点击显示 收起
1 From the Department of Clinical Sciences, Lund University, Malmö, Sweden (UE, ES, BG, and EW), and the Department of Clinical Sciences, Lund University, Lund, Sweden (HO)
2 Supported by grants from the Swedish Cancer Society, the Swedish Medical Research Council, the City of Malmö, The Albert Påhlsson Foundation, and the Swedish Research Council Formas.
3 Reprints not available. Address correspondence to U Ericson, Nutritional Epidemiology, Clinical Research Center, Building 60, floor 13, Malmö University Hospital, entrance 72, SE-205 02 Malmö, Sweden. E-mail: ulrika.pettersson{at}med.lu.se
See corresponding editorial on page 271.
ABSTRACT
Background: Epidemiologic studies of associations between folate intake and breast cancer are inconclusive, but folate and other plant food nutrients appear protective in women at elevated risk.
Objective: The objective was to examine the association between folate intake and the incidence of postmenopausal breast cancer.
Design: This prospective study included all women aged 50 y (n = 11699) from the Malmö Diet and Cancer cohort. The mean follow-up time was 9.5 y. We used a modified diet-history method to collect nutrient intake data. At the end of follow-up, 392 incident invasive breast cancer cases were verified. We used proportional hazard regression to calculate hazard ratios (HRs).
Results: Compared with the lowest quintile, the incidence of invasive breast cancer was reduced in the highest quintile of dietary folate intake (HR: 0.56; 95% CI: 0.35, 0.90; P for trend = 0.02); total folate intake, including supplements (HR: 0.56; 95% CI: 0.34, 0.91; P for trend = 0.006); and dietary folate equivalents (HR: 0.59; 95% CI: 0.36, 0.97; P for trend = 0.01).
Conclusion: A high folate intake was associated with a lower incidence of postmenopausal breast cancer in this cohort.
Key Words: Folate • breast cancer • obesity • body mass index • postmenopausal women • prospective study • diet
INTRODUCTION
Folate can be found in high concentrations in dark-green leafy vegetables, legumes, fruit, and liver (1). Folate is a coenzyme that carries one-carbon units and is thereby of great importance in the metabolism of amino acids and nucleotides (2). Two main mechanisms link folate deficiency to cancer development: a reduced synthesis of S-adenosyl methionine (SAM), which results in aberrations in DNA methylation, and a reduced synthesis of the pyrimidine thymidylate, which results in the misincorporation of uracil into DNA. A third possible mechanism is impaired purine synthesis and subsequent changes in DNA (3). Epidemiologic studies have indicated that high folate intake may protect against colorectal cancer (4, 5) but also against cancer at other sites (6-10). Current epidemiologic evidence of a relation between high folate intake and reduced breast cancer risk is, however, not conclusive. High folate intake was shown to be associated with a decreased risk of postmenopausal breast cancer in some studies (11-14), whereas this association was not confirmed in other studies (15-18).
Other B vitamins are involved in folate metabolism. Vitamin B-12 acts as a cofactor to the enzyme methionine synthase, and both vitamin B-6 and riboflavin serve as cofactors for folate-dependent enzymes (19, 20). A study in Shanghai showed that the association between a high folate intake and a decreased risk of breast cancer was stronger among women with high intakes of vitamin B-12 and vitamin B-6 (13). The modifying effect of vitamin B-12 is supported by other studies (11, 14).
Some studies indicate that the negative association between folate intake and breast cancer mainly appears among women at highest risk, ie, high consumers of alcohol. The detrimental influence of alcohol on folate bioavailability is probably the most important explanation for this observation (16-18). Obesity, however, is a well-known risk factor for breast cancer development, as is a high alcohol consumption (21, 22). Similar to many other studies, a previous report from the Malmö Diet and Cancer (MDC) study showed positive associations between BMI and breast cancer (23). Higher concentrations of endogenous sex hormones have been suggested to partly explain the increased risk of breast cancer among high consumers of alcohol (24). However, an elevated concentration of sex hormones is also a possible mechanism by which obesity can increase breast cancer risk (25, 26). In the Swedish mammography cohort, overweight women with high dietary intakes of ascorbic acid had a lower incidence of breast cancer (27), but the risk relation was reversed in lean women. Ascorbic acid, an antioxidant that may protect against DNA damage (28), is mainly found in fruit and vegetables, as is folate.
