Breast cancer is the second most prevalent cancer in the world and the most common cancer among women. Breast cancer still persists as an international health burden despite advances in its early diagnosis, treatment and also increased knowledge of its established risk factors.1 Although North American and Northern European countries have the highest rates of breast cancer incidence, the incidence of this disease is increasing in developing countries.2 Asia, the largest continent in the world, is a mixture of nations with different ethnics, religions, cultures and socioeconomic statuses. Incidence of breast cancer in developing Asian countries is different from developed countries in Asia and other parts of the world.3 Among the Asian countries, Iran has had one of the lowest incidence rates for breast cancer, but its increasing incidence during the last four decades has made it the most prevalent cancer among Iranian women (24 per 100,000), passing cancer of the uterine cervix.4 What is noteworthy is that due to the young age structure, mean age at the onset of breast cancer in Iranian women is about 10 years lower than that of their counterparts in developed countries.5 Breast cancer diagnosed in young patients tend to be more aggressive with worse prognosis than those detected in older women.6 Furthermore, breast cancer in young women is responsible for a greater potential years of life lost, namely, the number of life years which would be saved in the absence of disease.7 Age is one of the most important risk factors of breast cancer. Some of the recognized risk factors may differ relative to age at diagnosis and some have complicated age interactions.8, 9 The typical age incidence curve of breast cancer has a rapid rise until the age 40. After age 40, the rate of the increase slows; however, continuing to rise with increasing age until around 50, then it starts to decline, specifically in low risk populations.10 This inflection point around age 50, which is called the Clemmensen's hook, has been mainly attributed to menopause.11 Molecular, epidemiologic and clinical evidences suggested that Clemmensen's hook may reflect the superimposition of two different rate curves according to age at the onset, and it is the consequence of two main breast cancer types, which are mixed within the general population.12 The first type is early onset with its incidence peak about the age 50 years, associated with undifferentiated medulary and inflammatory breast carcinomas, and mostly estrogen receptor negative (ER−). The second type is late onset with incidence peak about the age 70 years, well differentiated papillary and mucious tumors, and mainly estrogen receptor positive (ER+).13, 14 The reproductive risk factors are more strongly associated with second type of breast cancer.15 Determination of breast cancer characteristics is very important for understanding the etiology of disease, its heterogeneity in the world and identifying high risk groups and health policies on prevention, early detection and treatment of breast cancer.16, 17
Epidemiological aspects of Iranian breast cancer are uncertain and no study has specifically addressed breast cancer risk factors in Iranian young women to date. The purpose of this study was to investigate the association of reproductive and anthropometric factors with risk of breast cancer on the Iranian young women.
Material and Methods
A hospital-based case–control study was performed to identify risk factors of breast cancer in young women of southern Iran.
Cases consist of women whose breast cancers were histologically confirmed and their data were collected in the Cancer Registry Center of Shiraz University of Medical Sciences. This center delivers oncology services to southern Iran, with a population of approximately four million. Controls were selected from female visitors of the hospital, who did not have history of breast cancer at enrollment, and were referred for conditions unrelated to known or likely risk factors for breast cancer, that is, gynecologic, hormonal, or neoplastic diseases.
Eligible cases included those women less than 50 years of age were interviewed between September 2005 and December 2008. Cases with unknown age and province of residence were excluded from the study. Cases were frequency-matched with controls on 5-year age groups and province of residence.
We used the standard questionnaire of cancer registry to collect data including information on the age, place of residence, marital status, educational level, family history of breast cancer (first- and second-degree relatives), age at menarche, occupation (employed or homemaker), parity, past use of oral contraceptives (OC) (ever or never use), age at first marriage, age at first pregnancy, menopausal status and lifetime duration of breast feeding. Women with no history of the menstrual period during last 12 months and women with history of hysterectomy or oophorectomy (not related to breast cancer treatment) were classified as postmenopausal. The questionnaire had been completed for all cases at the time of diagnosis, and we used the same questionnaire to collect information for the controls. Patients' data were provided to researchers without patients name and identification. All potential controls were informed about the protocol of the study and after having their oral consent; they were interviewed by a trained female interviewer at the hospital. Height and weight were taken at the end of the interview. Ethical approval of this study was obtained from the Research Ethics Committee of Shiraz University of Medical Sciences, Iran.
Univariate analyses were performed to calculate odds ratios (ORs) and 95% confidence intervals (CIs). Unconditional logistic regression was performed to estimate the predictive effect of each independent variable in risk of breast cancer when adjusting for other variables. All analyses were conducted using Stata 9.2 (Stata Corporation, College Station, TX).
A total of 521 women with breast cancer and 521 controls were studied. The mean age of cases and controls were 41.24 and 41.06, respectively (Table 1). In the univariate analysis, we found statistically significant differences between cases and controls for education, occupation, age at menarche, age at first marriage, age at first pregnancy, parity, breast feeding and family history of breast cancer, but no statistically significant differences were found between cases and controls for OC usage, marital status, menopausal status, height, weight and body mass index (BMI) (Table 2). After adjustment in the logistic model only occupation, parity, menopausal status, OC usage, family history of breast cancer and breast feeding revealed significant associations with the risk of breast cancer (Table 2).
Table 1. Age distribution of participants in the case control study in Southern Iran
Table 2. Distribution, odds ratios, and 95% confidence intervals (CI) for the factors investigated in the case control study in the Southern Iran
We had no information about family income of breast cancer patients. Therefore, we used level of education and occupation as proxies for socioeconomic status. Cases were more educated than controls (p < 0.001). Persons with university education were found more in cases than in controls (16.2% vs. 6.5%), and this difference was statistically significant. However, after adjustment for other variables, the effect of education declined. Employment rate was significantly higher among cases than in controls (20% vs. 8%) and this difference was significant (p < 0.001). After adjustment for other variables, employment was related to a higher risk of breast cancer.
In terms of marital status, married women were more in control group than cases (86% vs. 82%) and divorced/widow or never married persons were more among cases than controls but these differences were not significant (Table 2). Earlier age at menarche (<12 years) was associated with an increased risk of breast cancer compared with later age at menarche (≥15 years). After adjustment for other variables in multivariate analysis, women who had menarche before the age of 12 years were at a nonsignificantly increased risk of breast cancer. Age at first marriage showed a significant association with risk of breast cancer, and later age at first marriage was associated with an increased risk of breast cancer. Furthermore, those who never married compared with those who married under age 20 were at a higher risk of breast cancer. This relationship disappeared after adjustment for the other variables (Table 2). Age at first pregnancy showed an inverse relationship with risk of breast cancer, and those with first pregnancy over age 25 were at almost double risk of breast cancer (OR: 2.10; 95% CI: 1.47–3.02). Nulliparous women (this group contains both married and unmarried women) in comparison with those with first pregnancy under age 25 were at a higher risk of breast cancer. After adjustment for other variables, no significant relationship was observed between age at first pregnancy and the risk of breast cancer (Table 2). With respect to parity, a significant difference was seen between cases and controls (p < 0.001). Parity equal or more than three in comparison with parity 1–2 had a strong protective effect for developing breast cancer. Premenopausal women tended to be at a higher risk of breast cancer but this association was not significant. Lifetime duration of breast feeding was longer in controls than cases (mean ± standard deviation (SD) for controls and cases 57.4 ± 49.4 and 40.4 ± 39.7, respectively). Longer duration of breast feeding (37 months or more vs. less than 37 months) showed a protective effect on the risk of breast cancer. Cases more than controls used OC (66.3% vs. 62.9%). This difference was not significant in the univariate analysis but after adjustment for other variables, OC usage revealed a positive association with risk of breast cancer (Table 2).
There was no significant difference between cases and controls in terms of height and weight. For height, mean ± SD and median for controls and cases were 157.48 ± 6.3 and 157 versus 157.07 ± 6.2 and 157, respectively (Mann Whitney Test; p = 0.35), and for weight, mean ± SD and median for controls and cases were 67.3 ± 11.7 and 67 versus 67.2 ± 12.9 and 67, respectively (Mann Whitney Test; p = 0.27). Also, we found no significant association between height and weight and risk of breast cancer (Table 2). BMI in cases and controls was not significantly different, and mean ± SD and median for controls and cases were 27.1 ± 4.6 and 27.1 versus 27.2 ± 5.2 and 27.2, respectively (Mann Whitney Test; p = 0.78). In the analysis, we found no significant relationship between BMI and risk of breast cancer (Table 2).
This hospital-based case–control study, on young Iranian women found that family history of breast cancer, OC usage, low parity and employment were related to an increased risk of breast cancer, and longer time of breast feeding was associated with a decreased risk of breast cancer.
In this study, employed women were at a higher risk of breast cancer compared with those who were homemaker. Employed women usually are more educated and have more family income, so this relationship might be due to higher socioeconomic status; however, employed women might have exposure to other risk factors such as stress or carcinogens. Education showed a positive relationship with increased risk of breast cancer, but this association disappeared in multivariate analysis. It seems that association of higher education with risk of breast cancer is mostly due to other factors such as occupation, nulliparity, being older at first birth and long-term use of OC.
An early age at menarche appeared to be related to a higher risk of breast cancer. One explanation for this association is that women with earlier age at menarche are at exposure to endogenous estrogens for a longer time,18 but this is not clear whether this relationship is causally or because of correlation with other early life exposures.19 This study found a nonsignificant association between age at menarche and risk of breast cancer in the multivariate analysis, which is consistent with the other studies investigated this association in young or middle-aged women.20, 21
Younger age at first full-term pregnancy has been known as a protective factor in the lifetime risk of developing breast cancer. A plausible explanation for the protection effect is due to the higher degree of terminal differentiation of mammary epithelial cells at first birth, which makes the epithelium able to metabolize carcinogens and repair DNA damage more efficiently.22 Most studies,9, 21 but not all,23 have shown association between age at first pregnancy and breast cancer. Our results showed a doubled risk of breast cancer for women with age at first pregnancy 25 or more compared with younger than 25, but after adjustment for other variables in multivariate analysis, this association disappeared.
In our study, parity more than three had an inverse association with breast cancer risk and appeared to have the strongest protective effect on breast cancer risk among all other factors investigated. This protective effect of parity has also been shown in most other epidemiologic studies.24
We found that long-term breast feeding (life time duration more than 37 months) had an inverse association with risk of developing breast cancer, which is consistent with most other studies.25, 26 This relationship was persisted after adjustment for the potential confounding factors in the multivariate analysis such as parity, age at first pregnancy, age at menarche and education. Three studies have investigated the association of breast feeding with the risk of breast cancer in Iran, two of them found no significant association and the other found a weak inverse association between breast feeding and risk of breast cancer.27–29 Epidemiological studies conducted in other countries found that protective effect of breast feeding is confined to young premenopausal women.30, 31 Therefore, combining both younger and older women for analysis might hide or dilute the protective effect of breast feeding in young women, and this could be an explanation for the contradicting findings of the studies conducted in Iran.
The possibility of increasing risk of breast cancer due to OC usage has been researched extensively, but a lack of total consistency exists between studies. This inconsistency is attributed in part to changes in pattern of use, reductions in hormone dose and temporal considerations, which made it difficult to compare the many studies.32 This study consistent with other studies33 on young age groups showed OC usage is related to a higher risk of breast cancer in young women.
We found that family history of breast cancer in the close relatives was related to an increased risk of breast cancer, which suggest the existence of genetic or environmental factors that are shared among relatives and modify the risk of breast cancer. Our result was consistent with other epidemiological studies.34 Epidemiologic studies indicated that the risk of developing breast cancer related to genetic factors is greater in younger versus older women.34 Young age at onset of breast cancer is a powerful predictor of carrying BRCA1/2 mutations.35 Furthermore, the presence of BRCA1/2 genes mutations in young women is associated with the absence of estrogen and progesterone receptors and more aggressive breast cancer.36 A meta-analysis of 52 epidemiological studies on familial breast cancer showed that risk of breast cancer associated with family history, increased with decreasing age at onset of the disease in her relatives and with decreasing in her own age, especially for women younger than 50 years.34
We found no significant relationship between height, weight and BMI with risk of breast cancer in our study population. Height appeared to be related to slight increased risk of breast cancer in some studies37 but not in all.38 This inconsistent results in the association of height and the risk of breast cancer could be explained in some parts by variations in the socioeconomic status and malnourishment early in life,39 but may the association be confounded by other determinants of breast cancer risk.40 The association of overweight with risk of breast cancer in premenopausal women is investigated in many studies, but the results are inconsistent. Some studies found an inverse association between higher BMI and the risk of breast cancer,38 whereas others found a positive association.41 The association of BMI with the risk of breast cancer varies by race, menopausal status and hormone receptor status,42 and this can in some parts explain the controversial results between studies. BMI is clearly associated with a higher risk of estrogen receptor (ER) positive tumors than estrogen negative tumors.43 Also, breast cancer in young women (<50 years) is more likely to be ER negative.44 Younger age at onset of breast cancer in Iran and possibly more ER negative cases can be a reason for no association of BMI and breast cancer risk in our study.
During the last four decades, Iran had one of the highest population growth rates in the world with its pick in 1985 (3.94%), but from early 1990s to the present it has declined to about 1.3% because of nation-wide birth-control practices (Source: World Bank, world development indicators).The trend of decreasing parity and consequently, shorter lifetime duration of breast feeding along with a nation-wide increase in OC usage might be some of the underlying reasons for the rise of breast cancer incidence in Iranian young women. Another explanation could be changes in diet and lifestyle, which is in the process of rapid Westernization. On the other hand, Iran faces a health transition. In recent decades, an uptrend is observed in the life expectancy, age at marriage, age at pregnancy, OC usage, education level and employment rate among Iranian women, and national birth control policies caused severe reduction in parity.45 In addition, uptrend in prevalence of obesity made it a national problem in Iran.46 Reproductive factors have their strongest effects on postmenopausal breast cancer, and obesity is a risk factor for the old women.47 Iran is experiencing a youth bulge and will face aging in coming two decades. Therefore, although the incidence rate of breast cancer is increasing in young women, during the coming decades, we will face a more severe increase in the incidence of breast cancer in Iran, especially in postmenopausal women. One challenge is knowing whether the strategies in US and other developed countries in prevention, early detection and treatment of breast cancer will be effective or even possible in a developing country such as Iran with different ethnic, culture, socioeconomic status and religious beliefs. This situation should be noticed by health policy makers in prevention and early detection of breast cancer with regard to breast cancer characteristics in Iran and its features during the coming decades.
The strength of this study was sufficient sample size, which gave us opportunity to confine our study population to young women and for the first time focuses on this age group in Iran. This study was a hospital-based case–control study, so one weakness of this study is potential selection bias due to hospitalized controls. Recall bias might be another weakness, especially for information about family history of breast cancer in relatives. Another limitation of study is that we used current body weight of patients contrary to usual adult weight before the onset of the disease. It could be problematic if patients had weight loss; however, some studies have shown no weight reduction by stage of the disease.37, 48
Studies in developed countries with high prevalence of established risk factors showed that ∼50% of breast cancer risk is attributable to the established factors49 and this ratio is lower for young women.50 Therefore, further studies are necessary in young women to explore unknown risk factors. Also further studies are required to investigate possible causes of the rising trend of breast cancer incidence in Iran and other developing countries.
In conclusion, this was the first study focusing on risk factors of breast cancer in Iranian young women which showed association between family history, parity, occupation, breast feeding and OC usage with risk of breast cancer. The trend of decreasing parity and shortened duration of breast feeding along with OC usage may partly explain the increase of breast cancer incidence in Iranian young women.