Although the incidence of endometrial cancer (EC) is much lower in Asian than in Western countries, the incidence rate has been increasing and has almost doubled that of 1990 in Japan.[1, 2] Estrogens play a central role in the etiology of EC. When unopposed by progesterone, estrogens increase the mitotic activity of endometrial epithelial cells, increasing the likelihood of acquiring deleterious mutations that may result in EC.[3, 4] Events leading to the prolonged exposure of endometrium to unopposed estrogens are known risk factors of EC.[4, 5] Factors that influence estrogen concentrations might therefore be important in the prevention of EC.
Soy foods are an almost exclusive dietary source of a class of phytoestrogens called isoflavones, primarily represented by genistein and daidzein. Isoflavones are structurally similar to endogenous estrogen, and are of particular interest because of their ability to show both estrogenic and anti-estrogenic effects.[6, 7] Isoflavones from soy food intake have been found to decrease endogenous estrogen levels[8, 9] and stimulate the production of sex hormone-binding globulin in the liver, resulting in less free estradiol. They are also suggested to act as anti-estrogens by competing with the more potent endogenous estrogen for estrogen receptors. Moreover, in vitro studies have shown that isoflavones possess non-hormonal inhibitory properties, including inhibition of tyrosine protein kinases, cell growth, angiogenesis and induction of apoptosis, and antioxidant properties.[6, 10] Despite these possible cancer-protective properties, however, results of the few epidemiological studies of the relationship between dietary soy food and isoflavone and EC risk have been inconsistent.[12-18]
Because soy food intake varies widely among individuals in Asian countries, they are suitable venues in which to study the effects of dietary soy food and isoflavone intake. Regarding EC, however, only two case-control studies from an Asian country have been reported to date,[12, 16] and no prospective evaluation has been reported.
Here, we prospectively evaluated the association of soy food and isoflavone intake with EC risk in Japanese women in a large population-based cohort study.
The Japan Public Health Center-based Prospective Study (JPHC Study) was conducted in two cohorts, with cohort I initiated in 1990 and cohort II in 1993. The study population was defined as all registered Japanese inhabitants in 11 public health centre (PHC) areas, aged 40–59 years in cohort I and 40–69 years in cohort II. Study subjects were identified using population registries maintained by the local municipalities. Details of the study design have been described elsewhere.
The cohort participants were surveyed twice: the first survey was conducted at the time of initiation and the second was carried out 5 years later, in 1995 for cohort I and 1998 for cohort II. Participants were mailed a self-administered questionnaire on lifestyle and demographic characteristics and medical history, as well as a food frequency questionnaire (FFQ). Because information on food intake was more comprehensive in the 5-year follow-up survey than in the first survey, the present analysis was limited to women who responded to the 5-year follow-up survey, and the date of response to this survey was taken as the starting time. One PHC area was not included in the current analysis because of the lack of cancer incidence data (n = 4178). After excluding ineligible women (non-Japanese nationality, n = 20; late report of emigration occurring before the starting point, n = 73; incorrect birth date, n = 5; duplicate registration, n = 2), as well as those who moved out of the study area (n = 3542), died (n = 1090), were lost to follow up (n = 77), or whose EC was diagnosed before the starting point (n = 23), 62 688 women were found to be eligible for follow up, 52 416 of whom responded to the 5-year follow-up survey questionnaire, giving a participation rate of 83.6%. In the present analysis, we also excluded participants who had a history of surgery of the ovary or uterus (n = 1721), did not answer questions on soy food (n = 607) or who had energy consumption of <660 or >4450 kcal/day (n = 967), leaving a total of 49 121 women for analysis.
Dietary intake assessment
Dietary intake was assessed with a self-administered 138-item FFQ with pre-specified standard portion sizes and frequency of intake. Participants answered regarding the frequency of individual food items and the representative portion sizes relative to the standard portion size. Daily food intake was calculated by multiplying the frequency of food intake by the standard portion size and the relative portion size for each food item in the FFQ. The daily dietary intake of nutrients was calculated based on the 5th revised and enlarged edition of the Standard Tables of Food Composition in Japan.
The FFQ specifically inquired about the consumption of eight soy food items, namely, miso soup, soymilk, tofu for miso soup, tofu for other dishes, yushidofu (predrained tofu), koyadofu (freeze-dried tofu), aburaage (deep-fried tofu) and natto (fermented soybeans). Options for frequency and amount for miso soup was: almost never, 1–3 days/month, 1–2 days/week, 3–4 days/week, 5–6 days/week, or daily, and <1, 1, 2, 3, 4, 5, 6, 7–9 or >9 bowls, respectively. For soymilk, the FFQ contained questions on 10 frequency categories only: almost never, 1–3 times/month, 1–2 times/week, 3–4 times/week, 5–6 times/week, 1 glass/day, 2–3 glasses/day, 4–6 glasses/day, 7–9 glasses/day or >9 glasses/day. Options for frequency were: never, 1–3 times/month, 1–2 times/week, 3–4 times/week, 5–6 times/week, 1 time/day, 2–3 times/day, 4–6 times/day and ≥7 times/day, and the options for portion sizes were small (50% smaller than standard), medium (same as standard) and large (50% larger than standard) for other soy food items. The total dietary intake of genistein and daidzein was calculated by multiplying the amount of total food taken to the genistein and daidzein content of each food and summing across all the foods. The genistein and daidzein content of each food were taken from a specially developed food composition table for isoflavones in Japanese foods.[21, 22] Because the intake of genistein and daidzein were highly correlated, intake of genistein was taken as representative of isoflavone in the present analysis. The dietary intake of soy food estimated from the FFQ was validated in a subsample of both cohorts by comparison with intake in consecutive 14- or 28-day dietary record estimates. The Spearman correlation coefficient between energy-adjusted intake of genistein estimated from the FFQ and dietary records were 0.55 and 0.45 for cohort I and cohort II, respectively. The Spearman correlation coefficient for reproducibility between the two questionnaires for energy-adjusted genistein intake assessed 1 year apart was 0.69 and 0.41 for cohort I and cohort II, respectively.
Follow up and case ascertainment
Participants were followed from the date of completion of the 5-year follow-up survey questionnaire until 31 December 2009. Information on residential status, relocation and survival was obtained from the residential registry of each area. Information on cause of death was obtained from death certificates provided by the Ministry of Health, Labour and Welfare. Among the study participants, 3455 died, 1687 moved out of the study area and 153 women were lost to follow up during the study period. EC incidence was identified by active cancer patient notification through the major local hospitals in the study area and by data linkage with population-based cancer registries. Death certificates were used to supporting information on cancer incidence. The site and histology of EC was coded according to the International Classification of Diseases for Oncology, Third Edition (ICD-O-3), code C54.0–C54.9. During 596 444 person-years of follow up (mean follow up of 12.1 years), a total of 112 women were newly diagnosed with EC. EC was confirmed by histologic examination in 92.0% of cases (n = 103). The cases were histologically classified according to the WHO histological classification of endometrial tumours as adenocarcinoma in 93 (90.3%), other carcinoma in two (1.9%), other histologic types in five (4.9%), and undetermined in three (2.9%).
Person-years of follow up were calculated from the starting time until the date of EC diagnosis, date of moving out of a study area, date of death, or the end of 2009, whichever came first. Participants lost to follow up were censored on the last confirmed date of presence in the study area. Dietary intakes of soy food and isoflavones were adjusted for total energy intake by the residual method. Body mass index (BMI) was calculated as body weight in kilogrammes divided by the square of height in metres (kg/m2). Metabolic equivalent hours (MET-h) of physical activities were estimated by multiplying the reported time spent at each physical activity per day by its assigned MET intensity. Participants were divided into tertiles according to the intake of soy foods and isoflavone in all women. Hazard ratios (HR) and 95% confidence intervals (CI) were calculated by taking the lowest tertile of intake as the reference. Cox proportional hazards regression analyses were used to test the association of soy food and isoflavone intake with the incidence of EC. The models were adjusted for age (continuous), PHC, physical activity (tertile, MET-h/day), smoking (never-smoker, past/current smoker), alcohol consumption (non/occasional-drinker, regular drinker), BMI (<21, 21 to <24, 24 to <27, ≥27 kg/m2), past history of diabetes mellitus and cancer, age at menarche (<14, 14 to <16, ≥16), exogenous hormone use (yes, no), number of deliveries (nulliparous, 1, 2–3, ≥4), menopausal status and age at menopause (<50, 50 to <55, ≥55 years) for postmenopausal women, and coffee intake (none, 1–4 cups/week, 5–7 cups/week, ≥2 cups/day). Further adjustment was made for energy-adjusted intake of dietary fibre, fruits and vegetables (tertile categories, g/day). All P-values reported were two-sided and significance level was set at <0.05. All analyses were conducted with sas version 9.3 (SAS Institute Inc., Cary, NC, USA).
Baseline characteristics of the study participants according to tertile of intake of isoflavone are presented in Table 1. Women consuming more isoflavone were slightly older, less likely to be current smokers, alcohol and regular coffee drinkers or premenopausal, and more likely to have a history of diabetes mellitus. Participants with higher isoflavone intake also had a higher dietary intake of total fibre, fruits and vegetables, whereas there appeared to be little difference in terms of BMI, physical activity and energy consumption. Median intake of soy food ranged from 38.9 g/day in the lowest tertile to 129.6 g/day in the highest tertile.
Table 1. Baseline characteristics of participants according to tertile (T) categories of energy-adjusted isoflavone intake
|Age (year), mean (SD)||56.9 (8.4)||57.1 (7.8)||57.9 (7.6)|
|BMI (kg/m2), mean (SD)||23.4 (3.3)||23.4 (3.2)||23.6 (3.2)|
|Physical activity (MET-h/day), mean (SD)||24.9 (9.6)||25.2 (9.5)||25.1 (9.3)|
|Current smoker, %||7.8||5.6||5.2|
|Current alcohol drinker, %||13.5||12.5||11.0|
|Age at menarche (year), mean (SD)||14.6 (1.9)||14.7 (1.9)||14.8 (1.9)|
|Current exogenous female hormones use, %||2.6||2.4||2.6|
|Age at first delivery (year), mean (SD)||25.0 (3.6)||24.9 (3.3)||24.7 (3.3)|
|Menopausal status (premenopausal), %||25.5||22.3||18.0|
|Age at menopause (≥55 years old), %||3.8||3.9||4.3|
|History of diabetes mellitus, %||2.7||3.0||4.0|
|Coffee (≥1 cup/day), %||42.3||33.4||26.6|
| Dietary intake, mean (SD) |
|Total energy (kcal/day)||1866 (661)||1891 (635)||1835 (588)|
|Total fibre (g/day)a||11.9 (4.3)||13.7 (4.1)||16.1 (4.8)|
|Fruits (g/day)a||239 (201)||254 (175)||262 (177)|
|Vegetables (g/day)a||213 (137)||241 (133)||259 (144)|
Hazard ratios of EC by intake of isoflavone (genistein), soy food, tofu and miso soup are shown in Table 2. Non-dietary factor-adjusted HR and 95% CI estimates in the highest compared with the lowest tertile category for genistein (1.06; 0.62–1.82), soy food (1.11; 0.65–1.92), tofu (1.13; 0.65–1.93) and miso soup (0.89; 0.50–1.60) showed no association with EC. Further adjustment for the energy-adjusted dietary intake of fibre, fruits and vegetables did not change the estimates substantially (data not shown). Stratified analysis by menopausal status, BMI (≤25 or >25 kg/m2), median age (in two groups), and coffee consumption (< or ≥1 cup/day) showed no appreciable differences in the HR between strata (data not shown). Risk estimates after the omission of EC cases diagnosed in the first 2 years of follow up (n = 15), women with artificial menopause (four cases and 4533 non-cases), exogenous female hormone use (five cases and 1231 non-cases), and with history of diabetes mellitus (six cases and 1581 non-cases) were similar to those above.
Table 2. Hazard ratio (HR) and 95% confidence interval (CI) of endometrial cancer according to energy-adjusted isoflavone, soy food, tofu and miso soup intakea
| Isoflavone |
|No. of cases||37||32||43||0.74|
|HR (95% CI)||1.00 (ref.)||0.71 (0.41–1.25)||1.06 (0.62–1.82)|
|Per 15 mg/day|| || ||1.01 (0.84–1.22)|
| Soy food |
|No. of cases||34||32||46||0.63|
|HR (95% CI)||1.00 (ref.)||0.82 (0.47–1.44)||1.11 (0.65–1.92)|
|Per 25 g/day|| || ||1.02 (0.94–1.10)|
| Tofu |
|No. of cases||34||38||40||0.66|
|HR (95% CI)||1.00 (ref.)||1.00 (0.59–1.72)||1.13 (0.65–1.93)|
|Per 25 g/day|| || ||1.04 (0.94–1.14)|
| Miso soup |
|No. of cases||29||43||40||0.63|
|HR (95% CI)||1.00 (ref.)||1.14 (0.66–1.97)||0.89 (0.50–1.60)|
|Per 30 ml/day|| || ||0.99 (0.95–1.04)|
In this large study in Japanese women, we found no clear association of isoflavone, soy food, tofu, or miso soup intake with EC risk. To our knowledge, this is the first prospective study to investigate the association between a variety of dietary soy foods measures and the risk of EC in an Asian population. Our findings suggest that the intake of soy foods or isoflavones has no effect on the risk of EC.
Strengths and limitations
The present study had a number of strengths. It was a large population-based prospective study, with a reasonably high response rate (83.6%) and very little loss to follow up (0.3%). Information on dietary and lifestyle variables was collected prior to a diagnosis of EC, minimising the possibility of recall bias. Further, the FFQ used to estimate dietary intake data was validated and had a good correlation for validity, and the cancer registry was of sufficient quality to reduce the misclassification of outcomes.
A number of limitations should also be noted. Despite the large sample size and long follow-up period, the number of incident cancer cases was low, reflecting the low incidence rate of EC in Japan. Although our analysis on a continuous scale (per 15 mg/day increase in the genistein intake) showed no clear direction (HR = 1.01), the 95% CI was wide (0.84–1.22) and hence the possibility of a small decrease or increase in risk cannot be denied. Repeated measurement at regular intervals would likely provide a better estimate of exposure status. The present exposure was based on a single assessment, and therefore may not have correctly represented long-term intake. However, the correlation coefficient of reproducibility of isoflavone intake estimated from two questionnaires administered 5 years apart was reasonably high (0.61).[25, 26] Further, the overall estimated soy food intake in our study was comparable to that in a national nutritional survey, and the results of these national survey estimates have been fairly constant over the years, suggesting little change in soy food intake. The prevalence of hormone replacement therapy was very low, which avoided concerns over important confounding in the present study, albeit that it also limited the ability to investigate the effect of isoflavone in a high exogenous estrogen environment. Although we accounted for several measured potential confounders in the statistical model, we cannot fully exclude the effect of unmeasured or residual confounding. Also, the small number of cancer cases and limited analytical power prevented stratified analysis by other lifestyle or dietary factors. Isoflavone supplementation was not considered in the present study. However, a survey in 2006 found a very low prevalence (<1.6%) of isoflavone supplement use in Japan. Further, we also had no information on hysterectomy. Not accounting for hysterectomised women in the at-risk population would lead to the underestimation of incidence among women with intact uteri. However, when we analysed the data after excluding women with artificial menopause, possibly by hysterectomy, the results were virtually unchanged, and it is therefore unlikely that the findings were substantially affected by the lack of information on hysterectomy.
Previous studies of the relation between soy food measures and EC have not been consistent. A case-control study in non-Asian women in the San Francisco Bay area showed a statistically significant decrease in risk with increasing total isoflavone intake across all quartiles, with total isoflavone intake of 2.7 mg/day or more in the fourth quartile. In a recent prospective study in a multi-ethnic population, in contrast, a similar association was observed only in the highest intake group, with an isoflavone estimate of 7.82 mg/1000 kcal/day or higher. Although tofu and miso soup accounted for only about 17% of total isoflavone intake in this latter study, our present findings are in line with the lack of association seen in a subgroup with Japanese ethnicity in that study, in which more than 82% of isoflavone was reported to have originated from the consumption of these two foods; the results are also consistent with a case-control study in New Jersey. In contrast, a Chinese study which had consumption (median 42.5 mg/day) similar to that in the present study (35.5 mg/day) and higher consumption than the US-based studies[15, 17, 18] showed only a marginally significant inverse association with isoflavone. Soy food or tofu, the surrogate marker of isoflavone intake, was also found to be associated with EC in some studies,[13, 16] but not in others.[12, 14, 17, 18]As individual samples of the same soy food can have substantial variability in their isoflavone constituents, the difference in findings may in part be due to variation in the food item considered and the isoflavone database used. Such measurement error in exposure estimates attenuates the true association, and might partially explain the null finding in the present study. Nevertheless, it is unlikely that this null association was solely due to misclassification in exposure variable, as we have observed other associations using a similar analytic approach in the present cohort populations.[31, 32]
As isoflavone has higher affinity towards and appears to preferentially bind the estrogen receptor (ER)-β, its actions in carcinogenesis are thought to be regulated through an ER-β-mediated pathway.[33, 34] However, although both ER-α and ER-β are expressed in normal endometrium, levels of ER-α reportedly predominate.[34, 35] Unlike the case of colon or ovary cancer, in which ER-β expression is lost and the receptor is considered to function as a tumour suppressor, only a few studies have supported such a phenomenon in EC.[34-36]
Obesity is a known risk factor of EC, and the effect of soy intake on EC risk in lean and obese women is therefore of interest. The observed inverse association with a higher soy protein intake was confined to overweight women in the Chinese study. Horn-Ross et al. found that obese postmenopausal women consuming less isoflavone had an almost seven-fold higher risk than women with a BMI of <32.3 consuming ≥1500 μg/day of isoflavone. Obesity leads to increased estrogen concentrations from peripheral conversion of androgens to estrogens in adipose tissue by aromatase enzyme. One proposed mechanism for the effect of isoflavone in EC is the inhibition of this peripheral aromatase activity. If correct, the inverse findings of the US-based studies[15, 17] appear plausible, as approximately 50% of participants were overweight or obese; moreover, the lack of finding in present study is also unsurprising, given that participants had an average BMI of 23.4 kg/m2, and only 3.1% of women were over 30 kg/m2. Considering the differences in characteristics of the study populations, including wide range of soy food intakes, relatively lean body size, and low EC incidence, the results of our study might not extrapolate to other populations. Confirmation of these results awaits further studies in larger and more diverse populations. Note also that we evaluated the effect of habitual dietary consumption of traditional soy products, and our results might not therefore be comparable to those of isoflavone supplementation studies.
This study found no evidence to support the hypothesis that higher consumption of soy food and isoflavone is associated with the reduction of EC risk in Japanese women. Future studies with a greater number of cases and more precise assessment of exposure variables or use of related biomarkers are required to verify these findings.
Disclosure of interests
Contribution to authorship
SB and MI developed the concept, performed statistical analysis and prepared the first draft of the manuscript. MIw, NS, TY, TS, SS and MIn contributed to the collection and compilation of the data. ST was in charge of the design and implementation of the study. All authors contributed to the interpretation of the results and revision of the manuscript.
Details of ethics approval
The study protocol was approved by the institutional review board of the National Cancer Center, Tokyo, Japan (Approval no.: 13-012; date: 2013/08/06).
National Cancer Center Research and Development Fund (23-A31[toku]) (since 2011) and Grants-in-Aid for Cancer Research (from 1989 to 2010) and the Third-Term Comprehensive Ten-Year Strategy for Cancer Control from the Ministry of Health, Labour and Welfare of Japan.
We are indebted to the Aomori, Iwate, Ibaraki, Niigata, Osaka, Kochi, Nagasaki, and Okinawa Cancer Registries for providing their incidence data.
Members of the Japan Public Health Center-based Prospective Study (JPHC Study, Principal Investigator: S. Tsugane) Group are: S. Tsugane, S. Sasazuki, M. Iwasaki, N. Sawada, T. Shimazu, T. Yamaji, and T. Hanaoka, National Cancer Center, Tokyo; J. Ogata, S. Baba, T. Mannami, A. Okayama, and Y. Kokubo, National Cerebral and Cardiovascular Center, Osaka; K. Miyakawa, F. Saito, A. Koizumi, Y. Sano, I. Hashimoto, T. Ikuta, Y. Tanaba, H. Sato and Y. Roppongi, Iwate Prefectural Ninohe Public Health Center, Iwate; Y. Miyajima, N. Suzuki, S. Nagasawa, Y. Furusugi, N. Nagai, Y. Ito and S. Komatsu, Akita Prefectural Yokote Public Health Center, Akita; H. Sanada, Y. Hatayama, F. Kobayashi, H. Uchino, Y. Shirai, T. Kondo, R. Sasaki, Y. Watanabe, Y. Miyagawa, Y. Kobayashi, M. Machida, K. Kobayashi and M. Tsukada, Nagano Prefectural Saku Public Health Center, Nagano; Y. Kishimoto, E. Takara, T. Fukuyama, M. Kinjo, M. Irei, and H. Sakiyama, Okinawa Prefectural Chubu Public Health Center, Okinawa; K. Imoto, H. Yazawa, T. Seo, A. Seiko, F. Ito, F. Shoji and R. Saito, Katsushika Public Health Center, Tokyo; A. Murata, K. Minato, K. Motegi, T. Fujieda and S. Yamato, Ibaraki Prefectural Mito Public Health Center, Ibaraki; K. Matsui, T. Abe, M. Katagiri, M. Suzuki, K. and Matsui, Niigata Prefectural Kashiwazaki and Nagaoka Public Health Center, Niigata; M. Doi, A. Terao, Y. Ishikawa, and T. Tagami, Kochi Prefectural Chuo-higashi Public Health Center, Kochi; H. Sueta, H. Doi, M. Urata, N. Okamoto, and F. Ide and H. Goto, Nagasaki Prefectural Kamigoto Public Health Center, Nagasaki; H. Sakiyama, N. Onga, H. Takaesu, M. Uehara, and T. Nakasone, Okinawa Prefectural Miyako Public Health Center, Okinawa; F. Horii, I. Asano, H. Yamaguchi, K. Aoki, S. Maruyama, M. Ichii, and M. Takano, Osaka Prefectural Suita Public Health Center, Osaka; Y. Tsubono, Tohoku University, Miyagi; K. Suzuki, Research Institute for Brain and Blood Vessels Akita, Akita; Y. Honda, K. Yamagishi, S. Sakurai and N. Tsuchiya, University of Tsukuba, Ibaraki; M. Kabuto, National Institute for Environmental Studies, Ibaraki; M. Yamaguchi, Y. Matsumura, S. Sasaki, and S. Watanabe, National Institute of Health and Nutrition, Tokyo; M. Akabane, Tokyo University of Agriculture, Tokyo; T. Kadowaki and M. Inoue, The University of Tokyo, Tokyo; M. Noda and T. Mizoue, National Center for Global Health and Medicine, Tokyo; Y. Kawaguchi, Tokyo Medical and Dental University, Tokyo; Y. Takashima and Y. Yoshida, Kyorin University, Tokyo; K. Nakamura, Niigata University, Niigata; S. Matsushima and S. Natsukawa, Saku General Hospital, Nagano; H. Shimizu, Sakihae Institute, Gifu; H. Sugimura, Hamamatsu University School of Medicine, Shizuoka; S. Tominaga, Aichi Cancer Center, Aichi; N. Hamajima, Nagoya University, Aichi; H. Iso and T. Sobue, Osaka University, Osaka; M. Iida, W. Ajiki, and A. Ioka, Osaka Medical Center for Cancer and Cardiovascular Disease, Osaka; S. Sato, Chiba Prefectural Institute of Public Health, Chiba; E. Maruyama, Kobe University, Hyogo; M. Konishi, K. Okada, and I. Saito, Ehime University, Ehime; N. Yasuda, Kochi University, Kochi; S. Kono, Kyushu University, Fukuoka; S. Akiba, Kagoshima University, Kagoshima.