The aim of this study was to investigate whether folate intake is associated with postmenopausal breast cancer in women from the MDC cohort. This article also evaluates whether other nutrients influence the association between folate intake and breast cancer incidence. Finally, we wanted to examine whether the potential association between folate and breast cancer development was different among women who already had a higher breast cancer risk because they were overweight.
SUBJECTS AND METHODS
Study design
The MDC study is a prospective cohort study being conducted in Malmö, a city in the south of Sweden with 250 000 inhabitants. In 1991, the MDC source population was defined as all persons living in the City of Malmö and born between 1926 and 1945. However, in May 1995, the cohort was extended to include all women born between 1923 and 1950 and all men born between 1923 and 1945. With this extension, 74 138 persons constituted the source population. The MDC study was approved by the Ethical Committee at Lund University (LU 51-90). Details of the recruitment procedures and the cohort are described elsewhere (29). Briefly, the participants were invited by personal letters or came spontaneously after invitation by advertisement in local newspapers, in public places, or in primary health care centers. Inadequate Swedish language skills and mental incapacity were the only exclusion criteria. The participants visited the MDC screening center twice. During the first visit, groups of 6-8 participants were instructed on how to register meals in a menu book and how to fill out the diet questionnaire and the extensive questionnaire covering socioeconomic and lifestyle factors. Nurses drew blood samples, recorded blood pressures, and made anthropometric measurements. All questionnaires were completed at home. During the second visit, 10 d after the first visit, the socioeconomic questionnaire was checked and a dietary interview was conducted. In October 1996, when recruitment was closed, 28 098 participants had completed all baseline examinations.
Study population
This study includes women who completed the baseline examinations and were aged 50 y. An age criterion for menopause was used because of missing values on self-reported cessation of menses, imprecise cessation of menses in combination with information on menopausal hormone use among premenopausal women, and lack of detailed information on all women's medical history of hysterectomy. Fifty years was chosen as the definition for menopause (30), because the median natural age of menopause was 50.0 y in a subsample of 2898 women (without surgery and hormone therapy) from the cohort (31). Because the age at diagnosis was >55 y in 95% of the cases, we have reason to believe that very few women were premenopausal at diagnosis. All women with prevalent cancers at baseline, except those with cervix cancer in situ, were excluded. In total 11 699 women were included. The average follow-up time was 9.5 y.
Breast cancer case definition and ascertainment
The Swedish Cancer Registry and the Southern Swedish Regional Tumor Registry provided data on case definition and ascertainment until the end of follow-up (31 December 2003). Cases were women who received a diagnosis of invasive breast cancer during follow-up. Invasive cancer was defined as all cancer except in situ cancer. Because in situ cancer does not essentially progress into invasive cancer (32), inclusion of in situ cancer may obscure true associations between diet and serious disease. However, a sensitivity analysis including in situ cancers was performed. Information on vital status was obtained from the National Tax Board, which provides up-to-date information on vital status for all Swedish residents. Cases contributed person time from the date of enrollment until the time of diagnosis. Noncases contributed person time from the date of enrollment until death, migrating from Sweden or end of follow-up, whichever was the first. A total of 392 invasive breast cancers were documented during 110 925 y of follow-up.
Dietary data
The MDC study used an interview-based, modified diet-history method that combined 1) a 7-d menu-book for registration of lunch and dinner meals, cold beverages including alcohol, drugs, natural remedies, and nutrient supplements; 2) a 168-item questionnaire for assessment of meal pattern, consumption frequencies, and portion sizes of regularly eaten foods; and 3) a 45-min complementary interview. The consistency of the information provided was carefully checked so that the questionnaire and menu book did not overlap.
