SEARCH

SEARCH BY CITATION

Keywords:

  • breast cancer;
  • African-American women;
  • nutrition;
  • physical activity;
  • National Health and Nutrition Examination Survey;
  • menopause;
  • micronutrients;
  • Healthy Eating Index;
  • diet quality;
  • body mass index (BMI)

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES

BACKGROUND

Very little is known about the role of lifestyle in breast cancer risk, and even less is known about whether differences in lifestyle contribute to the disparities in this risk between African-American women and white women. In this study, the authors examined differences in diet and physical activity between African-American women and white women and discuss the research that is needed regarding the role of lifestyle in breast cancer risk.

METHODS

The authors used bivariate and multiple regression analyses to estimate the difference between African-American women and white women in body mass index (BMI), physical activity, the Healthy Eating Index, intake of selected nutrients, and serum levels of some micronutrients. Data were taken from the National Health and Nutrition Examination Survey III, 1988–1994.

RESULTS

African-American women had a higher BMI, and older African-American women were less active physically compared with white women. African-American women of all ages had a poorer quality diet, and they consumed more protein and cholesterol but less dietary fiber, folate, and vitamin A. African-American women had lower predicted serum levels of folate and vitamin A but higher predicted serum levels of lutein.

CONCLUSIONS

The current findings generally are consistent with the hypothesis that nutrition differences may contribute to the higher rate of breast cancer experienced by younger African-American women, although extensive research is needed. More longitudinal data and research on nutrition, genetics, and breast cancer among African-American women are needed. Studies should examine how to help African-American women make behavioral changes to reduce their risk of breast cancer. Cancer 2003;97(1 Suppl):280–8. © 2003 American Cancer Society.

DOI 10.1002/cncr.11020

According to the 1997 report from the American Institute for Cancer Research, Food, Nutrition, and the Prevention of Cancer: A Global Perspective,1 there is no convincing evidence of a link between diet and breast cancer. Some studies suggest a probable relationship between vegetables and fruit in reducing the risk. Conversely, high body mass, adult weight gain, and alcohol consumption probably increase the risk of breast cancer. Overall, we know very little about whether there is an association between diet and breast cancer, and we know even less about the dietary habits of African-American women that may put them at greater risk compared with white women.

Like other chronic diseases, the etiology and risk factors associated with the development of breast carcinoma are complex. McPherson and colleagues2 noted that established and probable risk factors for breast carcinoma include age, geography, reproductive history (age at menarche, age of first pregnancy, and menopause), family history, previous benign breast disease, disease in the other breast, socioeconomic group, diet, body weight, alcohol consumption, exposure to ionizing radiation, and exogenous hormones. Clearly, some of these factors are controllable, whereas others, such as race, ethnicity, and family history, are not.

Data show that women living in developed (Westernized) countries have higher mortality from breast cancer compared with women living in developing countries or East Asia. Ecologic studies conducted in the 1960s and 1970s found strong correlations between certain dietary factors and incidence of breast cancer.3–5 In 1975, Armstrong and Doll suggested that certain dietary components may contribute to increased risk of breast cancer and other malignancies.6 Moreover, countries with low fat intake, such as Japan, China, and India, exhibited low mortality from breast cancer compared with the United States, which has relatively higher intake of total fat and higher incidence of breast cancer.

Although genetics and family history play important roles in breast cancer risk, differences in environment, including diet and other lifestyle factors, also appear to be important. Hypotheses about the role of lifestyle in increasing the risk of breast cancer stemmed from the observations that low-risk migrant populations developed the same increased incidence of breast cancer and other malignancies as citizens in the newly adopted country.7 Several studies noted that Japanese and Polish migrants to the United States experienced an increased risk of breast cancer that nearly matched that of the U.S. population.8–12 More recently, Ziegler et al.13 found that Asian-American women with grandparents who were born in the West had a 50% greater risk of developing breast cancer compared with women who had grandparents who were born in the East. Moreover, the breast cancer risk among Asian-American women who were born in the East was determined by whether they were from rural or urban communities and by the number of years they lived in the West after migration. This further suggested that lifestyle may overcome genetic factors in influencing development of the disease.

