Cancer Epidemiology 48 (2017) 78–84

Contents lists available at ScienceDirect

Cancer Epidemiology
The International Journal of Cancer Epidemiology, Detection, and Prevention

journal homepage: www.cancerepidemiology.net

Hormonal and reproductive risk factors of papillary thyroid cancer: A

population-based case-control study in France
Emilie Cordina-Duvergera , Christophe Leuxa , Monica Neria , Catherine Tcheandjieua ,
Anne-Valérie Guizardb,c , Claire Schvartzd, Thérèse Truonga , Pascal Guénela,*
a
CESP (Center for Research in Epidemiology and Population Health), INSERM U1018, Cancer and Environment Team, Université Paris-Saclay, Université Paris-
Sud, Villejuif, France
b
Registre Général des tumeurs de Calvados, Centre François Baclesse, Caen, France
c
U1086 Inserm UCNB, Cancers and Prevention, Caen, France
d
Centre de Lutte contre le Cancer Jean Godinot, Reims, France

A R T I C L E I N F O A B S T R A C T

Article history:
Received 12 October 2016 The three times higher incidence of thyroid cancer in women compared to men points to a role of female
Received in revised form 30 March 2017 sex hormones in its etiology. However the effects of these factors are poorly understood. We analyzed the
Accepted 1 April 2017 association between thyroid cancer and hormonal and reproductive factors among women enrolled in
Available online xxx CATHY, a population-based case-control study conducted in France. The study included 430 cases of
papillary thyroid cancer and 505 controls frequency-matched on age and area of residence. The odds
Keywords: ratios for thyroid cancer increased with age at menarche (p trend 0.05). Postmenopausal women were at
Case-control studies increased risk, as compared to premenopausal women, particularly if menopause followed an
Hormone replacement therapy
ovariectomy, and for women with age at menopause < 55 years. In addition, use of oral contraceptives
Oral contraceptives
and menopausal hormone therapy reduced the association with thyroid cancer by about one third, and
Reproductive history
Thyroid neoplasms breastfeeding by 27%. Overall, these findings provide evidence that the risk of thyroid cancer increases
with later age at menarche and after menopause, and decreases with use of oral contraceptives and
menopausal hormone therapy. These findings confirm an implication of hormonal factors in papillary
thyroid cancer risk, whose mechanisms need to be elucidated.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction cancer were 12.6/100,000 in women and 4.1/100,000 in men

Incidence rates of thyroid carcinomas have been increasing
regularly over the last decades in most high-resource countries [1].
The rising incidence of thyroid cancer has been attributed mostly
to changes in medical practices, like the introduction of neck
ultrasonography in the 1980s that enhanced detection of small
dormant carcinomas [2]. Actually micropapillary lesions are
predominant in this increase, but several epidemiological studies
provided evidence of a growing incidence of larger-size carcinomas
as well, suggesting a true rise in thyroid cancer incidence [3,4]. In
France the estimated age standardized incidence rates of thyroid
Abbreviations: CI, confidence interval; MHT, menopausal hormone therapy; OC,
oral contraceptive; OR, odds ratio.
* Corresponding author at: Centre de Recherche en Épidémiologie et Santé des
Populations (CESP), Cancer and Environment Team, Inserm U1018 - Université Paris
Sud – Université Paris-Saclay, 16 Avenue Paul Vaillant-Couturier, 94807 Villejuif
Cedex, France.
E-mail address: pascal.guenel@inserm.fr (P. Guénel).
according to GLOBOCAN 2012 [5].
Exposure to ionizing radiation during childhood and obesity [6]
are recognized risk factors of thyroid cancer. The role of alcohol
drinking, tobacco smoking, iodine excess or deficiency, exposure to
endocrine disruptors is still unclear [7]. Female sex hormones are
likely to play a role, since thyroid cancer is generally three times
more frequent in women than in men (5), and the female-to-male
incidence ratio has been shown to be greater during reproductive
age for papillary thyroid cancer [8]. However, epidemiological
studies investigating hormonal and reproductive factors in women
in relation to thyroid cancer incidence have provided inconsistent
results. Recent meta-analyses showed a possible role for some
menstrual and reproductive factors in the etiology of thyroid
cancer [9–12], but their findings are difficult to interpret due to
heterogeneity between studies [13].
In the present paper we report on the association of hormonal
and reproductive factors with papillary thyroid cancer among
women who participated in a carefully designed population-based

http://dx.doi.org/10.1016/j.canep.2017.04.001
1877-7821/© 2017 Elsevier Ltd. All rights reserved.
 E. Cordina-Duverger et al. / Cancer Epidemiology 48 (2017) 78–84 79

case-control study conducted in France, one of the largest studies
on thyroid cancer risk conducted so far.
2. Subjects and methods
2.1. Case selection
The present research was based on the CATHY study, that
comprised all patients diagnosed with papillary or follicular
thyroid cancer between 2002 and 2007 and residing in the
Calvados, Marne or Ardennes, three French administrative areas
(départements). The data collection started in 2005. As the
prognosis of thyroid cancer is very good, incident cases diagnosed
in 2002–2004 could be contacted and included in the study, while
cases diagnosed in 2005–2007 were recruited prospectively. Cases
were identified by the cancer registries in these areas, and only
women were included in the present analysis.
The topography and morphology of thyroid cancers were coded
according to definitions and rules of the International Classifica-
tion of Diseases for Oncology, Third edition (ICDO-3). The thyroid
cancer cases were classified according to histology and size of the
largest cancerous nodule.
Out of 660 eligible cases of thyroid cancer in women, 177 (27%)
were not included because the subject refused to participate
(n = 122), had died at the time of interview (n = 15), could not be
contacted (n = 37), or was too ill to participate (n = 3). Only
papillary carcinomas (n = 430 cases) were included in the present
analysis, as their etiology may differ from follicular carcinoma
(n = 53 cases). Over 56% of the papillary carcinomas had a diameter
less than or equal to 10 mm.
2.2. Control selection
Controls were selected at random by a polling institute using a
telephone directory of all private homes in the study areas.
Controls reached by phone who accepted to participate were
assigned a year of reference at random from 2002 to 2007, and
were frequency-matched by 5-year age group and study area with
the cases in the corresponding year of diagnosis. In addition, in
order to control for potentially differential participation rates
across socio-economic status (SES) categories, the control group
was selected to reflect the distribution by SES of the underlying
female population of the same age in each study area. Any woman
in the household reached by phone was invited to participate in
the study, until the predefined number of controls in the strata
defined by study area, age, and SES was reached.
The polling institute identified 687 eligible female controls,
who initially accepted to be contacted by trained interviewers for
an in-person interview. Among them, 182 (26%) finally refused to
participate. The remaining 505 controls with completed ques-
tionnaires were included in the analyses.
Only variables determined at or before the year of reference (for
the controls) or the year of diagnosis (for the cases) underwent
subsequent analyses.

Table 1
Socio-demographic Characteristics and Lifestyle Habits of Cases and Controls. The CATHY Study, France, 2002–2007.

Cases (n = 430) Controls (n = 505) ORa 95% CI p trend

No. % No. %
Age (years)b
<30 23 5.3 37 7.3
30–39 62 14.4 84 16.6
40–49 91 21.2 126 25.0
50–59 148 34.4 128 25.3
60–69 73 17.0 86 17.0
70 33 7.7 44 8.7
Mean (SD) 51.2 (12.4) 50.0 (13.6)

Residence areac
Calvados 139 32.33 200 39.6
Ardennes 88 20.47 84 16.6
Marne 203 47.21 221 43.8

Education level (years)

5 117 27.2 130 25.8 1 Reference
6–9 164 38.1 166 33.0 1.16 [0.81–1.66]
 10 149 34.7 207 41.2 0.91 [0.63–1.33] 0.15

Body Mass Index (kg/m2)

 18.5 12 2.8 19 3.8 0.86 [0.41–1.84]
[18.5–25[ 209 48.9 271 54.1 1 Reference
[25–30[ 128 30.0 129 25.7 1.19 [0.87–1.63]
 30 78 18.3 82 16.4 1.10 [0.76–1.60] 0.19

Smoking status
Never 247 57.4 297 58.8 1 Reference
Current 70 16.3 102 20.2 0.85 [0.58–1.23]
Past 113 26.3 106 21.0 1.25 [0.90–1.73]

Alcohol consumption (glass/week)

Never 141 33.1 142 28.4 1 Reference
1–10 244 57.3 319 63.8 0.73 [0.54–0.98]
>10 41 9.6 39 7.8 1.04 [0.62–1.72] 0.57
a
Odds ratios adjusted for age and residence area.
b
Chi2 = 10.36 p = 0.07.
c
Chi2 = 5.86 p = 0.05.
80 E. Cordina-Duverger et al. / Cancer Epidemiology 48 (2017) 78–84

2.3. Data collection Table 2

Odds ratios of thyroid cancer associated with menstrual factors. The CATHY study,
France.
Trained interviewers conducted an in-person interview of cases
and controls at their home address using a structured question- Cases Controls ORa 95% CI p trend
(n = 430) (n = 505)
naire, after having obtained their written informed consent. The
interviewers collected information on socio-demographic charac- Age at menarche (years)b
11 68 89 1 reference
teristics, diet, alcohol drinking, tobacco smoking, anthropometric
12 100 124 1.08 [0.70–1.67]
factors, gynecological and reproductive history, medical condi- 13 104 125 1.23 [0.80–1.90]
tions, medical X-ray exposure, occupational and residential 14 84 95 1.27 [0.81–
history, and family history of thyroid cancer. Menstrual and 2.00]
reproductive factors and exogenous hormone use were the object 15 73 69 1.55 [0.96–2.50] 0.05

of the present investigation.

Irregular menstrual cycles
Women were considered postmenopausal if they reported no No 400 466 1 reference
menstruation for at least one year or used menopausal hormone Yes 30 39 0.83 [0.50–1.40]
therapy (MHT) (natural menopause), or if they had bilateral
Menopausal status
ovariectomy (artificial menopause). Women with unknown
Premenopausal 187 269 1 reference
menopausal status, because of hysterectomy before cessation of Postmenopausal 243 236 1.73 [1.06–2.83]
menstruations or unknown date of last menstruation, were Type of menopause
considered postmenopausal if they were 50 years old or more Natural 214 216 1.69 [1.03–2.77]
(the median age at menopause in women with natural meno- Artificialc 29 20 2.52 [1.16–5.46]
Age at menopaused
pause).
50 years 132 124 1.67 [1.00–2.78]
51–54 years 55 51 1.67 [0.89–3.13]
2.4. Statistical analysis 55 years 25 46 0.97 [0.46–2.06]

The odds ratios were calculated by unconditional logistic Ovariectomy

No 396 480 1 reference
regression, using SAS version 9.4. All odds ratios were adjusted for Yes 34 25 1.49 [0.83–2.67]
age (5-year age groups) and study area (Calvados, Marne and
Ardennes), the matching variables. We controlled for potential Hysterectomy
confounders by adjusting models for acknowledged or suspected No 348 437 1 reference
Yes 82 68 1.28 [0.87–1.90]
risk factors of thyroid cancer and for variables associated with
menstrual and reproductive characteristics, i.e. body mass index Years of menstrual cycling (OR for each additional year)e
(<18.5 kg/m2, 18.5–24.9 kg/m2, 25.0–29.9 kg/m2, 30 kg/m2), All women 396 484 0.97 [0.94–1.00]
smoking status (never, current, former), alcohol consumption (0, Premenopausal 186 266 1.04 [0.95–1.13]
1–10, >10 glasses per week) and menopausal status (pre, post) Postmenopausal 210 218 0.95 [0.92–0.99]

(Tables 2–4). Tests for trend were calculated by fitting models with
continuous exposure variables, assuming a log-linear relationship
between exposure and cancer risk.
3. Results
3.1. Characteristics of cases and controls
a
Adjusted for age, residence area, body mass index, smoking status, alcohol
consumption. Age at menarche, irregular menstrual cycles, ovariectomy and
hysterectomy are further adjusted for menopausal status.
b
Unknown age at menarche in 1 case and 3 controls.
c
Bilateral ovariectomy before cessation of menstruations.
d
Unknown age at menopause in 31 cases and 15 controls.
e
Subjects with unknown age at menarche and/or unknown age at menopause
are excluded.

The socio-demographic characteristics of cases and controls are
shown in Table 1. The distribution by the stratification variables, i.e.
age and residence area, was slightly different between the two
groups, because of lower participation rates among controls in the
age range 50–59 years and in the departments of Ardennes and
Marne. After adjustment for age and residence area, no difference
between cases and controls was observed as regards education,
body mass index and smoking status, while a statistically
significant inverse association was found between moderate
alcohol consumption and thyroid cancer.
3.2. Menstrual factors
Table 2 shows the association of menstrual factors with thyroid
cancer. A consistently increased risk was found with increasing age
at menarche (p trend 0.05). Borderline statistical significance was
reached in women who first menstruated at the age of 15 or later
(OR 1.55) as compared to women who menstruated before the age
of 12.
Menopause, either natural or artificial (i.e. following ovariec-
tomy before cessation of menstruations), was associated with
statistically significantly increased odds ratios (OR 1.69 and 2.52,
respectively). When women with known age at menopause were
compared to premenopausal women, a 67% increase in risk was
apparent in women who went into menopause before the age of
55 years, whereas the OR was close to one for women who went
into menopause at that age or later. Overall, these findings
suggest that the risk of thyroid cancer increases with late age at
menarche and early age at menopause, i.e. in women with shorter
lifetime duration of menstruations. A cumulative index of the
number of years with menstrual cycles was calculated, as the lag
time between age at menarche and age at last menstruation,
minus the total duration of pregnancies and breastfeeding
periods (Table 2). A 3% decrease of the OR was found for each
additional year of menstrual cycling (OR 0.97; 95% CI 0.94–1.00,
p = 0.04). This result was driven by postmenopausal women (OR
for each additional year of menstrual cycling 0.95; 95% CI 0.92–
0.99, p = 0.01).
The risk of thyroid cancer was not significantly modified among
women who did not menstruate regularly and among those who
underwent ovariectomy or hysterectomy.
3.3. Reproductive history
Table 3 shows the odds ratios for factors related to reproductive
life. Parity (yes or no) was not associated with risk of thyroid
cancer, but the odds ratio increased to 1.46 (not significant) in
women with 4 or more full-term pregnancies as compared to
 E. Cordina-Duverger et al. / Cancer Epidemiology 48 (2017) 78–84 81

Table 3
Odds ratios for thyroid cancer associated with reproductive factors. The CATHY study, France.

Cases Controls ORa 95% CI p-trend

(n = 430) (n = 505)
Ever had a full-term pregnancy
No 36 53 1 reference
Yes 394 452 1.15 [0.70–1.87]

Number of full-term pregnancies

0 36 53 1 reference
1 85 92 1.22 [0.70–2.11]
2 154 185 1.10 [0.65–1.85]
3 94 120 1.03 [0.59–1.79]
4 61 55 1.46 [0.78–2.72] 0.29

Ever had a miscarriage

No 330 401 1 reference
Yes 100 104 1.23 [0.89–1.71]

Ever had a voluntary abortion

No 364 446 1 reference
Yes 66 59 1.39 [0.94–2.08]

b
Age at first full-term pregnancy (years)
<20 61 58 1 reference
20–24 186 231 0.93 [0.61–1.42]
25–29 114 116 1.21 [0.75–1.96]
30 33 47 0.81 [0.44–1.49] 0.98

b
Time since last full-term pregnancy (years)
<5 33 53 1 reference
5–9 34 38 1.62 [0.76–3.46]
10 324 352 1.36 [0.62–2.99] 0.64

b
Breastfeeding and duration of breastfeeding
Never 194 190 1 reference
Ever 200 262 0.73 [0.55–0.97]
Duration of breastfeeding (months)
<5 129 158 0.80 [0.58–1.10]
6–11 42 64 0.63 [0.40–1.00]
12 29 40 0.64 [0.37–1.10] 0.67
a
Odds ratios adjusted for age, residence area, body mass index, smoking status, alcohol drinking and menopausal status.
b
Parous women only.

nulliparous women. A history of miscarriage or voluntary abortion
was not associated with thyroid cancer. No clear trend was found
with time since the last full-term pregnancy and with age at first
birth. Among parous women, those who breastfed showed
significantly decreased odds ratio compared to the others,
although no trend with duration could be demonstrated.
3.4. Exogenous hormone use
Odds ratios for thyroid cancer associated with use of exogenous
hormones are shown in Table 4. The odds ratio in women ever
using oral contraceptives (OC) was decreased by almost one third,
as compared to never users (odds ratio 0.69, 95% CI 0.48-1.01). The
odds ratio was almost halved among current users, with respect to
never OC users (statistically significant), although no dose-
response trend was seen with duration of OC use.
Among postmenopausal women, the use of MHT was associated
with a lowered risk of thyroid cancer reaching statistical
significance. However, there was no indication of a trend with
duration of use or with time since last use.
Finally, we fitted a multivariate model including all the main
hormonal and reproductive factors associated with thyroid cancer
in Tables 2–4 (age at menarche, menopausal status, parity,
breastfeeding, and OC use). The ORs were not substantially
modified, indicating that risk factors were independently associ-
ated with thyroid cancer (not shown).
No major variation in the association of hormonal reproductive
factors with thyroid cancer was found in stratified analyses based
on menopausal status (not shown).
4. Discussion
The present investigation is one of the largest study conducted
so far on thyroid cancer. Several hormonal and reproductive factors
in women were found to be associated with papillary thyroid
cancer. We found that late age at menarche and postmenopausal
status, particularly for women who went into menopause before
the age of 55 years, were associated with increased incidence of
thyroid cancer. Exposure to exogenous hormones (both OC and
MHT) as well as breastfeeding were inversely associated with
thyroid cancer.
A number of studies have been conducted since the 1980s on
the influence of female hormonal and reproductive factors on the
risk of thyroid cancer, giving conflicting results. Several meta-
analyses of published results and analyses of pooled datasets have
been conducted, but a comprehensive vision of these effects could
not be provided due to considerable heterogeneity between
studies [13]. Major differences exist among published studies with
respect to study design (case-control, cohort, nested case-control
studies), thyroid cancer subtypes investigated (all histological
types, papillary only, other groupings), ethnicity, age or other
population characteristics, definition and categorization of the
82 E. Cordina-Duverger et al. / Cancer Epidemiology 48 (2017) 78–84

Table 4
Odds ratios for thyroid cancer associated with exogenous hormone use. The CATHY study, France.

Cases Controls ORa [95%CI] p trend

(n = 430) (n = 505)
Oral contraceptives
Never 123 119 1 reference
Ever 303 384 0.69 [0.48–1.01]
Duration of OC use (years)
<10 131 142 0.79 [0.52–1.20]
10–19 93 147 0.56 [0.35–0.87]
 20 68 85 0.69 [0.42–1.13] 0.10
Time since last OC use (years)
20 102 93 0.84 [0.55–1.29]
10–19 65 98 0.58 [0.36–0.93]
1–9 57 67 0.77 [0.45–1.32]
Current use 61 105 0.55 [0.31–0.96] 0.24

Menopausal hormone therapyb

Never 143 116 1 reference
Ever 94 118 0.65 [0.44–0.96]
Duration of MHT use (years) b
<5 30 49 0.42 [0.24–0.73]
5-9 35 34 0.84 [0.47–1.51]
10 26 32 0.81 [0.44–1.50] 0.21
b
Time since last MHT use (years)
5 28 37 0.66 [0.36–1.19]
<5 23 27 0.64 [0.32–1.24]
Current use 41 51 0.63 [0.38–1.05] 0.82
a
Odds ratios adjusted for age, residence area, body mass index, smoking status, alcohol drinking and menopausal status.
b
Postmenopausal women only.

variables of interest. Small numbers have often been an additional
problem, making it difficult to draw definite conclusions.
Results of a pooled analysis combining individual data of 14
case-control studies were reported for exogenous hormone use
[14] and for menstrual and reproductive factors [15], but the
associations with thyroid cancer were weak and inconsistent
across studies. More recently, several meta-analyses have been
published [9–12].
4.1. Menstrual factors
Late menarche was associated with an increased risk of thyroid
cancer in our study. Similar results were shown in a meta-analysis
of case-control studies [10], but not in one of cohort studies [9],
which reported no association. A recently published case-control
study restricted to women aged up to 35 years found opposite
effects [16].
We also found that menopause was associated with an
increased risk of papillary thyroid cancer, a result in line with
the meta-analysis of case-control studies [10]. Slightly decreased
risks in postmenopausal women were observed in the meta-
analysis of cohort studies, but this result was based on only four
studies [9]. Among women with known age at menopause, our
results suggested that the increased risk in postmenopausal
women was restricted to those with onset of menopause before 55
years of age. In addition, we found a more than double risk in those
who went into menopause because of ovariectomy, suggesting a
deleterious effect of a sudden drop in estrogen levels, although a
role of enhanced medical surveillance of the thyroid gland in
women who underwent surgery cannot be ruled out. In the pooled
analysis, results were in line with ours, with higher risks for
artificial than for natural menopause [15], and recent individual
case-control studies also indicated increased risks of thyroid
cancer in ovariectomized women [17,18].
Taken together, these findings suggest that a shorter time
interval between menarche and menopause may be associated
with increased risk of papillary thyroid cancer. This pattern
suggests an unexpected relationship between the number of
menstrual cycles over a woman’s lifetime and thyroid cancer,
opposite to breast cancer, for which early menarche and late
menopause are well-established risk factors [19]. In our study, we
found that the odds ratio of thyroid cancer decreased by 5% for each
additional year of menstrual cycles among postmenopausal
women. This trend was not observed in premenopausal women.
Although we do not have a mechanistic explanation for this
finding, it is possible that the lifetime repeated exposure to
fluctuating levels of hormones had opposite effects in mammary
and thyroid gland, depending on the specific hormone receptors
that are present in the two organs and/or on other complex
interplays.
No influence of irregular menstrual cycles was observed in our
data, consistently with previous studies [13]. History of hysterec-
tomy was associated with a non-statistically significant 25%
increased risk, a finding consistent with the meta-analysis of
cohort studies [9] and with most previous case-control studies
[17,20–23]. If this association is real, it may be related to diseases
requiring hysterectomy. Uterine fibroids (the main therapeutic
indication for hysterectomy in our data) are frequently observed in
women with thyroid nodules [24], suggesting possible common
pathological mechanisms with thyroid cancer.
4.2. Reproductive factors
Odds ratios around one for parous vs. nulliparous women were
found in our study, as well as in all the meta-analyses. However, the
odds ratio increased to 1.5 (not significant) in women with a high
number of full-term pregnancies (4), similar to the meta-OR that
has been estimated for women with at least 3 pregnancies [11]. The
number of women with high parity may not be sufficient in
metropolitan France to detect an increased risk, as was the case in
New Caledonia [20] and French Polynesia [18]. This increased risk
might reflect repeated exposures to important hormonal changes
that occur during pregnancy, like high levels of estrogens, human
chorionic gonadotropin (hCG) and TSH.
We found that a history of breastfeeding was associated with a
27% decreased risk of thyroid cancer. Similarly, risks associated
 E. Cordina-Duverger et al. / Cancer Epidemiology 48 (2017) 78–84
with breastfeeding were lower than unity in a meta-analysis [10],
and a negative trend with duration of breastfeeding was shown in a
recent study of young women [16].
We reported null results for age at first full term pregnancy,
and we found no clear indication of an association between a
history of miscarriage or voluntary abortion and thyroid cancer.
Heterogeneous, inconclusive results were reported in the
literature in this regard. Time since last full term pregnancy
beyond 5 years was associated with non-significantly increased
odds ratios of thyroid cancer in our data. The risk was instead
significantly higher in women with a recent delivery in one of the
meta-analyses [11].
4.3. Exogenous hormone use
One of the most striking results in our data were the 30%
decreased risk of thyroid cancer in OC ever users, and the 45%
decrease in current users, compared to never users. Similarly, a
recent large case-control study found OC users to be at reduced risk
of thyroid cancer, particularly of the papillary type [16], and the
meta-analyses gave a clue in the same direction [9,10,12]. Previous
investigators have also reported an inverse association between OC
use and thyroid volume [25,26], and OC seems to exert protective
effects against goiter, a strong risk factor for thyroid cancer [20,27].
We found a one third decrease in risk among postmenopausal
women who had ever used MHT, although no clear trend with
duration of use emerged in our data. This finding is not consistent
with most previous studies, and odds ratios around unity for MHT
use were observed in both the meta-analyses of cohort and case-
control studies [9,10].
To date, the role of hormone-related factors in the occurrence of
thyroid cancer is poorly understood. The interpretation of findings
in previous studies has been hampered by great differences
between study populations in thyroid cancer incidence, medical
screening practices, and reproductive and menstrual factors.
Overall, our study is based on a relatively large sample of cases
and controls, and provides some clues on hormonal risk factors in
thyroid cancer. Our data suggest that early menarche and late
menopause, as well as breastfeeding, use of OC and of MHT can be
protective against thyroid cancer. With the exception of breast-
feeding, this pattern is remarkable as it clearly contrasts with that
of breast cancer, where early menarche and late menopause, recent
use of OC, MHT, and low parity are well-established determinants
of the disease.
4.4. Biological mechanisms
The well-known disparity in thyroid cancer incidence between
genders has been logically attributed to sex hormones differences,
but the role of female steroid hormones in thyroid carcinogenesis
remains unclear [28]. It has been found that estrogen levels were
higher and progesterone levels were lower in the serum of patients
with papillary thyroid carcinomas or other thyroid neoplasms than
in normal controls [29]. Estrogen Receptora and b (ERa, ERb) and
progesterone receptor (PR) were also found to be differentially
expressed in thyroid tumors as compared to normal thyroid tissue
[29,30]. Some studies suggested that estrogens may play different
roles in the development of thyroid cancer: they may promote cell
proliferation and growth through ERa and induce apoptosis and
other suppressive actions through ERb [31–33]. Determinants and
effects of tumor-specific sex-hormone receptor expression are still
unclear, but epidemiological findings are more consistent with an
effect of estrogens on ERb.
Estrogens may also exert indirect effects through the pituitary-
thyroid axis and/or through increased serum levels of thyroxin-
binding globulin (TBG) and subsequent changes in thyroid function
83
[34,35]. Studies conducted in women using contraceptive estro-
gens showed that this may result in changes in thyroid hormones
and TBG levels [34,36,37].
These results highlight the possibly contradictory interferences
between sex steroids and thyroid function and the need to further
investigate these biological mechanisms.
4.5. Strengths and limits
This study is one of the largest case-control studies ever
conducted on thyroid cancer to date. Other strengths include the
population-based design, with exhaustive identification of thyroid
cancer cases and a high participation rate among both cases and
controls. Detailed information on exposures during lifetime was
also available in our study.
Limitations include recall bias that is inherent to case-control
studies. However, questionnaires investigated a variety of factors,
including objective circumstances that can hardly be influenced by
the case or control status of the subject (e.g. number of children,
ovariectomy). In addition, a potential role for hormonal and
reproductive factors in thyroid cancer risk has been seldom
reported in popular media, making unlikely a strong recall bias for
these factors. Other sources of differential misclassification have
been minimized, since both groups of subjects were interviewed
with standardized questionnaires by trained interviewers.
A surveillance bias may partly explain the increased odds ratios
observed for women with a history of reproductive surgery due to
increased medical surveillance, however women who were under
hormonal treatment were at lower risk.
5. Conclusion
Our results provide indications on the role of hormonal factors
in the etiology of papillary thyroid cancer among women. In our
data, the risk of papillary thyroid cancer appeared to increase with
shorter lifetime duration of menstruations, i.e. with late age at
menarche and early age at menopause. In addition, the use of oral
contraceptives and menopausal hormone therapy, as well as
breastfeeding, lowered the risk. Interestingly, with the exception of
breastfeeding, these associations are in the opposite direction to
those observed for breast cancer incidence. Although our findings
may provide some clues, the biological mechanisms explaining the
role of hormonal factors in thyroid carcinogenesis are still poorly
understood and do not provide a clear explanation for the gender
disparity in thyroid cancer incidence. Further studies combining
epidemiological and biological data may provide better under-
standing on the role of female sex hormones in the etiology of
thyroid cancer.
Conflict of interest declaration
The authors declare that there are no known conflicts of interest
associated with the manuscript. There has been no significant
financial support for this work that could have influenced its
outcome.
Acknowledgements
This study was supported by Fondation de France; Agence
Française de Sécurité Sanitaire de l’Environnement et du Travail
(ANSES, ex-AFSSET), EDF-Commission d’Epidémiologie
References
[1] B.A. Kilfoy, T. Zheng, T.R. Holford, X. Han, M.H. Ward, A. Sjodin, Y. Zhang, Y. Bai,
C. Zhu, G.L. Guo, N. Rothman, Y. Zhang, International patterns and trends in
84 E. Cordina-Duverger et al. / Cancer Epidemiology 48 (2017) 78–84
thyroid cancer incidence 1973–2002, Cancer Causes Control 20 (5) (2009)
525–531.
[2] S. Vaccarella, S. Franceschi, F. Bray, C.P. Wild, M. Plummer, L. Dal Maso,
Worldwide thyroid-Cancer epidemic? the increasing impact of overdiagnosis,
N. Engl. J. Med. 375 (7) (2016) 614–617.
[3] M. Colonna, Z. Uhry, A.V. Guizard, P. Delafosse, C. Schvartz, A. Belot, P.
Grosclaude, F. network, Recent trends in incidence, geographical distribution,
and survival of papillary thyroid cancer in France, Cancer Epidemiol. 39 (4)
(2015) 511–518.
[4] S. Vaccarella, L. Dal Maso, M. Laversanne, F. Bray, M. Plummer, S. Franceschi,
The impact of diagnostic changes on the rise in thyroid cancer incidence: a
population-based study in selected high-resource countries, Thyroid 25 (10)
(2015) 1127–1136.
[5] L.A. Torre, F. Bray, R.L. Siegel, J. Ferlay, J. Lortet-Tieulent, A. Jemal, Global cancer
statistics, 2012, CA. Cancer J. Clin. 65 (2) (2015) 87–108.
[6] B. Lauby-Secretan, C. Scoccianti, D. Loomis, Y. Grosse, F. Bianchini, K. Straif, I.
for on Group, Body fatness and cancer — viewpoint of the IARC working group,
N. Engl. J. Med. 375 (8) (2016) 794–798.
[7] L. Davies, L.G. Morris, M. Haymart, A.Y. Chen, D. Goldenberg, J. Morris, J.B.
Ogilvie, D.J. Terris, J. Netterville, R.J. Wong, G. Randolph, A.E.S.S. Committee,
American association of clinical endocrinologists and american college of
endocrinology disease state clinical review: the increasing incidence of
thyroid cancer, Endocr. Pract. 21 (6) (2015) 686–696.
[8] B.A. Kilfoy, S.S. Devesa, M.H. Ward, Y. Zhang, P.S. Rosenberg, T.R. Holford, W.F.
Anderson, Gender is an age-specific effect modifier for papillary cancers of the
thyroid gland, Cancer Epidemiol. Biomarkers Prev. 18 (4) (2009) 1092–1100.
[9] S. Caini, B. Gibelli, D. Palli, C. Saieva, M. Ruscica, S. Gandini, Menstrual and
reproductive history and use of exogenous sex hormones and risk of thyroid
cancer among women: a meta-analysis of prospective studies, Cancer Causes
Control 26 (4) (2015) 511–518.
[10] Y. Cao, Z. Wang, J. Gu, F. Hu, Y. Qi, Q. Yin, Q. Sun, G. Li, B. Quan, Reproductive
factors but not hormonal factors associated with thyroid cancer risk: a
systematic review and meta-analysis, BioMed Res. Int. 2015 (2015) 103515.
[11] Y.Q. Zhou, Z. Zhou, M.F. Qian, T. Gong, J.D. Wang, Association of thyroid
carcinoma with pregnancy: a meta-analysis, Mol. Clin. Oncol/ 3 (2) (2015)
341–346.
[12] P. Wang, L. Lv, F. Qi, F. Qiu, Increased risk of papillary thyroid cancer related to
hormonal factors in women, Tumour Biol. 36 (7) (2015) 5127–5132.
[13] E. Peterson, P. De, R. Nuttall, BMI, diet and female reproductive factors as risks
for thyroid cancer: a systematic review, PLoS One 7 (1) (2012) e29177.
[14] C. La Vecchia, E. Ron, S. Franceschi, L. Dal Maso, S.D. Mark, L. Chatenoud, C.
Braga, S. Preston-Martin, A. McTiernan, L. Kolonel, K. Mabuchi, F. Jin, G.
Wingren, M.R. Galanti, A. Hallquist, E. Lund, F. Levi, D. Linos, E. Negri, A pooled
analysis of case-control studies of thyroid cancer. III. Oral contraceptives,
menopausal replacement therapy and other female hormones, Cancer Causes
Control 10 (2) (1999) 157–166.
[15] E. Negri, L. Dal Maso, E. Ron, C. La Vecchia, S.D. Mark, S. Preston-Martin, A.
McTiernan, L. Kolonel, Y. Yoshimoto, F. Jin, G. Wingren, M. Rosaria Galanti, L.
Hardell, E. Glattre, E. Lund, F. Levi, D. Linos, C. Braga, S. Franceschi, A pooled
analysis of case-control studies of thyroid cancer. II. Menstrual and
reproductive factors, Cancer Causes Control: CCC 10 (2) (1999) 143–155.
[16] C. Xhaard, C. Rubino, E. Clero, S. Maillard, Y. Ren, F. Borson-Chazot, G. Sassolas,
C. Schvartz, M. Colonna, B. Lacour, A. Danzon, M. Velten, A. Buemi, L. Bailly, E.
Marine Barjoan, M. Schlumberger, J. Orgiazzi, E. Adjadj, F. de Vathaire,
Menstrual and reproductive factors in the risk of differentiated thyroid
carcinoma in young women in France: a population-based case-control study,
Am. J. Epidemiol. 180 (10) (2014) 1007–1017.
[17] W.J. Mack, S. Preston-Martin, L. Bernstein, D. Qian, M. Xiang, Reproductive and
hormonal risk factors for thyroid cancer in Los Angeles County females, Cancer
Epidemiol. Biomarkers Prev. 8 (11) (1999) 991–997.
[18] P. Brindel, F. Doyon, F. Rachedi, 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 (2) (2008) 219–229.
[19] H. Samavat, M.S. Kurzer, Estrogen metabolism and breast cancer, Cancer Lett.
356 (2 Pt A) (2015) 231–243.
[20] T. Truong, L. Orsi, D. Dubourdieu, Y. Rougier, D. Hemon, P. Guenel, Role of goiter
and of menstrual and reproductive factors in thyroid cancer: a population-
based case-control study in New Caledonia (South Pacific), a very high
incidence area, Am. J. Epidemiol. 161 (11) (2005) 1056–1065.
[21] A. Memon, A. Varghese, A. Suresh, Benign thyroid disease and dietary factors in
thyroid cancer: a case-control study in Kuwait, Br. J. Cancer 86 (11) (2002)
1745–1750.
[22] M.A. Rossing, L.F. Voigt, K.G. Wicklund, J.R. Daling, Reproductive factors and
risk of papillary thyroid cancer in women, Am. J. Epidemiol. 151 (8) (2000)
765–772.
[23] R. Luoto, A. Auvinen, E. Pukkala, M. Hakama, Hysterectomy and subsequent
risk of cancer, Int. J. Epidemiol. 26 (3) (1997) 476–483.
[24] M.H. Kim, Y.R. Park, D.J. Lim, K.H. Yoon, M.I. Kang, B.Y. Cha, K.W. Lee, H.Y. Son,
The relationship between thyroid nodules and uterine fibroids, Endocr. J. 57 (7)
(2010) 615–621.
[25] X. Barrere, P. Valeix, P. Preziosi, M. Bensimon, B. Pelletier, P. Galan, S. Hercberg,
Determinants of thyroid volume in healthy French adults participating in the
SU.VI.MAX cohort, Clin. Endocrinol. (Oxf.) 52 (3) (2000) 273–278.
[26] N. Knudsen, I. Bulow, P. Laurberg, H. Perrild, L. Ovesen, T. Jorgensen, Low goitre
prevalence among users of oral contraceptives in a population sample of 3712
women, Clin. Endocrinol. (Oxf.) 57 (1) (2002) 71–76.
[27] N. Knudsen, T.H. Brix, Genetic and non-iodine-related factors in the aetiology
of nodular goitre, Best Pract. Res. Clin. Endocrinol. Metab. 28 (4) (2014) 495–
506.
[28] R. Rahbari, L. Zhang, E. Kebebew, Thyroid cancer gender disparity, Future
Oncol. 6 (11) (2010) 1771–1779.
[29] J. Liu, G. Chen, X.Y. Meng, Z.H. Liu, S. Dong, Serum levels of sex hormones and
expression of their receptors in thyroid tissue in female patients with various
types of thyroid neoplasms, Pathol. Res. Pract. 210 (12) (2014) 830–835.
[30] G. Vannucchi, S. De Leo, M. Perrino, S. Rossi, D. Tosi, V. Cirello, C. Colombo, G.
Bulfamante, L. Vicentini, L. Fugazzola, Impact of estrogen and progesterone
receptor expression on the clinical and molecular features of papillary thyroid
cancer, Eur. J. Endocrinol. 173 (1) (2015) 29–36.
[31] R. Chu, A. van Hasselt, A.C. Vlantis, E.K. Ng, S.Y. Liu, M.D. Fan, S.K. Ng, A.B. Chan,
Z. Liu, X.Y. Li, G.G. Chen, The cross-talk between estrogen receptor and
peroxisome proliferator-activated receptor gamma in thyroid cancer, Cancer
120 (1) (2014) 142–153.
[32] M.A. Cho, M.K. Lee, K.H. Nam, W.Y. Chung, C.S. Park, J.H. Lee, T. Noh, W.I. Yang, Y.
Rhee, S.K. Lim, H.C. Lee, E.J. Lee, Expression and role of estrogen receptor alpha
and beta in medullary thyroid carcinoma: different roles in cancer growth and
apoptosis, J. Endocrinol. 195 (2) (2007) 255–263.
[33] Y. Huang, W. Dong, J. Li, H. Zhang, Z. Shan, W. Teng, Differential expression
patterns and clinical significance of estrogen receptor-alpha and beta in
papillary thyroid carcinoma, BMC Cancer 14 (2014) 383.
[34] R. Tahboub, B.M. Arafah, Sex steroids and the thyroid, Best Practi. Res. Clin.
Endocrinol. Metab. 23 (6) (2009) 769–780.
[35] M.M. Abdel-Dayem, M.S. Elgendy, Effects of chronic estradiol treatment on the
thyroid gland structure and function of ovariectomized rats, BMC Res. Notes 2
(2009) 173.
[36] N. Sanger, S. Stahlberg, T. Manthey, K. Mittmann, U. Mellinger, E. Lange, H. Kuhl,
I. Wiegratz, Effects of an oral contraceptive containing 30 mcg ethinyl estradiol
and 2 mg dienogest on thyroid hormones and androgen parameters:
conventional vs. extended-cycle use, Contraception 77 (6) (2008) 420–425.
[37] I. Wiegratz, E. Kutschera, J.H. Lee, C. Moore, U. Mellinger, U.H. Winkler, H. Kuhl,
Effect of four oral contraceptives on thyroid hormones, adrenal and blood
pressure parameters, Contraception 67 (5) (2003) 361–366.