The mean daily intake of foods was calculated on the basis of frequency and portion-size estimates from the questionnaire and menu book. Food intake was converted to energy and nutrient intakes by using the MDC nutrient database, in which most of the nutrient information comes from PC-KOST2-93 from the National Food Administration in Uppsala, Sweden. The MDC method is described in detail elsewhere (33, 34). Nutrient intakes from supplements were calculated on the basis of the supplement consumption recorded in the menu book. Supplement consumption was converted to nutrient intakes by using the MDC supplement database. This database contains information about nutrient contents in medical drugs, herbal remedies, and supplements consumed in the MDC study (35). The relative validity of the MDC method was evaluated in 1984-1985 in a sample of Malmö residents (n = 105 women and 101 men aged 50-69 y) with the use of 18 d of weighed records, 3 d every second month during 1 y, as the reference method (36, 37). Pearson's correlation coefficients in women, adjusted for total energy, between the reference method and the MDC method were 0.75 for folate, 0.69 for dietary fiber, 0.70 for carotene, and 0.71 for ascorbic acid.
Dietary variables
This study examined daily dietary and total intakes (including supplements) of energy and the following nutrients: folate (µg), vitamin B-12 (µg), vitamin B-6 (mg), riboflavin (mg), dietary fiber (g), carotene (mg), and ascorbic acid (mg). Total folate intakes were calculated by adding folic acid intake from supplements to the intake from foods. In addition, dietary folate equivalents were calculated based on the assumption that the bioavailability of synthetic folic acid consumed in a meal is 1.7 times the bioavailability of food folate (38), ie, dietary folate equivalents = µg food folate + 1.7 x µg folic acid from supplements.
Energy-adjusted variables were obtained by regressing intakes of all nutrients on total energy intake (39). Quintiles of nutrient residuals were used as exposure categories. Consumption of folic acid-containing supplements (yes or no) was based on information about the current use of supplements from the menu book.
Other variables
In September 1994, the processing of dietary data was slightly altered (34). Method version (indicating data collection before or after 1 September 1994) and season of data collection were examined as potential confounders of dietary relations. Information on age was obtained from the personal identification number. Age was divided into 5-y categories. The smoking status of the participants was defined as smokers (including irregular smokers), exsmokers, and never-smokers. Information on total alcohol consumption was converted into a 4-category variable. Women reporting zero consumption in the menu book and indicating no consumption of any type of alcohol during the previous year were categorized as zero reporters. The other category ranges were <15 g alcohol/d (low), 15-30 g alcohol/d (medium), and >30 g alcohol/d (high). Alcohol intake was also dichotomized (4.7 or >4.7 g/d). Leisure-time physical activity was assessed by using a questionnaire adapted from the Minnesota Leisure Time Physical Activity Questionnaire (40, 41). The number of minutes per week of 18 different activities was multiplied by an activity-specific intensity coefficient, and an overall leisure-time physical activity score was created. The score was divided into tertiles and categorized as low, medium, or high. Household activities were estimated in hours per week and divided into 4 groups with cutoffs every 10 h (0-9, 10-19, 20-29, or 30).
The participants were divided into 4 categories according to their highest level of education (8 y, 9-10 y, 11-13 y, or university degree). Classification of the socioeconomic index was based on information on job title, tasks, and position at work. The procedure was adapted from that of the 1989 Swedish population census (42). In this study, the information was collapsed into 5 categories: blue-collar workers, white-collar workers (low, medium, or high), and self-employed. The retired and unemployed categories were classified according to their position before retirement or unemployment.
Weight was measured to the nearest 0.1 kg by using a balance-beam scale while the subjects were wearing light clothing and no shoes. Standing height was measured with a fixed stadiometer calibrated in centimeters. Body mass index (BMI; in kg/m2) was calculated from a direct measurement of weight and height and a 3-category variable was created (BMI 25, 25-29, or 30). BMI was also dichotomized (25 or >25).
Age at menarche and age at menopause were used as continuous variables. Lactation was reported as the total number of months for all children. Years with menstrual cycles was the time span between menarche and menopause accounting for interruptions for pregnancies and lactation. The duration of contraceptive pill use (in y) was divided into 4 categories, with zero consumption as the lowest category. Current hormone replacement therapy (HRT; yes or no) was based on the questionnaire item, "Which medications do you use on a regular basis? ", in combination with information on drug use from the 7-d menu book (43). Age at birth of first child was divided into 4 categories with an additional category for women with no children. Parity was the number of children with no children in the lowest category, and 4 or more in the highest.
Statistical analysis
The SPSS statistical computer package (version 11.5; SPSS Inc, Chicago, IL) was used for all statistical analyses. All nutrient and food variables were log transformed (10-log) to normalize the distribution before analysis. A very small amount (0.01) was added before transformation to handle zero intakes.
The baseline status of established risk factors and potential confounders was examined in cases and noncases with Cox proportional hazard regression or analysis of variance. All models were adjusted for age. Models including dietary data were also adjusted for method version and season. Energy-adjusted partial correlation coefficients between intakes (dietary and total) of folate and other nutrients were computed.
Cox proportional hazard regression examined quintiles of folate intake (regressed on total energy) in relation to hazard ratios (HRs) of invasive breast cancer. Adjustments were made for age category, method version, season, and energy with the residual method (39). A second model included adjustments for vitamin B-12, vitamin B-6, and riboflavin. In the third multivariate model, adjustments were also made for established risk factors and potential confounders (ie, weight, height, leisure-time physical activity, household work, smoking, alcohol, socioeconomic status, age at menopause, and HRT). These covariates were identified from the literature or previously indicated potential confounding of the folate-breast cancer association within the MDC cohort (44, 45). To evaluate the probability that folate is a genuine risk factor for breast cancer, independent of other plant food components, the full model was also repeated with adjustments for either intake of ascorbic acid, carotene, or dietary fiber. Finally, analysis with the full model was stratified on the dichotomous BMI variable (25 and >25) and on alcohol intake (4.7 and >4.7 g/d). Sensitivity analyses were made for age (55 y), for alcohol intake (15 g/d), and by inclusion of in situ cancers. All steps were repeated for dietary folate intakes (excluding folic acid from supplements), total folate intakes, and dietary folate equivalents.
RESULTS
Age-adjusted breast cancer HRs of the study cohort are presented in relation to lifestyle and socioeconomic factors in Table 1. Exsmokers had a higher breast cancer incidence than did never smokers, and women with an alcohol intake >30 g/d had a higher breast cancer incidence than did zero consumers. High white-collar workers had a higher incidence than did blue-collar workers. Women who spent much time on household work had a lower incidence (>30 h/wk compared with <10 h/wk), but the difference was no longer significant after adjustment for age at menopause and parity (data not shown).
View this table:
TABLE 1. Hazard ratios (HRs) of invasive breast cancer associated with lifestyle and socioeconomic factors among postmenopausal women in the Malmö Diet and Cancer cohort (1991-2003)1
Current HRT was also a significant risk factor for invasive breast cancer (Table 2). Use of contraceptive pills for 16 y was associated with a higher incidence of breast cancer. However, the increase was no longer significant after HRT and age at menopause were controlled for (data not shown). Age at menopause was higher in cases than in noncases (P for trend = 0.04), but when the total number of years with menstrual cycles was computed, a slightly stronger association was seen (P for trend = 0.03) (Table 3). Both height and weight were significantly higher in cases than in noncases. Total mean daily intakes of B vitamins, fiber, carotene, and ascorbic acid did not significantly differ between cases and noncases in a basic model with adjustments for age, method version, and season (Table 4). The total median intake of folate in the study population was 238 µg/d.
View this table:
TABLE 2. Hazard ratios (HRs) of invasive breast cancer associated with reproductive and anthropometric factors among postmenopausal women in the Malmö Diet and Cancer cohort (1991-2003)1
View this table:
TABLE 3. Anthropometric and reproductive characteristics among cases and noncases of postmenopausal breast cancer in the Malmö Diet and Cancer cohort (1991-2003)1
View this table:
TABLE 4. Total daily nutrient intakes at baseline in cases and noncases of postmenopausal breast cancer from the Malmö Diet and Cancer cohort (1991-2003)
Partial correlation coefficients between intakes of folate and other nutrients are shown in Table 5. Dietary folate intake was strongly correlated with dietary intakes of vitamin B-6, fiber, and ascorbic acid and was moderately correlated with dietary intakes of riboflavin and carotene. Total folate intake was moderately correlated with total intakes of vitamin B-6, riboflavin, fiber, and ascorbic acid.
View this table:
TABLE 5. Energy-adjusted correlation coefficients between mean daily intakes of folate and other nutrients in postmenopausal women (n = 11699) in the Malmö Diet and Cancer cohort
HRs for the incidence of invasive breast cancer by quintiles of folate intake are shown in Table 6, Table 7, and Table 8. An important reduction in the HR for breast cancer incidence was seen in the highest quintile of dietary folate intake (HR: 0.56; 95% CI: 0.35, 0.90), total folate intake (HR: 0.56; 95% CI: 0.34, 0.91), and dietary folate equivalents (HR: 0.59; 95% CI: 0.36, 0.97) compared with the lowest quintile of intake. This association was observed after simultaneous adjustments for other B vitamins (involved in folate metabolism) and established risk factors. Significant negative trends were seen across quintiles of dietary folate intake (P for trend = 0.02), total folate intake (P for trend = 0.01), and dietary folate equivalents (P for trend = 0.01).
View this table:
TABLE 6. Hazard ratios (and 95% CIs) of invasive breast cancer across quintiles of dietary folate intake in women aged 50 y from the Malmö Diet and Cancer cohort (1991-2003)1
View this table:
TABLE 7. Hazard ratios (and 95% CIs) of invasive breast cancer across quintiles of total folate intake in women aged 50 y from the Malmö Diet and Cancer cohort (1991-2003)1
View this table:
TABLE 8. Hazard ratios (and 95% CIs) of invasive breast cancer across quintiles of dietary folate equivalents in women aged 50 y from the Malmö Diet and Cancer cohort (1991-2003)1
However, no significant associations could be seen in the basic model when only season, method version, and energy were adjusted for or when the basic model was extended with other B vitamins. The same tendencies were seen when established risk factors replaced B vitamins in the model (data not shown).
The addition of ascorbic acid intake to the full multivariate model gave a slightly stronger trend across quintiles of total folate intake (P for trend = 0.003). The reduced HR in the highest quintile of folate intakes and the negative trend remained after adjustments for fiber intake. However, the observed trend was only significant for total folate intake (P for trend = 0.01) and dietary folate equivalents (P for trend = 0.01). The results did not substantially change when carotene replaced fiber in the multivariate model.
Nineteen percent of the women consumed folic acid-containing supplements. Consumption of folic acid-containing supplements was independently associated with a lower HR for invasive breast cancer (0.63; 95% CI: 0.43, 0.94) in the full multivariate model with adjustment for dietary folate. However, the association between dietary folate intake and breast cancer was not influenced by the consumption of folic acid-containing supplements (data not shown). Supplement consumption tended to be more frequent in women with high dietary folate intakes (22% in the highest quintile compared with 15% in the lowest quintile).
The analysis was repeated with the exclusion of women younger than 55 y at baseline. The significant reduction in the HR in the highest quintile of dietary folate equivalents was strengthened (HR: 0.48; 95% CI: 0.28, 0.84; P for trend = 0.004) in this subsample and did not substantially change after adjustment for fiber intake, carotene, and ascorbic acid.
The trend indicating a lower breast cancer incidence with a higher intake of total folate (P for trend = 0.03) and a higher intake of dietary folate equivalents (P for trend = 0.05) remained after women who consumed >15 g alcohol/d were excluded (full multivariate model including fiber intake). Unfortunately, the subjects who consumed >15 g alcohol/d were too few to allow a separate analysis. No differences in the folate-breast cancer association were observed between women above (P for trend = 0.10) and below (P for trend = 0.11) the median alcohol consumption in the study population (4.7 g/d). Both groups indicated a nonsignificant lower breast cancer incidence in the highest quintile of total folate intake.
The test for interaction was not significant (quintiles of total folate intake x BMI; P = 0.76) after the folate-breast cancer association in different strata of BMI was examined. In a sensitivity analysis including in situ cancers, the folate-breast cancer association was in the same direction as in the main analysis, but somewhat weaker (total folate intake; P for trend = 0.02).
DISCUSSION
Folate intake was negatively associated with the incidence of invasive breast cancer in the MDC cohort after adjustments for known risk factors, other B vitamins, and other potential confounders. The negative association remained after adjustment for intakes of fiber, carotene, or ascorbic acid.
The heterogeneous population and the prospective design, which minimized the problem of differential recall bias, were advantages of this study. Because Sweden has a well-established national cancer register, the time of follow-up was nearly complete. Other strengths of this study were the high relative validity and reliability of estimated dietary intakes (33, 36, 46). In addition to information about folate intake from foods, we had detailed information on folic acid intake from supplements. This reduced the risk of confounding by other unknown bioactive components in fruit and vegetables. Because a high intake of dietary fiber has been related to a lower risk of breast cancer in the MDC cohort (45), adjustments for fiber were made to further avoid the problem of confounding. We also included other B vitamins in the statistical model because of their involvement in folate metabolism. The synthesis of methionine from homocysteine is dependent on folate and also on vitamin B-12, which acts as a cofactor to the enzyme methionine synthase. Riboflavin is a cofactor for methylene tetrahydrofolate reductase. Vitamin B-6 serves as a cofactor when tetrahydrofolate is converted to 5,10-methylene tetrahydrofolate and when homocysteine is converted to cystathionine (19, 20).
Also, corrections for methionine intake would have been of great value, but we did not have such data. We could not correct for family history of breast cancer either, because this information was not available. This risk factor, however, is considered more important in younger age groups (47). Another limitation was that we were not able to explore whether the association between high folate intake and breast cancer incidence was stronger in women with a high consumption of alcohol, because the number of women with high alcohol intakes was too small. On the other hand, we showed that the association remained after the exclusion of women who reported high alcohol intakes. Still, results from self-reported alcohol consumption should be carefully interpreted. Although the relative validity and the reproducibility of alcohol intake were high in the MDC study, the intake was higher with the reference method (36, 46). This indicates that many women who consumed >15 g alcohol/d might be found among those reporting intakes <15 g/d.
In line with our hypothesis, an inverse association between folate intake and breast cancer incidence was observed in women with a BMI > 25, but no significant association was observed in normal-weight women. However, the power to detect an interaction between BMI and folate intake was low, and the test for interaction was not significant. Consequently, we cannot exclude that the observed differences between strata were due to chance.
Four case-control studies suggest that high folate intakes are associated with a decreased risk of postmenopausal breast cancer (11-13, 48), and a Scottish study indicated a nonsignificant negative association (49). In 2 case-control studies, folate intake could not be related to postmenopausal breast cancer (15, 50). High dietary folate intake was recently associated with a decreased risk of postmenopausal breast cancer in women from a large French cohort (14), and a nested case-control study from Denmark reported that total folate intakes <300 µg/d indicate an increased risk, albeit nonsignificant (51). Results from most other prospective studies do not support the overall inverse association between folate intake and postmenopausal breast cancer observed in this study (15-18, 52). In fact, 4 cohort found no significant relations (15-17, 52). In addition, a study from a multicenter cohort in the United States showed that high total intakes are related to a significantly increased risk (18). However, most cohort studies observed a decreased risk with high folate intakes in women with a moderate to high consumption of alcohol (16, 17, 52). In a Danish study, an increased breast cancer risk was only associated with alcohol consumption in women with low folate intakes (51). To the best of our knowledge, no study has previously investigated whether the relation between folate intake and breast cancer is different between overweight women and normal-weight women. Yet, in another Swedish study, a high intake of ascorbic acid was only related to a lower risk of breast cancer in overweight women (27).
Our findings of an association between high folate intake and lower breast cancer incidence confirm the biological hypothesis that low intakes of folate enhance the development of breast cancer. Folate in a reduced form, 5-MTHF, acts as a methyl donor in the synthesis of methionine, which is needed for the synthesis of SAM. SAM is the most important methyl donor in biological reactions, including DNA methylation (3). Changes in DNA methylation patterns are an early event in carcinogenesis and may influence gene silencing (53, 54). Folate is also involved in DNA synthesis and repair; folate deficiency has been related to elevated uracil incorporation to DNA and subsequent chromosome breaks, which may contribute to an increased risk of cancer (55). Overweight women might have an increased need for DNA repair in mammary epithelial cells because of an elevated proliferative activity (56) caused by higher concentrations of estrogens (26). Also, high concentrations of insulin may be responsible for the increased risk of breast cancer in overweight women (57). Because high folate intakes have been shown to be related to increased insulin sensitivity and reduced insulin concentrations (58), a protective effect of folate in overweight women is plausible. Although folate is involved in DNA repair, ascorbic acid is an antioxidant that may prevent DNA damage by scavenging free radicals (28). However, the antioxidative properties of ascorbic acid may also increase the stability of folate during food preparation and digestion (59).
It is possibly the rather low folate intake in this population that, unlike most other cohort studies, enables us to detect an overall inverse association between folate intake and postmenopausal breast cancer. In the Nurses' Health Study (NHS), the total folate intake in the first quintile was in range of that in this study, but the NHS used the second quintile as a reference category (17). In 3 published cohort studies, cutoffs between the first and second intake categories of total folate ranged from 210 to 335 µg/d, compared with 178 µg/d in this study population, and cutoffs between the highest intake groups ranged from 400 to 853 µg/d, compared with 349 µg/d in this study population (15, 18, 51). The study of French women by Lajous et al (14) had no information on total folate intake but detected an inverse association despite higher dietary intakes than in our study. However, the greatest risk reduction seemed to occur after the first quintile of dietary folate intake (14). In the present study, only in the highest quintile of total folate intake did the median intake (456 µg) exceed the Nordic recommended daily intake of 300 µg folate. Consequently, the negative associations observed in the present study were not achieved at unrealistic intakes. It might be possible, by several methods, to increase the folate intake in this study population. In Sweden, most supplements with folic acid contain 400 µg, which is in line with the amount related to lower breast cancer incidence in this study. In addition, Sweden has not yet come to a decision concerning the mandatory folic acid fortification of foods. With regard to the contradictory results from different studies, including an American study that indicated an increased risk at total intakes much higher than in our study (18), an increased consumption of fruit and vegetables would be a safer way to increase folate intakes. The consumption of fruit and vegetables is relatively low in Sweden compared with that in other European countries, but is similar to consumption in the United States (60, 61). On the other hand, a trend toward a higher consumption of fruit and vegetables has been observed (62, 63).
In summary, our results indicate that high folate intakes are associated with a lower incidence of postmenopausal breast cancer. These findings could not be explained by intakes of other nutrients found in the same foods as folate.
ACKNOWLEDGMENTS
We thank the staff and participants of the Malmö Diet and Cancer Study.
The authors' responsibilities were as follows—UE: wrote the manuscript, performed the data analysis, and contributed to the study design; ES: helped revise the manuscript; BG: gave statistical advice, contributed to the study design, and helped revise the manuscript; HO: gave advice on matters related to oncology and helped interpret the results and revise the manuscript; EW: contributed to the study design and helped interpret the results and revise the manuscript. No conflicts of interest were reported.
REFERENCES
Related articles in AJCN: