See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/228011612 Intake of selected micronutrients and the risk of breast cancer Article in International Journal of Cancer · January 1996 DOI: 10.1002/(SICI)1097-0215(19960117)65:2<140::AID-IJC3>3.0.CO;2-Z CITATIONS READS 97 26 12 authors, including: Renato Talamini Monica Ferraroni 576 PUBLICATIONS 21,195 CITATIONS 126 PUBLICATIONS 5,398 CITATIONS CRO Centro di Riferimento Oncologico di A… SEE PROFILE University of Milan SEE PROFILE Maurizio Montella Adriano Decarli 298 PUBLICATIONS 7,590 CITATIONS 368 PUBLICATIONS 15,222 CITATIONS Istituto Nazionale Tumori "Fondazione Pas… SEE PROFILE University of Milan SEE PROFILE Some of the authors of this publication are also working on these related projects: Mediterranean diet and composition of breast feeding milk View project All content following this page was uploaded by Adriano Decarli on 27 October 2014. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately. zyxwv zyxwvutsrqpo zyxwvutsrqp zyxwvuts zyxwvutsrq Int. J. Cancer: 65,140-144 (1996) 0 1996 Wiley-Liss, Inc. Publicationof the InternationalUnion Against Cancer Publicationde I’Union InternationaleContre le Cancer INTAKE OF SELECTED MICRONUTRIENTS AND THE RISK OF BREAST CANCER Eva NEGRI’I~, Carlo LA vECCHIA1%2,%]via FRANCESCH13, Barbara D’AVANZO~, Renato TALAMIN13, Maria PARPINEL3, Ettore cONT17 and Adriano DECARLI~ Monica FERRARONI~, Rosa FILIBERT14, Maurizio MONTELLA’, Fabio FALCINI~, ’Istitutod i Ricerche Farmacologiche “MarioNegn’”, 20157 Milan; 21stitutodi Statistica Medica e Biometria, Universita di Milano, Istituto Nazionale per lo Studio e la Cura dei Tumori, Milan; 3Sewiziodi Epidemiologia, Centro di Riferimento Oncologico,Aviano; 41stitutoNazionale per la Ricerca sul Cancro, Genoa; sIstitutoper lo Studio e la Cura dei Tumori “SenatorePascale”, Naples; %tituto OncologicoRomagnolo, Ospedale Pierantoni, Forli; and 71stitutoRegina Elena per lo Studio e la Cura dei Tumori, Rome, Italy. To investigate the relation between selected micronutrients and breast cancer risk, we conducted a case-control study of breast cancer between June 1991 and April 1994 in 6 Italian areas. The study included 2569 women admitted to the major teaching and general hospitals of the study areas with histologically confirmed incident breast cancer and 2588 control women with no history of cancer, who were admitted to hospitals in the same catchment areas for acute, non-neoplastic, nongynecological conditions unrelated to hormonal or digestive tract diseases or to long-term modifications of the diet. Dietary habits, including alcoholic beverage consumption, were investigated using a validated food frequency questionnaire, including 78 foods or food groups, several types of alcoholic beverages, some “fat intake pattern” questions and some open sections for foods consumed frequently by the subject and not reported in the questionnaire. To control for potential confounding factors, several multiple logistic regression models were used. When major correlates, energy intake and the mutual confounding effect of the various micronutrients were taken into account, beta-carotene, vitamin E and calcium showed a significant inverse association with breast cancer risk. The estimated odds ratios of the 5th quintile compared to the lowest one were 0.84 for beta-carotene, 0.75 for vitamin E and 0.81 for calcium. No significant association emerged for retinol, vitamin C, thiamin, riboflavin, iron and potassium. Our results suggest that a diet rich in several micronutrients, particularly beta-carotene, vitamin E and calcium, may be protectiveagainst breast cancer. o 1996 Wiley-Liss,Inc. Cases were 2569 women admitted to the major teaching and general hospitals of the study areas, who had histologically confirmed breast cancer first diagnosed within 1 year before interview and no history of cancer at the time of diagnosis. The age range was 23-74, and the median age was 55 years. Controls were 2588 women, with no history of cancer, admitted to hospitals in the same catchment areas for acute, non-neoplastic, nongynecological conditions unrelated to hormonal or digestive tract diseases, or to long-term modifications of the diet. Twenty-two percent of the controls were admitted for traumas, mostly fractures and sprains, 33% for other orthopedic diseases, 15% for surgical conditions, 18% for eye diseases and 12% for miscellaneous other diseases, such as ear, nose and throat, skin and dental conditions. The age range of the controls was 20-74 years, and their median age was 56 years. The distributions of cases and controls in terms of age and area of residence were similar, although cases and controls were not individually matched. Less than 4% of cases and controls refused to participate, The interviewers were centrally trained, and the same questionnaire and coding manual were used for all subjects. Each interview was centrally checked for reliability and consistency. The questionnaire included questions on sociodemographic characteristics, lifelong smoking habits, physical activity and anthropometric indicators at various ages, a problem-oriented medical history, family history of cancer, menstrual and reproductive history and history of use of sexual hormones for contraception, hormone replacement therapy or other indications. Dietary habits, including alcoholic beverage consumption, were investigated using a validated food frequency questionnaire. zyxwvut The breast is the leading cancer site in women, in terms of incidence and mortality in the world, and breast cancer mortality is increasing in many areas, including several developed countries. The identification of dietary correlates of breast cancer is therefore an important public health issue, because dietary habits are, at least in principle, modifiable. There is epidemiological evidence that several aspects of diet, including some micronutrients, may influence the risk of breast cancer (Rohan and Bain, 1987; Hunter and Willett, 1993). Some studies have suggested a protective effect of vitamins A, E and C, but published results are not totally consistent (Hunter and Willett, 1993). To further investigate the issue, we analyzed the relation between selected micronutrients and breast cancer risk using data from a case-control study conducted in various Italian regions, based on a validated food frequency questionnaire. Thefood Ji-equencyquestionnaire (FFQ) The FFQ was used to assess the usual diet in the 2 years before diagnosis, in order to estimate the mean daily intake of calories and of a number of macro- and micronutrients. Subjects had to report their average weekly consumption of 78 foods or food groups, including some of the most common recipes of the Italian diet (eg., pizza), and of several types of alcoholic beverages. Consumption of less than once per month was coded as 0, whereas 1-3 times per month was coded as 0.5. For a few types of vegetables and fruits subject to great seasonal variation, the consumption in season and the corresponding duration were elicited. At the end of each section, 1 or 2 open questions were used to report foods not included in the questionnaire but eaten at least once a week. For 40 food items, the portion was defined in “natural” units (e.g., 1 zyxw SUBJECTS AND METHODS We conducted a multicentric case-control study on breast cancer between June 1991 and April 1994 in 6 Italian areas: the provinces of Pordenone and Gorizia, the greater Milan area, the urban area of Genoa, the province of Forli, the province of Latina and the urban area of Naples. The general design of this investigation has already been described (La Vecchia et al., 1995). 8To whom correspondence and reprint requests should be sent, at Laboratory of Epidemiology, Istituto di Ricerche Farmacologiche “Mario Negri”, Via Eritrea 62, 20157 Milan, Italy. Fax: 39-023320-023 1. Received: June 12,1995 and in revised form September 21, 1995 z zy zyxwvuts zyxwvutsr zyxwvutsrq zyxwvuts MICRONUTRIENTS AND BREAST CANCER teaspoon of sugar, 1egg), while for the others it was defined as small, average or large with the help of pictures. Information on the type and quantity of fats used for seasoning raw and cooked vegetables, pasta or rice, meat and fried dishes was used in the calculation of nutrients. Dietary supplements were not considered, given their low levels of consumption by this population. To compute energy and nutrient intake, Italian food composition databases were used for about 80% of food items. These sources, however, had to be integrated and checked with other sources and with information from manufacturers, especially as regards carotene, retinol and vitamin E (Salvini et al., 1995). The results of the reproducibility and validity study of the present FFQ were satisfactory (Decarli et al., 1995; Franceschi et al., 1995). Statistical analysis To control for potential confounding factors, several multiple logistic regression models were used (Breslow and Day, 1980). All the regression equations included terms for age in quinquennia, area of residence, years of education ( < 717-11/ 12+), parity (0/1/2/3/4+), quintile of energy intake (alcohol excluded) and alcohol intake in 4 levels. Further inclusion in the models of measures of fat intake (not materially related in this study to breast cancer risk; La Vecchia et al., 1995) and of other breast cancer correlates, like body mass, age at menarche, menopausal status, age at menopause, age at first birth, history of benign breast disease and family history of breast cancer, did not appreciably modify any of the estimates. Quintiles were computed (i) directly on the nutrients and (ii) on the residuals of the regression of the nutrient on energy, following the method suggested by Willett (1990). Both analyses yielded similar results, and the latter were chosen for presentation. Tests for trends for the quintiles of nutrients were based on the likelihood ratio test between the models with and without a linear term for the nutrient's quintile. The nutrients were also introduced into the model as continuous variables. The unit measure for every nutrient was set as the difference between the upper cutpoint of the 4th quintile and that of the first. For example, the upper cutpoint of the 4th quintile for total vitamin A was 2361 kg/day, and that of the first was 706 pg/day (see Table 11), so the model with the continuous coefficient gives an estimate of the odds ratio relative to an increase of intake of total vitamin A of 2361 706 = 1655 pg/day, which is an indicator of the difference in risk between subjects with high intake and those with low intake in this population. To investigate deviation from linearity, the likelihood ratio test between the models with and without a further quadratic term was computed. The distribution of many nutrients was skewed to the right, so the continuous coefficient must be interpreted with caution. However, application of transformations to eliminate the skewness would impede the biological interpretation of the coefficients. RESULTS Table I gives the distribution of cases and controls according to age, education and parity. The age distribution of cases and controls was comparable, whereas cases were more educated than controls and tended to have lower parity. The distribution of cases and controls according to quintiles of intake of the micronutrients considered, and the upper quintile cutpoints are shown in Table 11. For most micronutrients, there were more cases in the 1st quintile than controls, while the opposite was true in the 5th one. Table I11 gives odds ratios of breast cancer according to subsequent quintiles of various micronutrients, as compared to the 1st (lowest) one. The risk tended to decrease with increasing quintile for all micronutrients, and the trend was 141 TABLE I -DISTRIBUTION' OF 2569 CASES OF BREAST CANCER AND 2588 CONTROLS BY AGE GROUP, YEARS OF EDUCATION AND PARITY, ITALY, 1991-1994 Characteristic Age group (years) < 35 35-44 45-54 55-64 2 65 Education (years)' <7 7-1 1 2 12 Paritv (number of births)' Ndliparae ' 1 2 3 24 Controls Cases Number % 87 383 772 799 528 3.4 14.9 30.1 31.1 20.6 140 332 694 802 620 5.4 12.8 26.8 31.0 24.0 1259 714 582 49.3 28.0 22.8 1569 642 354 61.2 25.0 13.8 401 587 968 406 207 15.6 22.8 37.7 15.8 8.1 380 496 909 489 314 14.7 19.2 35.1 18.9 12.1 Number % ~~ 'The sum may not add up to the total because of missing values. TABLE I1 - DISTRIBUTIONS OF 2569 CASES OF BREAST CANCER AND 2588 CONTROLS ACCORDING TO QUINTILES OF SELECTED MICRONUTRIENTS, ITALY, 1991-1994 1 (low) 2 3 4 Total vitamin A (pg/day) Upper limits 706 949 1303 2361 Cases 548 541 505 499 Controls 483 490 528 531 Retinol (pg/day) Upper limits 135 199 298 1598 Cases 530 533 517 505 Controls 502 497 515 526 Beta-carotene (pg/day) Upper limits 2778 3666 4566 5852 Cases 527 546 526 497 Controls 504 485 507 533 Vitamin E (mg/day) Upper limits 7.21 9.12 10.82 13.43 Cases 542 529 513 527 Controls 490 502 518 504 Vitamin C (mg/day) Upper limits 81 109 138 182 Cases 522 503 514 534 Controls 510 528 518 497 Thiamin (mg/day) Upper limits 0.73 0.87 1.00 1.16 Cases 528 507 507 518 Controls 504 523 526 513 Riboflavin (mg/day) Upper limits 1.23 1.51 1.76 2.11 Cases 528 534 522 498 Controls 503 498 510 533 Iron (mg/day) Upper limits 10.49 12.45 14.25 16.52 Cases 517 511 519 501 Controls 515 519 513 530 Calcium (mg/day) Upper limits 671 871 1050 1297 Cases 553 538 472 498 Controls 479 493 560 532 Potassium (mg/day) Upper limits 2614 3124 3569 4138 Cases 529 515 517 506 Controls 503 515 516 524 5 (high) 476 556 484 548 473 559 458 574 496 535 509 522 487 544 521 511 508 524 502 530 zyxwvu significant at the 5% level for total vitamin A, beta-carotene, vitamin E, riboflavin, iron, calcium and potassium. Likewise, odds ratios estimated using the continuous terms were below unity for all nutrients. They were 0.88 for total vitamin A, 0.94 for retinol, 0.82 for beta-carotene, 0.84 for vitamin E, 0.92 for zyxwvutsrq zyxwvuts zyxwvutsrqpon zyxw zyxwvutsrqponm zyxwvutsrqpon N E G R I ETAL 142 TABLE 111 - ODDS RATIOS' OF BREAST CANCER AND 95% CONFIDENCE INTERVALS ACCORDING TO QUINTILE OF INTAKE O F SELECTED MICRONUTRIENTS, ITALY, 1991-1994 X2 OR^ 7 1 4 5 trend continuous 1.00 (0.8-1.2) 1.00 (0.8-1.2) 1.06 10.9-1.3), ' 1.00 (0.8-1.2) 0.93 (0.8-1.1) 0.93 (0.8-1.1) 1.02 10.9-1.2) ' 0.94 ' (0.8-1.1) 0.93 (0.8-1.1) 0.97 (0.8-1.2) 0.86 (0.7-1.0) 1.08 (0.9-1.3) 0.94 (0.8-1.1) \ 0.89 (0.7-1.1) 0.96 (0.8-1.2) 0.91 (0.8-1.1) 0.96 (0.8-1.1'1 ' 0.94 ' (0.8-1.1) 0.73 (0.6-0.9) 0.93 (0.8-1.1) 0.83 (0.7-1.0) 1.08 (0.9-1.3) 0.83 (0.7-1.0) 0.93 (0.8-1.1) 0.99 (0.8-1.2) 0.94 (0.8-1.1) 0.86 (0.7-1.0) 0.80 (0.7-1.0) 0.77 (0.6-0.9) 0.86 (0.7-1.0) 0.73 (0.6-0.9) 0.86 (0.7-1.0) 0.74 (0.6-0.9) 0.69 (0.6-0.8) 0.81 (0.7-1.0) 0.88 (0.7-1.0) 0.81 10.7-1.0) 0.85 (0.7-1.0) 0.80 (0.7-1.0) 0.80 (0.7-1.0) 15.805 0.88 (0.81-0.95) 0.94 (0.86-1.02) 0.82 (0.76-0.89) 0.84 (0.78-0.91) 0.92 (0.85-0.99) 0.93 (0.85-1.01) 0.88 (0.81-0.95) 0.90 (0.80-0.98) 0.91 (0.84-0.99) 0.88 (0.80-0.95) Quintile of intake2,3 Nutrient Total vitamin A Retinol Beta-carotene Vitamin E Vitamin C Thiamin Riboflavin Iron Calcium Potassium 1.93 16.515 15.265 2.84 1.70 8.625 4.116 8.895 7.325 'Estimates from multiple logistic regression models including terms for a e, center, education, parity, energy and alcohol intake.-2Quintiles of calorie adjusted nutrients.- Quintile 1 (lowest) is the reference ~ategory-~Odds ratio for a difference in intake equal to the difference between the upper cutpoint of the 4th quintile and that of the lst.-5p < 0.01.-6p < 0.05. 8 zyxwvutsrqp zyx TABLE N - ODDS RATIOS' OF BREAST CANCER (AND 95% CONFIDENCE INTERVALS) ACCORDING TO QUINTILES OF INTAKE OF SELECTED MICRONUTRIENTS, ITALY, 1991-1994 XZ 0r4 2 ? 4 5 trend continuous 1.11 (0.9-1.3) 1.02 (0.9-1.2) 1.13 (0.9-1.4) 0.91 (0.8-1.1) 1.00 (0.8-1.2) 1.00 (0.8-1.2) 0.94 (0.8-1.1) 1.10 (0.9-1.3) 0.72 (0.6-0.9) 1.01 (0.8-1.2) 0.90 (0.7-1.1) 1.00 (0.8-1.2) 1.03 (0.8-1.3) 0.78 (0.6-1.0) 1.oo (0.8-1.2) 0.84 (0.7-1.0) 0.75 (0.6-0.9) 0.97 (0.8-1.2) 0.81 (0.7-1 .O) 1.04 (0.8-1.3) 4.875 0.86 (0.78-0.94) 0.86 (0.79-0.94) 0.94 (0.84-1.06) 0.94 (0.85-1.03) 1.00 (0.89-1.12) Quintile of Nutrient Beta-carotene Vitamin E Riboflavin Calcium Potassium 7.196 0.46 4.8OS 0.08 'Estimates from a multiple logistic regression models including terms for age, center, education, parity, and intake of energy, alcohol, and the above listed nutrients.-2Quintiles of calorie adjusted n~trients.-~Quintile1 (lowest) is the reference ~ategory-~Oddsratio for a difference in intake equal to the difference between the upper cutpoint of the 4th quintile and that of the lst.-5p < 0.05. -9< 0.01. vitamin C, 0.93 for thiamin, 0.88 for riboflavin, 0.90 for iron, 0.91 for calcium and 0.88 for potassium. Inclusion of a quadratic term to account for deviations from linearity did not significantly improve the fitting of the models for most micronutrients (p > 0.10), the only exceptions being total vitamin A (XI*= 4 . 0 5 , ~= 0.04) and thiamin (xI2 = 3 . 8 6 , ~= 0.05). Table IV gives the odds ratios of breast cancer according to intake quintile and for the continuous model for the micronutrients that presented a significant trend in risk (at the 1% level) in the overall analysis (Table 111), adjusted for the mutual confounding effect. The estimated odds ratios of the 5th quintile compared to the lowest one were 0.84 for betacarotene, 0.75 for vitamin E, 0.97 for riboflavin, 0.81 for calcium and 1.04 for potassium, the trend in risk being significant at the 5% level for beta-carotene, vitamin E and calcium. DISCUSSION This study suggests that breast cancer risk is inversely related to the intake of some specific micronutrients, particu- zyxw larly beta-carotene, vitamin E and calcium, o r to some correlate of their intake. A n important function of vitamin A is to regulate cell differentiation, and it has been suggested that a diet deficient in retinoids increases the occurrence of breast cancer and of other epithelial tumors in experimental animals (Moon et al., 1983). Furthermore, vitamin A inhibits the growth of human breast carcinoma cells in vitro (Fraker et al., 1984). Betacarotene is an important source of vitamin A, since it is partly converted into retinol by the human body. Results from epidemiological studies on the relation of dietary intake of vitamin A and breast cancer risk are, however, controversial. Ten case-control studies reported odds ratios below unity for subjects with a high intake of betacarotene or carotenoids with provitamin A activity compared to subjects with a low intake, although four did not (Hunter and Willett, 1993). The evidence from case-control studies with respect to retinol (preformed vitamin A) intake was less convincing, since out of 9 studies, 5 reported a relative risk below 1 and 4 greater than or equal to 1 (Hunter and Willett, zyxwvutsr zyxwvut zyxwvuts 143 MICRONUTRIENTS AND BREAST CANCER 1993). A combined analysis of 12 case-control studies found a significant relative risk of 0.85 for the highest quintile of beta-carotene intake as compared to the lowest one (Howe et al., 1990), and 1.04 for retinol. Four prospective studies found a moderate protection associated with both retinol and carotenoids with vitamin A activity or total vitamin A, although the association was generally weak, with relative risks around 0.8-0.9 for the highest category of consumption as compared to the lowest one (Hunter and Willett, 1993). In our study, total vitamin A and beta-carotene were inversely related to breast cancer, but the association with retinol was less convincing. A validation study of the food frequency questionnaire in our study suggested that it was able to correctly identify subjects with a very high intake of retinol (higher quintile), but a considerable misclassification emerged at lower levels of intake (Decarli et al., 1995). In fact, the odds ratio for subjects in the Sth quintile, compared to those in the 1st one, was below 1, and it is conceivable that the lack of dose-response relationship may be due to a substantial degree of misclassification in lower quintiles. Problems of validity in the estimation of retinol intake may also have affected other studies. We have therefore used beta-carotene rather than total vitamin A in subsequent analyses. Furthermore, retinol is stored in the liver, and in the absence of substantial dietary deficiencies, serum retinol levels are only weakly correlated to dietary intake (Peto et al., 1981). It is thus conceivable that dietary retinol does not materially affect breast cancer risk. For beta-carotene, dietary intake is directly correlated with serum levels (Peto et al., 1981). The effect of beta-carotene may be related not so much to its provitamin A activity as to its antioxidant properties. Beta-carotene, in fact, can quench electronically excited molecules, like singlet oxygen, and participates in some free radical reactions. Vitamin E also has antioxidant properties. The epidemiological evidence on the role of dietary vitamin E in breast cancer carcinogenesis is relatively recent and also not totally consistent. Three case-control studies reported a lower risk in the highest consumption category compared to the lowest one, but 2 others did not. Three cohort studies showed risk estimates between 0.9 and 1.0 (Hunter and Willett, 1993). A third powerful antioxidant for which anticancer properties have been suggested is vitamin C. Epidemiological studies of the relation between vitamin C and breast cancer, however, once again yielded inconclusive results. Two case-control studies found a relative risk below 1for high vs. low intake, and a review of 9 case-control studies yielded a statistically significant odds ratio of 0.69 for the highest quintile of intake compared with the lowest one (Howe et al., 1990). Other studies, however, found little evidence of protection (Hunter and Willett, 1993). In our study, the estimated relative risk for the highest quintile was 0.82 relative to the lowest one, but there was no evidence of a trend in risk. Consequently, our results do not support a specific effect of vitamin C on breast carcinogenesis. Calcium was associated with a moderate, but significant, decreased risk of breast cancer in this study. In 3 other case-control studies, women with high calcium consumption had a lower risk than women with low intake, the estimates ranging between 0.2 and 0.6 (Katsouyanni et al., 1988; Zaridze et al., 1991; Landa et al., 1994). These studies, however, were based on small numbers of cases, and their results were not significant. In a case-control study from western New York based on 439 post-menopausal cases, there was no difference in the mean daily calcium intake of cases and controls (Graham et al., 1991). Thus, a protective effect of calcium against breast cancer is possible, though far from established. Calcium has been suspected to be protective against colon cancer (Sorenson et aL, 1988), which may share some etiological factors, dietary ones in particular, with breast cancer (Willett, 1989).A possible mechanism of action is linked to the ability of calcium to bind fats and bile acids to form insoluble calcium soaps, thus reducing the exposure of intestinal mucosa to lipids. This would imply a local action in the lumen, but conceivably also some effect on intestinal lipid absorption. Some studies, however, have suggested a protective effect of vitamin D on colon cancer, and a protection has been postulated also for breast cancer (Garland et aL, 1990). Vitamin D promotes the transport of calcium from the lumen into the blood, so this would imply a systemic action (Sorenson et aL, 1988). This large data set allowed reasonably precise estimates, so the results are not appreciably affected by random variation. The food frequency questionnaire had satisfactory reproducibility (ie., r = 0.62 for beta-carotene, r = 0.48 for vitamin E and r = 0.65 for calcium; Franceschi et al., 1995) and validity (corresponding r = 0.49, 0.34 and 0.56; Decarli et aL, 1995). The use of hospital controls has been widely debated (Breslow and Day, 1980), and several strengths and weaknesses have been highlighted. Dietary habits of hospital controls may differ from those of the general population, but we took great care to exclude from the control group all diagnoses that might have involved any long-term modification of diet. On the other hand, the similar interview setting for cases and controls and the almost complete participation are reassuring. It has been suggested that cases may recall diet differently from any type of controls, though a study conducted in Canada did not find any appreciable difference between micronutrient intake estimated prospectively or retrospectively (Friedenreich et al., 1993). For all micronutrients considered, the estimated risk of breast cancer for the highest quintile of intake compared to the lowest one was below unity. This may reflect the high correlations between various micronutrients, which severely hamper the identification of any real protective factor(s). It is still possible that high intake of several micronutrients simply reflects a healthier diet, rich in fruit and vegetables, which has been shown to be protective against cancer at several sites (Block et al., 1992). The chief sources of both beta-carotene and vitamin E, estimated from our study questionnaire, were various raw vegetables, including vegetable oils used for seasoning. Such food items showed the strongest inverse associations with breast cancer in our study, whereas fruit, the chief source of vitamin C, was less clearly protective. Calcium was chiefly derived from dairy products, cheese firstly. Bearing these limitations in mind, our results suggest that a diet rich in several micronutrients, particularly beta-carotene, vitamin E and calcium, can contribute to decreasing the risk of breast cancer. zyxwvu zyxwvut ACKNOWLEDGEMENTS This work was conducted within the framework of the CNR (Italian National Research Council) Applied Projects “Clinical Applications of Oncological Research” (contracts 94.01268.PF39, 94.01119.PF39 and 94.01321.PF39) and “Risk Factors for Disease” (contract 95.00952.PF41) and with the contribution of the Italian Association for Research on Cancer and the Italian League against Tumors. The authors thank Mrs. P. Gnagnarella and Mrs. 0. Volpato for assistance with the nutritional database, the G. A. Pfeiffer Memorial Library staff and Mrs. J. Baggott and Ms. M. P. Bonifacino for editorial assistance. REFERENCES BLOCK,G., PATTERSON, B. and SUBAR, A,, Fruit, vegetables and cancer prevention: a review of the epidemiological evidence. Nutr. Cancer, 18, 1-29 (1992). BRESLOW, N.E. and DAY,N.E., Statistical methods in cancer research. 144 zyxwvutsrqpon zyxwvutsrq zyxwvutsr zyxwvutsrqp zyxwvutsrqpon zyxwv NEGRI ETAL Vol. 1. The analysis of case-control studies. IARC Scientific Publication KATSOUYANNI, K., WILLETT,W., TRICHOPOULOS, D., BOYLE,P., 32, International Agency for Research on Cancer, Lyon, France TRICHOPOULOU, A., VASILAROS, S., PAPADIAMANTIS, J. and MACMA(1980). HON,B., Risk of breast cancer among Greek women in relation to DECARLI,A., FRANCESCHI, S., FERRARONI, M., GNAGNARELLA, P., nutrient intake. Cancer, 61,181-185 (1988). PARPINEL, M.T., LA VECCHIA,C., NEGRI,E., SALVINI, S., FALCINI, F. LANDA,M.-C., FRAGO,N. and TRES,A., Diet and the risk of breast and GIACOSA,A., Validation of a food frequency questionnaire to cancer in Spain. Europ. J. Cancer Prev., 3,313-320 (1994). assess dietary intakes in cancer studies in Italy: results for specific LAVECCHIA, c.,NEGRI,E., FRANCESCHI, S., DECARLI, A., GIACOSA, A. nutrients. Ann. Epidemiol. (1995) (In press). L., Olive oil, other dietary fats and the risk of breast FRAKER,L.D., HALTER,S.A. and FORBES, J.T., Growth inhibition by and LIPWORTH, retinol of a human breast carcinoma cell line in vitro and in athymic cancer. Cancer Causes Control, 6,545-550 (1995). mice. Cancer Rex, 44,5757-5763 (1984). MOON, R.C., MCCORMICK, D.L. and MEHTA,R.G., Inhibition of FRANCESCHI, S., BARBONE, F., NEGRI,E., DECARLI,A,, FERRARONI, carcinogenesis by retinoids. Cancer Res., 43,2469s-24758 (1983). M., FILIBERTI,R., GIACOSA,A,, GNAGNARELLA, P., NANNI,O., J.D. and SPORN,M.B., Can dietary SALVINI, S. and LA VECCHIA,C., Reproducibility of an Italian food PETO, R., DOLL, R., BUCKLEY, frequency questionnaire for cancer studies. Results for specific nutri- beta-carotene materially reduce human cancer rates? Nature (Lond.), 290,201-208 (1981). ents. Ann. Epidemiol., 569-75 (1995). FRIEDENREICH. C.M.. HOWE.G.R. and MILLER.A.B.. Recall bias in the association of micronu<rient intake and breast cancer. J. clin. Epidemiol., 46,1009-1017 (1993). GARLAND. F.C.. GARLAND. C.F.. GORHAM.E.D. and YOUNG.J.F.. Geographic variation in breast cancer mortality in the United States: a hypothesis involving exposure to solar radiation. Prev. Med., 19, 614-622 (1990). GRAHAM, S., HELLMANN, R., MARSHALL, J., FREUDENHEIM, J., VENA, J., SWANSON, M., ZIELEZNY, M., NEMOTO,T., STUBBE,N. and RAIMONDO, T., Nutritional epidemiology of postmenopausal breast cancer in western New York. Amer. J. Epidemiol., 134,552-566 (1991). HOWE,G.R., HIROHATA, T., HISLOP, T.G., ISCOVICH,I.M., YUAN, J.-M., KATSOUYANNI, K., LUBIN,F., MARUBINI,E., MODAN,B., ROHAN, T., TONIOLO, P. and SHUNZHANG, Y., Dietary factors and risk of breast cancer: combined analysis of 12 case-control studies. J. nut. CancerInst., 82,561-569 (1990). HUNTER,D.J. and WILLETT,W.C., Diet, body size, and breast cancer. Epidemiol. Rev., 15,110-131 (1993). View publication stats ROHAN,T.E. and BAIN,C.J., Diet in the etiology of breast cancer. Epidemwl. Rev., 9,120-145 (1987). SALVINI, S., GNAGNARELLA, P., PARPINEL, M.T., BOYLE,P., DECARLI, A., FERRARONI, M., GIACOSA,A., LA VECCHIA,C., NEGRI,E. and FRANCESCHI, S., The food composition database for an Italian food frequency questionnaire. J. Food Compos. Anal. (1995). (In press), SORENSON, A.W., SLATTERY, M.L. and FORD, M.H., Calcium and colon cancer: a review. Nutr. Cancer, 11,135-145 (1988). WILLETT, W., The search for the causes of breast and colon cancer. Nature (Lond.),338,389-394 (1989). WILLETT,W., Nutritional epidemiology, Oxford University Press, New York (1990). ZARIDZE,D., LIFANOVA, Y., MAXIMOVITCH, D., DAY,N.E. and DUFFY, S.W., Diet, alcohol consumption and reproductive factors in a casecontrol study of breast cancer in Moscow. Int. J. Cancer, 48, 493-501 (1991).