| Japanese Journal of Clinical Oncology | Pages |
Risk Factors for Breast Cancer in Japan, with Special Attention to Anthropometric Measurements and Reproductive History
Introduction
Study Subjects And Method
Cases
Controls
Source of Information
Statistical Analysis
Categorization
Results
Discussion
Acknowledgments
References
Risk Factors for Breast Cancer in Japan, with Special Attention to Anthropometric Measurements and Reproductive History
Methods: Based on information from a self-administered questionnaire at Osaka Medical Center for Cancer and Cardiovascular Diseases, body mass index, body weight and height were compared between 376 cases and 430 controls, together with other factors such as age at menarche, age at first delivery and family history of breast cancer by menopausal status. Logistic regression analysis was employed for adjusting confounding factors and estimating odds ratios with their 95% confidence interval for breast cancer.
Results: A body mass index of >25 was significantly associated with the risk among post-menopausal women (age-adjusted odds ratio: 1.90, 95% confidence interval: 1.10-3.24) as compared with the risk for a body mass index of [le]20. A weight of [ge]58 kg showed significantly increased risk compared with a weight of [le]47 kg among post-menopausal women (1.83, 1.10-3.01), while height of [ge]159 cm showed a significantly elevated risk than height of [le]149 cm among pre-menopausal women (2.51, 1.17-5.39). Age at menarche of [le]13 years resulted in a higher risk of breast cancer among post-menopausal women, while age at first delivery of [ge]28 years was associated with the risk among pre-menopausal women. Family history of breast cancer was associated with the risk for breast cancer.
Conclusions: These results were all very consistent with findings observed in western countries.
INTRODUCTION
Breast cancer is one of the most common cancers among women. The highest rates (age-standard incidence rates are >70/100 000) (1) occur in North America, Israel and temperate South America (2). High rates (50-70/100 000) are observed in North Western Europe and Australia. Intermediate rates (35-49/100 000) are found in part of Eastern Europe, tropical South America, the Philippines and Singapore. Low rates (<35/100 000) are generally seen in Asian countries. There has been a marked increase in some Asian countries, including Japan, in recent years. In Osaka recently, the breast has become the leading site of cancer among women in terms of age-standardized incidence rate to the world population (24.5/100 000 in 1994) (3). The number of breast cancer incidence is expected to become the first place of cancer sites among Japanese women by the year 2000 (4). There were only a few studies on the risk factors for breast cancer among Japanese women. A case-control study was therefore conducted to evaluate the roles of obesity, height, earlier age at menarche, later age at first delivery, later age at menopause, smoking, drinking, physical activity and family history of breast cancer in the etiology of breast cancer among Japanese women.
STUDY SUBJECTS AND METHOD
Cases
Eligible cases were women with newly diagnosed breast cancer at Osaka Medical Center for Cancer and Cardiovascular Diseases (OMCC) between 1990 and 1995. A total of 808 breast cancer patients were admitted to OMCC during this period. Among them, 376 breast cancer patients participated in this study by filling out a self-administered questionnaire survey which asked for information on weight, height, reproductive history, lifestyle and educational level. One hundred and ninety cases were pre-menopausal and 182 cases were post-menopausal, while four cases were of unknown menopausal status.
Controls
Controls were female patients without any diagnosis of cancer, who were admitted to OMCC during the same period as the eligible cases. Selection of the control patients was based on the hospital information on primary diagnoses for inpatients, and the data from the self-administered questionnaire survey at the time of admission (see the next section on Source of information). A total of 562 patients were identified as candidates for the controls. However, 132 of them were excluded because of the inadequate conditions for this study (changes in weight, nutritional status and physical activity related to illness): 40 patients with endocrine, nutritional and metabolic disease (including 37 patients with diabetes), 65 with heart failure or cerebrovascular disease, 19 with Parkinsonism, five with neoplasms of uncertain behavior, two with tuberculosis and one with iron deficiency anemia. Thus, a total of 430 were obtained. Forty-three percent of control subjects were admitted for benign tumors, 29% for disease of the digestive system, 6% for disease of the circulation system, 5% for disease of the nervous system and sensory organs, 5% for disease of the respiratory system, 4% for disease of the musculoskeletal system and connective tissue, and 8% for other diseases, such as infectious and parasitic disease, injury and poisoning, disease of the blood and blood formation. One hundred and nineteen patients among the controls were pre-menopausal and 282 were post-menopausal (spontaneous menopause), while 29 were of unknown menopausal status or had undergone an artificial menopause.
Source of Information
This study was based on the self-administered questionnaire survey, the hospital information on primary diagnoses for inpatients and the hospital cancer registry of OMCC. The questionnaire survey started in 1990 and included items about smoking, drinking alcohol, dietary habits, weight, height, physical activities, educational level, reproductive histories (only for females), medical histories and family histories. The hospital cancer registry has been kept since 1963, and systematically registers information on diagnosis, treatment and prognosis among all diagnosed cancer patients. Identification of breast cancer patients was based on the database of the registry. Primary diagnoses of non-malignant disease patients were obtained from the hospital information on primary diagnoses for inpatients.
Statistical Analysis
Logistic regression analysis (5) was used to adjust for confounding variables and estimate odds ratios (ORs) with their 95% confidence interval (CI) for breast cancer. Student's t and [chi]2 tests were employed on appropriate occasions. P values of <0.05 were judged as being statistically significant. The statistical package software Stata (6) was used for statistical analyses. Stratification by menopausal status was employed for analysis, since the risk factors were reported to be frequently different between pre- and post-menopausal women.
Categorization
Ages at first delivery and menopause, and body weight, height and body mass index (BMI) were classified into several categories in order to have an almost equal number of study subjects in each category. In the case of age at menarche, we chose a cut-off point of 13 years, because in many observations the onset of menarche occurred most frequently at 13 years old. Educational levels were divided into three categories: low, intermediate and high. Those who finished only junior high or elementary school were classified as having a low educational level. Graduates from senior high or professional school were classified as having an intermediate educational level. Subjects who finished college or university level education were defined as highly educated. Smoking habits were classified into three categories. Smokers were defined as persons having smoked tobacco at least one year before the diagnosis. Ex-smokers were persons having given up smoking one year before the diagnosis. Non-smokers were persons having never smoked. A similar definition was employed for drinking habits. Drinkers were defined as persons having drunk alcohol at least one year before the diagnosis and they were all occasional drinkers. Ex-drinkers were persons having given up alcohol one year before the diagnosis. Non-drinkers were persons having rarely or never drunk. With family history, a `Yes' was defined as mother or sisters suffering from breast cancer. Data on family history of breast cancer had some limitation in this study, since the self-administered questionnaire survey did not always differentiate between those who had no family history of breast cancer and subjects who did not know about the disease history of their mothers and sisters. Since cancer patients usually lose weight in the development and progress of the disease, weight one year prior to the diagnosis was used for analysis of the association between body weight and breast cancer risk.
RESULTS
The age distributions at diagnosis among cases were younger than those among controls (Table 1) (mean age at diagnosis among cases was 51.6 years and among controls 54.5 years, P = 0.0001). Odds ratios for breast cancer were therefore estimated after adjustment for age at diagnosis. Table 2 shows the results for reproductive history and its related factors. The proportion of women who had menarche at the age of [le]13 years was rather high in the case group among all women (54.8%) and among post-menopausal women (48.1%). Moderate associations were shown between earlier age at menarche and the risk for breast cancer, especially among post-menopausal women (age-adjusted OR for women who had menarche at [le]13 years: 2.27, 95% CI: 0.28-0.66, as compared with the risk among women who had menarche at age [ge]16 years). Older age at first delivery ([ge]28 years) was also shown to be a moderate risk factor for breast cancer, especially among pre-menopausal women (age-adjusted OR: 2.97, 95% CI: 1.31-6.75), as compared with the risk among women who had their first delivery at [le]22 years old. A linear trend was demonstrated between age at first delivery and the risk of breast cancer among pre-menopausal women ([chi]2 for linear trend = 7.11; P = 0.008). Moderately increased risk for breast cancer was observed among pre-menopausal women without delivery, although the OR was less than that among women who had their first delivery at the age of [ge]28 years, and did not reach a statistically significant level. Weak associations were observed between later age at menopause ([ge]53 years) and the risk for breast cancer, although statistically not significant. There were no associations with marital status and the risk of breast cancer (data not shown).
Table 3 shows the associations between anthropometric measurements and the risk for breast cancer. The case group was heavier than the control group (26.9% of the cases were [ge]58 kg compared to 24.0% of the controls). A significantly increased risk for breast cancer was demonstrated among post-menopausal women who weighed [ge]58 kg as compared with the risk among women who were [le]47 kg (age-adjusted OR: 1.83, 95% CI: 1.10-3.01). A linear trend was also observed between weight and the risk of breast cancer among post-menopausal women ([chi]2 = 5.85; P = 0.016). Positive associations were also observed between height and the risk for breast cancer among pre-menopausal women. A significantly elevated risk was observed among pre-menopausal women with a height of [ge]159 cm as compared with the risk at [le]149 cm (age-adjusted OR: 2.51, 95% CI: 1.17-5.39).
Interactions between height and earlier age at menarche ([le]13 years) were examined after adjustment for age at diagnosis and menopausal status. When compared with the risk among heights of [le]154 cm and age at menarche of [ge]14 years, the OR for breast cancer among height of [ge]155 cm and earlier age at menarche was 2.14 (P = 0.001), among height of [le]154 cm and earlier age at menarche it was 1.34 (P = 0.162), and among height of [ge]155 cm and age at menarche of [ge]14 years it was 0.72 (P = 0.118). This suggests the possibilities of qualitative interactions (7) between height and age at menarche, i.e. a height of [ge]155 cm may be a risk factor for breast cancer among women with earlier age at menarche, although it may be preventive among women with age at menarche of [ge]14 years.
Those who had BMI > 25 showed a moderately increased risk for breast cancer among post-menopausal women (age-adjusted OR: 1.90, 95% CI: 1.10-3.24) in comparison with those with a BMI of [le]20. One of our major interests, the associations between obesity and the risk for breast cancer, were confirmed among post-menopausal women in this study.
Table 4 shows the associations between lifestyle factors and the risk for breast cancer. In the estimation of ORs, smoking and drinking habits were adjusted to each other, in addition to age at diagnosis. No associations were found between smoking habits and the risk for breast cancer. Neither were any associations observed between drinking habits and the risk for breast cancer. The proportion of subjects who participated in physical activity in the cases was almost equal to that in the controls, and no associations were found between physical activity and the risk for breast cancer in this study. Physical activity was further analyzed according to body weight, and the `Yes' category was divided into physical activity twice a week or more and twice a week or less. However, no associations were found.
Table 4 also presents the results of educational levels and the risk for breast cancer. Generally, educational levels among the cases were higher than those among the controls, although there were no great differences. Findings of such higher educational levels among breast cancer cases were consistent with other studies, and in Table 4 the results of family history of breast cancer are presented. The prevalence in the case group was almost three times as high in the control group. The age-adjusted OR for positive family history was 2.68 (95% CI: 1.44-5.00) in all women.
The adjusted ORs by multiple logistic regression analysis including seven variables (age at menarche and at first delivery, body weight, height, smoking, drinking, educational level) are shown in Tables 2-4 according to menopausal status. The ORs are not very different from the result after adjustment for age only.
DISCUSSION
This study confirmed that an early age at menarche ([le]13 years) and a later age at first delivery ([ge]28 years older) were both moderate risk factors for breast cancer among Japanese, and also showed that height ([ge]159 cm) and BMI (>25) were risk markers for pre- and post-menopausal breast cancer risk, respectively, among Japanese women. These findings were generally very consistent with results from other studies mainly conducted in western society (8-10).
Table 1.
| Age | Pre-menopause | Post-menopause | All women | |||
| Cases n = 190 (%) | Controls n = 119 (%) | Cases n = 182 (%) | Controls n = 282 (%) | Cases n = 376 (%) | Controls n = 430 (%) | |
| 29 | 6 (3.2) | 17 (14.3) | 6 (1.6) | 19 (4.4) | ||
| 30-39 | 37 (19.5) | 27 (22.7) | 37 (9.8) | 29 (6.7) | ||
| 40-49 | 120 (63.2) | 62 (52.1) | 20 (11.0) | 14 (5.0) | 141 (37.5) | 81 (18.8) |
| 50-59 | 27 (14.2) | 13 (10.9) | 70 (38.5) | 119 (42.2) | 99 (26.3) | 141 (32.8) |
| 60-69 | 68 (37.4) | 108 (38.3) | 69 (18.4) | 116 (27.0) | ||
| 70-79 | 22 (12.1) | 41(14.5) | 22 (5.9) | 44 (10.2) | ||
| 80- | 2 (1.1) | 0 (0.0) | 2 (0.5) | 0 (0.0) | ||
Table 2.
| Pre-menopause | ||||||
| Case n = 190 (%) | Control n = 119 (%) | OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Age at menarche | ||||||
| 16- | 31 (16.4) | 22 (18.6) | 1.00 | 1.00 | ||
| 14-15 | 44 (23.3) | 29 (24.6) | 1.39 | 0.77-2.50 | 1.37 | 0.72-2.63 |
| 13 | 114 (60.3) | 67 (56.8) | 1.61 | 0.84-3.23 | 1.59 | 0.78-3.23 |
| P for trend | 0.116 | 0.925 | ||||
| Age at first delivery | ||||||
| 22 | 28 (14.8) | 24 (20.5) | 1.00 | 1.00 | ||
| 23-25 | 44 (23.3) | 33 (28.2) | 1.12 | 0.54-2.28 | 1.19 | 0.54-2.52 |
| 26-27 | 31 (16.4) | 15 (12.8) | 1.63 | 0.71-3.73 | 1.32 | 0.54-3.26 |
| 28- | 44 (23.3) | 14 (12.0) | 2.97 | 1.31-6.75 | 2.48 | 1.03-6.02 |
| P for trend | 0.008 | 0.008 | ||||
| No delivery | 42 (22.2) | 31 (24.5) | 2.13 | 0.97-4.71 | 2.12 | 0.87-5.11 |
| Post-menopause | ||||||
| Case n = 186 (%) | Control n = 282 (%) | OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Age at menarche | ||||||
| 16- | 69 (38.1) | 142 (51.3) | 1.00 | 1.00 | ||
| 14-15 | 25 (13.8) | 59 (21.3) | 2.70 | 1.56-4.76 | 2.56 | 1.37-4.76 |
| 13 | 87 (48.1) | 76 (27.4) | 2.27 | 1.51-3.57 | 1.89 | 1.16-3.03 |
| P for trend | 0.006 | 0.001 | ||||
| Age at first delivery | ||||||
| 22 | 40 (22.6) | 78 (27.9) | 1.00 | 1.00 | ||
| 23-25 | 35 (19.8) | 63 (22.5) | 0.96 | 0.55-1.66 | 1.00 | 0.54-1.88 |
| 26-27 | 36 (20.3) | 50 (17.9) | 1.23 | 0.69-2.17 | 1.17 | 0.61-2.22 |
| 28- | 35 (19.8) | 52 (18.6) | 1.19 | 0.67-2.07 | 1.15 | 0.61-2.18 |
| P for trend | 0.265 | 0.175 | ||||
| No delivery | 31 (17.5) | 37 (13.2) | 1.48 | 0.80-2.66 | 1.38 | 0.69-2.72 |
| Age at menopause | ||||||
| 46 | 56 (30.8) | 104 (27.4) | 1.00 | |||
| 47-50 | 55 (30.2) | 140 (36.9) | 0.92 | 0.58-1.48 | ||
| 51-52 | 38 (20.9) | 61 (16.1) | 1.23 | 0.71-2.14 | ||
| 53- | 31 (17.0) | 54 (14.2) | 1.20 | 0.66-2.16 | ||
| P for trend | 0.373 | |||||
| All women | ||||||
| Case n = 376 (%) | Control n = 430 (%) | OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Age at menarche | ||||||
| 16- | 100 (26.7) | 167 (41.1) | 1.00 | 1.00 | ||
| 14-15 | 69 (18.5) | 93 (22.9) | 1.85 | 1.27-2.63 | 1.75 | 1.18-2.70 |
| 13 | 205 (54.8) | 146 (36.0) | 2.08 | 1.45-2.86 | 1.85 | 1.82-2.78 |
| P for trend | 0.001 | 0.001 | ||||
| Age at first delivery | ||||||
| 22 | 69 (18.6) | 104 (25.4) | 1.00 | 1.00 | ||
| 23-25 | 81 (21.9) | 101 (24.7) | 1.25 | 0.83-1.89 | 1.16 | 0.73-1.83 |
| 26-27 | 67 (18.1) | 67 (16.4) | 1.61 | 1.03-2.52 | 1.30 | 0.78-2.14 |
| 28- | 79 (21.4) | 67 (16.4) | 1.89 | 1.22-2.93 | 1.65 | 1.01-2.69 |
| P for trend | 0.014 | 0.007 | ||||
| No delivery | 74 (20.0) | 70 (17.1) | 1.52 | 0.97-2.37 | 1.34 | 0.81-2.21 |
Table 3.
| Pre-menopause | ||||||
| Case n = 190 (%) | Control n = 119 (%) | OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Weight (kg) | ||||||
| 47 | 35 (18.5) | 30 (25.2) | 1.00 | 1.00 | ||
| 48-52 | 44 (23.3) | 31 (26.0) | 1.00 | 0.50-2.02 | 1.07 | 0.42-2.29 |
| 53-57 | 65 (34.4) | 29 (24.4) | 1.63 | 0.82-3.19 | 1.47 | 0.68-3.16 |
| 58- | 45 (23.8) | 29 (24.4) | 1.05 | 0.52-2.12 | 0.9 | 0.38-2.08 |
| P for trend | 0.537 | 0.592 | ||||
| Height (cm) | ||||||
| 149 | 25 (13.2) | 19 (16.0) | 1.00 | 1.00 | ||
| 150-154 | 43 (22.6) | 25 (21.0) | 1.62 | 0.72-3.61 | 1.45 | 0.59-3.46 |
| 155-158 | 48 (25.3) | 38 (31.9) | 1.27 | 0.59-2.70 | 1.12 | 0.48-2.59 |
| 159- | 74 (39.0) | 37 (31.2) | 2.51 | 1.17-5.39 | 2.29 | 0.93-5.65 |
| P for trend | 0.032 | 0.248 | ||||
| BMI (kg/m2) | ||||||
| 20.0 | 40 (21.2) | 30 (25.2) | 1.00 | |||
| 20.1-23.0 | 69 (36.5) | 35 (29.4) | 1.12 | 0.57-2.14 | ||
| 23.1-25.0 | 37 (19.6) | 32 (26.9) | 0.58 | 0.28-1.19 | ||
| 25.1- | 43 (22.7) | 22 (18.5) | 0.98 | 0.46-2.06 | ||
| P for trend | 0.486 | |||||
| Post-menopause | ||||||
| Case n = 186 (%) | Control n = 282 (%) | OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Weight (kg) | ||||||
| 47 | 46 (25.3) | 105 (37.2) | 1.00 | 1.00 | ||
| 48-52 | 37 (20.3) | 53 (18.8) | 1.56 | 0.90-2.70 | 1.43 | 0.78-2.61 |
| 53-57 | 44 (24.2) | 57 (20.2) | 1.73 | 1.03-2.93 | 2.05 | 1.12-3.77 |
| 58- | 55 (30.2) | 67 (23.8) | 1.83 | 1.10-3.01 | 2.16 | 1.19-3.91 |
| P for trend | 0.016 | 0.033 | ||||
| Height (cm) | ||||||
| 149 | 73 (40.1) | 110 (39.0) | 1.00 | 1.00 | ||
| 150-154 | 50 (27.5) | 71 (25.2) | 0.99 | 0.62-1.61 | 0.93 | 0.54-1.61 |
| 155-158 | 35 (19.2) | 72 (25.5) | 0.69 | 0.41-1.15 | 0.62 | 0.34-1.11 |
| 159- | 24 (13.2) | 29 (10.3) | 1.14 | 0.60-2.15 | 0.91 | 0.84-0.98 |
| P for trend | 0.628 | 0.696 | ||||
| BMI (kg/m2) | ||||||
| 20.0 | 30 (16.5) | 65 (23.1) | 1.00 | |||
| 20.1-23.0 | 42 (23.1) | 64 (22.7) | 1.43 | 0.80-2.56 | ||
| 23.1-25.0 | 42 (23.1) | 75 (26.6) | 1.22 | 0.68-2.16 | ||
| 25.1- | 68 (37.4) | 78 (27.7) | 1.90 | 1.10-3.24 | ||
| P for trend | 0.037 | |||||
| All women | ||||||
| Case n = 376 (%) | Control n = 430 (%) | OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Weight (kg) | ||||||
| 47 | 81 (21.6) | 144 (33.5) | 1.00 | 1.00 | ||
| 48-52 | 82 (21.9) | 88 (20.5) | 1.55 | 1.03-2.34 | 1.45 | 0.93-2.25 |
| 53-57 | 111 (29.6) | 95 (22.1) | 1.95 | 1.32-2.89 | 1.93 | 1.24-3.01 |
| 58- | 101 (26.9) | 103 (24.0) | 1.68 | 1.13-2.46 | 1.75 | 1.11-2.75 |
| P for trend | 0.004 | 0.017 | ||||
| Height (cm) | ||||||
| 149 | 99 (26.3) | 135 (31.4) | 1.00 | 1.00 | ||
| 150-154 | 95 (25.3) | 104 (24.2) | 1.11 | 0.74-1.63 | 1.03 | 0.66-1.59 |
| 155-158 | 83 (22.1) | 118 (27.4) | 0.81 | 0.54-1.21 | 0.74 | 0.47-1.18 |
| 159- | 99 (26.3) | 73 (17.0) | 1.42 | 0.23-2.20 | 1.23 | 0.74-2.05 |
| P for trend | 0.392 | 0.692 | ||||
| BMI (kg/m2) | ||||||
| 20.0 | 70 (18.7) | 105 (24.4) | 1.00 | |||
| 20.1-23.0 | 111 (29.6) | 108 (25.1) | 1.58 | 1.05-2.37 | ||
| 23.1-25.0 | 82 (21.9) | 110 (25.6) | 1.21 | 0.79-1.84 | ||
| 25.1- | 112 (29.8) | 107 (24.88) | 1.73 | 1.15-2.62 | ||
| P for trend | 0.039 | |||||
Table 4.
| Pre-menopause | ||||||
| Case n = 190 (%) | Control n = 119 (%) | OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Smoking | ||||||
| Non-smoker | 152 (80.0) | 85 (75.2) | 1.00 | 1.00 | ||
| Ex-smoker | 17 (9.0) | 9 (8.0) | 0.89 | 0.37-2.16 | 0.82 | 0.32-2.09 |
| Smoker | 21 (11.0) | 19 (16.8) | 0.59 | 0.29-1.20 | 0.71 | 0.32-1.58 |
| Drinking | ||||||
| Non-drinker | 100 (52.6) | 68 (57.6) | 1.00 | 1.00 | ||
| Ex-drinker | 5 (2.6) | 4 (3.4) | 1.09 | 0.25-4.85 | 1.09 | 0.22-5.36 |
| Drinker | 85 (44.7) | 46 (39.0) | 0.68 | 0.40-1.13 | 0.73 | 0.41-1.25 |
| Activity | ||||||
| No | 125 (68.3) | 86 (72.3) | 1.00 | |||
| Yes | 58 (31.7) | 33 (27.7) | 1.19 | 0.72-1.99 | ||
| Educational level | ||||||
| Low | 13 (6.9) | 14 (12.1) | 1.00 | 1.00 | ||
| Intermediate | 103 (54.5) | 52 (44.8) | 2.69 | 1.16-6.25 | 2.24 | 0.88-5.76 |
| High | 73 (38.6) | 50 (43.1) | 2.43 | 1.01-5.88 | 1.58 | 0.58-4.31 |
| Family history | ||||||
| Yes | 22 (11.6%) | 6 (5.0%) | 2.21 | 0.86-5.72 | ||
| Post-menopause | ||||||
| Case n = 186 (%) |
Control n = 282 (%) |
OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Smoking | ||||||
| Non-smoker | 154 (84.6) | 200 (80.3) | 1.00 | 1.00 | ||
| Ex-smoker | 10 (5.5) | 21 (8.4) | 0.80 | 0.35-1.83 | 0.94 | 0.39-2.27 |
| Smoker | 18 (9.9) | 28 (11.2) | 0.83 | 0.44-1.59 | 0.97 | 0.47-1.98 |
| Drinking | ||||||
| Non-drinker | 130 (71.4) | 165 (67.4) | 1.00 | 1.00 | ||
| Ex-drinker | 6 (3.3) | 17 (6.9) | 0.48 | 0.17-1.33 | 0.43 | 0.15-1.26 |
| Drinker | 44 (24.2) | 63 (25.7) | 1.00 | 0.63-1.61 | 1.14 | 0.68-1.88 |
| Activity | ||||||
| No | 125 (72.2) | 186 (71.5) | 1.00 | |||
| Yes | 48 (27.8) | 74 (28.5) | 0.96 | 0.62-1.47 | ||
| Educational level | ||||||
| Low | 37 (21.4) | 71 (25.7) | 1.00 | 1.00 | ||
| Intermediate | 94 (54.3) | 157 (56.9) | 1.12 | 0.69-1.79 | 1.02 | 0.59-1.75 |
| High | 42 (24.3) | 48 (17.4) | 1.67 | 0.94-2.94 | 1.58 | 0.82-3.04 |
| Family history | ||||||
| Yes | 13 (7.0%) | 8 (2.8%) | 2.56 | 1.04-6.34 | ||
| All women | ||||||
| Case n = 376 (%) |
Control n = 430 (%) |
OR*1 | 95% CI*1 | OR*2 | 95% CI*2 | |
| Smoking | ||||||
| Non-smoker | 308 (81.9) | 302 (78.4) | 1.00 | 1.00 | ||
| Ex-smoker | 28 (7.5) | 34 (8.8) | 0.87 | 0.50-1.50 | 0.98 | 0.54-1.78 |
| Smoker | 40 (10.6) | 49 (12.8) | 0.73 | 0.47-1.15 | 0.90 | 0.55-1.49 |
| Drinking | ||||||
| Non-drinker | 233 (62.0) | 249 (64.7) | 1.00 | 1.00 | ||
| Ex-drinker | 11 (2.9) | 22 (5.7) | 0.47 | 0.21-1.02 | 0.42 | 0.19-0.95 |
| Drinker | 130 (34.6) | 114 (29.6) | 0.80 | 0.58-1.12 | 0.86 | 0.61-1.22 |
| Activity | ||||||
| No | 252 (70.0) | 294 (72.2) | 1.00 | |||
| Yes | 108 (30.0) | 113 (27.8) | 1.11 | 0.81-1.51 | ||
| Educational level | ||||||
| Low | 51 (13.9) | 93 (22.2) | 1.00 | 1.00 | ||
| Intermediate | 199 (54.4) | 223 (53.4) | 1.46 | 0.99-2.16 | 1.31 | 0.84-2.04 |
| High | 116 (31.7) | 102 (24.4) | 1.71 | 1.09-2.69 | 1.43 | 0.85-2.36 |
| Family history | ||||||
| Yes | 35 (9.3%) | 15 (3.5%) | 2.68 | 1.44-5.00 | ||
In Tables 5 and 6, our major study results are compared with others (11-15) from Japanese studies on risk factors for breast cancer, although there were some differences in the categories and methods. Some studies (11, 12) presented their results by menopausal status, while others (13-15) did not. Results on the age at menarche, age at first delivery and age at menopause were similar to each other (Table 5), although earlier age at menarche ([le]13 years) showed a stronger association in this study than in the others. Findings on body weight and BMI were consistent among all the studies (Table 6), while height was only associated with the risk for breast cancer among pre-menopausal women in this study, but only among post-menopausal women in the other two studies (11, 12).
Table 5.
| Reference and local | Number of cases/controls Type of controls | Age at menarche | Age at first delivery | Age at menopause | |||||||||
| Category | OR | Category | OR | Category | OR | ||||||||
| Pre- | Post- | Pre- | Post- | ||||||||||
| Current study Osaka | 376/430 Hospital controls | 16- | 1.00 | 1.00 | -22 | 1.00 | 1.00 | -46 | 1.00 | ||||
| 14-15 | 1.39 | 2.70* | 23-25 | 1.12 | 0.96 | 47-50 | 0.92 | ||||||
| -13 | 1.61 | 2.27* | 26-27 | 1.63 | 1.23 | 51-52 | 1.23 | ||||||
| 28- | 2.97* | 1.19 | 53- | 1.20 | |||||||||
| No birth | 2.13* | 1.48 | |||||||||||
| Hirose et al. (1995) (11) Nagoya | 1052/21 299 Hospital controls | Pre- | Post- | Pre- | Post- | ||||||||
| 16- | 1.00 | 1.00 | -23 | 1.00 | 1.00 | ||||||||
| 14-15 | 1.22* | 1.15 | 24-26 | 1.18 | 1.09 | ||||||||
| -13 | 1.22 | 1.19 | 27- | 1.63* | 1.34* | ||||||||
| Category | OR | Category | OR | Category | OR | ||||||||
| Wakai et al. (1994) (12) Tokyo | 300/900 Hospital controls | Pre- | Post- | Pre- | Post- | ||||||||
| 15- | 1.00 | 1.00 | -24 | 1.00 | 1.00 | -44 | 1.00 | ||||||
| 14 | 0.82 | 1.11 | 25-29 | 0.93 | 1.25 | 45-49 | 0.65 | ||||||
| 13 | 1.26 | 1.35 | 30-34 | 0.94 | 2.85* | 50-54 | 0.63 | ||||||
| -12 | 1.08 | 1.18 | 35- | 1.08 | 3.54* | 55- | 0.77 | ||||||
| Category | OR | Category | OR | Category | OR | ||||||||
| Yuasa et al. (1970) (13) | 861/2268 Hospital controls | 17- | 1.00 | -19 | 1.00 | -39 | 1.00 | ||||||
| 16 | 1.10 | 20-24 | 1.6 | 40-44 | 1.00 | ||||||||
| 15 | 1.30 | For all women | 25-29 | 2.1 | For all women | 45-49 | 1.00 | ||||||
| 14 | 1.30 | 30-34 | 2.4 | 50-54 | 1.40 | ||||||||
| 13 | 1.40 | 35- | 3.6 | 55- | 1.20 | ||||||||
| -13 | 2.20 | ||||||||||||
| Category | OR | Category | OR | ||||||||||
| Hirohata et al. (1985) (14) Fukuoka | 212/424 Neighborhood controls (n = 212) Hospital controls(n = 212) | 18- | 1.00 | -19 | 1.00 | ||||||||
| 16-17 | 1.37 | For all women | 20-24 | 1.56 | For all women | ||||||||
| 14-15 | 1.45 | 25-29 | 2.37 | ||||||||||
| 13- | 2.00 | 30-34 | 2.07 | ||||||||||
| 35- | 5.04 | ||||||||||||
| No birth | 5.55* | ||||||||||||
| OR | OR | ||||||||||||
| Category | Age at diagnosis | Category | Age at diagnosis | ||||||||||
| Hamajima et al. (1995) (15) Nagoya | 1052/21 299 Hospital controls | 30-39 | 40-49 | 50-59 | 60-69 | 30-39 | 40-49 | 50-59 | 60-69 | ||||
| 16- | 1.00 | 1.00 | 1.00 | 1.00 | -23 | 1.00 | 1.00 | 1.00 | 1.00 | ||||
| 14-15 | 0.83 | 1.21 | 1.43* | 1.10 | 24-26 | 1.18 | 1.18 | 1.00 | 1.16 | ||||
| -13 | 0.69 | 0.90 | 1.45 | 1.18 | 27- | 1.67 | 1.52* | 1.36 | 1.57 | ||||
| No birth | 1.84 | 1.72* | 1.74* | 1.37 | |||||||||
Table 6.
| Reference and local | Number of cases/controls Type of controls | Weight | Height | BMI | ||||||||
| Category | OR | Category | OR | Category | OR | |||||||
| Pre- | Post- | Pre- | Post- | Pre- | Post- | |||||||
| Current study Osaka | 376/430 Hospital controls | -47 | 1.00 | 1.00 | -149 | 1.00 | 1.00 | -20 | 1.00 | 1.00 | ||
| 48-52 | 1.00 | 1.56 | 150-154 | 1.62 | 0.99 | 20.1-23.0 | 1.12 | 1.43 | ||||
| 53-57 | 1.63 | 1.73* | 155-158 | 1.27 | 1.69 | 23.1-25.0 | 0.58 | 1.22 | ||||
| 58- | 1.05 | 1.83* | 159- | 2.51* | 1.14 | 25.1- | 0.98 | 1.90* | ||||
| Hirose et al. (1995) (11) Nagoya | 1052/21299 Hospital controls | -48 | 1.00 | 1.00 | -150 | 1.00 | 1.00 | -20.0 | 1.00 | 1.00 | ||
| 49-56 | 0.91 | 1.17 | 151-158 | 0.84 | 1.01 | 20.1-26.4 | 1.05 | 1.77* | ||||
| 57- | 0.81 | 2.05* | 159- | 0.89 | 1.54* | 26.5- | 1.44 | 3.60* | ||||
| Wakai et al. (1994) (12) Tokyo | 300/900 Hospital controls | -49 | 1.00 | 1.00 | -149 | 1.00 | 1.00 | |||||
| 50-59 | 1.77* | 0.77 | 150-154 | 0.77 | 1.81 | |||||||
| 60-69 | 1.59 | 1.09 | 155-159 | 0.71 | 1.51 | |||||||
| 70- | 2.76 | 4.82* | 160- | 1.08 | 1.46 | |||||||
| Category | Age at diagnosis | |||||||||||
| Hamajima et al. (1995) (15) Nagoya | 1052/21299 Hospital controls | 30-39 | 40-49 | 50-59 | 60-69 | |||||||
| -19 | 1.00 | 1.00 | 1.00 | 1.00 | ||||||||
| 20-24 | 1.16 | 0.92 | 0.85 | 1.05 | ||||||||
| 25- | 1.62 | 0.68* | 1.12 | 2.17* | ||||||||
One of the limitations of this study is the use of hospital controls, which might be inadequate to evaluate the roles of smoking and drinking habits in the etiology of breast cancer, since these patients' diseases might be related to smoking and drinking habits, too. We believe that if healthy controls had been used, the relationship between alcohol drinking and breast cancer risk might have been detected. Although there were no differences in smoking and drinking habits between cases and controls in this study, we need to reserve our conclusions about those habits and the risk for breast cancer. Another limitation is that among a total of 808 new breast cancer cases during this period, only 376 patients received the questionnaire and enrolled in this study. This was because, at the beginning, a feasibility test of the questionnaire survey had been conducted in some wards of OMCC. In addition, the questionnaire was sometimes not delivered to inpatients because of unspecified clerical reasons. Although we included all new breast cancer cases who responded to the questionnaire in our study, it was impossible to count how many patients were eligible for this study during 1990-95. Mean age at diagnosis was 51.6 years for breast cancer cases enrolled in this study and 52.3 years for the remaining breast cancer patients who did not participate. Although there is no evidence, it is unlikely that this situation distorted our results seriously.
