IJC International Journal of Cancer

Transcription

1 IJC International Journal of Cancer Reproductive and menstrual factors and risk of differentiated thyroid carcinoma: The EPIC study Raul Zamora-Ros 1 *, Sabina Rinaldi 1 *, Carine Biessy 1, Anne Tjïnneland 2, Jytte Halkjær 2, Agnes Fournier 3,4, Marie-Christine Boutron-Ruault 3,4, Sylvie Mesrine 3,4, Kaja Tikk 5, Renee T. Fortner 5, Heiner Boeing 6, Jana F orster 6, Antonia Trichopoulou 7,8, Dimitrios Trichopoulos 7,8,9, Eleni-Maria Papatesta 7, Giovanna Masala 10, Giovanna Tagliabue 11, Salvatore Panico 12, Rosario Tumino 13, Silvia Polidoro 14, Petra H.M. Peeters 15,16, H.B(as) Bueno-de-Mesquita 16,17,18, Elisabete Weiderpass 19,20,21,22, Eiliv Lund 19, Marcial Arg uelles 23, Antonio Agudo 24, Esther Molina-Montes 25,26, Carmen Navarro 26,27,28, Aurelio Barricarte 26,29, Nerea Larra~naga 26,30, Jonas Manjer 31, Martin Almquist 32, Maria Sandstr om 33, Joakim Hennings 34, Konstantinos K. Tsilidis 35,36, Julie A. Schmidt 35, Kay-Thee Khaw 37, Nicholas J. Wareham 38, Isabelle Romieu 1, Graham Byrnes 1, Marc J. Gunter 16, Elio Riboli 16 and Silvia Franceschi 1 1 International Agency for Research on Cancer (IARC), Lyon, France 2 Danish Cancer Society Research Center, Copenhagen, Denmark 3 Inserm, Centre for Research in Epidemiology and Population Health (CESP), U1018, Villejuif, France 4 Paris South University, UMRS Inserm 1018 Team 9, Villejuif, France 5 Division of Cancer Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany 6 Department of Epidemiology, German Institute of Human Nutrition Potsdam-Rehbr ucke, Nuthetal, Germany 7 Hellenic Health Foundation, Athens, Greece 8 Bureau of Epidemiologic Research, Academy of Athens, Athens, Greece 9 Department of Epidemiology, Harvard School of Public Health, Boston, MA 10 Molecular and Nutritional Epidemiology Unit, Cancer Research and Prevention Institute ISPO, Florence, Italy 11 Lombardy Cancer Registry Unit, Fondazione IRCCS Istituto Nazionale Tumori, Milano, Italy 12 Dipartimento di Medicina Clinica e Chirurgia, Federico II University, Naples, Italy 13 Cancer Registry and Histopathology Unit, Civic M.P. Arezzo Hospital, ASP Ragusa, Italy 14 Laboratory of Molecular Epidemiology, Human Genetics Foundation (HuGeF), Torino, Italy 15 Department of Epidemiology, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands 16 School of Public Health, Imperial College London, London, United Kingdom 17 National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands 18 Department of Gastroenterology and Hepatology, University Medical Center Utrecht, Utrecht, The Netherlands 19 Department of Community Medicine, University of Tromsï, Tromsï, Norway 20 Cancer Registry of Norway, Oslo, Norway 21 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden 22 Samfundet Folkh alsan, Helsinki, Finland 23 Public Health Directorate, Asturias, Spain 24 Unit of Nutrition, Environment and Cancer, Cancer Epidemiology Research Program, Catalan Institute of Oncology, Barcelona, Spain 25 Andalusian School of Public Health, Granada, Spain Key words: reproductive factors, menstrual factors, hormone use, differentiated thyroid carcinoma, EPIC Conflict of interest: Nothing to report *R.Z.-R. and S.R. contributed equally to this work Grant sponsors: This work was supported by the European Commission: Public Health and Consumer Protection Directorate 1993 to 2004; Research Directorate-General 2005; the French National Cancer Institute (L Institut National du Cancer; INCA); Grant Number: ; Grant sponsors: Ligue contre le Cancer, Institut Gustave Roussy, Mutuelle Generale de l Education Nationale, Institut National de la Sante et de la Recherche Medicale (INSERM) (France); German Cancer Aid; German Cancer Research Center (DKFZ); German Federal Ministry of Education and Research; Danish Cancer Society; Health Research Fund (FIS) of the Spanish Ministry of Health (RTICC (DR06/0020/ 0091); the participating regional governments from Asturias, Andalucıa, Murcia, Navarra and Vasco Country and the Catalan Institute of Oncology of Spain; Cancer Research UK; Medical Research Council, UK; the Stroke Association, UK; British Heart Foundation; Department of Health, UK; Food Standards Agency, UK; the Wellcome Trust, UK; the Hellenic Health Foundation; Italian Association for Research on Cancer; Compagnia San Paolo, Italy; Dutch Ministry of Public Health, Welfare and Sports; Dutch Ministry of Health; Dutch Prevention Funds; LK Research Funds; Dutch ZON (Zorg Onderzoek Nederland); World Cancer Research Fund (WCRF); Statistics Netherlands (The Netherlands); Swedish Cancer Society; Swedish Scientific Council; Regional Government of Skane, Sweden; Nordforsk Centre of Excellence programme; Some authors are partners of ECNIS, a network of excellence of the 6FP of the EC. DOI: /ijc History: Received 29 Apr 2014; Accepted 11 June 2014; Online 8 July 2014 Correspondence to: Dr. Sabina Rinaldi, International Agency for Research on Cancer, Lyon, France, rinaldis@iarc.fr

