Endocrine DOI 10.1007/s12020-014-0342-7
ORIGINAL ARTICLE
The association of menstrual and reproductive factors with thyroid nodules in Chinese women older than 40 years of age Kun Wang • Yu Yang • Yang Wu • Jie Chen Danyu Zhang • Chao Liu
•
Received: 8 April 2014 / Accepted: 17 June 2014 Ó Springer Science+Business Media New York 2014
Abstract The purpose of the study was to explore the association of menstrual and reproductive factors with thyroid nodules in Chinese women older than 40 years of age. A questionnaire was completed by 6,571 women aged 40 years or older in a community-based epidemiological investigation of thyroid nodules conducted from June to November 2011 in Nanjing City. Thyroid nodules were measured by ultrasound. The Thyroid Imaging Reporting and Data System score was used to differentiate between benign and possibly malignant nodules. Menopausal age ([55 vs. \50 years: RR = 1.17, 95 % CI 1.00–1.34) and number of reproductive years ([40 vs. \35 years: RR = 1.12, 95 % CI 1.01–1.24) increased the risk of thyroid nodules, but were not associated with suspected malignant nodules. Women who experienced more pregnancies (C5 vs. B1: RR = 2.09, 95 % CI 1.79–2.40) and abortions (C3 vs. 0: RR = 1.61, 95 % CI 1.41–1.81) were prone to development of thyroid nodules, and more likely to form suspected malignant nodules (pregnancies, RR = 3.59, 95 % CI 1.60–7.20; abortions, RR = 2.36, 95 % CI 1.31–4.06). Furthermore, higher risks of thyroid nodules (RR = 1.36, 95 % CI 1.14–1.59) and suspected malignant nodules (RR = 2.80, 95 % CI 1.08–6.53) were Kun Wang and Yu Yang contributed equally to this study. K. Wang Y. Yang J. Chen D. Zhang C. Liu (&) Department of Endocrinology, The First Clinical College, Nanjing University of Chinese Medicine, Jiangsu Province Hospital on Integration of Chinese and Western Medicine, 138 Xianlin Dadao Road, Nanjing 210023, China e-mail:
[email protected] Y. Wu Department of Endocrinology, The First People’s Hospital of Changzhou, Third Affiliated Hospital of Suzhou University, 185 Juqian Road, Changzhou 213003, China
observed in women who had undergone artificial compared with natural abortion. Periods of elevated estrogen and progesterone levels in women, such as pregnancy, were the key occasions for occurrence of both benign and suspiciously malignant thyroid nodules, while longer lifetime length of exposure to female sex hormones might promote the growth of thyroid nodules. Keywords Thyroid nodules Menopause Reproductive factors TIRADS score
Introduction Thyroid nodules are more frequently observed in women [1, 2], although the mechanism responsible for this phenomenon is not fully understood. The popular view is that thyroid nodules are attributable to multiple factors [3], such as high estrogen levels, autoimmune disorders, stress, environmental endocrine disruptors, excessive iodine intake [4], and other factors. Thus, estrogen may contribute to the higher incidence in females, as evidenced by the increased prevalence in pregnancy [5] and the slow growth of nodules after menopause [6]. However, asymptomatic and impalpable thyroid nodules were in the majority [7] and most nodules were undetected until they grew into palpable lesions during perimenopause [8], which may be responsible for the higher morbidity of thyroid nodules than those that occur in pregnancy [9]. Thyroid nodules found during perimenopause or postmenopause might have existed for several years, while the long-term process could cause changes in nodules including increased volume, hypoechoic lesions, irregular margins, intranodular vascularization, and calcification, which are regarded as signs that are suspicious of malignancy [10]. Although the occurrence of these signs
123
Endocrine
does not definitely indicate malignancy, even some suspected malignant nodules prove benign by fine-needle aspiration biopsy [11], these signs cause stress in patients because they are characteristic of cancer. Thyroid nodules suspected of being malignant are more often found during perimenopause and postmenopause rather than during premenopause when estrogen is relatively sufficient. This suggests that menopause might play an important role in the development of thyroid nodules. If so, then high estrogen levels alone are an inadequate explanation of the risk of thyroid cancer, even though several studies indicate that high estrogen levels increase the risk of thyroid cancer [12]. Because some studies show that the risk of thyroid cancer might be related to chronic exposure to estrogen [13], this raises the possibility that lifetime length of exposure to female sex hormones (menopausal age, number of reproductive years, pregnancies, and live births) might affect the development of perimenopausal thyroid nodules. To our knowledge, most studies have focused on the association of premenopausal endogenous hormones with thyroid cancer. Data on the association of menstrual and reproductive factors on the occurrence and growth of thyroid nodules, especially on cancer-prone nodules, are still limited. To explain the phenomena mentioned above, we used data from the Risk Evaluation of Cancers in Chinese Diabetic Individuals, a community-based longitudinal (REACTION) study [14], to investigate the association of menstrual and reproductive factors, including age at menarche, age at menopause, number of reproductive years, number of live births, and number of pregnancies on the risk of development and growth of thyroid nodules in perimenopausal and postmenopausal women.
Patients and methods Study population The REACTION study was designed to address the effects of different courses and treatments of diabetes on tumor risk in type 2 diabetes patients. As a multicenter, prospective, observational study, the baseline participants were screened from the general population aged 40 years or older in 25 communities of China. Details of the scientific rationale, eligibility requirements, and participant characteristics of the REACTION study have been published elsewhere [14, 15]. In addition to diabetes, each subcenter was authorized to perform additional research with regard to tumor or endocrine disease according to need. As one of the subcenters in Jiangsu province, we chose thyroid ultrasound in the Maigaoqiao community of
123
Nanjing City between June 1, 2011, and November 31, 2011 as an additional criterion to investigate the association between various metabolic parameters and thyroid disease, Nanjing City is not an iodine-deficient area, and the median urinary iodine of school-age children in this area is stable at 100–300 lg/L. Those who suffered from severe diseases such as chronic renal failure, hepatic cirrhosis, evident cardiac insufficiency, or abdominal ascites were excluded. Participants in our baseline study included 6,571 females and 3,479 males who were required to provide a recent medical examination report and complete a comprehensive face-to-face questionnaire survey on medical history, lifestyle, and psychological factors. The ethics committee of Shanghai Jiaotong University approved the study design. Study procedures were overseen by a research team from Ruijing Hospital, which is affiliated with Shanghai Jiaotong University School of Medicine. All participants gave informed and signed consent. The following participants were excluded from the original cohort of 6,571 for this analysis: 57 women who had history of thyroidectomy at baseline, including 26 thyroid cancers and 21 other diseases; 12 women who joined but provided incomplete information; 14 women who had no thyroid ultrasonography. After exclusions, 6,503 women remained for further analysis. Data collection At enrollment, participants completed a questionnaire to collect information on medical, family, menstrual, and reproductive history. In the medical history questionnaire, women were asked whether they had histories of benign or malignant tumor and if so, the age of onset and treatment including date of surgery and surgical site; whether they had histories of thyroid disease, including hyperthyroidism, hypothyroidism, thyroid nodules, thyroid cancer, or other thyroid disorders; and whether they had a history of diabetes or other metabolic disorders. In addition, women were also asked to provide valid documents, such as medical records, and radiological, laboratory, or pathology reports issued by trained physicians. In the menstrual history questionnaire, women were asked about age at menarche, age at menopause, and whether they experienced artificial menopause, which was defined as the age when the participant underwent removal of the uterus or bilateral ovarian cystectomy, and the reason for surgical removal including uterine fibroids, ovarian tumor, endometrial diseases, or other diseases. In addition, we estimated the number of reproductive years, defined as the difference between age at menopause and age at menarche. Women were asked about number of live births, number of pregnancies lasting at least 50 days, and number
Endocrine
of abortions, defined as termination of pregnancy at over the 50th day of gestation, including natural and artificial abortion. Furthermore, women were also asked whether they had ever taken oral contraceptives or menopausal hormones or both as well as the age and duration used. Measurements Neck ultrasound was performed to measure the morphology of the thyroid gland. We used a high frequency ultrasound (LOGIQ e portable ultrasound scanner, GE Healthcare Bio-Sciences, USA) with a 12 MHz transducer. Ultrasound exams were performed by the same radiologist with the individual in a supine position and with all clothing removed from around the neck, which was kept slightly extended. A thyroid nodule was defined by a lesion with a diameter larger than 3 mm; the size, location, morphology, echotexture, acoustic halo, blood flow, and calcification of thyroid nodules were documented. The echotexture was classified into hyperecho, isoecho, hypoecho, and non-echo; the acoustic halos were divided into non-halo, regular thin halos, and irregular thick halos; blood flow included intranodular and peripheral blood flow; calcifications were categorized into micro- and macrocalcifications, bright and granular calcification echo points less than 2 mm in diameter were defined as microcalcifications, and all other calcifications with a maximum diameter larger than 2 mm as well as irregular calcifications were considered to be macrocalcifications. All thyroid nodules were graded by the thyroid imaging reporting and data system (TIRADS) [16, 17], according to the ultrasonic characteristics. TIRADS was used to evaluate the risk of malignant thyroid nodules and was classified as follows: TIRADS 1, normal thyroid gland; TIRADS 2, benign lesions; TIRADS 3, probably benign lesions; TIRADS 4a: unclear lesion; TIRADS 4b–c: suspected lesions; TIRADS 5: probably malignant lesions; TIRADS 6: biopsy-proven malignancy. The details of the evaluation method are available in Refs. 16 and 17. Because fine-needle aspiration biopsy was not performed, thyroid nodules with a score of 6 were based on medical documentation. To differentiate more clearly between benign- and cancer-prone nodules, and analyze their association with menopausal and reproductive factors, we divided thyroid nodule scores into three categories according to the TIRADS score: no thyroid nodules (Score 1), probable benign nodules (Score 2–4a) and suspected malignant nodules (Score 4b–6). Statistical analysis Baseline characteristics and possible confounding factors by TIRADS Score are presented in Tables 1 and 2. The
Chi-square test was used to evaluate differences for categorical covariates, and one-way analysis of variance (ANOVA) was used for continuous variables. Multivariate logistic regression analysis was performed to estimate the odds ratios and 95 % confidence intervals for thyroid nodules (Score 2–6) by age at menarche and menopause, number of reproductive years, number of live births, number of pregnancies, number of abortions, and type of abortion in Table 3. Table 4 shows the regression coefficients and coefficient standard errors of variables for adjustment to estimate the weight of each covariant, when there were significant differences in the odds ratios and 95 % confidence intervals between before and after adjustment for control variables. The association of menstrual and reproductive exposure with suspected malignant nodules (Score 4b–6) is shown in Tables 5 and 6. In addition, we used the following formula to convert the odds ratio to the estimated relative risk [18]: RR ¼ OR=½ð1P0 Þ þ ðP0 ORÞ: In this formula, ‘‘P0’’ is the incidence of the outcome in the nonexposed population, ‘‘OR’’ is the odds ratio from a logistic regression equation, and ‘‘RR’’ is the estimated relative risk. The lower and upper confidence limits of the RR are also corrected by the 95 % CIs of the adjusted OR with this formula. All statistical evaluations were performed using SPSS software (version 13.0, SPSS, Chicago, IL, USA).
Results Characteristics of participants The characteristics of the 6,503 female participants are shown in Table 1. The 1,965 cases of thyroid nodule occurred in 30.2 % of women, and 172 cases of suspected malignant nodules occurred in 8 % of all women with thyroid nodules. The median ages for thyroid nodules cases and non-nodules subjects were 56.8 and 61.6 years, respectively; nonsignificant differences in age were found between participants with probable benign and suspected malignant nodules (P = 0.661). In general, for characteristics such as education level and history of cancer, there were significant differences between women with and without thyroid nodules, but no significant differences were seen between women with probable benign and suspected malignant nodules (P [ 0.05). Women with diabetes or elevated HbA1c were more prone to the development of thyroid nodules (P \ 0.001), and the HbA1c value of women with suspected malignant nodules was also higher than those who had benign nodules (P = 0.037); the median BMI of women with thyroid nodules was 24.79 kg/m2, greater than that of
123
Endocrine Table 1 Baseline characteristics by TIRADS Score Characteristics
Total
Thyroid nodule (by TIRADS score)a Pe
No
Yes
Score 1 n = 4,538 N (%)b
Score 2–6 n = 1,965 N (%)b
(2–4a) (n = 1,793) N (%)c
(4b–6) (n = 172) N (%)c
Pd
n = 6,503 N (%)b 58.25 ± 10.32
56.82 ± 10.26
61.59 ± 9.67
61.56 ± 9.64
61.90 ± 10.03
0.661
\0.001
Less than Primary School
1,656 (25.5)
1,140 (25.1)
516 (26.3)
477 (26.6)
39 (22.7)
Primary School Graduate
1,149 (17.7)
811 (17.9)
338 (17.2)
315 (17.6)
23 (13.4)
Junior Middle School Graduate
2,052 (31.6)
1,482 (32.7)
570 (29.0)
517 (28.8)
53 (30.8)
High School Graduate
1,332 (20.5)
927 (20.4)
405 (20.6)
358 (20.0)
47 (27.3)
College Graduate
314 (4.8)
178 (3.9)
136 (6.9)
126 (7.0)
10 (5.8)
0.126
\0.001
Never
6,291 (96.7)
4,387 (96.7)
1,904 (96.9)
1,739 (97.0)
165 (95.9)
Former Current
114 (1.8) 98 (1.5)
80 (1.6) 71 (1.8)
34 (1.4) 27 (1.7)
31 (1.7) 23 (1.3)
3 (1.7) 4 (2.3)
0.532
0.841
Never
6,417 (98.7)
4,480 (98.7)
1,973 (98.6)
1,766 (98.5)
171 (99.4)
Ever
0.634
Age at baseline (years), Mean ± SD Education level
Cigarette smoking status
Oral contraceptive use 86 (1.3)
58 (1.3)
28 (1.4)
27 (1.5)
1 (0.6)
0.328
Body mass index (kg/m2), Mean ± SD
24.63 ± 3.47
24.56 ± 3.49
24.79 ± 3.43
24.78 ± 3.96
24.97 ± 3.07
0.084
0.015
HbA1c (%), Mean ± SD
5.98 ± 0.95
5.93 ± 0.92
6.09 ± 0.99
6.08 ± 0.99
6.24 ± 0.99
0.037
\0.001
No
5,183 (79.7)
3,717 (81.9)
1,466 (74.6)
1,350 (75.3)
116 (67.4)
Yes (pre-existing and newly diagnosed)
1,320 (20.3)
821 (18.1)
499 (25.4)
443 (24.7)
56 (32.6)
0.024
\0.001
No
6,197 (95.3)
4,348 (95.8)
1,849 (94.1)
1,687 (94.1)
162 (94.2)
Yes
306 (4.7)
190 (4.2)
116 (5.9)
106 (5.9)
10 (5.8)
0.958
0.003
Diabetes diagnosed
History of cancer
Histories of other thyroid disease Thyroid cancer with thyroidectomyf
26
Thyroid cancer with conservative treatment
2 (0.03)
0
2 (0.1)
0
2 (1.2)
Hyperthyroidism
107 (1.6)
102 (2.2)
5 (0.3)
5 (0.3)
0
Hypothyroidism
164 (2.5)
98 (2.2)
66 (3.4)
58 (3.2)
8 (4.7)
Other
172 (2.6)
119 (2.6)
53 (2.7)
50 (2.8)
3 (1.7)
\0.001 \0.001
a
TIRADS, Thyroid Imaging Reporting and Data System; thyroid nodule ultrasound images were classified into three categories (TIRADS score: 1, 2–4a, 4b–6); score 1 (no thyroid nodules), score 2–6 (all thyroid nodules), score 2–4a (probable benign nodules), score 4b–6 (suspected malignant nodules)
b
Percent of all female participants
c
Percent of all female participants with thyroid nodules
d
P value for comparisons between score 2–4a and score 4b–6
e
P value for comparisons between score 0 and score 2–6
f
Excluded from statistical data
women without nodules (P = 0.015), and there were nonsignificant differences between subjects with benign and suspected malignant nodules in relation to oral contraceptive use (P = 0.328). Potentially confounding menstrual and reproductive factors are shown in Table 2. Menstruation was still occurring in
123
1,547 (23.8 %), while 4,565 (76.2 %) were menopausal. Compared with women without nodules, those with thyroid nodules were significantly more likely to have histories of uterine fibroids (P \ 0.001), but no difference was seen between women with probable benign and suspected malignant nodules (P = 0.932). Furthermore, women with histories
Endocrine Table 2 Possible confounding factors adjusted by TIRADS score Characteristics
Thyroid nodule (by TIRADS score)a
Total
Pe
No
Yes
n = 6,503 N (%)b
Score 1 n = 4,538 N (%)b
Score 2–6 n = 1,965 N (%)b
(2–4a) (n = 1,793) N (%)c
No
6,261 (96.3)
4,390 (96.7)
1,871 (95.2)
1,707 (95.2)
Yes
242 (3.7)
148 (3.3)
Premenopause
1,547 (23.8)
1,271 (28.0)
276 (14.0)
257 (14.3)
19 (11.0)
Natural menopause
4,565 (70.2)
2,995 (66.0)
1,570 (79.9)
1,429 (79.7)
141 (82.0)
391 (6.0)
272 (6.0)
119 (6.1)
107 (6.0)
12 (7.0)
3,055 (47.0)
2,352 (51.8)
703 (35.8)
36 (20.9)
667 (37.2)
3,448 (53.0)
2,186 (48.2)
1,262 (64.2)
136 (79.1)
1,226 (62.8)
Uterine fibroids
164 (2.5)
103 (2.3)
61 (3.1)
56 (3.1)
5 (2.9)
Oophorectomy
83 (1.3)
63 (1.4)
20 (1.0)
19 (1.1)
1 (0.6)
(4b–6) (n = 172) N (%)c
Pd
History of uterine fibroids 94 (4.8)
86 (4.8)
164 (95.3) 8 (4.7)
0.932
\0.001
0.455
\0.001
\0.001
\0.001
0.207
0.159
Menopause status
Artificial menopause Abortion experiences Never Ever Artificial menopause
Endometrial diseases
67 (1.0)
51 (1.1)
16 (0.8)
15 (0.8)
1 (0.6)
Other
77 (1.2)
55 (1.2)
22 (1.1)
17 (0.9)
5 (2.9)
a
TIRADS, Thyroid Imaging Reporting and Data System; thyroid nodule ultrasound images were classified into three categories (TIRADS score: 1, 2–4a, 4b–6); score 1 (no thyroid nodules), score 2–6 (all thyroid nodules), score 2–4a (probable benign nodules), score 4b–6 (suspected malignant nodules)
b
Percent of all female participants
c
Percent of all female participants with thyroid nodules
d
P value for comparisons between score 2–4a and score 4b–6
e
P value for comparisons between score 0 and score 2–6
Table 3 Adjusted and unadjusted relative risks and 95 % confidence intervals (CI) for TIRADS scores from 2 to 6 (all thyroid nodules) by reproductive and menstrual factors Exposure
TIRADS score ranged from 2 to 6 No. of casesa
RRb
95 % CI
RRc
95 % CI
\12 years 12–13 years
592/1,951 794/2,600
1.00 1.01
Reference 0.92, 1.10
1.00 0.99
(Reference)d,e,h (0.91, 1.09)d,e,h
[13 years
579/1,952
0.98
0.89, 1.07
0.98
(0.88, 1.08)d,e,h
\50 years
648/2,098
1.00
Reference
1.00
(Reference)d,e,h
50–55 years
893/2,462
1.17
1.08, 1.27
1.13
(1.04, 1.23)d,e,h
148/396
1.21
1.05, 1.38
1.17
(1.00, 1.34)d,e,h
Never menopause
276/1,547
1.00
Reference
1.00
(Reference)d,e,h
Natural menopause
1,570/4,565
1.93
1.75, 2.11
1.20
(1.03, 1.38)d,e,h
Artificial menopause
119/391
1.71
1.42, 2.02
0.93
(0.71, 1.20)d,e,h
562/1,816
1.00
Reference
1.00
(Reference)d,e,h
Menstrual exposure Age at menarche
Age at menopause
[55 years Type of menopause
1
Age at natural menopausei \50 years
123
Endocrine Table 3 continued Exposure
TIRADS score ranged from 2 to 6 No. of casesa
RRb
95 % CI
RRc
95 % CI
50–55 years
865/2,369
1.18
1.08, 1.28
1.15
(1.05, 1.25)d,e,h
[55 years
143/380
1.22
1.05, 1.39
1.19
(1.02, 1.37)d,e,h
\35 years
510/1,643
1.00
Reference
1.00
(Reference)d,e,h
35–40 years
723/2,049
1.14
1.04, 1.24
1.09
(0.99, 1.19)d,e,h
[40 years
456/1,264
1.16
1.05, 1.28
1.12
(1.01, 1.24)d,e,h
No
1,871/6,261
1.00
Reference
1.00
(Reference)d,h
Yes
94/242
1.30
1.10, 1.53
1.29
(1.07, 1.52)d,h
0
27/88
1.00
Reference
1.00
(Reference)e,g,h
1
846/3,345
0.82
0.57, 1.14
1.01
(0.71, 1.37)e,g,h
2 C3
574/1,658 518/1,412
1.13 1.20
0.81, 1.49 0.87, 1.56
0.91 0.69
(0.63, 1.26)e,g,h (0.33, 1.01)e,g,h
0–1
302/1,597
1.00
Reference
1.00
(Reference)e,g,h
2–3
963/3,203
1.58
1.43, 1.75
1.44
(1.26, 1.64)e,g,h
4
376/960
2.06
1.84, 2.30
1.89
(1.63, 2.17)e,g,h
C5
324/750
2.28
2.03, 2.53
2.09
(1.79, 2.40)e,g,h
0
703/3,055
1.00
Reference
1.00
(Reference)e,g
1
609/1,752
1.51
1.39, 1.64
1.49
(1.36, 1.63)e,g
2
441/1,159
1.65
1.51, 1.80
1.61
(1.46, 1.77)e,g
212/537
1.72
1.52, 1.92
1.61
(1.41, 1.81)e,g
Natural
105/344
1.00
Reference
1.00
(Reference)e,f,g
Artificial
1,157/3,104
1.22
1.03, 1.44
1.36
(1.14, 1.59)e,f,g
Number of reproductive yearsj
History of uterine fibroids
Reproductive exposure Number of live births
Number of pregnancies
Number of abortions
C3 k
Type of abortion
a
TIRADS score 2–6 cases to score 1–6 cases
b
Unadjusted
c
Adjusted for age at baseline, education, body mass index, HbA1c, diabetes, histories of thyroid disease, and histories of cancer
d
Additionally adjusted for number of pregnancies
e
Additionally adjusted for history of uterine fibroids
f
Additionally adjusted for number of abortions
g
Adjusted for age at menopause
h
Additionally adjusted for abortion experience
i
Restricted to female participants with natural menopause
j
Excludes female participants who were still menstruating
k
Restricted to female participants with histories of abortion; natural: restricted to those with only natural abortions, artificial: those with only artificial abortions or both natural and artificial abortions
1
Regression coefficients (B) and coefficients of standard error (S.E.) for each variable are shown in Table 4
of hypothyroidism were more likely to develop both thyroid nodules and suspected malignant nodules, while an inverse risk relation was observed in women with histories of hyperthyroidism (P \ 0.001). In addition, the number of artificial
123
menopause resulting from uterine fibroids was far more than artificial menopause from other causes (41.9 %, P \ 0.001).
Endocrine Table 4 Variables in the model for associations between thyroid nodules (TIRADS score 2–6) and type of menopause, with their regression coefficients (B), coefficient standard errors (SE), and adjusted relative risk (RR) Variables
SE
RR
P
Baseline variables to be controlled Age at baseline
0.050
0.005
1.05
\0.001
Education level
0.149
0.026
1.16
\0.001
Body mass index (BMI)
0.016
0.008
1.02
0.045
HbA1c
0.012
0.035
1.01
0.734
Diabetes diagnosed
-0.084
0.081
0.92
0.030
Histories of thyroid disease
-0.025
0.027
0.98
0.034
Histories of cancer
-0.242
0.131
0.78
0.035
-0.795 -0.034
0.169 0.034
0.61 0.97
\0.001 0.031
0.253
0.039
1.29
\0.001
1.00
0.037
Confounding variables to be controlled
Abortion experience
Pregnancies and abortions
TIRADS score 2–6 B
History of uterine fibroids Number of pregnancies
those with less than 35 years (RR = 1.12, 95 % CI 1.01–1.24).
As shown in Table 3, a twofold risk of thyroid nodules was observed in women with pregnancies of 5 or more versus that of 1 or less (RR = 2.09, 95 % CI 1.79–2.40), after additionally adjusting for abortion experience. Compared with women who never had an abortion, those who had three or more abortions had a 1.6-fold risk of thyroid nodules, adjusted for the number of pregnancies (RR = 1.61, 95 % CI 1.41–1.81). Women who had experienced artificial abortions had a higher thyroid nodule risk than those had only undergone natural abortions, after taking abortions number into consideration (RR = 1.36, 95 % CI 1.14–1.59). There were no evidence showing that live births numbers were associated with risk of thyroid nodules (C3 vs. 0, RR = 0.69, 95 % CI 0.33–1.01).
Independent variables Type of menopause Never menopause Natural menopause Artificial menopause
0.118
0.099
1.20
0.020
-0.215
0.162
0.93
0.197
B regression coefficient, SE coefficient of standard error, RR relative risk
Association of menstrual and reproductive exposures with development of thyroid nodules Menarche and menopause As shown in Table 3, there was a nonsignificant increased risk of thyroid nodule at an older menarcheal age after adjustment for significant control variables (C 14 vs. B11 years: RR = 0.98, 95 % CI 0.88–1.08). A significant increased risk of thyroid nodules was observed in women older than 55 years at natural menopause versus those whose natural menopausal age younger than 50 years, after taking the number of pregnancies into consideration (RR = 1.17, 95 % CI 1.39–s1.34). Compared with women who were still menstruating naturally menopausal women had a 1.2-fold risk of developing thyroid nodules (RR = 1.20, 95 % CI 1.03–1.38), which in artificial menopausal women increased to 1.7-fold (RR = 1.71, 95 % CI 1.42–2.02), but was reduced to 0.9-fold (RR = 0.93, 95 % CI 0.71–1.20) after adjusting for a history of uterine fibroids (The variables in the model for associations between thyroid nodules and type of menopause, with their regression coefficients, coefficient standard errors, and adjusted relative risk are shown in Table 4). Women with more than 40 reproductive years had a higher thyroid nodule risk than
Association of menstrual and reproductive exposures with development of suspected malignant thyroid nodules Menarche and menopause As shown in Table 5, menarcheal age, number of reproductive years, and type of menopause were not associated with a risk of suspected malignant thyroid nodules; however, compared with women of menopausal age younger than 50 years, a 1.6-fold risk of suspicious malignant thyroid nodules was found in those whose menopausal age was between 50 and 55 years (RR = 1.59, 95 % CI 1.13–2.23), while a nonsignificant risk was observed in those with a menopausal age older than 55 years (RR = 1.40, 95 % CI 0.77–2.46) after adjustment for control variables (The variables in the model for associations between suspected malignant thyroid nodules and age at menopause, with their regression coefficients, coefficient standard errors, and adjusted RR are shown in Table 6). Pregnancies and abortions As shown in Table 5, there was no evidence that the number of live births were associated with a risk of nodules suspected to be malignant. A 3.5-fold risk of suspected malignant nodules was observed in women with pregnancies of 5 or more versus those of 1 or less (RR = 3.59, 95 % CI 1.60–7.20), after additionally adjusting for abortion experience. Compared with women who had no abortion experience, those with abortions of three or more had a 2.3-fold risk of suspected malignant nodules, after adjusting for number of pregnancies (RR = 2.36, 95 % CI
123
Endocrine Table 5 Adjusted and unadjusted relative risk and 95 % confidence intervals (CI) for TIRADS score from 4b to 6 (suspected malignant thyroid nodules) by reproductive and menstrual factors in women with thyroid nodules Exposure
TIRADS score ranged from 4b to 6 No. of casesa
RRb
95 % CI
RRc
95 % CI
\12 years 12–13 years
45/592 76/794
1.00 0.86
Reference 0.58, 1.27
1.00 1.26
(Reference)d,e,h (0.88, 1.78)d,e,h
C14 years
51/579
1.09
0.77, 1.52
1.15
(0.78, 1.68)d,e,h
\50 years
43/648
1.00
Reference
1.00
(Reference)d,e,h
50–55 years
96/893
0.70
0.38, 1.26
1.59
(1.13, 2.23)d,e,h
[55 years
14/148
1.14
0.65, 1.94
1.40
(0.77, 2.46)d,e,h
Never menopause
19/276
1.00
Reference
1.00
(Reference)d,e,h
Natural menopause
141/1,570
1.30
0.82, 2.03
1.32
(0.74, 2.27)d,e,h
12/119
1.46
0.73, 2.80
1.51
(0.64, 3.27)d,e,h
\50 years
38/562
1.00
Reference
1.00
(Reference)d,e,h
50–55 years
89/865
1.26
1.12, 1.42
1.48
(1.02, 2.10)d,e,h
14/143
1.32
1.06, 1.62
1.42
(0.78, 2.50)d,e,h
37/510
1.00
Reference
1.00
(Reference)d,e,h
35–40 years
64/723
1.52
1.06, 2.14
1.17
(0.78, 1.71)d,e,h
[40 years
52/456
1.44
0.80, 2.51
1.49
(0.99, 2.20)d,e,h
No
164/1,871
1.00
Reference
1.00
(Reference)d,h
Yes
8/94
1.30
1.10, 1.53
1.23
(0.58, 2.52)d,h
0
2/27
1.00
Reference
1.00
(Reference)e,g,h
1
70/846
1.12
0.28, 3.78
1.00
(0.24, 3.53)e,g,h
2
52/574
1.22
0.30, 4.08
1.16
(0.28, 3.94)e,g,h
C3
48/518
1.25
0.31, 4.15
1.31
(0.30, 4.35)e,g,h
0–1
13/302
1.00
Reference
1.00
(Reference)e,g,h
2–3
74/963
1.79
1.01, 3.07
1.55
(0.77, 3.00)e,g,h
4 C5
35/376 50/324
2.16 3.58
1.18, 3.83 2.05, 5.94
1.99 3.59
(0.89, 4.21)e,g,h (1.60, 7.20)e,g,h
0
36/703
1.00
Reference
1.00
(Reference)e,g
1
48/609
1.54
1.01, 2.30
1.33
(0.85, 2.04)e,g
2
49/441
2.17
1.44, 3.20
1.66
(1.02, 2.63)e,g
39/212
3.59
2.38, 5.23
2.36
(1.31, 4.06)e,g
4/105
1.00
Reference
1.00
(Reference)e,f,g
Menstrual exposure Age at menarche
Age at menopause1
Type of menopause
Artificial menopause Age at natural menopause
i
[55 years j
Number of reproductive years \35 years
History of uterine fibroids
Reproductive exposure Number of live births
Number of pregnancies
Number of abortions
C3 k
Type of abortion Natural
123
Endocrine Table 5 continued Exposure
TIRADS score ranged from 4b to 6
Artificial a
No. of casesa
RRb
95 % CI
RRc
95 % CI
132/1,157
2.99
1.13, 7.94
2.80
(1.08, 6.53)e,f,g
TIRADS score 4b–6 cases to score 2–6 cases
b
Unadjusted
c
Adjusted for age at baseline, education, body mass index, HbA1c, diabetes, histories of thyroid disease, and histories of cancer
d
Additionally adjusted for number of pregnancies
e
Additionally adjusted for history of uterine fibroids
f
Additionally adjusted for number of abortions
g
Additionally adjusted for age at menopause
h
Additionally adjusted for abortion experience
i
Restricted to female participants with natural menopause Excludes female participants who was still menstruating
j k
Restricted to female participants with histories of abortion; natural: restricted to those with only natural abortions, artificial: those with only artificial abortions or both natural and artificial abortions
1
Regression coefficients (B) and coefficients of standard error (SE) for each variable are shown in Table 6
Table 6 Variables in the model for association between suspected malignant thyroid nodules (TIRADS score 4b–6) and age at menopause, with their regression coefficients (B), coefficient of standard errors (SE), and adjusted relative risk (RR) Variables
TIRADS score 2–6 B
Baseline variables to be controlled Age at baseline 0.830 Education level
0.161
SE
RR
P
0.114
1.97
\0.001
0.073
1.17
0.028
Body mass index (BMI)
0.018
0.025
1.02
0.467
HbA1c
0.466
0.192
1.97
0.003
Diabetes diagnosed
0.348
0.022
1.78
0.044
Histories of thyroid disease
0.435
0.075
1.04
0.037
Histories of cancer
0.081
0.373
1.08
0.829
Confounding variables to be controlled History of uterine fibroids
0.403
0.148
1.50
0.006
Number of pregnancies
0.233
0.087
1.14
0.026
-0.051
0.398
0.95
0.899
1.00
0.038
0.202 0.372
0.197 0.33
1.59 1.40
0.011 0.259
Abortion experience Independent variables Age at menopause \50 years 50–55 years [55 years
B regression coefficient, SE coefficient of standard error, RR relative risk
1.31–4.06). Women who had ever experienced artificial abortions had a higher risk of suspected malignant nodules than those had only undergone natural abortions, after taking the number of abortions into consideration (RR = 2.80, 95 % CI 1.08–6.53).
Discussion We found that older menopausal age and a greater number of reproductive years might increase the risk of thyroid nodules, but were not associated with suspected malignant nodules; women who had experienced more pregnancies and abortions were prone to development of thyroid nodules and suspected malignant nodules. Higher risks of thyroid nodules and suspected malignant nodules were also observed in women who had ever undergone artificial abortion, compared with those who had only experienced natural abortion. Estrogen might increase the proliferation of both benign and malignant human thyroid cells [19]; Xu et al. [20] found that estrogen was a potent stimulator of growth in thyroid stem and progenitor cells, which might be the origin of thyroid nodules in females. Meanwhile clinical data suggest that estrogen could induce an increase in serum thyroxine-binding globulin concentration, which might lead to elevated levels of total circulating thyroid hormones and corresponding decreased levels of free triiodothyronine and free thyroxine, while decreased levels often resulted in very small but significant increases in serum TSH concentrations [21]; the elevated TSH was proved to be associated with goiter and thyroid nodule. As evidence for an association between high levels of estrogen and thyroid nodules, an increased prevalence of thyroid nodules during pregnancy has been reported [5]. In fact, levels of both estrogen and progesterone are increased significantly throughout pregnancy. Due to the age limit of participants in our study, we used the number of pregnancies to evaluate the periods of high female sex hormone levels experienced in a lifetime; a higher risk of thyroid
123
Endocrine
nodules was observed in women with more pregnancies and abortions. Meanwhile, we also found that a longer length of exposure to estrogen increased the risk of thyroid nodules, while both menopausal age and number of reproductive years showed the effect of chronic exposure to estrogen in women. We suggest that periods of high estrogen levels, such as during pregnancy, are the key occasions for occurrence of thyroid nodules, while chronic exposure to estrogen in a lifetime might promote growth of nodules until menopause. The number of live births seemed not to be associated with a risk of thyroid nodules in our study. To explain these results, we had to take the population structure of our study into consideration. Due to the family planning policy that was implemented in China from the 1970s, the number of women with live births less than or equal to one accounted for 60.2 % of those whose baseline age was younger than 70 years. Similarly, those who had given birth to one child were obliged to perform an induced abortion in the event of pregnancy; therefore, the percentage of women who had an artificial abortion accounted for 48.4 % of all participants. Most of the induced abortions were performed at about the 50th day of gestation. Women who experienced artificial abortion were more prone to development of thyroid nodules than those who had only undergone natural abortion. This phenomenon might be explained by the different changes in female sex hormone levels in natural versus artificial abortion [22]: in cases of artificial abortion, levels of estrogen and progesterone were increased significantly during pregnancy and fell abruptly after induced abortion was performed and then returned to normal levels within 2 weeks, while with natural abortion, levels of estrogen and progesterone were increased slightly and fell gradually until miscarriage happened spontaneously. Furthermore, several studies indicate the potential role of estrogen receptor genetic variation in spontaneous abortion [23], which could impair the effect of estrogen on thyroid cells. A higher risk of development of thyroid nodules in women with artificial menopause was observed in our study compared with those with natural menopause regardless of other factors, but this seemed to become a nonsignificant association with thyroid nodules after taking a history of uterine fibroids into consideration. Further analysis showed that uterine fibroids accounted for 41.9 % of causes of hysterectomy in our study, which was found to have a close relationship with thyroid nodules in both our analysis and in other studies [6], while a higher level of estrogen was regarded as playing a pivotal role in the occurrence of both uterine fibroids and thyroid nodules. The development of thyroid cancer was proved to have a close relationship with estrogen. Several studies demonstrated the association of menstrual and reproductive factors with a risk of thyroid cancer [24, 25]. However, to our
123
knowledge, information on the effects of the two factors on cancer-prone thyroid nodules is still limited. In our study, we found that a greater number of pregnancies and abortions had been associated with a risk of suspected malignant nodules, and like thyroid nodules, artificial abortion was more likely to increase the risk of suspected malignant nodules more than natural abortion, suggesting that more lifetime periods of elevated estrogen and progesterone levels contributed to the development of cancer-prone thyroid nodules; in addition, women of a menopausal age from 50 to 55 years were more prone to development of suspected malignant nodules. Meanwhile, experimental research on the effects of estrogen and progesterone receptor to thyroid cells might indicate the role of female sex hormones in the development of thyroid cancer. From another perspective, studies have shown that elevated expression of estrogen receptor in thyroid cells and its interaction with estrogen were responsible for increased thyroid cancer risk, and progesterone receptors had also been identified on normal and malignant thyroid cells, which might play a key role in promoting development and growth of thyroid cancer [26]. Nevertheless, we focused on suspected malignant thyroid nodules, which were only morphologically analogous to thyroid cancer, and did not mean definite malignancy. Some proved benign by fine-needle aspiration biopsy. In our study, suspected malignant thyroid nodules were defined by a TIRADS score 4b–6 by ultrasonography, which were identified with ultrasonic echo, acoustic halo, blood flow, and calcification of nodules [16, 17], while estrogen might promote the formation of hypoechoic, an irregular thick halo, and vascularization in thyroid nodules. The hypoecho was one of the features in malignant thyroid nodules, the formation of which was related to the structure of nodule tissue [27]. Xu et al. [20] revealed that estrogen stimulated the growth and simultaneously inhibited the differentiation of thyroid nodule-derived stem and progenitor cells, suggesting that estrogen could increase the development of incomplete differential cells in thyroid nodule tissues, which might appear as a hypoecho in ultrasonography due to loss of synthesis and secretion function. An irregular thick margin of the nodule was regarded as a key sign of a malignant thyroid nodule; the formation of a thick margin was related to the peripheral vascularization of the nodule, and an irregular margin was considered to be formed by the vascular invasion into nodules [28, 29]. Intranodular vascularization in thyroid nodules was another important feature of malignancy, although the diagnostic specificity and sensitivity for thyroid cancer are still controversial [30]. The expression of estrogen receptor in the endothelium of thyroid vessels has been reported [31], and several studies also indicated that estrogen directly modulated angiogenesis via its effects on
Endocrine
endothelial cells [32], while excessive angiogenesis and an increased density of microvessels caused by angiogenesis were regarded as evidence of pre-malignant lesions [33]. This suggested that vascularization in thyroid nodules might be related to increased levels of estrogen in women, and were probably an indicator of cancer-prone thyroid nodules, but these opinions need to be supported by further studies. In addition, a number of studies have shown a close relationship between calcifications and thyroid cancer [34]. As Shi et al. [35] reported, the incidence of calcification in thyroid carcinoma (65.9 %) was higher than that in nodular goiter (35.3 %), but the mechanism of calcification in thyroid nodules had been reported to be irrelevant to female sex hormones [36]. As a community-based non-interventional study, a limitation of our study is that a fine-needle aspiration biopsy was not performed to confirm the pathologic diagnosis of suspected malignant thyroid nodules. Thus, we could not definitively show the association of menstrual and reproductive factors with thyroid cancer. Another limitation is that we defined abortion as termination of pregnancy at over the 50th day of gestation, but did not further classify abortion by mid- and late termination of pregnancy due to a limitation of the questionnaire design. As a result, we lacked details on the association between abortion and development of thyroid nodules. In conclusion, it is commonly thought that high levels of estrogen may be related to both thyroid nodules and thyroid cancer. We demonstrated that periods of elevated estrogen and progesterone levels, such as during pregnancy, were the key occasions for occurrence of thyroid nodules, while lifetime chronic exposure to female sex hormones might promote the growth of thyroid nodules, including intranodular structure changes and vascularization of nodules, until menopause. To further support this possibility, longterm follow-up is needed. Acknowledgments The reaction study is supported by the grants from the Chinese Society of Endocrinology, and this work was supported in part by the Special Foundation for Science and Technology Service Platform of Jiangsu Province (BM2012064). We thank all the participants in this study, along with Dr Tao Chen from Nanjing Medical University for assistance with the statistical analysis.
3.
4.
5.
6.
7.
8. 9. 10.
11.
12.
13.
14.
15.
16.
Disclosure The authors declare that they have no conflict of interest. The study complies with all laws relating to scientific experimentation in China. 17.
References 1. R. Malboosbaf, F. Hosseinpanah, M. Mojarrad, S. Jambarsang, F. Azizi, Relationship between goiter and gender: a systematic review and meta-analysis. Endocrine 43, 539–547 (2013) 2. H. Guo, M. Sun, W. He, H. Chen, W. Li, J. Tang, W. Tang, J. Lu, Y. Bi, G. Ning, T. Yang, Y. Duan, The prevalence of thyroid
18.
19.
nodules and its relationship with metabolic parameters in a Chinese community-based population aged over 40 years. Endocrine 45, 230–235 (2014) N. Knudsen, P. Laurberg, H. Perrild, I. Bu¨low, L. Ovesen, T. Jørgensen, Risk factors for goiter and thyroid nodules. Thyroid 12, 879–888 (2002) M. Kocak, C. Erem, O. Deger, M. Topbas, H.O. Ersoz, E. Can, Current prevalence of goiter determined by ultrasonography and associated risk factors in a formerly iodine-deficient area of Turkey. Endocrine. (2014) [Epub ahead of print] A.W. Kung, M.T. Chau, T.T. Lao, S.C. Tam, L.C. Low, The effect of pregnancy on thyroid nodule formation. J. Clin. Endocrinol. Metab. 87, 1010–1014 (2002) G. Costante, U. Crocetti, E. Schifino, O. Ludovico, C. Capula, M. Nicotera, F. Arturi, S. Filetti, Slow growth of benign thyroid nodules after menopause: no need for long-term thyroxine suppressive therapy in post-menopausal women. J. Endocrinol. Investig. 27, 31–36 (2004) V. Virmani, I. Hammond, Sonographic patterns of benign thyroid nodules: verification at our institution. AJR Am. J. Roentgenol. 196, 891–895 (2011) E.N. Pearce, Thyroid dysfunction in perimenopausal and postmenopausal women. Menopause Int. 13, 8–13 (2007) E.L. Mazzaferri, Evaluation and management of common thyroid disorders in women. Am. J. Obstet. Gynecol. 176, 507–514 (1997) E. Papini, R. Guglielmi, A. Bianchini, A. Crescenzi, S. Taccogna, F. Nardi, C. Panunzi, R. Rinaldi, V. Toscano, C.M. Pacella, Risk of malignancy in nonpalpable thyroid nodules: predictive value of ultrasound and color-Doppler features. J. Clin. Endocrinol. Metab. 87, 1941–1946 (2002) I.S. Nam-Goong, H.Y. Kim, G. Gong, H.K. Lee, S.J. Hong, W.B. Kim, Y.K. Shong, Ultrasonography-guided fine-needle aspiration of thyroid incidentaloma: correlation with pathological findings. Clin. Endocrinol (Oxf) 60, 21–28 (2004) B.E. Henderson, R.K. Ross, M.C. Pike, J.T. Casagrande, Endogenous hormones as a major factor in human cancer. Cancer Res. 42, 3232–3239 (1982) C. Iribarren, T. Haselkorn, I.S. Tekawa, G.D. Friedman, Cohort study of thyroid cancer in a San Francisco Bay area population. Int. J. Cancer 93, 745–750 (2001) G. Ning, Reaction study group, Risk Evaluation of cAncers in Chinese diabeTic Individuals: a lONgitudial (REACTION) study. J. Diabetes 4, 172–173 (2012) Y. Bi, J. Lu, W. Wang, Y. Mu, J. Zhao, C. Liu, L. Chen, L. Shi, Q. Li, Q. Wan, S. Wu, T. Yang, L. Yan, Y. Liu, G. Wang, Z. Luo, X. Tang, G. Chen, Y. Huo, Z. Gao, Q. Su, Z. Ye, Y. Wang, G. Qin, H. Deng, X. Yu, F. Shen, L. Chen, L. Zhao, J. Zhang, J. Sun, M. Dai, M. Xu, Y. Xu, Y. Chen, S. Lai, Z.T. Bloomgarden, D. Li, G. Ning, Cohort profile: risk evaluation of cancers in Chinese diabetic individuals: a longitudinal (REACTION) study. J. Diabetes 6, 147–157 (2014) J.Y. Kwak, K.H. Han, J.H. Yoon, H.J. Moon, E.J. Son, S.H. Park, H.K. Jung, J.S. Choi, B.M. Kim, E.K. Kim, Thyroid imaging reporting and data system for US features of nodules: a step in establishing better stratification of cancer risk. Radiology 260, 829–892 (2011) M. Friedrich-Rust, G. Meyer, N. Dauth, C. Berner, D. Bogdanou, E. Herrmann, S. Zeuzem, J. Bojunga, Interobserver agreement of Thyroid Imaging Reporting and Data System (TIRADS) and strain elastography for the assessment of thyroid nodules. PLoS ONE 8, e77927 (2013) J. Zhang, K.F. Yu, What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 280, 1690–1691 (1998) M.A. Gantus, L.M. Alves, J. Stipursky, E.C. Souza, A.J. Teodoro, T.R. Alves, D.P. Carvalho, A.M. Martinez, F.C. Gomes, L.E.
123
Endocrine
20.
21.
22.
23.
24.
25.
26.
27.
28.
Nasciutti, Estradiol modulates TGF-b1 expression and its signaling pathway in thyroid stromal cells. Mol. Cell. Endocrinol. 337, 71–79 (2011) S. Xu, G. Chen, W. Peng, K. Renko, M. Derwahl, Oestrogen action on thyroid progenitor cells: relevant for the pathogenesis of thyroid nodules? J. Endocrinol. 218, 125–133 (2013) C.L. Westhoff, K.A. Petrie, S. Cremers, Using changes in binding globulins to assess oral contraceptive compliance. Contraception 87, 1761–1781 (2013) D.B. Nelson, J.A. Grisso, M.M. Joffe, C. Brensinger, L. Shaw, E. Datner, Does stress influence early pregnancy loss? Ann. Epidemiol. 13, 223–229 (2003) Z. Kozinszky, K. Boda, G.Y. Ba´rtfai, Determinants of abortion among women undergoing artificial termination of pregnancy. Eur. J. Contracept. Reprod. Health Care 6, 145–152 (2001) J. Hu, J. Wang, H. Xiang, Z. Li, B. Wang, Y. Cao, X. Ma, Association of polymorphisms in the estrogen receptor b (ESR2) with unexplained recurrent spontaneous abortion (URSA) in Chinese population. Matern. Fetal. Neonatal. Med. 25, 1727–1729 (2012) P. Brindel, F. Doyon, F. Rache´di, J.L. Boissin, J. Sebbag, L. Shan, V. Chungue, L.Y. Sun, F. Bost-Bezeaud, P. Petitdidier, J. Paoaafaite, J. Teuri, F. de Vathaire, Menstrual and reproductive factors in the risk of differentiated thyroid carcinoma in native women in French Polynesia: a population-based case-control study. Am. J. Epidemiol. 167, 219–229 (2008) E. Kansakar, Y.J. Chang, M. Mehrabi, V. Mittal, Expression of estrogen receptor, progesterone receptor, and vascular endothelial growth factor-A in thyroid cancer. Am. Surg. 75, 785–789 (2009) D. Berker, S. Isik, U. Ozuguz, Y.A. Tutuncu, K. Kucukler, G. Akbaba, Y. Aydin, S. Guler, Prevalence of incidental thyroid cancer and its ultrasonographic features in subcentimeter thyroid nodules of patients with hyperthyroidism. Endocrine 39, 13–20 (2011) K.J. Clark, J.J. Cronan, F.H. Scola, Color Doppler sonography: anatomic and physiologic assessment of the thyroid. J. Clin. Ultrasound 23, 215–223 (1995)
123
29. B.K. Chan, T.S. Desser, I.R. McDougall, R.J. Weigel, R.B. Jeffrey, Common and uncommon sonographic features of papillary thyroid carcinoma. J. Ultrasound Med. 22, 1083–1090 (2003) 30. P. Trimboli, R. Guglielmi, S. Monti, I. Misischi, F. Graziano, N. Nasrollah, S. Amendola, S.N. Morgante, M.G. Deiana, S. Valabrega, V. Toscano, E. Papini, Ultrasound sensitivity for thyroid malignancy is increased by real-time elastography: a prospective multicenter study. Clin. Endocrinol. Metab. 97, 4524–4530 (2012) 31. F. Magri, V. Capelli, M. Rotondi, P. Leporati, L. La-Manna, R. Ruggiero, A. Malovini, R. Bellazzi, L. Villani, L. Chiovato, Expression of estrogen and androgen receptors in differentiated thyroid cancer: an additional criterion to assess the patient’s risk. Endocr. Relat. Cancer 19, 463–471 (2012) 32. C. Pe´queux, I. Raymond-Letron, S. Blacher, F. Boudou, M. Adlanmerini, M.J. Fouque, P. Rochaix, A. Noe¨l, J.M. Foidart, A. Krust, P. Chambon, L. Brouchet, J.F. Arnal, F. Lenfant, Stromal estrogen receptor-a promotes tumor growth by normalizing an increased angiogenesis. Cancer Res. 72, 3010–3019 (2012) 33. A.R. Gadbail, A. Hande, M. Chaudhary, A. Nikam, M. Gawande, S. Patil, S. Tekade, S. Gondivkar, Tumor angiogenesis in keratocystic odontogenic tumor assessed by using CD-105 antigen. J. Oral Pathol. Med. 40, 263–269 (2011) 34. C.W. Wu, G. Dionigi, K.W. Lee, P.J. Hsiao, M.C. Paul Shin, K.B. Tsai, F.Y. Chiang, Calcifications in thyroid nodules identified on preoperative computed tomography: patterns and clinical significance. Surgery 151, 464–470 (2012) 35. C. Shi, S. Li, T. Shi, B. Liu, C. Ding, H. Qin, Correlation between thyroid nodule calcification morphology on ultrasound and thyroid carcinoma. J. Int. Med. Res. 40, 350–357 (2012) 36. Y.S. Lee, S.W. Hong, H.S. Chang, C.S. Park, Scattered psammomatous calcifications around papillary thyroid carcinoma. World. J. Surg. (2014) [Epub ahead of print]