Qualitative interactions between height and age at menarche were suggested in this study (in Results). Height was only associated with risk of breast cancer in women who have early menarche. Generally, both age at menarche and height are dependent on ovarian function, the onset of which depends, in turn, on nutritional status (16). However, height is expected to be determined by multiple factors other than nutritional status, such as heredity, physical activity and social background. In case of women with later menarche ([ge]14 years) and height of [ge]155 cm, the height, in part, might be explained by physical activity or other factors. These factors had preventive effects on breast cancer. This might be an explanation that height show the different effects on the risk of breast cancer among women with later menarche.
Currently in Japan, remarkable changes have occurred in the reproductive situation. Women are getting married at a later age (average age at marriage: 23 years in the 1950s and 27.5 years in 1995) and having fewer children (average number of children: 3.7 in the 1950s and 1.4 in 1995) (17). The improvement in nutritional status has resulted in taller heights in childhood and earlier age at menarche (mean age at menarche: 15.0 years in 1950-1952 and 12.5 years in 1980) (18). Furthermore, due to the high caloric intake and less physical activities, overweight and obesity have become prevalent in middle-aged women in Japan. Since breast cancer risk is well determined by total cumulative exposure to estrogen of breast tissue and androgen (chiefly androstenedione through peripheral conversion in adipose tissue) is the main source of estrogen in post-menopausal women (19), it might result in an increased risk for breast cancer among post-menopausal women in Japan. These factors seem to have contributed to the increase in the incidences of breast cancer in Japan.
In this study, we do not have sufficient data for a study about dietary factors. Diets high in saturated fat possibly increase the risk of breast cancer, while diets high in unsaturated or vegetable fats, or cholesterol, probably have no relationship with the risk of breast cancer (20). Vegetables and fruits are reported to have protective effects for breast cancer risk. The evidence is more abundant and consistent for vegetables, particularly green vegetables, than for fruits. In Japan, analysis of dietary factors will be needed to clarify in more detail about the amount of food stuff or caloric intake.
Among the risk factors, obesity seems to be very important in regard to cancer prevention, since it is easier to modify than the others. Population-attributable risk percent for a BMI of [ge]20 was estimated as 28.8% of post-menopausal breast cancer, based on the OR and distribution of BMI among post-menopausal control women. Although we did not find any associations between physical activity and the breast cancer risk among Japanese women, we still believe that maintaining an ideal weight can help in the prevention of breast cancer, especially among middle-aged women. Further studies will be necessary to evaluate the risk factors for breast cancer among Japanese women, especially the roles of smoking and drinking, which have become popular among Japanese young women in recent years, and to elucidate protective factors against breast cancer, including weight control and physical activity.
Acknowledgments
We are grateful to Mr John Hoyle for correction of our English. This study was supported in part by grant for New 10-year Strategy for Cancer Control, Prevention of Cancer, from the Ministry of Health and Welfare. H.T.T. was supported by a Uehara Memorial Foundation Fellowship, Japan.
References
This article has been cited by other articles:
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 19 Mar 1999
Copyright© 1999 Foundation for Promotion of Cancer Research.
CiteULike 
Connotea 
Del.icio.us What's this?
C. Nagata, T. Mizoue, K. Tanaka, I. Tsuji, K. Wakai, M. Inoue, S. Tsugane, and for the Research Group for the Development and Eva
Alcohol Drinking and Breast Cancer Risk: An Evaluation Based on a Systematic Review of Epidemiologic Evidence among the Japanese Population
Jpn. J. Clin. Oncol.,
August 18, 2007;
(2007)
hym062v1.
[Abstract]
[Full Text]
[PDF]
K. B. Michels, K. L. Terry, and W. C. Willett
Longitudinal Study on the Role of Body Size in Premenopausal Breast Cancer
Arch Intern Med,
November 27, 2006;
166(21):
2395 - 2402.
[Abstract]
[Full Text]
[PDF]
C. Nagata, T. Mizoue, K. Tanaka, I. Tsuji, K. Wakai, M. Inoue, S. Tsugane, and Research Group for the Development and Evaluation
Tobacco Smoking and Breast Cancer Risk: An Evaluation Based on a Systematic Review of Epidemiological Evidence among the Japanese Population
Jpn. J. Clin. Oncol.,
June 1, 2006;
36(6):
387 - 394.
[Abstract]
[Full Text]
[PDF]
Y. Miyoshi, T. Funahashi, S. Kihara, T. Taguchi, Y. Tamaki, Y. Matsuzawa, and S. Noguchi
Association of Serum Adiponectin Levels with Breast Cancer Risk
Clin. Cancer Res.,
November 15, 2003;
9(15):
5699 - 5704.
[Abstract]
[Full Text]
[PDF]
Y. Miyoshi, Y. Tanji, T. Taguchi, Y. Tamaki, and S. Noguchi
Association of Serum Estrone Levels with Estrogen Receptor-positive Breast Cancer Risk in Postmenopausal Japanese Women
Clin. Cancer Res.,
June 1, 2003;
9(6):
2229 - 2233.
[Abstract]
[Full Text]
[PDF]
This Article 
Abstract
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (14)
Request Permissions
Google Scholar
Articles by Tung, H.
Articles by Koyama, H
Search for Related Content
PubMed
PubMed Citation
Articles by Tung, H.
Articles by Koyama, H
Social Bookmarking
What's this?