2 Zamora-Ros et al CIBER Epidemiologıa y Salud Publica (CIBERESP), Madrid, Spain 27 Department of Epidemiology, Murcia Regional Health Council, Murcia, Spain 28 Department of Health and Social Sciences, Universidad de Murcia, Murcia, Spain 29 Public Health Institute of Navarra, Pamplona, Spain 30 Public Health Division of Gipuzkoa, Regional Government of the Basque Country, San Sebastian, Spain 31 Deparment of Surgery, Malm o University Hospital, Malm o, Sweden 32 Department of Surgery, University Hospital Lund, Lund, Sweden 33 Department for Radiation Sciences, Oncology Umeå University, Umeå, Sweden 34 Department of Surgical and Perioperative Sciences, Umeå University, Umeå, Sweden 35 Cancer Epidemiology Unit, Nuffield Department of Population Health, University of Oxford, Oxford, United Kingdom 36 Department of Hygiene and Epidemiology, University of Ioannina School of Medicine, Ioannina, Greece 37 Department of Public Health and Primary Care, University of Cambridge, Cambridge, United Kingdom 38 MRC Epidemiology Unit, University of Cambridge, Cambridge, United Kingdom Differentiated thyroid carcinoma (TC) is threefold more common in women than in men and, therefore, a role of female hormones in the etiology of differentiated TC has been suggested. We assessed these hypotheses in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort. Among 345,157 women (mean age 51) followed for an average of 11 years, 508 differentiated TC cases were identified. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazards regression models. No significant associations were observed between differentiated TC risk and number of pregnancies, breast feeding, menopausal status, and age at menarche and at menopause. Significant associations were found with history of infertility problems (HR 1.70; 95% CI ), a recent pregnancy (HR for 5 vs. >5 years before recruitment 3.87; 95% CI ), menopause type (HR for surgical vs. natural menopause: 2.16; 95% CI ), oral contraceptive (OC) use at recruitment (HR: 0.48; 95% CI ) and duration of OC use (HR for 9 vs. 1 year: 0.66; 95% CI: ). An increased risk was also found with hormone replacement therapy use at recruitment (HR , 95% CI ), but this was not significant after adjustment for type of menopause (HR , 95% CI ). Overall, our findings do not support a strong role of reproductive and menstrual factors, and female hormone use in the etiology of differentiated TC. The few observed associations may be real or accounted for by increased surveillance in women who had infertility problems, recent pregnancies or underwent surgical menopause. What s new? Differentiated thyroid carcinoma (TC) is three-fold more common in women than in men. This raises the question: Do female hormones play a role in the pathogenesis of TC? In this large, prospective study, the authors did not find any strong association between reproductive or menstrual factors and TC risk. They did, however, identify positive associations between infertility, recent pregnancy, or surgical menopause and TC risk, and also detected an inverse association with prolonged use of oral contraceptives. Thyroid carcinoma (TC) is the most common cancer of the endocrine system and the second most common cancer (after breast) among young women in high-income countries. 1,2 TC incidence has increased markedly worldwide over the last decades, and this increase is only partially explained by increased surveillance of thyroid nodules. Thus, TC has now become an important clinical and economic burden. 2 The only well-known risk factors for TC are ionizing radiation exposure, especially during childhood, and benign thyroid disease. Evidence from large epidemiological studies also shows that height and weight may play a role. 3,4 The incidence of TC is about threefold higher in women than in men, 1 and the excess in TC risk in women is the largest during women s reproductive years. 5,6 Therefore, factors related to reproductive and menstrual history have been suggested to be relevant to TC etiology. However, results from case control 6 15 and prospective studies showed weak and inconsistent results on the associations between pregnancy, parity, menstrual cycle regularity, exogenous hormone use and menopausal status and TC risk. The aim of our study was to investigate prospectively the association between reproductive and menstrual factors, and exogenous hormones use with the risk of differentiated TC in women who participated in a large European cohort, the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Because of its large sample size and the high number of cases cumulated, we were able to evaluate these relationships in all women combined and separately for young women.

3 1220 Reproductive factors and differentiated thyroid carcinoma in women Material and Methods Subjects and study design The EPIC project is a European network of prospective cohorts that was set up to examine relationships between diet and environmental factors and cancer risk. The EPIC cohort includes about half a million men and women, recruited from 23 regional and national research centers located in ten western European countries: Denmark, France, Germany, Greece, Italy, Norway, Spain, Sweden, The Netherlands and United Kingdom. The rationale and study design were previously described in detail. 23 All participants provided written informed consent, and the Internal Review Boards of the International Agency of Research on Cancer and all boards from recruitment centers approved the EPIC project. The present study is based on data from a population of 345,157 women after excluding those with prevalent cancer other than non-melanoma skin cancer at baseline. Participants were also excluded if they had missing lifestyle (n 5 573) or anthropometry data (n 5 819). Identification and selection of thyroid cancer cases In most of the EPIC countries (Denmark, France, Italy, Norway, Spain, Sweden, The Netherlands and United Kingdom), data on vital status were ascertained by record linkage with regional and/or national mortality registries. In Greece, Germany and Naples (Italy) data on vital status were continuously collected through active follow-up. 23 In all centers except those in Greece, Germany and France, incident cancer cases were identified through record linkage with regional cancer registries. In France, Germany and Greece, follow-up for cancer incidence was based on a combination of methods, including the use of health insurance records, contacts with cancer and pathology registries and active follow-up through study participants and their next-of-kin. Closure dates for our study were defined as the latest date of complete followup for both cancer incidence and vital status. The follow-up spans ended between December 2006 and December 2009, depending on the EPIC center. A total of 537 first primary incident TC cases (codes C73 as per the tenth Revision of the International Statistical Classification of Diseases, Injury and Causes of Death) in women were identified within the EPIC cohort at the time of analyses. TC cases with medullary cancer (n 5 22), undifferentiated cancers (n 5 4, anaplastic), lymphoma (n 5 1) and other specific morphologies (n 5 2, squamous cell and large cell carcinomas) were excluded from our study. Overall, 508 female first primary incident differentiated TC cases were included in our study. Reproductive and lifestyle characteristics Baseline questionnaire data (including detailed questions about current habitual diet, menstrual and reproductive history, lifetime history of tobacco smoking, consumption of alcoholic beverages and physical activity) and anthropometric measurements were collected from study participants in the period ,24 Information on age at menarche and at menopause, age at first and last full-term pregnancy, spontaneous or induced abortion, breast feeding, exogenous hormones use [oral contraceptives (OC) and hormone replacement therapy (HRT)] and fertility problems were recorded in all centers (except in Sweden, where information on surgical menopause was not collected). The number of full-term pregnancies was calculated as the sum of live born children and stillbirths. Except from Norway and Umeå (Sweden), where information about twin pregnancies was available, the number of full-term pregnancies is overestimated, as twin pregnancies counted twice. Women were considered premenopausal if reporting menstrual periods over the past 12 months before recruitment, or when they were below the age of 46 years, and had either a hysterectomy or missing information on menopausal status. Women were considered as postmenopausal when they reported (i) no menses over the past 12 months before recruitment or (ii) bilateral ovariectomy or (iii) either a hysterectomy or missing information on menopausal status at age 55 years or above. Other women were considered as perimenopausal/unknown menopause status, and are referred to as perimenopausal throughout the article. Statistical analyses Cox proportional hazard models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CI) of TC for each hormonal, reproductive and menstrual factor. Tests and graphs based on Schoenfeld residuals were used to assess the proportional hazards assumption, which was satisfied. Age was used as underlying primary dependent time variable, with entry time defined as the subject s age at recruitment and exit time as age at TC diagnosis, death or censoring date (lost or end of follow-up), whichever occurred first. All statistical models were stratified by study center and by age at recruitment (1-year category). Potential confounders included educational level (no formal education, primary, technical/professional school, secondary, university and not specified), smoking status (never, former and current smokers), alcohol intake (0, >0 15, > and 30 g/day), Cambridge physical activity index (inactive, moderately inactive, moderately active, active and unknown) 24 and body mass index (<18.5, , and 30). None of these alone or combined variables introduced a >10% change in HRs; therefore, all presented models were unadjusted. Trend tests across levels of exposure categories were performed on continuous variables, whereas for categorical variables the test has been computed by assigning consecutive scores to the categories as an ordinal variable. All associations were examined for differentiated TC in all women and separately in women below age 50 and for papillary TC only. All statistical analyses were performed using the Statistical Analysis System (SAS) software package, Version 9.3 (SAS Institute).

4 Zamora-Ros et al Table 1. Distribution of baseline characteristics of the study population All Denmark France Germany Greece Italy Norway Spain Sweden The Netherlands United Kingdom Cases n n 5 16 n n 5 66 n 5 20 n 5 67 n 5 31 n 5 49 n 5 21 n 5 11 n 5 27 Person-years 3,775, , , , , , , , , , ,903 Mean age recruitment (SD) (years) 50.9 (9.9) 56.8 (4.4) 52.7 (6.6) 49.1 (9.0) 53.3 (12.5) 50.6 (8.1) 48.1 (4.3) 48.3 (8.4) 52.0 (10.7) 51.0 (11.6) 48.2 (14.4) Mean length of follow-up (SD) (years) 11.0 (2.7) 11.0 (2.4) 10.4 (2.6) 9.9 (2.7) 9.8 (3.1) 11.2 (2.3) 9.7 (1.4) 12.1 (2.1) 13.2 (3.4) 11.8 (2.5) 11.2 (2.4) No. of cases (papillary) Number of full-term pregnancies (%) 1 None Age at first full-term pregnancy (years) (%) 1,2 < Years since last pregnancy (%) > Use of oral contraceptives (%) 1 Ever Never Use of hormone replacement therapy (%) 1 Ever Never Menopausal status (%) Premenopausal Postmenopausal Perimenopausal Type of menopause (%) 3 Natural NA Surgical NA

5 1222 Reproductive factors and differentiated thyroid carcinoma in women Table 1. Distribution of baseline characteristics of the study population (Continued) All Denmark France Germany Greece Italy Norway Spain Sweden The Netherlands United Kingdom Cases n n 5 16 n n 5 66 n 5 20 n 5 67 n 5 31 n 5 49 n 5 21 n 5 11 n 5 27 Person-years 3,775, , , , , , , , , , ,903 Age at menarche (years) (%) 1 < Age at menopause (years) (%) 1,3 < Body mass index (kg/m 2 ) (%) < Alcohol consumption (g/day) (%) > Smoking status (%) 1 Never Former Current Missing values: 14,688 (4.3%) for parous women; 9,073 (2.8%) among parous women for number of full-term pregnancies; 1,157 (0.4%) among parous women for age at first full-term pregnancy; 260,193 (89.7%) among parous women for years since last pregnancy; 8,972 (2.6%) for use of oral contraceptives; 22,737 (6.6%) for use of hormone replacement therapy; 11,559 (3.4%) for age at menarche; 33,405 (20.9%) among postmenopausal women for age at menopause; 8,416 (2.4%) for smoking status. 2 Among parous women only. 3 Among postmenopausal women only. Abbreviations: SD: standard deviation; NA: data not available.

6 Zamora-Ros et al Table 2. Hazard ratios of differentiated thyroid carcinoma in women according to reproductive variables at recruitment in EPIC 1 Differentiated thyroid carcinoma Women <50 at All women recruitment Papillary carcinoma py Cases HR (95% CI) 2 Cases HR (95% CI) 2 Cases HR (95% CI) 2 Overall 3,775, Full-term pregnancy No 547, (ref) (ref) (ref) Yes 3,061, ( ) ( ) ( ) Number of full-term pregnancies None 547, (ref) (ref) (ref) 1 541, ( ) ( ) ( ) 2 1,437, ( ) ( ) ( ) 3 970, ( ) ( ) ( ) p-trend Per 1-pregnancy increase 0.96 ( ) 0.96 ( ) 0.95 ( ) Age at first full-term pregnancy (years) 3 <23 960, (ref) (ref) (ref) , ( ) ( ) ( ) 26 1,183, ( ) ( ) ( ) p-trend Per 5-year increase 1.00 ( ) 0.92 ( ) 0.99 ( ) Ever breast feeding 3,4 No 432, (ref) (ref) (ref) Yes 2,407, ( ) ( ) ( ) Duration of breast feeding (months) 3,4 <1 570, (ref) (ref) (ref) , ( ) ( ) ( ) , ( ) ( ) ( ) , ( ) ( ) ( ) >24 301, ( ) ( ) ( ) p-trend Per 5-month increase 0.96 ( ) 1.05 ( ) 0.95 ( ) Number of abortions and miscarriages None 1,959, (ref) (ref) (ref) 1 644, ( ) ( ) ( ) 2 234, ( ) ( ) ( ) 3 128, ( ) ( ) ( ) p-trend Per 1-abortion/miscarriage increase 1.07 ( ) 1.10 ( ) 1.11 ( ) Infertility problems No 1,968, (ref) (ref) (ref) Yes 76, ( ) ( ) ( ) Outcome first pregnancy Full-term birth 1,817, (ref) (ref) (ref) Miscarriage 193, ( ) ( ) ( )

7 1224 Reproductive factors and differentiated thyroid carcinoma in women Table 2. Hazard ratios of differentiated thyroid carcinoma in women according to reproductive variables at recruitment in EPIC (Continued) Differentiated thyroid carcinoma Women <50 at All women recruitment Papillary carcinoma py Cases HR (95% CI) 2 Cases HR (95% CI) 2 Cases HR (95% CI) 2 Abortion 116, ( ) ( ) ( ) Never pregnant 475, ( ) ( ) ( ) Years since last pregnancy 5 16, ( ) ( ) ( ) >5 323, (ref) (ref) (ref) Never pregnant 475, ( ) ( ) ( ) 1 The sums may not add up to the total because of some missing values. 2 Derived from Cox regression models using age as the underlying time variable, and stratified by center and age at recruitment. 3 Among parous women only. 4 This information is only available for the first three full-term pregnancies and the last one. Abbreviations: HR: hazard ratio; CI: confidence interval; py: person-years. Results Baseline characteristics of the study women are presented in Table 1. Women were followed for a mean of 11 years for a total of 3,775,943 person-years. During this follow-up period, 508 cases of differentiated TC were identified, 218 of which were in women <50 years old at recruitment (42.9%) and 402 were classified as papillary TC (79.1%). In all countries, the majority of women had had pregnancies and had more than one child. More than 45% of women in Italy, the Netherlands and the United Kingdom were older than 26 years at their first full-term pregnancy. Use of OC varied widely between countries, with the highest percentage of never users in Greece (90.5%) and the lowest in Germany (18.8%). Large differences across countries were also observed for HRT use, with the highest percentage of never users in Greece (95.5%) and the lowest in Denmark (55.2%). On an average, <16% of women had their menarche after the age of 15 years and 46.5% were postmenopausal at recruitment. All countries had a similar percentage of women with surgical menopause (between 6.5 and 9.6%), except Spain (13.9%) and Norway (1.4%). Table 2 shows the associations of selected reproductive variables with differentiated TC risk in all women, and separately for women who were <50 years old at recruitment, or had papillary TC. No significant associations were observed with full-term pregnancies, age at first full-term pregnancy, breast feeding, abortions and miscarriages or with the outcome of the first pregnancy. Women who reported infertility were at significantly higher risk in relation to those with no history of infertility (HR , 95% CI ). Moreover, women who had their last pregnancy <5 years before recruitment were also at increased risk (HR for 5 vs. >5 years , 95% CI ). However, for both associations, categories at risk included small numbers of cases. Table 3 shows the associations of menstrual variables and female hormone use with differentiated TC risk in all women, and separately for women who were <50 years old at recruitment, or had papillary TC. Age at menarche, menopausal status and age at menopause were not significantly associated with the risk of differentiated TC. Postmenopausal women who had surgical menopause were at higher risk compared to those who had natural menopause (HR , 95% CI in all women). OC use (ever vs. never) was not significantly associated with differentiated TC risk but the duration of OC use was inversely associated (HR for 9 vs. 1 year5 0.66, 95% CI ). OC use at recruitment was rare but it was also associated with a significantly decreased risk (HR , 95% CI ). HRT use at recruitment was associated with a statistically significant increase risk of differentiated TC (HR , 95% CI ) (Table 3), but this association attenuated and no longer significant when adjusted for type of menopause (HR , 95% CI ). After stratifying by HRT type, a significant association between estrogen use alone at recruitmentanddifferentiatedtcwasobserved(hr5 1.67, 95% CI ), but this also disappeared after the adjustment for type of menopause (HR , 95% CI ). No association was found with other HRT use at recruitment (progestin alone or progestin and estrogen combined). Generally, associations observed for risk of papillary TC were similar to those observed for differentiated TC overall. Discussion In this prospective cohort study, the largest to date on differentiated TC risk and reproductive and menstrual factors, and female hormone use, no strong association was observed but positive associations were found with infertility problems, a recent pregnancy and surgical menopause. Prolonged OC use and OC use at recruitment were inversely associated with differentiated TC risk. In our study, parity, number of births, abortions and breast feeding were not significantly associated with differentiated TC risk. Our results support the hypothesis that reproductive factors do not substantially affect differentiated TC risk. 5,10 Our study provides some support to previous observations 12,19 of higher differentiated TC risk in women who became pregnant <5 years previous to recruitment.

8 Table 3. Hazard ratios of differentiated thyroid carcinoma in women according to exogenous hormone use and menstrual variables at recruitment in EPIC 1 Differentiated thyroid carcinoma Women <50 at All women recruitment Papillary carcinoma py Cases HR (95% CI) 2 Cases HR (95% CI) 2 Cases HR (95% CI) 2 Overall 3,775, Menstrual variables: Age at menarche (years) <13 1,322, (ref) (ref) (ref) ,731, ( ) ( ) ( ) , ( ) ( ) ( ) p-trend Per 1-year increase 1.00 ( ) 0.94 ( ) 1.00 ( ) Menopausal status Premenopausal 1,322, ( ) ( ) ( ) Perimenopausal 708, ( ) ( ) ( ) Postmenopausal 1,744, (ref) (ref) (ref) Type of menopause 3 Natural 1,636, (ref) (ref) Surgical 107, ( ) ( ) Age at menopause 3 <50 710, (ref) (ref) , ( ) ( ) , ( ) ( ) Unknown 344, p-trend Per 5-year increase 1.01 ( ) 1.00 ( ) Exogenous hormone use: OC use Never 1,519, (ref) (ref) (ref) Ever 2,145, ( ) ( ) ( ) Duration taken OC continuously (years) 1 1,917, (ref) (ref) (ref) , ( ) ( ) ( ) , ( ) ( ) ( ) 9 669, ( ) ( ) ( ) p-trend Per 1-year increase 0.98 ( ) 1.00 ( ) 0.98 ( ) OC use at recruitment No 3,425, (ref) (ref) (ref) Yes 188, ( ) ( ) ( ) HRT use Never 2,346, (ref) (ref) (ref) Ever 832, ( ) ( ) ( ) HRT use at recruitment No 2,643, (ref) (ref) (ref) Yes 548, ( ) ( ) ( ) 1 The sums may not add up to the total because of some missing values. 2 Derived from Cox regression models using age as the underlying time variable, and stratified by center and age at recruitment. 3 Among postmenopausal women only (with natural or surgical menopause). Abbreviations: HR: hazard ratio; CI: confidence interval; py: person-years; OC: oral contraceptives; HRT: hormone replacement therapy.

9 1226 Reproductive factors and differentiated thyroid carcinoma in women Pregnancy causes major changes in the levels of thyroid hormones (mainly tri-iodothyronine and thyroxine), 25 which have been associated with differentiated TC risk. 26,27 However, the number of women who had been pregnant during the 5 years before recruitment in our study was too small to draw firm conclusions on this issue (n 5 5). Women who reported infertility problems were at higher risk of differentiated TC. Infertility may be positively associated with previous thyroid problems as thyroid dysfunction has also been linked to reduced fertility. 28 Women who had infertility problems may have also taken fertility drugs (e.g., clomiphene), which have been directly associated with the risk of developing TC in a Danish cohort study. 18 An alternative explanation is that a recent pregnancy and infertility problems may have led to increased surveillance of thyroid function and nodules. A pooled analysis of 14 case control studies showed weak associations between TC risk and menstrual factors, particularly later age of menarche, which appeared to be stronger in young women. 10,19 Our results do not support the presence of associations with menopausal status, and age at menarche and menopause. Only surgical menopause was positively associated with a twofold of differentiated TC risk. This last association may again reflect enhanced medical surveillance of women who underwent surgical interventions because of symptoms of sudden ovarian failure. 10 An inverse association of differentiated TC with prolonged OC use or use at recruitment was observed, as reported in some, 12,17 but not in all previous studies. 7 HRT use showed a positive association with differentiated TC risk in agreement with the results of two large cohort studies, 19,29 particularly after estrogen-only use. However, the association with HRT use was a result of the positive correlation between HRT use and surgical menopause, as women after surgical menopause are preferentially prescribed estrogens without progestins. This is confirmed in our study population, where 24% of women after surgical menopause used estrogens-only vs. 4%inthosewith natural menopause. It is worth bearing in mind that, although differentiated TC incidence rates in women are nearly three times those in men, increases were observed in both genders in the last decades. 2,5 The earlier increases in women and the especially large gender difference below age 45 can be, however, explained by agerelated differences in the use of health services. Intense surveillance of thyroid nodules and thyroid function tend to occur in young and middle-age women due to events related to reproduction and perimenopausal and postmenopausal symptoms. References Conversely, men make more frequent medical visits after middle age because of various chronic conditions. 2,5 The results of our study do not support reproductive or menstrual factors as particularly important in influencing differentiated TC risk in young women. 19 To replicate the analyses of previous studies which had defined young women as those below 45 years of age 6,19 we repeated our analyses using that cutoff point (data not shown). The number of cases below 45 years of age was low in our study (n 5 108) but HRs were very comparable to those reported in women below 50 years (Tables 2 and 3). The similarity between overall findings and those for papillary TC is not surprising as the other histological type of differentiated TC, mainly follicular TC, only accounted for 20.9% of all female cases. Strengths of our study include its prospective design, the substantial number of women with differentiated TC and the availability of standardized information on reproductive and menstrual factors and female hormone use, as well as a wide range of potential lifestyle confounding factors. The most important limitation of our study is the availability of information only at cohort recruitment. This weakness is especially concerning for variables that are likely to have changed thereafter (e.g., menopausal status and HRT use). Conversely, information on reproductive factors is likely to be complete on account of the age distribution of women in EPIC (mean age 5 51 years). Other weaknesses include the lack of information on the type of OC, familiar susceptibility for TC and history of benign thyroid disease, iodine deficiency and past radiation exposure, which are well-known TC risk factors. 5 However, severe iodine deficiency is rare in EPIC countries and past heavy exposure to ionizing radiation should not be a problem as individuals who reported a history of cancer other than nonmelanoma skin cancer and therefore potentially received radiotherapy were excluded. In conclusion, our study provided little evidence of a relationship between differentiated TC risk and reproductive or menstrual factors. The few observed associations may be real or accounted for by increased surveillance in women who had infertility problems, recent pregnancies or underwent surgical menopause. Further research is needed on the strong female male difference in the risk of differentiated TC including additional studies of gender differences in circulating thyroid-stimulating hormone and thyroid hormones 30 and sex hormones, and the influence of genetic variants in genes involved in hormone pathways. 31,32 Acknowledgement The authors thank Mr. Bertrand Hemon for his precious help with the EPIC database. 1. Ferlay J, Soerjomataram I, Ervik M, et al. GLO- BOCAN 2012 v1.0, cancer incidence and mortality worldwide: IARC CancerBase No. 11 [Internet]. Lyon, France: International Agency for Research on Cancer, Kilfoy BA, Zheng T, Holford TR, et al. International patterns and trends in thyroid cancer incidence, Cancer Causes Control 2009;20: Rinaldi S, Lise M, Clavel-Chapelon F, et al. Body size and risk of differentiated thyroid carcinomas: findings from the EPIC study. Int J Cancer 2012; 131:E1004 E ZhaoZG,GuoXG,BaCX,etal.Overweight, obesity and thyroid cancer risk: a meta-analysis of cohort studies. JIntMedRes2012;40: Dal Maso L, Bosetti C, La Vecchia C, et al. Risk factors for thyroid cancer: an epidemiological review focused on nutritional factors. Cancer Causes Control 2009;20: Brindel P, Doyon F, Rachedi F, et al. Menstrual and reproductive factors in the risk of differentiated thyroid carcinoma in native women in French Polynesia: a population-based

10 Zamora-Ros et al case-control study. Am J Epidemiol 2008;167: La Vecchia C, Ron E, Franceschi S, et al. A pooled analysis of case-control studies of thyroid cancer. III. Oral contraceptives, menopausal replacement therapy and other female hormones. Cancer Causes Control 1999;10: Mack WJ, Preston-Martin S, Bernstein L, et al. Reproductive and hormonal risk factors for thyroid cancer in Los Angeles County females. Cancer Epidemiol Biomarkers Prev 1999;8: Memon A, Darif M, Al-Saleh K, et al. Epidemiology of reproductive and hormonal factors in thyroid cancer: evidence from a case-control study in the Middle East. Int J Cancer 2002;97: Negri E, Dal Maso L, Ron E, et al. A pooled analysis of case-control studies of thyroid cancer. II. Menstrual and reproductive factors. Cancer Causes Control 1999;10: Rossing MA, Voigt LF, Wicklund KG, et al. Reproductive factors and risk of papillary thyroid cancer in women. Am J Epidemiol 2000;151: Sakoda LC, Horn-Ross PL. Reproductive and menstrual history and papillary thyroid cancer risk: the San Francisco Bay Area thyroid cancer study. Cancer Epidemiol Biomarkers Prev 2002;11: Truong T, Orsi L, Dubourdieu D, et al. 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 2005;161: Zivaljevic V, Vlajinac H, Jankovic R, et al. Casecontrol study of female thyroid cancer menstrual, reproductive and hormonal factors. Eur J Cancer Prev 2003;12: Fernandez E, Gallus S, Bosetti C, et al. Hormone replacement therapy and cancer risk: a systematic analysis from a network of case-control studies. Int J Cancer 2003;105: Akslen LA, Nilssen S, Kvale G. Reproductive factors and risk of thyroid cancer. A prospective study of 63,090 women from Norway. Br J Cancer 1992;65: Dorjgochoo T, Shu XO, Li HL, et al. Use of oral contraceptives, intrauterine devices and tubal sterilization and cancer risk in a large prospective study, from 1996 to Int J Cancer 2009;124: Hannibal CG, Jensen A, Sharif H, et al. Risk of thyroid cancer after exposure to fertility drugs: results from a large Danish cohort study. Hum Reprod 2008;23: Horn-Ross PL, Canchola AJ, Ma H, et al. Hormonal factors and the risk of papillary thyroid cancer in the California Teachers Study cohort. Cancer Epidemiol Biomarkers Prev 2011;20: Meinhold CL, Ron E, Schonfeld SJ, et al. Nonradiation risk factors for thyroid cancer in the US Radiologic Technologists Study. Am J Epidemiol 2010;171: Navarro Silvera SA, Miller AB, Rohan TE. Risk factors for thyroid cancer: a prospective cohort study. Int J Cancer 2005;116: Braganza MZ, de Gonzalez AB, Schonfeld SJ, et al. Benign breast and gynecologic conditions, reproductive and hormonal factors, and risk of thyroid cancer. Cancer Prev Res (Phila) 2014;7: Riboli E, Hunt KJ, Slimani N, et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr 2002;5: Wareham NJ, Jakes RW, Rennie KL, et al. Validity and repeatability of a simple index derived from the short physical activity questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Public Health Nutr 2003;6: De Groot LJ, Mestman J. Detecting and treating thyroid nodules and cancer before, during, and after pregnancy: a patient s guide. J Clin Endocrinol Metab 2012;97:37A 38A. 26. McLeod DS, Watters KF, Carpenter AD, et al. Thyrotropin and thyroid cancer diagnosis: a systematic review and dose-response meta-analysis. J Clin Endocrinol Metab 2012;97: Stefaneanu L, Tasca C. Immunohistochemical localization of thyroglobulin (Tg), thyroxine (T4), and triiodothyronine (T3) in thyroid diseases. Acta Histochem 1986;78: Krassas GE, Poppe K, Glinoer D. Thyroid function and human reproductive health. Endocr Rev 2010;31: Schonfeld SJ, Ron E, Kitahara CM, et al. Hormonal and reproductive factors and risk of postmenopausal thyroid cancer in the NIH-AARP Diet and Health Study. Cancer Epidemiol 2011;35:e85 e Rinaldi S, Plummer M, Biessy C, et al. Thyroidstimulating hormone, thyroglobulin, and thyroid hormones and risk of differentiated thyroid carcinoma: the EPIC study. J Natl Cancer Inst 2014; 13:dju Gudmundsson J, Sulem P, Gudbjartsson DF, et al. Discovery of common variants associated with low TSH levels and thyroid cancer risk. Nat Genet 2012;44: Schonfeld SJ, Neta G, Sturgis EM, et al. Common genetic variants in sex hormone pathway genes and papillary thyroid cancer risk. Thyroid 2012; 22:151 6.