Cancer Risks of Patients with Graves’ Disease Who Received Antithyroid Drugs as Initial Treatment: A Nationwide Population-Based Analysis

Abstract

Background:

Population-based studies that examine the associations between hyperthyroidism and cancer risk have yielded inconsistent results. It remains unclear whether the risks of different cancers increase in patients with Graves’ disease (GD) who received antithyroid drugs (ATDs) as initial treatment. We aimed to determine whether cancer risk increases in patients with GD, compared with controls.

Methods:

This nationwide retrospective cohort study utilized data from the National Health Information Database of South Korea. We included 29,502 patients aged >20 years with GD, who received ATDs as initial treatment, and 57,173 age- and sex-matched controls. The primary outcome was the incidence of various types of cancers. Hazard ratios (HRs) with confidence intervals (CIs) for cancer risk were estimated using Cox proportional hazards models. We also analyzed HR by follow-up period since the diagnosis of GD, accounting for surveillance effect.

Results:

The risk of biliary tract and pancreatic cancers (HR: 1.41, CI: 1.24–1.60), thyroid cancer (HR: 15.51, CI: 12.29–19.57), prostate cancer (HR: 1.48, CI: 1.28–1.71), and ovarian cancer (HR: 1.31, CI: 1.13–1.52) was elevated in the GD group than in the control group even after the first year of follow-up was excluded. The increased risk of these cancers persisted after a follow-up period of more than 5 years. The risk of thyroid cancer in patients with GD was higher during the initial follow-up period (1 to <2 years) (HR: 19.35, CI: 7.66–48.87) compared with that in the follow-up period exceeding 2 years. The cancer risk estimates remained significant after excluding patients with GD who underwent subsequent radioactive iodine therapy.

Conclusion:

In this large-scale population-based study, GD was associated with increased risks of biliary tract and pancreatic, prostate, ovarian, and thyroid cancers. The increased risk of thyroid cancer, particularly during the initial follow-up period, may be a surveillance effect.

Introduction

Graves’ disease (GD) is an autoimmune thyroid disease, accounting for the most common underlying disease of hyperthyroidism in iodine-sufficient populations.¹ The initial treatment options for GD consist of antithyroid drugs (ATDs), radioactive iodine (RAI) therapy, and thyroid surgery.² Historically, initial ATD treatment has been preferred in European and East Asian countries,³ but RAI has been preferred in the United States.⁴ However, ATDs currently constitute the most common initial treatment for GD worldwide,⁵ with the recent decrease in the use of RAI in the United States.⁶

Thyroid hormones play major roles in various physiological processes including growth, differentiation, development, and metabolism.⁷ They also modulate the various pathways implicated in tumorigenesis. In vitro and in vivo experiments have revealed effects of thyroid hormones on cancer cell proliferation. These hormones influence the oncogenic PI3K pathway, angiogenesis, and immunoreactivity.^8

–13 Given the biologically plausible roles of thyroid hormones in tumorigenesis, previous population-based studies have sought associations between hyperthyroidism and cancer development.^{14

–30} Some such studies have reported increased risks of breast,^{14,16,22,23,30} lung,^14,18,30 prostate,^17,18,30 gastrointestinal tract,²⁰ ovarian,²⁴ and thyroid cancers^15,16,19,25 in patients with hyperthyroidism but the results have been inconsistent.^21,26

–29 These studies also had certain limitations in that they examined the cancer risks of only a few specific organs and did not comprehensively adjust for confounding factors including smoking and alcohol consumption status, obesity, or comorbidities.^{15,16,18
–20,22,23,25} In addition, earlier studies have investigated cancer risks in patients who were primarily treated with RAI, excluding those treated with ATDs,^31
–33 which are the preferred treatments for GD worldwide. Thus, a comprehensive evaluation of the cancer risk in patients with GD who receive ATDs as primary treatment is required. This large-scale, retrospective cohort study aimed to investigate the risks of various cancers in patients with GD receiving ATDs as first-line treatment, compared with controls.

Methods

This study received approval from the Institutional Review Board of Myongji Hospital (IRB approval no. 2021–11–020–001). The requirement of informed patient consent was waived because the study was retrospective, and all data were deidentified and anonymized.

Data source

This study was based on the data recorded from 2003 to 2022 in the National Health Information Database (NHID) of South Korea. The NHID is a public database of the National Health Insurance Service (NHIS) of South Korea, that includes sociodemographic data, the responses to lifestyle questionnaires, national health screening examination information, diagnosis based on the International Classification of Disease (ICD) codes and information on any treatment provided to the entire Korean population.³⁴ Our cohort was established after considering all medical records in this nationally representative NHIS database.

Study population

Data from patients with GD diagnosed between January 2005 and December 2012 were extracted from the NHID. We defined patients with GD as those who had received ATDs for over 60 days, with ICD codes E050 or E058. This definition aligns with that used in previous epidemiological studies of GD in Korea.^35,36

A 2-year washout period (January 2003 to December 2004) was applied to exclude any previous history of GD or treatment for GD; thus, we identified newly diagnosed cases of GD. The exclusion criteria during the entire study period were as follows: patients aged <20 years or ≥80 years; those with thyroiditis, hypothyroidism, or any other thyroid disorder (E01–4, E051–4, E06); those for whom demographic data were missing; those with any history of cancer (‘C’ codes) before the GD index date; and those on continuous ATDs therapy for <60 days as primary treatment.

The control cohort was also extracted from the NHID using a random number generator. For each patient with GD, two GD-free individuals with available national health screening examination data were randomly selected and matched by age, sex, and index year. To prevent bias from difference in follow-up periods when tracking cancer occurrences, we matched the follow-up periods (index years) of both groups. We applied the identical exclusion criteria to the control cohort before matching them to the GD cohort.

Definitions of outcomes and covariates

Various cancer types were identified based on the diagnostic codes assigned by medical doctors that reflected the possible outcomes. A cancer ‘C’ code greatly affects the insurance cover and available health care benefits; the code is assigned after careful consideration, and the assignment is reliable. The primary outcomes were as follows: esophageal cancer (C15); stomach cancer (C16); lung cancer (C33–4); liver cancer (C22); biliary tract and pancreatic cancers (C24–5); renal cancer (C64, C67); thyroid cancer (C73); colorectal cancer (C19–21); hematological cancers including lymphoma, myeloma, and leukemia (C81–6, C90–5); prostate cancer (C61); breast cancer (C50); and ovarian, uterine, or cervical cancer (C51–8).

The common covariates for all cancers, including demographic and lifestyle data, and basic comorbidities, were considered. These included age (20–44 or 45–79 years), sex, income (lower, middle lower, middle upper, or upper), smoking status (none, ex, or current), alcohol consumption (none, mild-to-moderate, or heavy), the Charlson comorbidity index (0, 1, or ≥2), physical activity status (none, mild, or moderate), body mass index, systolic blood pressure, fasting glucose level, total cholesterol level, and RAI treatment. Anthropometric measurements and laboratory data were based on the national health screening examination data from the NHID and were obtained within 2 years of index date. The covariates were used to construct the final regression model. Follow-up was censored at the end of the study (December 31, 2022).

Statistical analysis

All data manipulation and statistical analyses were performed by a single independent analyst (J.Y.L) who had undergone specialized training by the Health Insurance Review and Assessment Institute focusing on the handling of big data.³⁷ Continuous variables were compared between the groups using the Wilcoxon rank sum test, and the proportions of all group variables were compared employing the chi-square test. The stratified log-rank test was used to compare the incidence rates of all cancers between the GD and control groups. Cox proportional hazards regression models were used to compute adjusted hazard ratios (HRs) when comparing the GD and control groups. Stratified analyses were performed according to younger/older age (<45 years, ≥45 years) and sex (male or female). In the sub-analyses, comparisons between patients with GD who did not undergo RAI therapy and control group were conducted to support the principal results. The risk of cancer was further analyzed by the duration of ATDs (<2 years, ≥2 years) to assess the possible association between exposure of ATDs and risk of cancer in GD cohort.

The primary outcomes were subjected to sensitivity testing based on the follow-up period since the diagnosis of GD (≥1 year, 1 to <2 years, 2 to <5 years, and ≥5 years) to account for the possibility of surveillance effect. All statistical analyses were conducted using SAS Enterprise Guide version 6.1 software (SAS Inc., Cary, NC, USA). p-Values of <0.05 were considered statistically signiﬁcant.

Results

Baseline characteristics

In total, 29,502 patients with GD and 57,173 age-, sex-, and index year-matched controls were included during the enrollment period (Fig. 1). Baseline characteristics of the subjects are shown in Table 1. The mean age of the subjects in both groups was 45.8 ± 11.9 years. The mean follow-up period for the entire cohort was 12.4 years.

FIG. 1.

Identifying process and selection of the study cohort. GD, Graves’ disease; ATD, antithyroid drugs.

Table 1.

Baseline Characteristics of Cohort in Each Study Group

Variable	Graves’ disease (n = 29,502)	Control (n = 57,173)	p-Value
Age	45.8 ± 11.9	45.8 ± 11.9	<0.125
Male sex	12,094 (41.0%)	23,519 (41.1%)	<0.689
Income
Lower	4368 (14.8%)	12,843 (22.5%)	<0.001
Middle lower	5686 (19.3%)	14,446 (25.3%)
Middle upper	7388 (25%)	13,736 (24.0%)
Upper	11,349 (38.5%)	15,247 (26.7%)
NA^a	711 (2.4%)	901 (1.5%)
Smoking			<0.001
None	19,126 (64.8%)	37,579 (65.7%)
Ex-smoker	3457 (11.7%)	4516 (7.9%)
Current smoker	6070 (20.6%)	13,675 (23.9%)
NA	849 (2.9%)	1403 (2.5%)
Alcohol			<0.001
None	18,020 (61.1%)	30,687 (53.7%)
Mild-to-moderate	8837 (30.0%)	18,408 (32.2%)
Heavy	1854 (6.3%)	6812 (11.9%)
NA	791 (2.7%)	1266 (2.2%)
CCI			<0.001
0	19,597 (66.4%)	43,493 (76.1%)
1	5357 (18.2%)	7803 (13.6%)
≥2	4548 (15.4%)	5877 (10.3%)
Physical activity			<0.001
None	4645 (15.7%)	9042 (15.8%)
Mild	20,480 (69.4%)	38,623 (67.6%)
Moderate	3557 (12.1%)	8175 (14.3%)
NA	820 (2.8%)	1333 (2.3%)
BMI	22.9 ± 3.0	23.6 ± 3.3	<0.001
Systolic BP (mmHg)	121.9 ± 14.8	122.1 ± 15.9	<0.169
Glucose (mg/dL)	97.9 ± 22.3	96.5 ± 24.4	<0.001
Cholesterol (mg/dL)	177.5 ± 44.0	195.6 ± 37.5	<0.001

CCI, Charlson comorbidity index; BMI, body mass index; BP, blood pressure.

Not available due to nonresponder during questionnaire.

Risk of cancer in patients with GD versus controls

During the follow-up period, 4880 (16.5%) cancer cases were reported in the GD group, and 5566 (9.7%) cancer cases were reported in the control group. In the GD group, there were 610 new liver cancers, 582 biliary tract and pancreatic cancers, 957 thyroid cancers, 416 prostate cancers, and 404 ovarian cancers (Table 2). The incidence rate per 100,000 person-years was 168.46 for liver cancer, 159.89 for biliary tract and pancreatic cancers, 265.02 for thyroid cancer, 281.75 for prostate cancer, and 188.86 for ovarian cancer in the GD group compared with 123.73, 95.61, 12.69, 177.42, and 140.01, respectively, in the control group. In multivariate Cox regression analysis, the risks of these cancers were significantly higher in the GD group.

Table 2.

Cancer Risk in Patients with Graves’ Disease Compared with Age- and Sex-Matched Controls in Whole Cohort

Cancer	Graves’ disease			Control			HR^a (CI)
Cancer	No.^b	PYs	IR^c	No.^b	PYs	IR^c	HR^a (CI)
Stomach	255	365,726	69.72	437	705,564	61.94	0.95 (0.80–1.12)
Esophageal	9	368,258	2.44	33	709,333	4.65	0.46 (0.21–1.02)
Liver	610	362,104	168.46	868	701,549	123.73	1.26 (1.13–1.41)
Lung	324	365,514	88.64	538	705,159	76.29	1.00 (0.85–1.15)
Biliary tract/pancreas	582	364,010	159.89	674	704,921	95.61	1.48 (1.32–1.67)
Hematological	75	367,669	20.40	115	708,622	16.23	1.03 (0.75–1.41)
Renal	203	366,749	55.35	319	707,443	45.09	1.18 (0.97–1.42)
Colorectal	108	367,211	29.41	224	707,396	31.67	0.84 (0.66–1.07)
Thyroid	957	361,110	265.02	90	708,957	12.69	18.45 (14.76–23.08)
Prostate	416	147,648	281.75	511	288,014	177.42	1.56 (1.35–1.79)
Ovarian	404	213,917	188.86	578	412,842	140.01	1.23 (1.07–1.41)
Breast	266	215,008	123.72	422	413,967	101.94	0.97 (0.83–1.15)
Uterus	223	215,988	103.25	337	415,655	81.08	1.11 (0.92–1.33)

Adjusted by age, sex, income, smoking, alcohol, CCI, physical activity, body mass index, BP, glucose level, and cholesterol level.

Number of events.

Incidence rate was calculated per 100,000 persons.

No., number; IR, incidence rate; PY, person-year; HR, hazard ratio; CI, confidence interval.

In the sex-stratified analysis, the risk estimates were generally similar to those of the entire population and the increased risk remained significant for both sexes (Table 3). The risk of liver and biliary tract and pancreatic cancers associated with GD was higher among females, whereas the risk of thyroid cancer associated with GD was higher among males.

Table 3.

Cancer Risk in Patients with Graves’ Disease and Controls According to Sex

Cancer	Group	Men				Women
Cancer	Group	No.^b	PYs	IR^c	HR (CI)	No.^b	PYs	IR^c	HR (CI)
Stomach	Control^a	229	289,709	79.04	1	208	415,856	50.02	1
	GD	133	149,854	88.75	1.00 (0.79–1.27)	122	215,871	56.52	0.91 (0.71–1.15)
Esophageal	Control	23	291,652	7.89	1	10	417,681	2.39	1
	GD	5	151,084	3.31	0.33 (0.11–0.99)	4	217,173	1.84	0.80 (0.24–2.68)
Liver	Control	480	287,336	167.05	1	388	414,213	93.67	1
	GD	270	148,230	182.15	1.12 (0.95–1.32)	340	213,873	158.97	1.44 (1.23–1.68)
Lung	Control	274	289,615	94.61	1	264	415,543	63.53	1
	GD	163	149,573	108.98	1.00 (0.76–1.17)	161	215,941	74.56	1.03 (0.84–1.27)
Biliary tract/pancreas	Control	320	289,562	110.51	1	354	415,359	85.23	1
	GD	222	149,399	148.60	1.2 (1.00–1.46)	360	214,611	167.75	1.73 (1.48–2.02)
Hematological	Control	50	291,421	17.16	1	65	417,200	15.58	1
	GD	41	150,730	27.20	1.26 (0.80–2.00)	34	216,939	15.67	0.86 (0.56–1.34)
Renal	Control	145	290,838	49.86	1	174	416,606	41.77	1
	GD	90	150,465	59.81	1.13 (0.85–1.51)	113	216,284	52.25	1.24 (0.97–1.61)
Colorectal	Control	113	290,799	38.86	1	111	416,597	26.64	1
	GD	53	150,575	35.20	0.88 (0.62–1.24)	55	216,636	25.39	0.80 (0.57–1.12)
Thyroid	Control	16	291,736	5.48	1	74	417,220	17.74	1
	GD	294	148,937	197.40	35.9 (20.84–61.84)	663	212,174	312.48	15.02 (11.74–19.21)
Prostate	Control	511	288,014	177.42	1	—	—	—	—
	GD	416	147,648	281.75	1.56 (1.35–1.79)	—	—	—	—
Ovarian	Control	—	—	—	—	578	412,842	140.01	1
	GD		—	—	—	404	213,917	188.86	1.23 (1.07–1.41)
Breast	Control	—	—	—	—	422	413,967	101.94	1
	GD		—	—	—	266	215,008	123.72	0.97 (0.83–1.15)
Uterus	Control	—	—	—	—	337	415,655	81.08	1
	GD		—	—	—	223	215,988	103.25	1.11 (0.92–1.33)

Control cohort was reference group.

Number of events.

Incidence rate was calculated per 100,000 persons.

GD, Graves’ disease.

Stratified analysis by age

Risk of cancer stratified by age group is presented in Supplementary Table S1. The risk of liver, biliary tract and pancreatic, thyroid, and prostate cancers associated with GD was increased in both age groups (<45 years, ≥45 years). However, the association between GD and ovarian cancer was significant only in the younger age group.

Sensitivity testing and sub-analysis

For cancers that were significant in primary analysis, we assessed the risks of developing those cancers based on the follow-up period since the diagnosis of GD (≥1 year, 1 to <2 years, 2 to <5 years, and ≥5 years) (Table 4). After excluding the first year of follow-up, GD-associated risk of biliary tract and pancreatic cancer (HR: 1.41, confidence interval [CI]: 1.24–1.60), thyroid cancer (HR: 15.51, CI: 12.29–19.57), prostate cancer (HR: 1.48, CI: 1.28–1.71) and ovarian cancer (HR: 1.31, CI: 1.13–1.52) remained significant, but the risk of liver cancer did not reach statistical significance. The HR for thyroid cancer was 19.35 (CI: 7.66–48.87) during the early follow-up period (1 to <2 years), but 14.38 (CI: 9.18–22.52) during 2 to <5 years period, and 15.86 (11.93–21.07) after ≥5 years. The increased risk for biliary tract and pancreatic, prostate, and ovarian cancers associated with GD persisted after >5 years of follow-up.

Table 4.

Cancer Risk in Patients with Graves’ Disease and Controls According to Follow-up Period Since the Diagnosis of Graves’ Disease

	GD		Control		HR (CI)
Follow-up period	No.	IR	No.	IR	HR (CI)
Liver
≥1 year	459	126.77	767	109.33	1.08 (0.95–1.22)
1 to < 2 years	39	67.70	71	63.16	1.03 (0.67–1.59)
2 to <5 years	149	103.43	251	89.25	1.07 (0.86–1.34)
≥5 years	271	75.15	445	63.70	1.09 (0.92–1.29)
Biliary tract and pancreatic
≥1 year	505	138.74	617	87.53	1.41 (1.24–1.60)
1 to < 2 years	37	64.07	31	27.47	2.22 (1.34–3.71)
2 to <5 years	109	75.42	157	55.58	1.08 (0.83–1.40)
≥5 years	359	99.00	429	61.08	1.48 (1.28–1.72)
Thyroid
≥1 year	765	211.87	85	11.99	15.51 (12.29–19.57)
1 to < 2 years	72	126.19	5	4.39	19.35 (7.66–48.87)
2 to <5 years	202	141.28	22	7.72	14.38 (9.18–22.52)
≥5 years	491	136.63	58	8.21	15.86 (11.93–21.07)
Prostate
≥1 year	380	257.39	491	170.48	1.48 (1.28–1.71)
1 to <2 years	22	94.70	38	82.80	1.02 (0.58–1.78)
2 to <5 years	99	169.72	101	87.85	1.93 (1.43–2.63)
≥5 years	259	176.18	352	122.64	1.41 (1.19–1.68)
Ovary
≥1 year	374	174.84	505	122.33	1.31 (1.13–1.52)
1 to <2 years	31	91.57	31	47.06	1.78 (1.01–3.13)
2 to <5 years	67	79.12	116	70.33	0.99 (0.71–1.37)
≥5 years	276	129.6	358	87.11	1.39 (1.17–1.64)

No., number of events; IR, incidence rate per 100,000 person-years.

In the analysis that excluded all cases who received RAI (n = 1214), the results were very similar to the primary analysis (Supplementary Table S2). The risks of liver cancer (HR: 1.27, CI: 1.13–1.41), biliary tract and pancreatic cancer (HR 1.49, CI: 1.32–1.67), prostate cancer (HR 1.55, CI: 1.35–1.78), and ovarian cancer (HR: 1.22, CI: 1.07–1.41) were significantly higher in the GD group than in the control group. The risk of thyroid cancer was slightly higher in analyses that excluded RAI (HR: 20.83, CI: 16.47–26.34) compared with the overall analysis (HR: 18.45, CI: 14.76–23.08).

In the analysis of HR based on the duration of treatment with ATDs, the risk of cancer was generally similar between patients who were treated for <2 years and those treated for ≥2 years in the GD group (Supplementary Table S3).

Discussion

In this nationwide, population-based, retrospective cohort study, we demonstrated elevated risks of biliary tract and pancreatic, thyroid, prostate, and ovarian cancers in patients with GD compared with controls. The increased cancer risk was also significant in most subgroup analyses by age and follow-up duration and remained significant even after excluding patients who received subsequent RAI.

Previous population-based studies have investigated the associations between hyperthyroidism and cancer risk, but the results have been inconsistent.^{14

–30,38} Possible reasons for these inconsistencies include differences in the definition of hyperthyroidism, characteristics of study participants, adjustments for various confounding factors, and types of cancers evaluated. In several prospective cohort studies based on measurement of thyroid hormones, hyperthyroid state was at overall higher risk of solid cancer,¹⁸ lung cancer,¹⁸ breast cancer,²³ and prostate cancer,^17,18 compared with those in a euthyroid state. In a Rotterdam prospective cohort study, a positive relationship was observed between free thyroxine levels and the risk of overall solid cancers, as well as lung and breast cancer.¹⁴ These studies propose a causal link between hyperthyroidism and cancer risk. In questionnaire-based studies, self-reported hyperthyroidism was associated with mortality of breast cancer³⁸ or risk of ovarian cancer.²⁴ However, other studies found no significant association between hyperthyroidism and risk of breast²¹ or liver cancer.²⁷ Similar to our study, previous register-based studies have shown an increased risk of solid cancer,³⁰ thyroid cancer,^15,16,19,25 breast cancer,^16,22,30 and gastrointestinal tract cancer²⁰ in patients with hyperthyroidism. Nonetheless, some found no such association between hyperthyroidism and thyroid cancer²⁸ or breast cancer,²⁹ and others even observed a decreased risk of colorectal cancer in patients with hyperthyroidism.²⁶ However, aforementioned studies have limitations, such as including only selected populations^{14,15,21,23,28} (predominantly middle aged to older or postmenopausal patients or only hospitalized patients), focusing on specific organ,^{20,21,23,25,27,28} using sample cohort data,¹⁹ and inadequate adjustment for confounding factors and comorbidities.^{15,16,18,20,22,25} The present study investigated the various types of cancer in patients with GD among an entire population using a long cohort design and comprehensively considered related factors.

The increased risk of cancer during the early follow-up period of GD likely reflects surveillance effect, while long-term elevated risks may indicate causal association between GD and cancer. In our study, the risk of liver cancer increased in patients with GD during the entire follow-up period, but the association was not significant after excluding the first year of follow-up. Diagnostic workup for accompanying symptoms such as weight loss and abnormal liver function test, as well as for side effects of ATDs may be attributable to the increased detection of liver cancer during the early follow-up period. However, the increased risk of biliary tract and pancreatic, prostate, and ovarian cancers remained significant even after excluding the first year of follow-up and continued to be significant for >5 years of follow-up period. These findings suggest that the association between GD and cancers cannot be solely attributed to the surveillance effect.

Notably, a strong increase in the risk of thyroid cancer was observed in patients with GD compared with that of other cancers, particularly during the early follow-up period similar to findings from previous register-based studies.^15,16,25 Our findings regarding risk of thyroid cancer should be interpreted with caution, considering the epidemiological features related to overdiagnosis of thyroid cancer. In Korea, a national screening program for other common malignancies implemented in the early 2000s has contributed to increased opportunistic examinations for thyroid cancer.³⁹ Thus, the increase in thyroid cancer incidence in many countries, including Korea, may be attributable to the increased use of ultrasound and imaging tools, enhancing the detection of indolent tumors.⁴⁰ For patients with GD, intensive examinations of the thyroid gland are likely performed given the thyroid-related symptoms and goiter, particularly in the early stages of disease. This heightened scrutiny may have contributed to a surge in the incidence of thyroid cancer in patients with GD.

The potential relationship between breast cancer and thyroid dysfunction has long been investigated. Although in vitro studies have shown that thyroid hormones stimulate breast cancer cells via nuclear receptor αvβ3 and interaction with estradiol.^1,41 The results of population-based studies and meta-analyses remain controversial.^{16,17,21
–23,42,43} In a 2022 meta-analysis, hyperthyroidism was associated with 1.2-fold higher risk of breast cancer, with the risk being greater in postmenopausal women.⁴² However, the included studies in the meta-analysis were heterogenous in terms of measure of risk estimate, treatment modality, number of patients, and the extent of adjusting clinical variables. Our study does not support an association between GD and breast cancer, regardless of age group (<45 years and ≥45 years). Further studies are needed that integrate both pathophysiological and clinical data, such as the use of female hormones, to establish a causal relationship between these diseases.

Regarding cancer risk according to treatment modalities in GD, previous studies have reported an elevated risk of cancer incidence and/or mortality in patients who were treated with RAI,^31,33 but not all studies have found a significant association.⁴⁴ Recent meta-analyses have shown that there was no increase in solid cancer mortality in patients across various treatment groups.⁴⁵ In patients treated with RAI, a dose–response relationship was observed, but the risk increased only at higher cumulative doses.⁴⁵ In this study, cancer risk estimates remained significant even after excluding patients who received subsequent RAI. In addition, cancer risks were generally comparable between those treated with ATDs for >2 years and those treated for <2 years. Taken together, the increased cancer risk in patients with GD is more likely due to the underlying disease rather than an effect of treatment modality.

This study has several strengths lacking in previous studies. First, we included patients with GD aged >20 years who received ATDs as primary treatment. This cohort is representative of patients with GD encountered in current real-world practice. Second, we derived the risks of various cancers in patients with GD unlike previous studies focusing on specific organs. Third, the various stratified analyses ensured the validity of data and reduced the risk of reverse causality. However, our work had several limitations. First, individual data on thyroid hormone tests were lacking; it was not possible to analyze the associations between changes in thyroid function and cancer incidence. Second, reproductive factors such as menopause or fertility status were not adjusted, although we attempted to overcome this limitation by performing an age-stratified analysis (<45 years, ≥45 years) to explore whether the risk varied by menopausal status. The possibility that unmeasured confounders were in play cannot be entirely ruled out, as is true of most observational studies. Third, we performed sensitivity analysis by follow-up period to account for surveillance effect and demonstrated long-term cancer risk; however, further studies that differentiate incidentally detected indolent and more advanced cancers, including molecular profiles and tumor stage, are required to confirm our findings. Finally, this study was performed primarily on a Korean population, our results may have limited generalizability to other population with different genetic susceptibility and environmental factors such as iodine intake.

In conclusion, patients with GD who received ATDs as initial treatment exhibited elevated risks of biliary tract and pancreatic, prostate, ovarian, and thyroid cancers. The increased risk remained significant after excluding the first year of follow-up and continued to be significant 5 years after the diagnosis of GD. The increased risk of thyroid cancer, particularly during the initial follow-up period, may reflect a surveillance effect.

Footnotes

Authors’ Contributions

J.Y.L.: Conceptualization, methodology, formal analysis, interpretation of data, and writing (original draft, review, editing). M.K.L.: Methodology, interpretation of data, and writing (review, editing). J.H.L.: Methodology, interpretation of data, and writing (review, editing). K.K.: Methodology, interpretation of data, and writing (review, editing). K.B.: Conceptualization, interpretation of data, and writing (review, editing). S.Y.S.: Conceptualization, data acquisition, methodology, interpretation of data, and writing (original draft, review, editing).

Role of the Funder

The funder was not involved in the design and conduct of the study.

Data Sharing Statement

Sharing the data underlying this study is restricted for authors because Korean National Health Insurance Service (NHIS) owns the data. Researchers can request access on the NHIS website (https://nhiss.nhis.or.kr). Further details about this process and a provision guide are available at ().

Author Disclosure Statement

The authors have declared no conflicts of interest.

Funding Information

J.Y.L. was supported by faculty grant from Myongji hospital (Grant no. 2201–03-04) and K.B. was supported by the Seoul National University Hospital contribution grant (Grant no. 3020230260).

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

References

Taylor

, Albrecht

, Scholz

, et al. Global epidemiology of hyperthyroidism and hypothyroidism. Nat Rev Endocrinol, 2018; 14(5):301–316; doi: 10.1038/nrendo.2018.18

Ross

, Burch

, Cooper

, et al. 2016 American thyroid association guidelines for diagnosis and management of hyperthyroidism and other causes of thyrotoxicosis. Thyroid, 2016; 26(10):1343–1421; doi: 10.1089/thy.2016.0229

Bartalena

, Burch

, Burman

, et al. A 2013 European survey of clinical practice patterns in the management of Graves’ disease. Clin Endocrinol (Oxf), 2016; 84(1):115–120; doi: 10.1111/cen.12688

Burch

, Burman

, Cooper

. A 2011 survey of clinical practice patterns in the management of Graves’ disease. J Clin Endocrinol Metab, 2012; 97(12):4549–4558; doi: 10.1210/jc.2012-2802

Lee

, Pearce

. Hyperthyroidism: A review. JAMA, 2023; 330(15):1472–1483; doi: 10.1001/jama.2023.19052

Brito

, Schilz

, Singh Ospina

, et al. Antithyroid drugs-the most common treatment for Graves’ disease in the United States: A nationwide population-based study. Thyroid, 2016; 26(8):1144–1145; doi: 10.1089/thy.2016.0222

Hulbert

. Thyroid hormones and their effects: A new perspective. Biol Rev Camb Philos Soc, 2000; 75(4):519–631; doi: 10.1017/s146479310000556x

Moeller

, Fuhrer

. Thyroid hormone, thyroid hormone receptors, and cancer: A clinical perspective. Endocr Relat Cancer, 2013; 20(2):R19–R29; doi: 10.1530/ERC-12-0219

Pinto

, Soares

, Ribatti

. Thyroid hormone as a regulator of tumor induced angiogenesis. Cancer Lett, 2011; 301(2):119–126; doi: 10.1016/j.canlet.2010.11.011

10.

Muller

, Giani

, Zhang

, et al. Does thyroid peroxidase provide an antigenic link between thyroid autoimmunity and breast cancer? Int J Cancer, 2014; 134(7):1706–1714; doi: 10.1002/ijc.28493

11.

Davis

, Goglia

, Leonard

. Nongenomic actions of thyroid hormone. Nat Rev Endocrinol, 2016; 12(2):111–121; doi: 10.1038/nrendo.2015.205

12.

Carmona-Cortes

, Rodriguez-Gomez

, Wangensteen

, et al. Effect of thyroid hormone-nitric oxide interaction on tumor growth, angiogenesis, and aminopeptidase activity in mice. Tumour Biol, 2014; 35(6):5519–5526; doi: 10.1007/s13277-014-1726-2

13.

Mousa

, Davis

, Mohamed

, et al. Pro-angiogenesis action of thyroid hormone and analogs in a three-dimensional in vitro microvascular endothelial sprouting model. Int Angiol, 2006; 25(4):407–413.

14.

Khan

, Chaker

, Ruiter

, et al. Thyroid function and cancer risk: The rotterdam study. J Clin Endocrinol Metab, 2016; 101(12):5030–5036; doi: 10.1210/jc.2016-2104

15.

Shu

, Ji

, Li

, et al. Cancer risk in patients hospitalised for Graves’ disease: A population-based cohort study in Sweden. Br J Cancer, 2010; 102(9):1397–1399; doi: 10.1038/sj.bjc.6605624

16.

Chen

, Lin

, Chang

, et al. Cancer risk in patients with Graves’ disease: A nationwide cohort study. Thyroid, 2013; 23(7):879–884; doi: 10.1089/thy.2012.0568

17.

Chan

, Knuiman

, Divitini

, et al. Lower TSH and higher free thyroxine predict incidence of prostate but not breast, colorectal or lung cancer. Eur J Endocrinol, 2017; 177(4):297–308; doi: 10.1530/EJE-17-0197

18.

Hellevik

, Asvold

, Bjoro

, et al. Thyroid function and cancer risk: A prospective population study. Cancer Epidemiol Biomarkers Prev, 2009; 18(2):570–574; doi: 10.1158/1055-9965.EPI-08-0911

19.

Choi

, Han

, Cho

, et al. Cancer and mortality risks of Graves’ disease in South Korea based on national data from 2010 to 2019. Clin Epidemiol, 2023; 15:535–546; doi: 10.2147/CLEP.S406361

20.

Kirkegard

, Farkas

, Jorgensen

JOL

, et al. Hyper- and hypothyroidism and gastrointestinal cancer risk: A Danish nationwide cohort study. Endocr Connect, 2018; 7(11):1129–1135; doi: 10.1530/EC-18-0258

21.

Simon

, Tang

, Bernstein

, et al. Do thyroid disorders increase the risk of breast cancer? Cancer Epidemiol Biomarkers Prev, 2002; 11(12):1574–1578.

22.

Sogaard

, Farkas

, Ehrenstein

, et al. Hypothyroidism and hyperthyroidism and breast cancer risk: A nationwide cohort study. Eur J Endocrinol, 2016; 174(4):409–414; doi: 10.1530/EJE-15-0989

23.

Tosovic

, Becker

, Bondeson

, et al. Prospectively measured thyroid hormones and thyroid peroxidase antibodies in relation to breast cancer risk. Int J Cancer, 2012; 131(9):2126–2133; doi: 10.1002/ijc.27470

24.

Ness

, Grisso

, Cottreau

, et al. Factors related to inflammation of the ovarian epithelium and risk of ovarian cancer. Epidemiology, 2000; 11(2):111–117; doi: 10.1097/00001648-200003000-00006

25.

Kitahara

, K Rmendiné Farkas

, Jørgensen

JOL

, et al. Benign thyroid diseases and risk of thyroid cancer: A nationwide cohort study. J Clin Endocrinol Metab, 2018; 103(6):2216–2224; doi: 10.1210/jc.2017-02599

26.

L’Heureux

, Wieland

, Weng

, et al. Association between thyroid disorders and colorectal cancer risk in adult patients in Taiwan. JAMA Netw Open, 2019; 2(5):e193755; doi: 10.1001/jamanetworkopen.2019.3755

27.

Hassan

, Kaseb

, Li

, et al. Association between hypothyroidism and hepatocellular carcinoma: A case-control study in the United States. Hepatology, 2009; 49(5):1563–1570; doi: 10.1002/hep.22793

28.

Balasubramaniam

, Ron

, Gridley

, et al. Association between benign thyroid and endocrine disorders and subsequent risk of thyroid cancer among 4.5 million U.S. male veterans. J Clin Endocrinol Metab, 2012; 97(8):2661–2669; doi: 10.1210/jc.2011-2996

29.

Bach

, Kostev

, Schiffmann

, et al. Association between thyroid gland diseases and breast cancer: A case-control study. Breast Cancer Res Treat, 2020; 182(1):207–213; doi: 10.1007/s10549-020-05675-6

30.

Riis

, Bonnema

, Brix

, et al. Hyperthyroidism and the risk of non-thyroid cancer: A Danish register-based long-term follow-up study. Eur Thyroid J, 2024; 13(2); doi: 10.1530/ETJ-23-0181

31.

Metso

, Auvinen

, Huhtala

, et al. Increased cancer incidence after radioiodine treatment for hyperthyroidism. Cancer, 2007; 109(10):1972–1979; doi: 10.1002/cncr.22635

32.

Franklyn

, Maisonneuve

, Sheppard

, et al. Cancer incidence and mortality after radioiodine treatment for hyperthyroidism: A population-based cohort study. Lancet, 1999; 353(9170):2111–2115; doi: 10.1016/S0140-6736(98)12295-X

33.

Ryodi

, Metso

, Jaatinen

, et al. Cancer incidence and mortality in patients treated either with RAI or thyroidectomy for hyperthyroidism. J Clin Endocrinol Metab, 2015; 100(10):3710–3717; doi: 10.1210/jc.2015-1874

34.

Cheol Seong

, Kim

, Khang

, et al. Data resource profile: The national health information database of the National Health Insurance Service in South Korea. Int J Epidemiol, 2017; 46(3):799–800; doi: 10.1093/ije/dyw253i

35.

Kwon

, Jung

, Han

, et al. Prevalence and annual incidence of thyroid disease in Korea from 2006 to 2015: A nationwide population-based cohort study. Endocrinol Metab (Seoul), 2018; 33(2):260–267; doi: 10.3803/EnM.2018.33.2.260

36.

Sohn

, Ahn

, Lee

, et al. Risk of non-thyroidal autoimmune diseases in patients with Graves’ disease: A nationwide retrospective cohort study. Rheumatology (Oxford), 2024; doi: 10.1093/rheumatology/kead708

37.

Lee

, Lee

, Park

, et al. Cohort profile: The National Health Insurance Service-National Sample Cohort (NHIS-NSC), South Korea. Int J Epidemiol, 2017; 46(2):e15; doi: 10.1093/ije/dyv319

38.

Journy

NMY

, Bernier

, Doody

, et al. Hyperthyroidism, hypothyroidism, and cause-specific mortality in a large cohort of women. Thyroid, 2017; 27(8):1001–1010; doi: 10.1089/thy.2017.0063

39.

Ahn

, Kim

, Welch

. Korea’s thyroid-cancer “epidemic”–screening and overdiagnosis. N Engl J Med, 2014; 371(19):1765–1767; doi: 10.1056/NEJMp1409841

40.

Pizzato

, Li

, Vignat

, et al. The epidemiological landscape of thyroid cancer worldwide: GLOBOCAN estimates for incidence and mortality rates in 2020. Lancet Diabetes Endocrinol, 2022; 10(4):264–272; doi: 10.1016/S2213-8587(22)00035-3

41.

Flamini

, Uzair

, Pennacchio

, et al. Thyroid hormone controls breast cancer cell movement via integrin alphaV/beta3/SRC/FAK/PI3-Kinases. Horm Canc, 2017; 8(1):16–27; doi: 10.1007/s12672-016-0280-3

42.

Tran

, Kitahara

, Leenhardt

, et al. The effect of thyroid dysfunction on breast cancer risk: An updated meta-analysis. Endocr Relat Cancer, 2023; 30(1); doi: 10.1530/ERC-22-0155

43.

Fang

, Yao

, Sun

, et al. Does thyroid dysfunction increase the risk of breast cancer? A systematic review and meta-analysis. J Endocrinol Invest, 2017; 40(10):1035–1047; doi: 10.1007/s40618-017-0679-x

44.

Gronich

, Lavi

, Rennert

, et al. Cancer risk after radioactive iodine treatment for hyperthyroidism: A cohort study. Thyroid, 2020; 30(2):243–250; doi: 10.1089/thy.2019.0205

45.

Kitahara

, Preston

, Sosa

, et al. Association of radioactive iodine, antithyroid drug, and surgical treatments with solid cancer mortality in patients with hyperthyroidism. JAMA Netw Open, 2020; 3(7):e209660; doi: 10.1001/jamanetworkopen.2020.9660