Lifestyle that contributes to obesity may influence breast cancer risk. For example, Galanis et al.14 found that women in the highest quintile of body mass index (BMI) were at significantly higher risk of developing breast cancer compared with their leaner counterparts. The association between BMI and postmenopausal breast cancer risk was strongest among women age ≥ 65 years. However, other studies showed that higher BMI during late adolescence was associated with lower breast cancer incidence before and after menopause.15, 16 In addition, Huang et al.17 found that waist circumference was not related to premenopausal risk of breast cancer but was associated positively with risk among postmenopausal women.

Dietary factors influence the metabolism or bioavailability of endogenous estrogens, which may affect the risk of breast cancer. Case–control studies indicate that higher levels of estrone, estradiol, and bioavailable estradiol may quadruple the risk of breast cancer among postmenopausal women.18 Goldin and his colleagues19 found that premenopausal, vegetarian women had greater intakes of dietary fiber but less total and saturated fat compared with omnivorous women. The vegetarians consequently had increased fecal output, leading to greater estrogen excretion and reduced concentration of plasma estrogen. Other studies suggest that dietary fiber, especially wheat bran fiber, may protect against breast cancer by inhibiting enterohepatic recirculation of estrogens.20, 21 Foods such as soy may provide protection by mimicking estrogens and binding the estrogen receptors on breast tissue.

Nonetheless, how much do we know about the role of diet in the risk of breast cancer? We know very little, and we know even less about the diets of minority women and whether diet plays a different role in increasing the risk of breast cancer compared with white women.22

Therefore, we examined the nutrient intake, serum levels, diet quality (as measured by the Healthy Eating Index [HEI]), body mass index (BMI), and physical activity of African-American, women of other races, and white women in the nationally representative National Health and Nutrition Examination Survey III, 1988–1994 (NHANES III).

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES

The data for our analyses were taken from all women who were included in the adult file of NHANES III. In addition, we analyzed separately the subsets of women age ≤ 50 years (premenopausal) and women age > 50 years (presumed postmenopausal). NHANES III does not have reliable direct measures of whether a woman is premenopausal or postmenopausal; thus, we chose these cut-off ages to correspond with data reports from the Surveillance, Epidemiology, and End Results (SEER) Program of the Centers for Disease Control and Prevention. Sample sizes for each analysis varied slightly because of missing data. For the subset of women age ≤ 50 years, sample sizes ranged from 4639 to 4323 white women and from 888 to 783 African American women. For the subset of women age > 50 years, sample sizes ranged from 2528 to 2267 white women and from 294 to 243 African-American women. We present our results for the following variables:

  • BMI (body weight in kilograms divided by height in meters squared);

  • cumulative physical activity index, which was calculated by the authors by multiplying each activity in NHANES III by its intensity rating and summing all of the products;

  • nutrient intake, including protein, total fat, saturated fat, cholesterol, carbohydrates, fiber, alcohol, vitamin A (retinol equivalent [RE]), carotenes (RE), vitamin C, folacin, and lycopene;

  • serum levels for vitamin A, α-carotene, β-carotene, vitamin C, and folate; and

  • the HEI, including overall score, grains, fruits, vegetables, dairy, meat, total fat, saturated fat, cholesterol, sodium, and variety.

A set of preliminary analyses estimated the difference between the means of these variables for African-American women and white women. A difference-between-means t test was used to determine whether these differences were statistically significant.

A more sophisticated analysis estimated the differences between African-American women and white women using multiple regression to control for income, region, and residence in a city, because these variables may be related to both race and the outcome variables. The models for serum levels also controlled for total cholesterol and BMI. All models used standard errors that were corrected for effects of the sample design.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES

The tables presented below report the mean values of the variables discussed above for African-American women and white women. They also report the difference between the means and whether the difference is statistically significant at the 0.05 confidence level. The difference between means test did not assume that African-American and white subpopulations had identical variances.

These bivariate analyses showed that the diet and physical activity of African-American women may place them at greater risk for developing breast cancer. The data presented in Table 1 show that African-American women of all ages, age ≤ 50 years (premenopausal), and age > 50 years (postmenopausal) had significantly higher BMIs (P < 0.05) and lower levels of physical activity (P < 0.05) compared with white women. Lower levels of physical activity were particularly striking among older African-American women.

Table 1. Body Mass Index and Physical Activity: Mean Differences between African-American Women and White Women
BMI and activityWeighted sample mean valuesa
African AmericansWhitesDifference
  • BMI: body mass index.

  • a

    The numbers in the columns are the weighted sample mean values for African-American women and white women and the difference between the two.

  • b

    The difference was statistically significant at the 0.05 confidence level.

All ages   
 BMI28.4425.882.56b
 Physical activity78.39101.12− 22.73b
Age ≤ 50 yrs   
 BMI28.0125.162.85b
 Physical activity87.22106.92− 19.70b
Age > 50 yrs   
 BMI29.7327.202.53b
 Physical activity53.4291.25− 37.83b

African-American women of all ages consumed more total fat and cholesterol but less dietary fiber, carotenes, folate, and alcohol compared with whitewomen (Table 2). When the subset of African Americans was segmented into premenopausal and postmenopausal age groups, African-American women consumed as much total fat as white women, and premenopausal African-American women consumed more vitamin C (98.94 mg vs. 89.27 mg, respectively).

Table 2. Intake of Selected Nutrients: Mean Differences between African-American Women and White Women
Nutrient intakeMeans and estimated differences by race and age groupa
All agesAge ≤ 50 yrsAge > 50 yrs
African AmericansWhitesDifferenceAfrican AmericansWhitesDifferenceAfrican AmericansWhitesDifference
  • RE: retinol equivalent.

  • a

    The numbers in the columns are the weighted sample mean values for African-American women and white women and the difference between the two.

  • b

    The difference was statistically significant at the 0.05 confidence level.

Protein (g)67.1966.111.0870.2368.881.3557.6360.68−3.06
Total fat (g)71.7368.922.81b77.1774.802.3754.6157.39−2.78
Saturated fat (g)23.4123.290.1225.1625.46−0.3017.8919.03−1.15
Cholesterol (mg)256.56216.7239.84b266.48229.6336.86b225.35191.4333.92b
Carbohydrates (g)222.80224.49−1.68234.25237.75−3.49186.76198.48−11.73b
Fiber (g)12.3014.38−2.08b12.1014.01−1.91b12.9015.10−2.20b
Vitamin A (RE)853.84948.31−94.74759.75866.18−106.431150.081109.3240.77
Carotenes (RE)401.28482.88−81.60b336.87434.28−97.41b603.83578.1325.70
Vitamin C (mg)98.4394.064.3798.9489.279.68b96.83103.47−6.64
Folate (μg)209.61249.26−39.65b201.98242.06−40.08b233.65263.39−29.74b
Alcohol (g)4.726.55−1.83b5.847.82−1.98b1.194.05−2.86b

African-American women of all ages and the premenopausal subsegment of African-American women had a poorer diet quality, as measured by HEI scores, on all the components except saturated fat and sodium (Table 3). In general, African-American women tended to consume fewer grains, fruits, vegetables, and dairy foods and had less variety in their diets.

Table 3. Selected Serum Levels: Mean Differences Between African-American and White Women
Serum levelsMeans and estimated differences by race and age groupa
All agesAge ≤ 50 yrsAge > 50 yrs
African AmericansWhitesDifferenceAfrican AmericansWhitesDifferenceAfrican AmericansWhitesDifference
  • a

    The numbers in the columns are the weighted sample mean values for African-American women and white women and the difference between the two.

  • b

    The difference was statistically significant at the 0.05 confidence level.

Folate (ng/mL)5.487.87−2.39b5.026.77−1.75b6.909.89−2.99b
Vitamin C (mg/dL)0.650.85−0.20b0.650.78−0.13b0.630.97−0.34b
Vitamin A (μg/dL)50.4956.41−5.92b47.5252.85−5.33b59.5862.92−3.34b
Alpha carotene (μg/dL)3.835.20−1.37b3.434.74−1.30b5.066.06−1.00b
Beta carotene (μg/dL)20.2922.40−2.1116.1419.20−3.06b32.9828.264.73
Lutein (μg/dL)23.8920.743.16b21.8319.072.75b30.2223.786.44b
Lycopene (μg/dL)22.3322.770.4423.8524.27−0.4317.7020.03−2.33b

The poorer diet quality and lower intake of vegetables and fruits appeared to be reflected in significantly lower serum levels of folate, vitamin C, vitamin A, and α-carotene but significantly higher serum levels of lutein among postmenopausal and African-American women of all ages (Table 4). Premenopausal African-American women also had lower serum levels of β-carotene.

Table 4. Healthy Eating Index and Components: Mean Differences Between African-American Women and White Womena
HEI componentsMeans and estimated differences by race and age groupb
All agesAge ≤ 50 yrsAge > 50 yrs
African AmericansWhitesDifferenceAfrican AmericansWhitesDifferenceAfrican AmericansWhitesDifference
  • HEI: Health Eating Index.

  • a

    The HEI is constructed so that higher numbers indicate a better diet.

  • b

    The numbers in the columns are the weighted sample mean values for African-American women and white women and the difference between the two.

  • c

    The difference was statistically significant at the 0.05 confidence level.

Overall score59.9664.90−4.94c58.5762.52−3.95c64.3569.56−5.21c
Grains5.736.33−0.60c5.706.22−0.52c5.806.53−0.72c
Fruits3.303.99−0.69c2.863.20−0.34c4.685.54−0.86c
Vegetables5.196.13−0.94c5.015.83−0.82c5.786.73−0.95c
Dairy4.646.40−1.76c4.736.54−1.81c4.366.12−1.76c
Meat7.096.260.83c7.186.280.90c6.816.220.59c
Total fat6.136.41−0.28c5.946.21−0.276.726.80−0.08
Saturated fat6.426.190.246.295.900.39c6.846.740.09
Cholesterol7.898.52−0.63c7.738.32−0.60c8.408.91−0.51c
Sodium6.947.11−0.176.536.67−0.148.237.970.26
Variety6.637.56−0.93c6.607.35−0.74c6.737.99−1.26c

Multivariate Statistics

Because some of the differences that we observed in the bivariate analyses may have been due to differences in income (as a percent of the poverty line), regionality, or urban living, we performed multivariate regression analyses to control for those variables. Tables 5–8 present the results of our multivariate analyses. The tables show the estimated coefficients for the differences between African-American women and white women in the outcome variables discussed above. The models for serum levels also controlled for BMI and total serum cholesterol levels. The following general findings were based on the multivariate regression models:

  • After controlling for income, region, and living in a city, several of the differences observed in the bivariate analysis were no longer statistically significant. Premenopausal African-American women had the same level of physical activity, the same consumption of carotenes and alcohol, and the same serum levels of α-carotene, β-carotene, and vitamin C as similarly situated white women. Postmenopausal African-American women had the same folate consumption and serum levels of α-carotene and lycopene as similarly situated white women.

  • The multivariate analysis confirmed the findings in the bivariate model with regard to BMI. Controlling for income, region, and urban dwelling, premenopausal, postmenopausal and African-American women of all ages had higher predicted BMIs, but only postmenopausal African-American women were less physically active compared with white women of similar age (Table 5).

  • African-American women of all ages consume more protein and cholesterol, less dietary fiber, but the same amount of total and saturated fat as their white counterparts. However, the difference in protein consumption is not different when the women were divided into in premenopausal and postmenopausal age groups (Table 6).

  • Premenopausal and postmenopausal African-American women had poorer overall diet quality and less variety compared with similarly situated whitewomen (Table 7). On the individual components, premenopausal and postmenopausal African-American women had a worse predicted HEI score for grains, fruits, vegetables, dairy, total fat, cholesterol, and variety; they had a better predicted score for the meat component only.

  • African-American women had lower predicted serum levels of folate and vitamin A but higher predicted serum levels of lutein compared with white women. Older African-American women had lower serum vitamin C levels and higher β-carotene levels (Table 8).

Table 5. Body Mass Index and Physical Activity: Predicted Differences between African-American Women and White Women Controlling for Income, Region, and Residence in a City
VariableEstimated difference (95% CI)a
All agesAge ≤ 50 yrsAge > 50 yrs
  • 95% CI: 95% confidence interval; BMI: body mass index.

  • a

    The numbers in the columns are the estimated coefficients for the difference between African-American women and white women. The 95% CI is presented in parentheses.

  • b

    The estimated coefficient was statistically significant at the 0.05 confidence level.

BMI2.37 (1.74, 3.01)b2.49 (1.73, 3.25)b2.62 (1.69, 3.55)b
Physical activity−6.92 (−16.58, 2.74)−2.98 (−14.89, 8.94)−23.98 (−34.42, −13.53)b
Table 6. Nutrient Intakes: Predicted Differences between African-American Women and White Women Controlling for Income, Region, and Residence in a City
Nutrient intakeEstimated difference (95% CI)a
All agesAge ≤ 50 yrsAge > 50 yrs
  • 95% CI: 95% confidence interval; RE: retinol equivalent.

  • a

    The numbers in the columns are the estimated coefficients for the difference between African-American women and white women. The 95% CI is presented in parentheses.

  • b

    The estimated coefficient was statistically significant at the 0.05 confidence level.

Protein (g)2.11 (0.05, 4.17)b2.51 (−0.36, 5.39)−2.10 (−5.15, 0.94)
Total fat (g)2.54 (−0.02, 5.11)2.03 (−1.02, 5.09)−1.95 (−6.49, 2.58)
Saturated fat (g)−0.19 (−1.18, 0.81)−0.67 (−1.85, 0.51)−0.91 (−2.57, 0.76)
Cholesterol (mg)33.46 (20.05, 46.88)b29.29 (14.88, 43.71)b30.76 (8.81, 52.71)b
Carbohydrates (g)−5.3 (−13.95, 3.36)−8.34 (−19.92, 3.23)−11.39 (−22.02, −0.77)b
Fiber (g)−1.53 (−2.23, −0.84)b−1.42 (−2.28, −0.56)b−1.53 (−2.67, −0.40)b
Vitamin A (RE)−37.82 (−140.66, 65.01)−54.68 (−191.25, 81.90)97.79 (−111.3, 306.9)
Carotenes (RE)−33.42 (−89.83, 22.99)−53.39 (−117.81, 11.04)80.46 (−5.63, 166.54)
Vitamin C (mg)9.48 (2.79, 16.18)b13.92 (5.17, 22.67)b−1.89 (−10.65, 6.87)
Folate (μg)−27.94 (−41.09, −14.79)b−27.83 (−43.39, −12.28)b−18.94 (−40.57, 2.68)
Alcohol (g)−0.60 (−2.12, 0.91)−0.53 (−2.54, 1.48)−1.89 (−2.77, −1.02)b
Table 7. Healthy Eating Index and Components: Predicted Differences between African-American Women and White Women Controlling for Income, Region, and Residence in a Citya
HEI componentEstimated difference (95% CI)b
All agesAge ≤ 50 yrsAge > 50 yrs
  • 95% CI: 95% confidence interval; HEI: Healthy Eating Index.

  • a

    The HEI is constructed so that higher numbers indicate a better diet.

  • b

    The numbers in the columns are the estimated coefficients for the difference between African-American women and white women. The 95% confidence interval is presented in parentheses.

  • c

    The estimated coefficient was statistically significant at the 0.05 confidence level.

Overall score−3.46 (−4.58, −2.33)c−2.61 (−3.90, −1.31)c−3.93 (−5.25, −2.61)c
Grains−0.53 (−0.76, −0.30)c−0.44 (−0.72, −0.16)c−0.71 (−1.04, −0.39)c
Fruits−0.33 (−0.69, 0.02)0.05 (−0.33, 0.44)−0.72 (−1.16, −0.28)c
Vegetables−0.69 (−0.91, −0.46)c−0.67 (−0.97, −0.38)c−0.50 (−0.82, −0.17)c
Dairy−1.53 (−1.74, −1.33)c−1.51 (−1.76, −1.25)c−1.66 (−2.03, −1.29)c
Meat0.87 (0.69, 1.05)c0.87 (0.65, 1.09)c0.72 (0.40, 1.04)c
Total fat−0.28 (−0.58, 0.02)−0.29 (−0.65, 0.07)−0.08 (−0.52, 0.37)
Saturated fat0.31 (0.02, 0.60)c0.44 (0.10, 0.78)c0.12 (−0.39, 0.63)
Cholesterol−0.49 (−0.70, −0.28)c−0.46 (−0.74, −0.17)c−0.38 (−0.74, −0.01)c
Sodium−0.16 (−0.41, 0.09)−0.12 (−0.47, 0.22)0.18 (−0.12, 0.47)
Variety−0.61 (−0.92, −0.30)c−0.48 (−0.87, −0.09)c−0.88 (−1.19, −0.58)c
Table 8. Serum Levels: Predicted Differences between African-American Women and White Women Controlling for Income, Region, Residence in a City, Body Mass Index, and Total Serum Cholesterol
Serum levelsEstimated difference (95% CI)a
All agesAge ≤ 50 yrsAge > 50 yrs
  • 95% CI: 95% confidence interval.

  • a

    The numbers in the columns are the estimated coefficients for the difference between African-American and white women. The 95% CI is presented in parentheses.

  • b

    The estimated coefficient was statistically significant at the 0.05 confidence level.

Folate (ng/mL)−1.46 (−1.85, −1.07)b−0.77 (−1.16, −0.38)−2.47 (−3.25, −1.69)b
Vitamin C (mg/dL)−0.09 (−0.15, −0.04)b−0.01 (−0.07, 0.04)−0.26 (−0.34, −0.18)
Vitamin A (μg/dL)−4.33 (−5.63, −3.02)b−4.39 (−5.69, −3.10)b−2.35 (−4.48, −0.23)b
Alpha carotene (μg/dL)−0.42 (−0.96, 0.13)−0.44 (−0.93, 0.05)0.05 (−1.13, 1.23)
Beta carotene (μg/dL)1.79 (−0.16, 3.75)0.75 (−0.97, 2.48)8.26 (3.60, 12.93)b
Lutein (μg/dL)5.6 (4.33, 6.88)b4.87 (3.65, 6.10)b8.51 (6.27, 10.75)b
Lycopene (μg/dL)1.01 (0.01, 2.01)b0.45 (−0.72, 1.61)−0.07 (−1.33, 1.18)

Women who consumed diets rich in vegetables and fruits seemed to have a lower risk of breast cancer.23–25 The details of the protection vary from study to study, but a protective effect has been documented in India, Greece, Italy, the United States, Sweden, and Europe. Our study shows that poor diet quality, lower intake of certain micronutrients associated with fruit and vegetables, and lower serum levels of vitamin A and folate by African-American women may contribute to greater risk of breast cancer. This should be viewed with extreme caution, however, because our findings did not test directly the relation between nutrition and breast cancer.

Foods that are high in dietary fiber and antioxidants (primarily grains, fruits, and vegetables) also contain many other nutrients that may reduce the risk of breast cancer. The best current evidence indicates that diets high in carotenoids and, to a lesser extent, vitamin A reduce the risk of breast cancer, particularly for premenopausal women who may be predisposed to breast cancer because of family history or moderate-to-high alcohol consumption.24, 25 Diets high in vitamins C or E do not appear to reduce the risk of breast cancer. Serum and plasma β-carotene studies have not been consistent.24

Diets high in dietary fiber appear to reduce the risk of breast cancer, but there is debate over the actual mechanism.26, 27 Although the African-American women in our study consumed significantly less dietary fiber compared with white women (1.5 grams less), it is unlikely that this difference is of biologic significance.

Moderate-to-high alcohol consumption (> 15–20 g [about 1 drink] per day) appears to increase the risk of breast cancer.28 Folate consumption may reduce the risk of breast cancer among women who consume > 15 g of alcohol per day.29–31 In the current study, folate consumption was reduced significantly in premenopausal and African-American women of all ages, but not among the older group. Again, it is unknown whether 28 μg fewer of folate are biologically significant enough to place these women at greater risk for breast cancer compared with their white counterparts. This reduced folate intake may be especially important, because alcohol consumption among premenopausal African-American women is the same for white women after controlling for income, region, and living in the city.

Overweight or obese women have a greater risk of developing breast cancer after menopause.32 This increased risk appears to be even more pronounced among women who gain weight throughout adulthood. Conversely, premenopausal overweight women have less risk of developing breast cancer.33 The finding that younger African-American women had a higher BMI suggests that they should be at lower risk of breast cancer. However, data from the SEER Program show that this group is at greater risk of developing breast cancer compared with white women or women of other races.34 Thus, our results are inconsistent with the previous findings. Other factors probably explain why this group is at greater risk of breast cancer.

Physical activity appears to reduce the risk of breast cancer.33 the results of the current study demonstrated that, after controlling for income, region, and city living, premenopausal African-American women were as active physically as similarly situated white women. Conversely, older African-American women were less active physically but were at less risk of breast cancer. These findings also are inconsistent with other hypotheses about the role of physical activity in breast cancer risk.

Very little research has investigated the nutritional and physical activity differences between African-American women and white women as they apply to breast cancer. We do not know how the diets of African-American women and white women differ with regard to nutrients believed to be related to breast cancer, and we do not know whether diet has the same effect on the risk of breast cancer for African-American women and white women. Two ongoing studies—the Los Angeles, CA (LAC)-Hawaii Multi-Ethnic Cohort Study and the Women's Health Initiative—should help to address some of these issues as the data become available.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES

The current study findings generally are consistent with the hypothesis that nutrition differences may contribute to the higher rate of breast cancer experienced by younger African-American women. Conversely, our findings are not consistent with the lower rate of breast cancer experienced by older African-American women. That is, nutrition differences generally do not seem to contribute to the lower rate of breast cancer among older African-American women.

Even after controlling for income, region, and residence in a city, African-American women have poorer nutrient profiles that may place them at greater risk of breast cancer. African-American women of all ages consume fewer servings of vegetables and consume less dietary fiber, and African-American women age > 50 years consume fewer servings of fruit. African-American women of all ages have a higher BMI, which may put older African-American women at greater risk of breast cancer. African-American women age < 50 years consume about the same amount of alcohol as their white counterparts, but they consume less folate and have lower levels of serum folate. Overall, there are so few data about diet, nutrient intakes, and corresponding serum levels among African-American women that we can draw no conclusions about breast cancer risk in this population.

Research Needs

The expert round table discussion on nutrition and breast cancer at the 2000 Summit Meeting on Breast Cancer among African-American Women identified several areas that require further research.

Data on Nutrition and Breast Cancer among African-American Women

Our analysis highlights the need for more research on race specific factors and breast cancer risk. Many of the existing breast cancer studies, especially prospective studies, are comprised primarily of white women and do not have the necessary statistical power to evaluate race by risk factor interactions. Because breast cancer is relatively rare, it is important for researchers to develop end points prior to cancer diagnosis. Studies also should include enhanced biomarkers for long-term nutrition and early signs of cancer.

A major longitudinal, prospective study would be ideal, but it would be expensive and slow to yield research results. Alternatively, a meta-analysis pooling several existing longitudinal studies with cohorts of African-American women may be possible if the data from these studies are compatible. Studies that may be suitable for pooling include the Women's Health Initiative, the Los Angeles Multiethnic Cohort Study, and the Nurses Health Study. Case–control studies may be a more cost-effective strategy for studying these linkages, but it will be difficult to examine the long-term impact of diet and physical activity on breast cancer using case–control studies.

Participants in the round table discussion noted that studies including African-American women must be sensitive to the culture and special needs of the population. Developing the trust of the African-American community is critical in a health study of breast cancer.

Finally, we encourage the broader availability of existing longitudinal data on nutrition and breast cancer. The breast cancer research community would benefit if more researchers could access ongoing longitudinal studies to conduct original research projects and to reproduce existing findings.

The interaction of genetics and nutrition is a key area for future research. It is likely that specific genetic polymorphisms interact with aspects of the diet to affect the risk of breast cancer. Identifying these gene-diet interactions holds the promise of introducing early, preventative diet modifications to reduce the risk of breast cancer among individuals with certain genetic predispositions.

Many questions remain about the definitions and measurement of dietary components themselves. The participants of the round table discussion highlighted the issues of particular importance: dietary fat, dietary fiber, certain micronutrients, and other nonnutrient components of the diet.

For example, the relation between dietary fat and the risk of breast cancer is complex, and several aspects of the relation between dietary fat and the risk of breast cancer require further investigation. The consensus of the expert panel was that measurement of total dietary fat is too crude, and better methods need to be developed that include biomarkers capable of assessing long-term fat consumption.

Specific components of fat are likely to interact with genetic polymorphisms and hormones to affect a woman's risk of breast cancer. For instance, the olive oil used in Southern Europe is associated with protection from breast cancer; however, we do not know whether it is the monounsaturated fat in olive oil or other components that offer the protection. Epidemiologic research on the link between dietary fat and breast cancer needs to reflect these complexities, and animal studies need to develop further our understanding of the relations between specific components of fat and breast cancer risk.

Research also is needed to develop targeted and culturally sensitive programs to help African-American women change their behaviors to reduce the risk of breast cancer. Initiatives from the National Heart, Lung, and Blood Institute may offer a useful model for evaluating a number of pilot programs designed to produce dietary and physical activity behavioral changes in the African-American community. Multiple pilot programs should be initiated and, more importantly, evaluated for their effectiveness. After evaluation, the successful programs can be launched to a broader population.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES

The authors gratefully acknowledge the assistance of Rachel Kerestes and Patricia Smith in the preparation of this article.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES