (0.56−1.60) 19 86 0.26 (0.04−1.50) >149 105 106 1.25 (0.76−2.03) 20 106 0.57 (0.14−2.39) P **(within the normal range) trend 0.13 0.38 P trend**(overall) 0.021 0.38 Total T4***(μg/dL) <5 1 1 1.20 (0.07−21.85) 0 1 ⎯ 5−7.7 53 70 1.14 (0.67−1.96) 7 70 0.25 (0.06−1.09) 7.8−8.8 65 91 1.00 17 91 1.00 8.9−10.6 83 113 1.12 (0.69−1.79) 21 113 0.63 (0.19−2.14) >10.6 72 66 1.55 (0.91−2.65) 21 66 1.03 (0.27−3.90) P trend**(within the normal range) 0.90 0.34 P trend**(overall) 0.13 0.34 Free T4***(ng/dL) <0.80 32 2.35 (0.34−16.07) 0 2 ⎯ 0.80−1.16 105 144 1.09 (0.72−1.65) 25 144 0.96 (0.30−3.04) 1.17−1.29 81 112 1.00 22 112 1.00 1.30−1.80 81 81 1.43 (0.89−2.29) 17 81 1.92 (0.49−7.54) >1.80 5 1 2.80 (0.27−28.91) 21 ⎯ P trend**(within the normal range) 0.32 0.24 P 0.2 0.24 trend**(overall)

0.00 1.00 2.00 3.00 4.00 5.00 6.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00

Figure 4. Risk of PTC associated with serum concentrations of TSH and thyroid hormones among females, stratiﬁed by histological subtypes. , Conditional logistic regression, adjusted for BMI and branch of military service. , Estimated by continuous variables. , Additionally adjusted for serum concentration of TSH.

PTC was only observed among classical PTC cases. TSH levels 520,000 healthy individuals of ages 35 to 69 years when recruited showed a stronger association with PTC with larger tumor size. A between 1992 and 1998 in 10 European countries. A total of 357 suggestive inverse association between higher TT3 levels and risk incident thyroid cancer cases (57 men and 300 women) diag- of PTC was observed among men. nosed during 1992 to 2009 and 767 matched controls were The inverse trends between TSH levels and risk of PTC observed included in the analyses. Blood samples were collected at enroll- in the present study were in accordance with results from a nested ment. This European study found an inverse dose–response case–control study within a large population-based prospective relationship between overall TSH levels and risk of differentiated cohort in Europe (40). The cohort consisted of approximately thyroid cancer. The years between sample collection and thyroid

Tumor size ≤10 mm (n = 115) Tumor size >10 mm (n = 262)

Cases Controls OR* (95% CI) Cases Controls OR* (95% CI)

TSH (μU/mL) <0.30 2 4 2.16 (0.17−27.82) 2 4 1.34 (0.15−12.03) 0.30−1.19 41 109 1.60 (0.74−3.49) 88 109 1.27 (0.79−2.05) 1.20−1.93 29 123 1.00 84 123 1.00 1.94−4.20 34 145 0.99 (0.47−2.12) 64 145 0.57 (0.35−0.92) >4.20 9 19 2.61 (0.71−9.54) 24 19 2.05 (0.91−4.62) P **(within the normal range) trend 0.21 0.0003 P **(overall) trend 0.76 0.57 Total T3***(ng/dL) <79 1 2 1.24 (0.05−32.13) 2 2 2.89 (0.25−33.68) 79−117 29 101 0.62 (0.28−1.39) 62 101 1.07 (0.64−1.76) 118−132 42 119 1.00 85 119 1.00 133−149 26 101 0.79 (0.37−1.69) 81 101 1.22 (0.75−1.97) >149 17 77 0.38 (0.13−1.11) 32 77 0.72 (0.38−1.35) P trend**(within the normal range) 0.93 0.24 P 0.19 trend**(overall) 0.46 Total T4***(μg/dL) <5 1 2 0.61 (0.03−14.40) 2 2 1.20 (0.06−22.60) 5−7.7 50 126 1.27 (0.63−2.56) 93 126 1.50 (0.96−2.34) 7.8−8.8 31 141 1.00 81 141 1.00 8.9−10.6 25 111 1.10 (0.45−2.66) 78 111 1.12 (0.69−1.81) >10.6 8 20 6.24 (0.78−49.92) 8 20 0.71 (0.25−1.99) P trend**(within the normal range) 0.51 0.2 P trend**(overall) 0.56 0.13 Free T4***(ng/dL) <0.80 10 ⎯ 2 0 ⎯ 0.80−1.16 37 99 1.44 (0.56−3.69) 69 99 0.96 (0.57−1.62) 1.17−1.29 29 121 1.00 72 121 1.00 1.30−1.80 45 177 0.99 (0.45−2.20) 118 177 0.97 (0.61−1.55) >1.80 3 3 ⎯ 13 0.16 (0.01−2.46) P trend**(within the normal range) 0.41 0.71 P 0.42 0.49 trend**(overall) 0.00 1.00 2.00 3.00 4.00 5.00 6.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00

Figure 5. Risk of PTC associated with serum concentrations of TSH and thyroid hormones among males, stratiﬁed by tumor size. , Conditional logistic regression, adjusted for BMI and branch of military service. , Estimated by continuous variables. , Additionally adjusted for serum concentration of TSH.

OF6 Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 Cancer Epidemiology, Biomarkers & Prevention

TSH, Thyroid Hormone, and PTC

Tumor size ≤ 10 mm (n = 120) Tumor size > 10 mm (n = 200)

Cases Controls OR* (95% CI) Cases Controls OR* (95% CI)

TSH (μU/mL) <0.30 77 3.62 (0.85−15.48) 14 7 4.98 (1.30−19.06) 0.30−1.19 44 121 2.47 (1.10−5.55) 94 121 1.30 (0.79−2.14) 1.20−1.93 29 107 1.00 57 107 1.00 1.94−4.20 34 91 1.60 (0.76−3.36) 28 91 0.53 (0.28−1.02) >4.20 6 15 0.87 (0.28−2.72) 6 15 1.25 (0.33−4.73) P **(within the normal range) trend 0.66 0.027 P trend**(overall) 0.55 0.26 Total T3***(ng/dL) <79 1 0 ⎯ 0 0 ⎯ 79−117 20 76 1.04 (0.47−2.30) 36 76 0.58 (0.28−1.18) 118−132 24 73 1.00 48 73 1.00 133−149 33 86 1.31 (0.56−3.05) 40 86 0.64 (0.32−1.27) >149 42 106 1.84 (0.85−4.02) 76 106 0.77 (0.41−1.42) P trend**(within the normal range) 0.55 0.032 P 0.17 0.12 trend**(overall) Total T4***(μg/dL) <5 1 1 ⎯ 0 1 ⎯ 5−7.7 17 70 0.59 (0.24−1.47) 39 70 1.06 (0.57−2.00) 7.8−8.8 36 91 1.00 46 91 1.00 8.9−10.6 39 113 0.99 (0.47−2.10) 55 113 0.89 (0.50−1.57) >10.6 27 66 0.94 (0.40−2.22) 59 66 1.85 (0.97−3.55) P trend**(within the normal range) 0.70 0.87 P trend**(overall) 0.56 0.097 Free T4***(ng/dL) <0.80 32 ⎯ 3 2 ⎯ 0.80−1.16 110 144 1.24 (0.64−2.38) 130 144 0.93 (0.55−1.55) 1.17−1.29 107 112 1.00 103 112 1.00 1.30−1.80 176 81 1.41 (0.69−2.90) 98 81 1.25 (0.68−2.30) >1.80 4 1 0.32 (0.01−8.96) 71 ⎯ P trend**(within the normal range) 0.34 0.61 P 0.6 0.26 trend**(overall)

0.00 1.00 2.00 3.00 4.00 5.00 6.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00

Figure 6. Risk of PTC associated with serum concentrations of TSH and thyroid hormones among females, stratiﬁed by tumor size. , Conditional logistic regression, adjusted for BMI and branch of military service. , Estimated by continuous variables. , Additionally adjusted for serum concentration of TSH.

cancer diagnosis were similar between the European study and between those nodules and TSH concentrations has not yet our study. However, as compared with the European study, our been determined (18). population was younger and healthier (45), with participants of TSH plays an important role in regulating thyroid function, ages 17 to 56 years at blood samples collection. In addition, our including increasing number, size, and secretory activity of thyr- study had a larger number of the male cases than the European ocytes, increasing thyroid blood flow, and increasing thyroid study, which provides sufficient power to examine the associa- hormone production and secretion from follicular thyroid cells tions among men. The present study observed inconsistent asso- (10). Classical TSH actions are mainly mediated through the Gas- ciations between TSH levels and risk of PTC among women as adenylyl cyclase-protein kinase A-cyclic adenosine monopho- compared to men, whereas the European study reported similar sphate (cAMP) pathway, which is associated with production of associations among men and women. There were 2 another thyroid hormones and proliferation of thyroid epithelial cells prospective studies with smaller sample size that investigated the (47). However, somatic mutations in thyroid epithelial cell can association between TSH and risk of thyroid cancer (38, 39). also activate the cAMP pathway, which facilitate the cell growth Although no significantly inverse association was observed in and clonal expansion, leading to the formation of an autono- these studies, both reported lower TSH levels among thyroid mously functioning thyroid adenoma. The adenoma can synthe- cancer cases than controls. size and secrete thyroid hormones autonomously, thereby sup- A previous meta-analysis showed that higher TSH levels were pressing TSH secretion (48). Therefore, constitutive activation of associated with an increased thyroid cancer risk (46). However, the cAMP pathway could be associated with an increased carci- all 22 studies included in the meta-analysis were cross-sectional nogenic potential and a decreased TSH level. Because of the studies and measured TSH levels after treatment of thyroid deprivation of TSH stimulation, the extranodular tissue would cancer began. The cross-sectional design could not clarify become quiescent. Depending on the iodine intake, growth whether elevated TSH levels preceded thyroid cancer diagnosis potential, and other factors, it may take months to a decade or or were effects of treatment. The low levels of thyroid hormones longer for an adenoma to grow large enough to cause hyperthy- due to dysfunction of the thyroid gland among thyroid cancer roidism (49). patients could cause the pituitary gland to release more TSH. In While the underlying mechanism of lower TSH levels increas- addition, higher TSH levels may promote the growth of already ing the risk of PTC is currently unclear, two genome-wide asso- initiated thyroid cancer, making the cancer larger and more ciation studies have found that 5 common variants [rs965513[A] easily diagnosed. Therefore, the positive association seen in the on 9q22.23, rs944289[T] and rs116909374[T] on 14q13.3, cross-sectional studies could be due to ascertainment bias (10). rs966423[C] on 2q35, and rs2439302[G] on 8p12] were associ- On the other hand, controls in these studies were always ated with both an increased risk of thyroid cancer and low TSH patients with thyroid nodules or patients undergoing surgical level (50, 51). According to Gudmundsson and colleagues, the 5 treatment for a suspicious thyroid tumor. Some nodules can variants could refer to genes FOXE1, NKX2-1, DIRC3, and NRG1. produce high levels of thyroid hormones, thus lowering TSH The FOXE1 gene can regulate the transcription of thyroglobulin levels (40). Many patients with thyroid cancer also had addi- and thyroperoxidase genes, and together with the NKX2-1 gene, tional benign thyroid nodules, and the mutual influence plays an essential role in thyroid gland formation, differentiation,

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 OF7

Huang et al.

and function (52). Although the function of the DIRC3 gene is information on several potential confounding factors, such as unknown, it is presumed to have tumor suppressor activity (53). ionizing radiation exposure, history of benign thyroid disease, The gene NRG1 encodes a signaling protein which mediates cell– family history of thyroid cancer, and smoking status. There were cell interactions and plays a critical role in the growth and also a high percentage of participants with missing BMI data, development of thyroid gland. The carriers of these 5 variants which may have led to insufficient adjustment for BMI. The lack of may be characterized by lower concentrations of TSH. The con- data on medication use precludes us from carrying out sensitivity sequence of the lower TSH levels may result in less differentiation analyses excluding people who were taking thyroid hormones. of the thyroid epithelium, causing a higher predisposition to Furthermore, the subgroup analyses, which were stratified by malignant cell transformation (51). years between samples collection and diagnosis, histologic sub- The present study observed positive associations in a dose– type, and tumor size, may have yielded unstable results due to the response manner between serum concentrations of TT3 and risk small subgroup counts. of PTC only among women. The observed inverse association In conclusion, the present study showed a significantly between TSH levels and risk of PTC among women still held after increased risk of PTC associated with TSH levels lower than the excluding those 50 years old. These findings suggest that women normal range among women and higher than the normal range may be more sensitive to the effect of TSH and thyroid hormones among men. The observed associations varied by histologic as compared with men. Possible explanations including effect subtype and tumor size. These results could have significant modification of estrogen and different exposure to endocrine- clinical implications for physicians who are managing patients disrupting chemicals (e.g., birth control pills and personal care with abnormal thyroid functions and those with thyroidectomy. products) by gender that need to be explored in future studies. Future studies are warranted to further understand these The present study also noted different associations of TSH and associations. thyroid hormones on risk of PTC by histologic subtype and tumor size. Lower TSH levels showed an association with increasing risk Disclosure of Potential Conflicts of Interest of the follicular variant of PTC and PTC > 10 mm, whereas higher R. Udelsman is a director at the Endocrine Neoplasia Institute (Miami Cancer TSH levels were associated with a decreased risk of classical PTC Institute). No potential conflicts of interest were disclosed by the other authors.. and PTC microcarcinoma. These associations may support the hypothesis that follicular variant of PTC and papillary microcar- Authors' Contributions cinoma are unique clinical entities with different etiologic profiles Conception and design: J. Rusiecki, R. Udelsman, Y. Zhang Development of methodology: J. Rusiecki, R. Udelsman, Y. Zhang (54, 55). fi Acquisition of data (provided animals, acquired and managed patients, In addition, the present study failed to nd a clear pattern provided facilities, etc.): J. Rusiecki, Y. Chen, Y. Zhang between TSH levels and risk of PTC with timing of the serum Analysis and interpretation of data (e.g., statistical analysis, biostatistics, samples drawn. The relevant time window in which TSH exerts computational analysis): H. Huang, J. Rusiecki, S. Ma, H. Yu, R. Udelsman, influence on development of thyroid cancer needs to be further Y. Zhang studied. Writing, review, and/or revision of the manuscript: H. Huang, J. Rusiecki, S. Ma, H. Yu, M.H. Ward, R. Udelsman, Y. Zhang The present study has several strengths. It included a relatively fi Administrative, technical, or material support (i.e., reporting or organizing large number of male cases, providing suf cient statistical power data, constructing databases): H. Huang, N. Zhao, Y. Zhang to investigate and compare the associations by gender, which is Study supervision: J. Rusiecki, Y. Zhang important because women are much more likely to develop thyroid cancer than men. The study population was composed Disclaimer entirely of U.S. active-duty military personnel, minimizing the The content of this publication is the sole responsibility of the authors(s) and potential differences in effects of TSH and thyroid hormones does not necessarily reflect the views or policies of the Uniformed Services among healthy people and people with high risk of thyroid University of the Health Sciences or the Department of Defense. cancer. Potential selection bias from difference in access to med- Grant Support ical care was also minimized for our study population. The serum This research was supported by the National Institutes of Health grant concentrations of TSH and thyroid hormones were prospectively fl R01ES020361 (to Y. Zhang and J. Rusiecki) and American Cancer Society grant assessed and were not in uenced by the disease process or RSGM-10-038-01-CCE (to Y. Zhang). treatment, which provided an opportunity to estimate potentially causal relationships between TSH, thyroid hormones, and thyroid Received October 25, 2016; revised January 11, 2017; accepted March 23, cancer. A limitation of this study is that there was a lack of 2017; published OnlineFirst April 4, 2017.

References 1. Chen AY, Jemal A, Ward EM. Increasing incidence of differentiated 4. Wartofsky L.Increasing world incidence of thyroid cancer: increased thyroid cancer in the United States, 1988–2005. Cancer 2009;115: detection or higher radiation exposure? Hormones (Athens) 2010; 3801–7. 9:103–8. 2. Howlader N, Noone AM, Krapcho M, Garshell J, Miller D, Altekruse SF, 5. Imaizumi M, Usa T, Tominaga T, Neriishi K, Akahoshi M, Nakashima E, et al. SEER cancer statistics review, 1975–2011. Bethesda, MD: National et al. Radiation dose-response relationships for thyroid nodules and Cancer Institute; 2014. Available from: http://seer.cancer.gov/csr/ autoimmune thyroid diseases in Hiroshima and Nagasaki atomic bomb 1975_2011/, based on November 2013 SEER data submission, posted to survivors 55–58 years after radiation exposure. JAMA 2006;295: the SEER web site, Apr 2014. 1011–22. 3. Meza R, Chang JT. Multistage carcinogenesis and the incidence of thyroid 6. Preston-Martin S, Franceschi S, Ron E, Negri E. Thyroid cancer pooled cancer in the US by sex, race, stage and histology. BMC Public Health analysis from 14 case-control studies: what have we learned? Cancer Causes 2015;15:789. Control 2003;14:787–9.

OF8 Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 Cancer Epidemiology, Biomarkers & Prevention

TSH, Thyroid Hormone, and PTC

7. Iribarren C, Haselkorn T, Tekawa IS, Friedman GD. Cohort study of thyroid 29. Boelaert K, Horacek J, Holder RL, Watkinson JC, Sheppard MC, Franklyn JA. cancer in a San Francisco Bay area population. Int J Cancer 2001;93: Serum thyrotropin concentration as a novel predictor of malignancy in 745–50. thyroid nodules investigated by fine-needle aspiration. J Clin Endocrinol 8. Kitahara CM, Platz EA, Freeman LEB, Hsing AW, Linet MS, Park Y, et al. Metab 2006;91:4295–301. Obesity and thyroid cancer risk among US men and women: a pooled 30. Kim HK, Yoon JH, Kim SJ, Cho JS, Kweon SS, Kang HC. Higher TSH level is analysis of five prospective studies. Cancer Epidemiol Biomarkers Prev a risk factor for differentiated thyroid cancer. Clin Endocrinol 2013; 2011;20:464–72. 78:472–7. 9. Rinaldi S, Lise M, Clavel-Chapelon F, Boutron-Ruault M-C, Guillas G, 31. Gerschpacher M, Gobl C, Anderwald C, Gessl A, Krebs M. Thyrotropin Overvad K, et al. Body size and risk of differentiated thyroid carcinomas: serum concentrations in patients with papillary thyroid microcancers. findings from the EPIC study. Int J Cancer 2012;131:E1004–14. Thyroid 2010;20:389–92. 10. McLeod DS. Thyrotropin in the development and management of differ- 32. Kim KW, Park YJ, Kim EH, Park SY, Park do J, Ahn SH, et al. Elevated risk of entiated thyroid cancer. Endocrinol Metab Clin North Am 2014;43: papillary thyroid cancer in Korean patients with Hashimoto's thyroiditis. 367–83. Head Neck 2011;33:691–5. 11. Franco AT, Malaguarnera R, Refetoff S, Liao XH, Lundsmith E, Kimura S, 33. Petric R, Perhavec A, Gazic B, Besic N. Preoperative serum thyroglobulin et al. Thyrotrophin receptor signaling dependence of Braf-induced thyroid concentration is an independent predictive factor of malignancy in follic- tumor initiation in mice. Proc Natl Acad Sci USA 2011;108:1615–20. ular neoplasms of the thyroid gland. J Surg Oncol 2012;105:351–6. 12. Jonklaas J, Sarlis NJ, Litofsky D, Ain KB, Bigos ST, Brierley JD, et al. 34. Polyzos SA, Kita M, Efstathiadou Z, Poulakos P, Slavakis A, Sofianou D, Outcomes of patients with differentiated thyroid carcinoma following et al. Serum thyrotropin concentration as a biochemical predictor of initial therapy. Thyroid 2006;16:1229–42. thyroid malignancy in patients presenting with thyroid nodules. J Cancer 13. Moeller LC, Fuhrer D. Thyroid hormone, thyroid hormone receptors, and Res Clin 2008;134:953–60. cancer: a clinical perspective. Endocr Relat Cancer 2013;20:R19–29. 35. Azizi G, Malchoff CD. Autoimmune thyroid disease: a risk factor for 14. Ye ZQ, Gu DN, Hu HY, Zhou YL, Hu XQ, Zhang XH. Hashimoto's thyroid cancer. Endocr Pract 2011;17:201–9. Thyroiditis, microcalcification and raised thyrotropin levels within normal 36. Castro MR, Espiritu RP, Bahn RS, Henry MR, Gharib H, Caraballo PJ, et al. range are associated with thyroid cancer. World J Surg Oncol 2013;11;56. Predictors of malignancy in patients with cytologically suspicious thyroid 15. Moon SS, Lee YS, Lee IK, Kim JG. Serum thyrotropin as a risk factor for nodules. Thyroid 2011;21:1191–8. thyroid malignancy in euthyroid subjects with thyroid micronodule. Head 37. Maia FFR, Matos PS, Silva BP, Pallone AT, Pavin EJ, Vassallo J, et al. Role of Neck 2012;34:949–52. ultrasound, clinical and scintigraphyc parameters to predict malignancy in 16. Gul K, Ozdemir D, Dirikoc A, Oguz A, Tuzun D, Baser H, et al. Are thyroid nodule. Head Neck Oncol 2011;3:17. endogenously lower serum thyroid hormones new predictors for thyroid 38. Hrafnkelsson J, Tulinius H, Kjeld M, Sigvaldason H, Jonasson JG. Serum malignancy in addition to higher serum thyrotropin? Endocrine 2010; thyroglobulin as a risk factor for thyroid carcinoma. Acta Oncol 2000; 37:253–60. 39:973–7. 17. Jonklaas J, Nsouli-Maktabi H, Soldin SJ. Endogenous thyrotropin and 39. Thoresen SO, Myking O, Glattre E, Rootwelt K, Andersen A, Foss OP. Serum triiodothyronine concentrations in individuals with thyroid cancer. Thy- thyroglobulin as a preclinical tumor-marker in subgroups of thyroid- roid 2008;18:943–52. Cancer. Br J Cancer 1988;57:105–8. 18. Zafon C, Obiols G, Mesa J. Preoperative TSH level and risk of thyroid cancer 40. Rinaldi S, Plummer M, Biessy C, Tsilidis KK, stergaard JN, Overvad K, et al. in patients with nodular thyroid disease: nodule size contribution. Endo- Thyroid-stimulating hormone, thyroglobulin, and thyroid hormones and crinol Nutr 2015;62:24–8. risk of differentiated thyroid carcinoma: the EPIC study. J Natl Cancer Inst 19. Zimny M, Selkinski I, Blasius S, Rink T, Schroth HJ, Grunwald F. Risk of 2014;106:dju097. malignancy in follicular thyroid neoplasm: predictive value of thyrotropin. 41. Rubertone MV, Brundage JF. The Defense Medical Surveillance System and Nuklearmedizin 2012;51:119–24. the Department of Defense serum repository: glimpses of the future of 20. Chiu HK, Sanda S, Fechner PY, Pihoker C. Correlation of TSH with the risk public health surveillance. Am J Public Health 2002;92:1900–4. of paediatric thyroid carcinoma. Clin Endocrinol (Oxf) 2012;77:316–22. 42. McGlynn KA, Quraishi SM, Graubard BI, Weber JP, Rubertone MV, Erick- 21. Dorange A, Triau S, Mucci-Hennekinne S, Bizon A, Laboureau-Soares S, son RL. Polychlorinated biphenyls and risk of testicular germ cell tumors. Illouz F, et al. An elevated level of TSH might be predictive of differentiated Cancer Res 2009;69:1901–9. thyroid cancer. Ann Endocrinol (Paris) 2011;72:513–21. 43. Perdue CL, Cost AA, Rubertone MV, Lindler LE, Ludwig SL. Description and 22. Haymart MR, Repplinger DJ, Leverson GE, Elson DF, Sippel RS, Jaume utilization of the United States Department of Defense serum repository: a JC, et al. Higher serum thyroid stimulating hormone level in thyroid review of published studies, 1985-2012. PLoS One 2015;10:e0114857. nodule patients is associated with greater risks of differentiated thyroid 44. Enewold L, Zhou J, McGlynn KA, Devesa SS, Shriver CD, Potter JF, et al. cancer and advanced tumor stage. J Clin Endocrinol Metab 2008; Racial variation in tumor stage at diagnosis among Department Df defense 93:809–14. beneficiaries. Cancer 2012;118:1397–403. 23. Fiore E, Rago T, Provenzale MA, Scutari M, Ugolini C, Basolo F, et al. L- 45. Bollinger MJ, Schmidt S, Pugh JA, Parsons HM, Copeland LA, Pugh MJ. thyroxine-treated patients with nodular goiter have lower serum TSH and Erosion of the healthy soldier effect in veterans of US military service in Iraq lower frequency of papillary thyroid cancer: results of a cross-sectional and Afghanistan. Popul Health Metr 2015;13:8. study on 27 914 patients. Endocr Relat Cancer 2010;17:231–9. 46. McLeod DSA, Watters KF, Carpenter AD, Ladenson PW, Cooper DS, Ding 24. Zafon C, Obiols G, Baena JA, Castellvi J, Dalama B, Mesa J. Preoperative EL. Thyrotropin and thyroid cancer diagnosis: a systematic review and thyrotropin serum concentrations gradually increase from benign thyroid dose-response meta-analysis. J Clin Endocrinol Metab 2012;97:2682–92. nodules to papillary thyroid microcarcinomas then to papillary thyroid 47. Kimura T, Van Keymeulen A, Golstein J, Fusco A, Dumont JE, Roger PP. cancers of larger size. J Thyroid Res 2012;2012:530721. Regulation of thyroid cell proliferation by TSH and other factors: a critical 25. Shi L, Li Y, Guan H, Li C, Shan Z, Teng W. Usefulness of serum thyrotropin evaluation of in vitro models. Endocr Rev 2001;22:631–56. for risk prediction of differentiated thyroid cancers does not apply to 48. Paschke R, Ludgate M. The thyrotropin receptor in thyroid diseases. N Engl J microcarcinomas: results of 1,870 Chinese patients with thyroid nodules. Med 1997;337:1675–81. Endocr J 2012;59:973–80. 49. Sandrock D, Olbricht T, Emrich D, Benker G, Reinwein D. Long-term 26. Jin J, Machekano R, McHenry CR. The utility of preoperative serum thyroid- follow-up in patients with autonomous thyroid adenoma. Acta Endocrinol stimulating hormone level for predicting malignant nodular thyroid (Copenh) 1993;128:51–5. disease. Am J Surg 2010;199:294–7. 50. Gudmundsson J, Sulem P, Gudbjartsson DF, Jonasson JG, Sigurdsson A, 27. Kim ES, Lim DJ, Baek KH, Lee JM, Kim MK, Kwon HS, et al. Thyroglobulin Bergthorsson JT, et al. Common variants on 9q22.33 and 14q13.3 pre- antibody is associated with increased cancer risk in thyroid nodules. dispose to thyroid cancer in European populations. Nat Genet 2009; Thyroid 2010;20:885–91. 41:460–4. 28. Nixon IJ, Ganly I, Hann LE, Lin O, Yu C, Brandt S, et al. Nomogram for 51. Gudmundsson J, Sulem P, Gudbjartsson DF, Jonasson JG, Masson G, He H, predicting malignancy in thyroid nodules using clinical, biochemical, et al. Discovery of common variants associated with low TSH levels and ultrasonographic, and cytologic features. Surgery 2010;148:1120–7. thyroid cancer risk. Nat Genet 2012;44:319–22.

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 OF9

Huang et al.

52. Zhuang Y, Wu W, Liu H, Shen W. Common genetic variants on FOXE1 54. Yu XM, Schneider DF, Leverson G, Chen H, Sippel RS. Follicular variant of contributes to thyroid cancer susceptibility: evidence based on 16 studies. papillary thyroid carcinoma is a unique clinical entity: a population-based Tumour Biol 2014;35:6159–66. study of 10,740 cases. Thyroid 2013;23:1263–8. 53. Kohler A, Chen B, Gemignani F, Elisei R, Romei C, Figlioli G, et al. Genome- 55. Zhang Y, Chen Y, Huang H, Sandler J, Dai M, Ma S, et al. Diagnostic wide association study on differentiated thyroid cancer. J Clin Endocrinol radiography exposure increases the risk for thyroid microcarcinoma: a Metab 2013;98:E1674–81. population-based case-control study. Eur J Cancer Prev 2015;24:439–46.

OF10 Cancer Epidemiol Biomarkers Prev; 26(8) August 2017 Cancer Epidemiology, Biomarkers & Prevention

Thyroid-Stimulating Hormone, Thyroid Hormones, and Risk of Papillary Thyroid Cancer: A Nested Case−Control Study

Huang Huang, Jennifer Rusiecki, Nan Zhao, et al.

Cancer Epidemiol Biomarkers Prev Published OnlineFirst April 4, 2017.

Updated version Access the most recent version of this article at: doi:10.1158/1055-9965.EPI-16-0845

Supplementary Access the most recent supplemental material at: Material http://cebp.aacrjournals.org/content/suppl/2017/04/04/1055-9965.EPI-16-0845.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cebp.aacrjournals.org/content/early/2017/07/12/1055-9965.EPI-16-0845. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.