Reproductive Risk Factors for Thyroid Cancer: A Prospective Cohort Study in Khon Kaen, Thailand

Thyroid cancer is the most common cancer of the endocrine glands. The global age-standardized incidence rate (ASR) is 3.9 in males, and 12.3 in females per 100,000 person-years (Sassolas et al., 2009). In Thailand, it is among the top ten leading cancers in females, with an ASR of 1.0 in males and 3.7 in females (Khuhaprema et al., 2012). Moreover, its incidence has been growing during the last three-decade around the world (Pellegriti et al., 2013). The reason for this increasing incidence is unclear but it has been suggested that it may be due to more frequent use of sensitive diagnostic tools, such as droppler ultrasonography and PET scans; or a genuine increase. In fact, it can well be due to a combination of both factors (Pellegriti et al., 2013). The thyroid is a radiosensitive organ in a position potentially exposed to the radioactive substances in the environment. Thus, it is known that exposure to ionizing radiation, especially in childhood, increases the risk of developing thyroid cancer (Silverman, 1984; IARC, 2001; Zielinski et al., 2009; Furukawa et al., 2013) However, it is still not clear why thyroid cancer is markedly predominant in females. Previous observational studies on the reproductive system have produced conflicting results (Akslen et al., 1992; Pham et al., 2009; Horn-Ross et al., 2011; Kabat et al., 2012). We, therefore, investigated potential risk factors for thyroid cancer using data from


Introduction
Thyroid cancer is the most common cancer of the endocrine glands. The global age-standardized incidence rate (ASR) is 3.9 in males, and 12.3 in females per 100,000 person-years (Sassolas et al., 2009). In Thailand, it is among the top ten leading cancers in females, with an ASR of 1.0 in males and 3.7 in females (Khuhaprema et al., 2012). Moreover, its incidence has been growing during the last three-decade around the world (Pellegriti et al., 2013). The reason for this increasing incidence is unclear but it has been suggested that it may be due to more frequent use of sensitive diagnostic tools, such as droppler ultrasonography and PET scans; or a genuine increase. In fact, it can well be due to a combination of both factors (Pellegriti et al., 2013).
The thyroid is a radiosensitive organ in a position potentially exposed to the radioactive substances in the environment. Thus, it is known that exposure to ionizing radiation, especially in childhood, increases the risk of developing thyroid cancer (Silverman, 1984;IARC, 2001;Zielinski et al., 2009;Furukawa et al., 2013) However, it is still not clear why thyroid cancer is markedly predominant in females. Previous observational studies on the reproductive system have produced conflicting results (Akslen et al., 1992;Pham et al., 2009;Horn-Ross et al., 2011;Kabat et al., 2012). We, therefore, investigated potential risk factors for thyroid cancer using data from the Khon Kaen Cohort Study, with the main objective of determining the effect of reproductive factors on the occurence of thyroid cancer.

Materials and Methods
A prospective cohort study was conducted based on the Khon Kaen Cohort Study (KKCS). A total of 24,528 subjects were recruited between 1990 and 2001. The details of the cohort have been described in a previous publication (Sriamporn et al., 2005). The inclusion criteria for this study were females in the age group 30-69. The resulting number of subjects in the analysis was 10,767 subjects.
The data on determinants, mainly reproductive factors; i.e. the age of onset of the menarche, gravidity, oral contraceptive use, and major risk factors for head and neck cancer, i.e. tobacco smoking and alcohol drinking, were obtained by questionnaire interview during the cohort recruitment phase.
The cohort was followed up until December 31st, 2011 to detect the occurrence of thyroid cancer (ICD-O: C73). The 13-digit ID number was used to link the cohort to the Khon Kaen population-based cancer registry and double-checked with the subject's name, birth date, and address in questionable cases. The medical records of all detected cases were reviewed to confirm the diagnosis.
The incidence of thyroid cancer was calculated in relation to the person-time of the population at risk. This was also stratified by potential risk factors related to thyroid cancer.

Results
A total of 10,767 female subjects were included in the analysis, with a mean age of 50.3 years old (SD=8.7). There were 17 cases of thyroid cancer diagnosed during 152,508.41 person-years, giving an incidence of 11.2 per 100,000 person-years. The most common histopathology  was papillary carcinoma (12 cases, 70.6%), followed by follicular carcinoma (3 cases, 17.6%) and undifferentiated carcinoma (2 cases, 11.8%).
Thyroid cancer tended to occur in non-farmers with the higher incidence in the younger age group (Table 1). Although the incidence was high in the young, there was only one case found in a single female. The body mass index (BMI) between 18.5-22.9 was found to have the lowest incidence of thyroid cancer. Tobacco smoking and alcohol drinking slightly increased the risk of thyroid cancer, especially after excluding undifferentiated carcinoma.
For reproductive factors, the incidence of thyroid cancer was found to be apparently higher among those with an early age at menarche and nulligravida women (Table 2). Moreover, the incidence of thyroid cancer was lower in those with higher parity. No difference was found among different ages of first pregnancy. The incidence of well-differentiated thyroid carcinoma was also higher among oral contraceptive users, especially for those who had used the pill for five years or more.

Discussion
The incidence of thyroid cancer in this cohort was 11.2 per 100,000 with papillary carcinoma being the most common type found. The pattern of disease occurrence in this cohort is comparable to that from the population-based cancer registry of the same province (Khuhaprema et al., 2012). Unlike cancer of the upper aerodigestive tract in the same cohort, thyroid cancer tended to occur in relatively high socioeconomic status individuals (Kampangsri et al., 2013). As the number of cases of tobacco smokers and alcohol drinkers were limited, we could not make any conclusions about the relevance of these factors. The previous pooled analysis, nevertheless, showed that smoking associated with reducing the risk of thyroid cancer (Mack et al., 2003), due to its anti-estrogenic effect, and its tendency to lower TSH (Wiersinga, 2013).
Our study showed an apparently high incidence of thyroid cancer in those with an early age at menarche. Likewise, the previous study found a long reproductive period increased risk of papillary thyroid cancer (Akslen, 1992). However, the California Teacher Study showed later age of menarche was associated with an increased risk of thyroid cancer in young women, but not in the elderly (Horn-Ross et al., 2011). Those with less reproductive life (ie: never been pregnant or limited gravidity) were more likely to have thyroid cancer in our study. This reflects a trend of thyroid cancer to be associated with longer exposure to estrogen, agreeing with a previous study that revealed that estrogen significantly increased proliferation, migration and invasive properties in thyroid cancer cell lines (Rajoria et al., 2010).
We also found a higher incidence of thyroid cancer in oral contraceptive users with a dose-response relationship. However, the association between oral contraceptives and thyroid cancer requires further clarification, as the lower risk of thyroid cancer is related to a higher progesteroneto-estrogen ratio (Zivaljevic et al., 2003). In addition, the progesterone-to-estrogen ratio was different among  various populations, i.e. that in African Americans is higher than that in White Americans (Potischman et al., 2005). Exposure to high levels of estrogen, not progesterone, is likely to increase the risk of thyroid cancer. However, there was no conclusive evidence in relationship to the risk of any reproductive factor (Pham et al., 2009;Kabat et al., 2012).
This study is the first cohort study in Thailand to determine the incidence of thyroid cancer according to its risk factors with long term follow up. It was conducted in a province where a population-based cancer registry is well established. A few limitations, however, have to be addressed including: 1) the case number in the cohort is insufficient to use Cox's regression for analysis of the association, 2) thyroid cancer occurrence is likely to be under estimated, because of asymptomatic undiagnosed cases, due to the dormant nature of this cancer, 3) some data were collected at the initial assessment during the recruitment of the subjects in the cohort without followup information, resulting in potential misclassification.
In conclusion, this study showed a tendency for early menarche and low gravidity to be associated with an increased risk of thyroid cancer. It is still unclear how reproductive factors relate to thyroid cancer. Further research is needed to explain the greater incidence of thyroid cancer among females