Similar and Dissimilar Findings in Japan-Indonesia Case-Control Study on Breast Cancer : Two Phases Study

Three folds Japan-Indonesia joint study on breast cancer has been conducted, i,e, epidemiological, clinical and histopathological aspects. Case-control study was used throughout as standard method. Hisopathologically confirmed breast cancer cases were women aged 25-69 years with primary tumor. The study was performed in two batches, i.e. during the first batch 300 cases and 600 controls, and during the second batch 226 cases and 552 controls were collected. The epidemiological study revealed that in premenopausal breast cancer increased risk was related to breast trauma, contraceptive use, milk and fresh fruits intakes ; while decreased risk was related to cooked vegetable intake. In postmenopausal breast cancer increased risk was found in age at menarche of 15 years or over, regular menstruation after 30 years, and milk consumption ; while decreased risk was related to divorced or widowed status, and number of live births or children breast-fed. The clinicopathological findings showed that the Indonesian breast cancer had already occured more frequently at younger age group (10-29 years) and peaked at 40-49 years group. The majority was at advanced stage, the stage I and II was 2% and 15.3% respectively. In contrast, the majority of Japanese cases were at early stage. The majority (88.33%) of Indonesian breast cancers was the invasive carcinoma, while 1.33% was the noninvasive carcinoma and 9.67% was the special type. Medullary carcinoma was found more frequently among the Indonesian females as compared to the Japanese females. Further nutritional analysis showed that certain micronutrients increased the risk such as fat, animal protein, mineral, calcium, phosphate, ferrum. Retinol, thiamin and ascorbic acid of the postmarriage intakes might contribute to the risk, while such indication was not shown by the premarriage intakes. The importance of the interaction of certain micronutrients instead of interpreting the effect of individual nutrient separately has to be considered. J Epidemiol,1996 ; 6 : S175-S180.


Similar and Dissimilar
Findings in Japan-Indonesia Case-Control Study on Breast Cancer : Two Phases Study Santoso Cornain1, Yoshiyuki Ohno2, Joedo Prihartono1, Goi Sakamoto3, Gunawan Tjahjadi1, Didid Tjindarbumi1, Kenji Wakai2, Muchlis Ramli1, Esti Sutrisno1, Setyawati Budiningsih1, Sadao Suzuki2, Idral Darwis1, Susumu Watanabe3, Endang S. Roostini1, Drupadi S. Dillon1, and Susilowati Herman1 Three folds Japan-Indonesia joint study on breast cancer has been conducted, i,e, epidemiological, clinical and histopathological aspects.Case-control study was used throughout as standard method.Hisopathologically confirmed breast cancer cases were women aged 25-69 years with primary tumor.The study was performed in two batches, i.e. during the first batch 300 cases and 600 controls, and during the second batch 226 cases and 552 controls were collected.The epidemiological study revealed that in premenopausal breast cancer increased risk was related to breast trauma, contraceptive use, milk and fresh fruits intakes ; while decreased risk was related to cooked vegetable intake.In postmenopausal breast cancer increased risk was found in age at menarche of 15 years or over, regular menstruation after 30 years, and milk consumption ; while decreased risk was related to divorced or widowed status, and number of live births or children breast-fed.The clinicopathological findings showed that the Indonesian breast cancer had already occured more frequently at younger age group (10-29 years) and peaked at 40-49 years group.The majority was at advanced stage, the stage I and II was 2% and 15.3% respectively.In contrast, the majority of Japanese cases were at early stage.The majority (88.33%) of Indonesian breast cancers was the invasive carcinoma, while 1.33% was the noninvasive carcinoma and 9.67% was the special type.Medullary carcinoma was found more frequently among the Indonesian females as compared to the Japanese females.Further nutritional analysis showed that certain micronutrients increased the risk such as fat, animal protein, mineral, calcium, phosphate, ferrum.Retinol, thiamin and ascorbic acid of the postmarriage intakes might contribute to the risk, while such indication was not shown by the premarriage intakes.The importance of the interaction of certain micronutrients instead of interpreting the effect of individual nutrient separately has to be considered.J Epidemiol,1996 ; ' National Cancer Institute , Tokyo, Japan.

Address for correspondence
: Santoso Cornain, Faculty of Medicine, University of Indonesia, Salemba 6 Jakarta, 10430 Indonesia.Semarang 7).Several aspects of breast cancer, including the epidemiological, clinical and pathological studies have become the subjects of studies in order to elucidate the factors which might play a role in such a difference.Both similarity and dissimilarity found will be valuable in understanding the biological behaviour of breast cancer and further application to the disease control.The present paper will discuss the results of a Japan-Indonesia Joint Study on Etiology and Clinicopathology of Breast Cancer.
A number of epidemiologic studies have been performed in breast cancer 8-13), which revealed a variety of risk factors including genetic / hormonal factors 10-12), reproductive status 1a 12,13) , life style, dietary and environmental factors 3.8,11)_

MATERIALS AND METHODS
A Japan-Indonesia collaborative study on breast cancer has been performed by using a case-control study in three folds, i.e. epidemiological, clinical and histopathological aspects.Cases were women with primary breast cancer, histopathologically confirmed and aged ranging from 25 to 69 years.Two controls were selected from females without lump, matched by age (circa 3 years) and socioeconomical status, time of visit / admission to the Cipto Mangunkusumo National Central General Hospital at Jakarta.The study was performed in two batches, namely during 3 years of the first batch 300 cases and 600 controls were studied, which was extended with the second batch for further nutritional analysis on 226 cases and 452 controls.Epidemiological data were collected by using well designed questionnaire, which composed of identification and diagnosis categorization, demographic characteristics, obstetrical history, contraception history, smoking / drinking history, cancer history in family hospitalization history and dietary history.Two nurses were trained for at least two weeks for the details of questionnaire and the manual.Clinicopathological studies used standard clinical procedures, chest X-ray, skeletal X-ray, liver function test and clinical stag- ing by TNM classification 14).The operable cases were operated on using radical-simple-mastectomy where approprite, plus radiation / chemotherapy according to standard procedures.Histological typing was performed according to the Japanese Breast Classification 15).The extension of the study was aimed for further nutritional analysis using the guidelines published elsewhere 16).Food Frequency Amount questionnaires were designed using 98 food items in Indonesian diet.During preliminary phase, 4 different periods were probed, namely before marriage versus after marriage, and five years versus ten years prior to the disease.Interviewing were done by three well trained nutritionists, using food model.The micronutrients were calculated using South East Asian Food Coversion Table .The nutritional data were analysed using t-test and quartile cat-egorization117) for comparing the difference between the cases and controls, while paired t-test was performed in comparing nutritional intake before marriage and after marriage.One case was tested against two matched controls.Three hundreds triplets were determined statistically (a =0.05; B=0.20, exposure rate of risk factor=40 %).Stepwise statistical analyses were initiated with univariate cross tabulation and chisquare test.The Relative Risk (R.R) was estimated by computing the Odds Ratio (age adjusted or adjusted for confounders by regression analysis) 8.18).The significance of the R.R. was determined based on 95 % Confidence Interval (95 % C.I.).

Epidemiological Study
The epidemiological study on all breast cancer cases reveled that the the most prominent risk factors were induced menopause, short period (less than 4 months) breast feeding (R.R. of 5.96 and 5.44 respectively).Less strong risk factors were shown by underweight, among close genetic traits, fatty food consumption (R.R. of 2.2, 2.85 and 2.63 respectively), followed by urban life, underweight, breast trauma and infrequent (1-2) pregnancy (R.R of 2.22, 2.2, 1.88, 1.51 respectively).The summary completed with 95 % C.I. are given in Table 1.
Some factors showed decreased risk of breast cancer, name- ly education level (senior high school), separated or widowed status, overweight, irregular menstrual cycle, indirect X-ray exposure (R.R. of 0.41, 0.37, 0.68, 0.32, 0.73 0.08 respectively).See also Table 2 for the corresponding 95 % C.I..In order to elucidate further about the categorization of the risk factor, they were evaluated according to their menopausal status.Thus the factors were analysed separately for premenopausal group and postmenopausal group respectively.Crude R.R. were compared to adjusted (for age, socioeconomic status, residence and other variables listed by unconditional logistic regression analysis).The results of the premenopausal status are given in Table 3. Breast trauma and pill (contraceptive) use were confirmed to increase the risk of breast cancer.The effect of milk consumption was significantly shown as risk factor, which was also significant for trend test (P < .05).Similarly for fresh fruits consumption; which should be prudently interpreted.Cooked vegetables seemed to show protective effect if it was consumed daily.
The results of the risk factor analysis in post-menopausal status are given in Table 4. Marital status, i.e. separated or widowed caused decreased risk of breast cancer.Weak effect of height in crude R.R. was not confirmed in adjusted R.R., in spite of significant test for trend (P < .05).The same for waist-hip ratio.Age 15 or above at menarche showed significant risk for breast cancer.Regular menstruation also caused increased risk.Higher number of children breast fed (3 or higher) seemed to decrease the risk, and was highly significant in test for trend.Smoking habit did not show significant risk effect.Milk intake showed a strong evidence as risk factor with highly significant test for trend.

Clinicopathological Study
Age distribution of breast cancer cases indicated that the disease had occurred more frequently at younger age group (10-29 years) and peaked at 40-49 years group.The majority were at advanced stage, the stage I and II was 2 % and 15.3 % respectively.In contrast, the majority of Japanese cases were at early stage.The tumor site was mostly found at the upper outer quadrant.
Histopathological study revealed that the majority (88.33 %) of Indonesian breast cancers was the invasive carcinoma, while 1.33 % was the non-invasive carcinoma and 9.67 % was the special type.Medullary carcinoma known to have better prognosis, was found more frequently among the Indonesian females as compared to the Japanese females.

Further Nutritional Analysis
During the second batch of study further nutritional analysis has calculated the micronutrients and they were evaluated by comparing cases and controls for determining the risk for breast cancer.In addition, the data was also analysed between two periods, namely before married versus after marriage.The data of either before marriage or after marriage was analysed separately according to 'quartile analysis'.The results of the significant risk factors are given in Tables 5 dan 6 Paired t-test on breast cancer cases, comparing premarriage intake versus postmarrage intake, revealed that the majority of micronutrients tested, i.e. calory, protein, fat, carbohydrate, mineral, calcium, ferrum, retinol thiamin, ascorbic acid, animal Table 5. Pre-marriage nutritional intake and breast cancer risk (Quartile Analysis).
protein, vegetable protein and carotene were significantly different.The premarriage intakes were lower that postmarriage ones except for phosphate, animal fat and vegetable fat without significant diference.

DISCUSSION
The present study showed that a given risk factor with high R.R.might be not representing its real mode before being analyzed in the appropriate group.For example, the previously known as the most prominent risk factor, namely induced menopause, in the general assessment turned to be unconfirmed during subgroup analysis.In addition, the risk factor seemed to be better defined in regard with the adjusted R.R. and trend test.The evidence that fatty food was considerably adequate has been confirmed by further micronutrient analysis.In particular both animal fat and vegetable fat was significant in quartile analysis as well.The data also confirm earlier reports 9-13.19) concerning various risk factors both reproductive or non-reproductive related or dietary non-dietary related.In respect to be aware of recall bias, we consider that our study has been away from such trouble since the preliminary trial has tested that the recall was matched to the weighing method and the probing on two different period, pre-and post marriage period, has revealed comparable data.The clinicopatholgoical study has clarified the contrast between the Indonesian cases being the majority in advanced stages and the Japanese cases being the majority in early stage.It might also raise the question whether any difference in its biological behaviour, which has become the interest of comparing the American and the Japanese 20).

Table 1 .
Factors showing increased risk of breast cancer.

Table 2 .
Factors showing decreased risk of breast cancer.

Table 3 .
Risk factor analysis of breast cancer in premenopausal group.

Table 4 .
Risk factor analysis of breast cancer in post-menopausal group.

Table 6 .
Post-marriage nutritional intake and breast cancer risk (Quartile Analysis).