Intake of whole grain products and risk of breast cancer by hormone receptor status and histology among postmenopausal women

Authors


Abstract

No clear relationship between whole grain products and risk of breast cancer has been established. In a large prospective cohort study, we investigated the association between intake of whole grain products and risk of breast cancer by tumour receptor status [oestrogen receptor (ER) and progesterone receptor (PR)] and tumour histology (ductal/lobular). It was further investigated whether the association differed by use of hormone replacement therapy (HRT). The study included 25,278 postmenopausal women participating in the Danish Diet, Cancer and Health cohort study (1993–1997). During a mean follow-up time of 9.6 years, 978 breast cancer cases were diagnosed. Associations between intake of whole grain products and the breast cancer rate were analysed using Cox's regression model. A higher intake of whole grain products was not associated with a lower risk of breast cancer. Per an increment in intake of total whole grain products of 50 g per day the adjusted incidence rate ratio (95% confidence interval) was 1.01 (0.96–1.07). Intake of rye bread, oatmeal and whole grain bread was not associated with breast cancer risk. No association was observed between the intake of total or specific whole grain products and the risk of developing ER+, ER−, PR+, PR−, combined ER/PR status, ductal or lobular breast cancer. Furthermore, there was no interaction between intake of whole grain products and use of HRT on risk of breast cancer. In conclusion, intake of whole grain products was not associated with risk of breast cancer in a cohort of Danish postmenopausal women. © 2008 Wiley-Liss, Inc.

It has been hypothesized that a high intake of whole grains may reduce the risk of breast cancer.1 Whole grains are sources of several compounds, including dietary fibres and lignans, which might be of importance in breast cancer development.2 It has been suggested that dietary fibres reduce endogenous oestrogen levels by interruption of the enterohepatic circulation of oestrogens3 through several proposed mechanisms.4–8 Lignans and/or their metabolite enterolactone may exert a possible protective effect by competitive binding with oestrogen receptors (ERs) and in this way reduce the more potent effects of endogenous oestrogens9, 10 Other properties of the lignans, like antioxidative effects, increasing plasma sex hormone–binding globulin levels, alteration of growth factor actions, inhibition of 5α-reductase and 17β-hydroxysteroid dehydrogenase and inhibition of aromatase have recently been summarized.10 Additionally, whole grains might indirectly reduce the risk of breast cancer by preventing obesity, a breast cancer risk factor in postmenopausal women.

Only few epidemiological studies have examined the hypothesis that a high intake of whole grains reduces risk of breast cancer. Results from case–control studies have been inconsistent11, 12 and the only prospective study published to date based on data from the Iowa Women's Health Study do not report any protective role of whole grains towards development of breast cancer.13 This finding may reflect a low and less varied whole grain product intake in the Iowa Women's cohort, thus an association between higher intake of whole grains and risk of breast cancer cannot be excluded. One of the characteristics of the diet in the Nordic countries is the fundamental position of whole grain products. This means that the Nordic populations are characterized by high intake of whole grain products and considerable variation in intake. Therefore, Nordic populations are well suited for investigations aiming at exploring potential cancer reducing effects of whole grain products. We are unaware of any previous studies evaluating the association between intake of whole grain products and risk of breast cancer using data from Nordic populations.

The level of circulating endogenous oestrogens could have an impact on the effects of lignans/enterolactone in relation to risk of breast cancer, thus it is relevant to consider whether use of hormone replacement therapy (HRT) interferes with the possible benefit of whole grain product intake. This hypothesis was addressed in the only prospective study on whole grains and breast cancer, in which an increasing risk of breast cancer with an increasing whole grain intake was indicated among current oestrogen users, whereas this was not the case among nonusers or former oestrogen users.13

Breast cancer can be classified according to oestrogen (ER) and progesterone (PR) receptor status.14 Epidemiological studies have shown that the effect of reproductive15 and anthropometric factors16 might vary by hormone receptor status, indicating that receptor-positive breast cancers are aetiologically different from receptor-negative breast cancers. It has been indicated that dietary factors also have different effects on the risk of breast cancers characterized by hormone receptor status17–20; however, the only prospective study focusing on whole grains found that the effect of whole grains did not differ by ER status.21 Also, the effect of breast cancer risk factors may vary according to histological subtype,22–24 though dietary factors have only been investigated to a limited degree25 and no previous study has focused on whole grains.

The aim of the present study was to evaluate the association between intake of whole grain products, including specific types, and risk of breast cancer among postmenopausal women within the Diet, Cancer and Health cohort study. Also, it was investigated whether this association differed not only by ER status, but also PR status and tumour histology, and whether use of HRT modifies the association between consumption of whole grain products and risk of breast cancer.

Abbreviations

BMI, body mass index; CI, confidence interval; ER, oestrogen receptor; ER/PR status, oestrogen receptor/progesterone receptor status; HRT, hormone replacement therapy; IRR, incidence rate ratio; OR, odds ratio; PR, progesterone receptor.

Material and methods

Study population

The prospective Diet, Cancer and Health cohort study was initiated in 1993–1997, where 160,725 Danish men and women were invited to participate. The inclusion criteria were age 50–64 years, living in the greater Copenhagen and Aarhus, born in Denmark and not registered with a previous cancer diagnosis in the Danish Cancer Registry.26 Subjects were identified by the unique 10-digit identification number, which is allocated to every Danish citizen by the Danish Central Population Registry. Among the invited women, 29,875 did participate.

The Diet, Cancer and Health study and the present study were approved by the Regional Ethical Committees on human studies in Copenhagen and Aarhus and by the Danish Data Protection Agency.

Data collection

Dietary information was obtained by a 192-item semiquantitative food frequency questionnaire completed by the participants at baseline. A description of the development and validation of the food frequency questionnaire has been published previously.27–30 Consumption was assessed in 12 categories of predefined responses, ranging from “never” to “8 times or more per day.” Daily intakes of foods and nutrients were calculated for each participant by means of the software programme FoodCalc (www.ibt.ku.dk/jesper/foodcalc/) using population specific standardized recipes and portion sizes. The intake of whole grain products was calculated from responses to 3 specific questions concerning intake of rye bread, whole grain bread and oatmeal, and was measured in grams per day. Data on lifestyle was collected at baseline through a lifestyle questionnaire, including questions about social factors, health status, reproductive factors and lifestyle habits. From this questionnaire the following information was available: years of school education, parity, age at first birth, use of HRT, duration of HRT use, breast feeding and physical activity. Anthropometrical measurements (weight and height) were obtained by trained professionals.

Case ascertainment

All cohort members were linked to the Central Population Registry for information on vital status and emigration. Each cohort member was followed up for breast cancer from date of entry, i.e., date of visit to the study center until the date of diagnosis of any cancer (except nonmelanoma skin cancer), date of death, date of emigration or April 27, 2006. Up to December 31, 2003 information on cancer incidence was obtained by linkage to the Danish Cancer Registry using the personal identification number of each participant. All individuals in Denmark that are diagnosed with cancer are registered herein by means of the Central Population Registry (CPR) system.26 From 2004 to April 27, 2006 information on cancer occurrence was obtained through record linkage to the Danish Pathology Databank (www.patobank.dk) because of a delay in the update of the Danish Cancer Registry. The Danish Pathology Databank is a national databank that via online computer systems collects and stores all histological and cytological examinations that have taken place at pathology departments in Denmark. Information on hormone receptor status and tumour histology was ascertained from the Danish Breast Cancer Cooperative Group,31 which holds records on a range of details for about 90% of breast cancers diagnosed in Denmark. Linkage was performed by use of the personal identification number.

Statistics

The analyses of the relation between the exposure variables and breast cancer rates were based on the Cox proportional hazard model using age as the time scale to ensure that the estimation procedure was based on comparisons of individuals at the same age, which allowed for age adjustment to prevent confounding by age. Time under study was included as a time-dependent variable modelled by a linear spline with a boundary at 1, 2 and 3 years after entry into the study cohort to allow for a possible “healthy-participants” effect. A linear spline was used because this allows a steady increase in the rate during the 1st year of follow-up.32 In the analyses considering ER, PR status and histological status the 2 types of breast cancer defined by the receptor status or histological status were treated as competing causes of failure, that is, in separate analyses in which breast cancers of the opposite type were censored at the age at the cancer diagnosis.

Adjustments were performed in 2 steps. First, all models were adjusted for baseline values of established breast cancer risk factors, such as parity (entered as two variables; parous/nulliparous (categorical) and number of births (n; continuous), age at first birth (years; continuous), school education [low(≤7 years), medium (8–10 years), high (≥11 years)], duration of HRT use (years, continuous), HRT status (never, past or current), body mass index (weight [kg]/height [m]2; continuous) and alcohol intake (g/day; continuous). In a second adjustment step, models were additionally adjusted for baseline values of potential breast cancer risk factors, such as intake of red meat (g/day; continuous), processed meat (g/day; continuous), fruits (g/day; continuous), vegetables (g/day; continuous), vitamin C from diet (mg/day; continuous), total energy intake (kJ/day; continuous), smoking status (never, past or current), cumulative breastfeeding (months; continuous) and total leisure time physical activity (active vs. nonactive; continuous).

All quantitative variables were entered linearly in the Cox model because this is biologically more reasonable than the step functions corresponding to categorization and, furthermore, increases the power of the analyses.33 The linearity of the associations was evaluated graphically by linear splines with 3 boundaries placed at the quartiles among cases. We found no significant departure from linearity; therefore, all quantitative variables could be entered linearly in the model. The incidence rate ratios (IRRs) of the associations of the linear whole grain product variables were presented as the IRRs associated with a higher intake of 50 or 20 g per day based on an evaluation of the interquartile range among cases. The continuous analyses were further supplemented with categorical analyses for illustrative purposes. In these categorical analyses, the whole grain product variables were categorized into quartiles based on the distribution among cases. We investigated a possible interaction between the intake of whole grain products and HRT status by testing whether the slopes for the intake of whole grain products differed significantly in the three categories of HRT status (never, past or current). This model was compared with a model estimating a single common slope for the intake of whole grain products by using the likelihood ratio test. This means that the model included the main effect of the indicator variables for the 3 categories of HRT status and the 3 variables representing the whole grain product intake in the 3 HRT status groups.

Two-sided 95% confidence intervals (CI) for the IRRs were calculated based on Wald's test of the Cox regression parameter, i.e., on the log rate ratio scale. The procedure PHREG in SAS (release 9.0; SAS Institute, Cary, NC) on a TextPad platform was used for the statistical analyses.

Results

Women were excluded from the analysis if they had a previous cancer diagnosis before baseline (n = 335), did not fill in the lifestyle questionnaire (n = 8), reported no menstruation (n = 9) or were premenopausal (n = 2,429). From the resulting group of 27,094 postmenopausal women, we finally excluded 1,816 participants because of missing information on confounder or exposure variables leaving 25,278 postmenopausal women for analysis. The median age at entry into the cohort of the 25,278 postmenopausal women was 56 years (range, 50–64). Nine hundred and seventy-eight incident cases of breast cancer were identified during an average (5th to 95th percentile) length of follow-up of 9.6 (5-11) years.

Baseline characteristics of the breast cancer cases and the total cohort according to intake of whole grain products and established breast cancer risk factors are shown in Table I. The distribution of the established risk factors for breast cancer for the entire cohort and for those with breast cancer during follow-up differed in the expected directions. More cases were nulliparous, had a higher age at first birth, were current users of HRT at baseline and had a longer duration of HRT use, had a higher intake of alcohol and had a higher education.

Table I. Baseline Characteristics of 978 Breast cancer Cases and 25,278 Postmenopausal Women in the Danish Diet, Cancer and Health Study
 Cases (n = 978)Cohort (n = 25,278)
  • BMI, body mass index; HRT, hormone replacement therapy.

  • 1

    Among ever users of hormone replacement therapy.

  • 2

    Among parous.

Age at baseline, years  
 Median (5th–95th percentile)57 (50–64)56 (50–64)
Current use of HRT, n (%)467 (48)7,972 (32)
Past use of HRT, n (%)124 (13)3,813 (15)
Never use of HRT, n (%)387 (40)13,493 (53)
Duration of HRT use1, years5 (1–20)4 (1–18)
 Median (5th–95th percentile)  
Parous, n (%)824 (84)22,129 (88)
Age at first birth, years  
 Median (5th–95th percentile)24 (18–32)23 (18–31)
Number of births2  
 Median (5th–95th percentile)2 (1–4)2 (1–4)
BMI, kg/m2  
 Median (5th–95th percentile)25 (20–34)25 (20–34)
Alcohol intake, g/day  
 Median (5th–95th percentile)11 (1–44)10 (1–42)
Low school education, n (%)292 (30)8,156 (32)
Medium school education, n (%)486 (50)12,647 (50)
High school education, n (%)200 (20)4,475 (18)
Whole grain products, g/day  
 Median (5th–95th percentile)113 (32–230)112 (31–224)
Rye bread, g/day  
 Median (5th–95th percentile)63 (11–113)63 (11–113)
Oatmeal, g/day  
 Median (5th–95th percentile)7 (1–50)7 (1–50)
Whole grain bread, g/day  
 Median (5th–95th percentile)40 (1–100)40 (1–100)

Table II shows the relationship between intake of whole grain products and breast cancer IRRs. In continuous analyses, higher intake of whole grain products was not associated with the incidence rate of breast cancer in the unadjusted analysis. Per a 50-g increment in daily intake of whole grain products the IRR was 1.01 (95% CI, 0.96–1.06). Adjustments for established breast cancer risk factors did not alter the association between intake of whole grain products and risk of breast cancer [IRR = 1.01 (95% CI, 0.96–1.07)]. The association did not change with further adjustment for intake of red meat, processed meat, fruits, vegetables, vitamin C from diet, energy, smoking status, physically activity and cumulative breastfeeding (data not shown).

Table II. Incidence Rate Ratios (IRR) and 95% Confidence Intervals (CI) for Breast Cancer According to Intake of Whole Grain Products, the Danish Diet, Cancer and Health Study
 Whole grain product intake (g/day)Whole grain product intake as continuous variable
≤7272 to ≤112112 to ≤163>163Per each additional 50 g/day
  • IRR, incidence rate ratio; CI, confidence interval; ER, oestrogen receptor; PR, progesterone receptor.

  • 1

    Unadjusted analysis.

  • 2, 1

    Analysis adjusted for parity (parous/nulliparous and number of births), age at first birth, education, duration of hormone replacement therapy use, use of hormone replacement therapy, intake of alcohol and body mass index.

Total breast cancer     
 No. of cases243240280215978
 Crude IRR (95% CI)11.000.99 (0.83–1.18)1.01 (0.85–1.20)1.03 (0.86–1.23)1.01 (0.96–1.06)
 Adjusted IRR (95% CI)21.000.98 (0.82–1.17)1.00 (0.85–1.19)1.03 (0.85–1.24)1.01 (0.96–1.07)
ER+ cases     
 No. of cases175169193157694
 Crude IRR (95% CI)11.000.97 (0.79–1.20)0.97 (0.79–1.19)1.04 (0.84–1.29)1.00 (0.94–1.07)
  Adjusted IRR (95% CI)21.000.96 (0.78–1.19)0.96 (0.78–1.18)1.04 (0.84–1.30)1.00 (0.94–1.07)
ER− cases     
 No. of cases39445238173
 Crude IRR (95% CI)11.001.13 (0.74–1.74)1.17 (0.77–1.77)1.13 (0.72–1.77)1.06 (0.94–1.19)
 Adjusted IRR (95% CI)21.001.13 (0.73–1.74)1.16 (0.77–1.76)1.13 (0.72–1.77)1.06 (0.94–1.19)
PR+ cases     
 No. of cases71594848226
 Crude IRR (95% CI)11.000.81 (0.57–1.14)0.63 (0.44–0.91)0.80 (0.55–1.15)0.94 (0.84–1.05)
 Adjusted IRR (95% CI)21.000.79 (0.56–1.12)0.62 (0.43–0.90)0.79 (0.55–1.15)0.94 (0.84–1.06)
PR− cases     
 No. of cases32425238164
 Crude IRR (95% CI)11.001.28 (0.81–2.02)1.51 (0.97–2.34)1.39 (0.87–2.23)1.12 (0.99–1.26)
 Adjusted IRR (95% CI)21.001.26 (0.79–1.99)1.48 (0.95–2.31)1.36 (0.85–2.19)1.11 (0.99–1.26)
ER+/PR+ cases     
 No. of cases69564746218
 Crude IRR (95% CI)11.000.79 (0.55–1.12)0.64 (0.44–0.92)0.79 (0.54–1.14)0.94 (0.84–1.06)
 Adjusted IRR (95% CI)21.000.78 (0.54–1.10)0.63 (0.43–0.91)0.79 (0.54–1.15)0.95 (0.85–1.06)
ER−/PR− cases     
 No. of cases1726311993
 Crude IRR (95% CI)11.001.49 (0.81–2.75)1.69 (0.94–3.05)1.31 (0.68–2.52)1.10 (0.94–1.29)
 Adjusted IRR (95% CI)21.001.44 (0.78–2.66)1.61 (0.89–2.91)1.23 (0.64–2.39)1.08 (0.92–1.28)
Ductal cases     
 No. of cases176172182153683
 Crude IRR (95% CI)11.000.95 (0.77–1.17)0.96 (0.78–1.18)1.02 (0.82–1.27)1.01 (0.95–1.07)
 Adjusted IRR (95% CI)21.000.95 (0.77–1.17)0.96 (0.78–1.19)1.04 (0.83–1.29)1.01 (0.95–1.08)
Lobular cases     
 No. of cases26254130122
 Crude IRR (95% CI)11.000.94 (0.54–1.62)1.47 (0.90–2.40)1.35 (0.80–2.29)1.08 (0.94–1.25)
 Adjusted IRR (95% CI)21.000.89 (0.51–1.54)1.39 (0.85–2.28)1.26 (0.74–2.15)1.07 (0.93–1.23)

ER and PR status were available on 867 (89%) and 390 (40%) of cases, respectively. Of these, 694 were ER+ and 226 were PR+. We had information on combined ER/PR status for 389 (40%) of cases, with 218 being ER+/PR+ and 93 being ER−/PR−. We do not report result for cases with mixed ER/PR status because of small case numbers (ER−/PR+ = 7 and ER+/PR− = 71).

In continuous analyses, higher intake of whole grain products was not associated with either ER+, ER−, PR+ or PR− breast cancer incidence (Table II). When we divided cases according to combined ER/PR status, a higher intake of whole grain products was not significantly associated with the incidence rate of ER+/PR+ breast cancer or ER−/PR− breast cancer (Table II).

Information on tumour histology was obtained for 866 (89%) cases, with 683 of the tumours being ductal and 122 of the tumours being lobular. Information on histology was not obtained for the remaining 112 women. When we divided breast cancer cases according to histology, there was no overall association between intake of whole grain products and risk of ductal or lobular breast cancer (Table II).

We also examined the individual whole grain products in relation to overall breast cancer risk and by hormone receptor status and histology. However, none of the individual whole grain products (rye bread, oatmeal and whole grain bread) were significantly associated with overall risk of breast cancer or risk of different subtypes of breast cancer (data not shown).

The modifying effects of HRT status on the association between intake of whole grain products and risk of breast cancer were examined and results are shown in Table III. No significant interaction between intake of whole grain products and HRT status was observed (pinteraction = 0.10)

Table III. Incidence Rate Ratios (IRR) and 95% Confidence Interval (CI) for Breast Cancer of the Combined Effect of Whole Grain Products Per a 50 Grams Increment in Daily Intake and Hormone Replacement Therapy Status, the Danish Diet, Cancer and Health Study
 Hormone replacement therapy statusPinteraction
Never IRR (95% CI)Past IRR (95% CI)Current IRR (95% CI)
  • IRR, incidence rate ratio; CI, confidence interval.

  • 1

    Per each additional 50 g/day.

  • 2

    Unadjusted analysis.

  • 3

    Analysis adjusted for parity (parous/nulliparous and number of births), age at first birth, education, duration of hormone replacement therapy use, intake of alcohol and body mass index.

Whole grain products1    
 Model 121.01 (0.94–1.10)1.15 (1.00–1.31)0.97 (0.90–1.05)0.12
 Model 231.02 (0.94–1.10)1.15 (1.00–1.31)0.97 (0.90–1.05)0.12

Discussion

In a relatively large prospective study, no associations between overall intake of whole grain products or individual whole grain products and risk of breast cancer in postmenopausal women were observed. Furthermore, no associations were found when breast cancer cases were analysed according to ER and PR status or tumour histology. There was no interaction between intake of whole grain products and HRT status on the risk of breast cancer.

The strength of the present study includes its large size with 978 breast cancer cases. The linkage by the unique personal identification number to the Danish Cancer Registry or the Danish Pathology Databank ensures a complete and valid ascertainment of incident cases of breast cancer. Exposure information was obtained through validated questionnaires.34–36 The prospective design of the study precludes potential recall and selection biases. The intake of whole grain products was high and the variation in intake was large, which increases the possibility of finding an association between intake of whole grain products and risk of breast cancer, if one exists. The comprehensive data material made it possible to adjust for various potential confounders for breast cancer, with the exception of family history of breast cancer, which was not available. Control for other dietary and lifestyle factors besides established breast cancer risk factors did not change the estimates substantially; thus, it seems unlikely, that the results are affected by residual confounding from these variables. Yet, the study has limitations. It cannot be excluded that results may be biased by unmeasured confounders. In addition, measurement errors introduced by either the participants or the questionnaires may have affected the results. These potential misclassifications are probably nondifferential and may therefore only have biased results towards unity. Of the invited women only 37% agreed to participate in the study,37 which may limit generalization of the results reported in this study.

Despite a number of, and perhaps overlapping, plausible biological mechanisms4–10 supporting a protective role of whole grains towards breast cancer, we found no association between intake of whole grain products or individual whole grain products and risk of breast cancer in the present study. Nevertheless, in a previous nested case–control study within the present cohort higher plasma levels of enterolactone were associated with lower risk of breast cancer.38 It has been hypothesized that high levels of plasma enterolactone are related to a decreased risk of breast cancer, although the epidemiological evidence is still conflicting.10 Our null result on overall breast cancer is consistent with results from the only prospective cohort study published evaluating the whole grain–breast cancer association.13 In that study, it was although shown that women in the highest quintile of whole grain intake had a nonsignificant 20% higher risk of breast cancer compared with women who rarely had whole grains.13 Few case–control studies have evaluated the association between intake of whole grains and risk of breast cancer. Chatenoud et al.12 found an odds ratio (OR) of 0.9 for Italian pre- and postmenopausal women eating whole grain foods more than 3 days/week compared with those with no or rare consumption in a large case–control study of 3,412 cases and 4,770 controls. In the much smaller German case–control study with 310 cases and 353 controls, pre- and postmenopausal women combined, the intake of whole grain products was inversely associated with breast cancer risk. Women in the highest quartile of intake had an significant OR of 0.57 compared with women in the lowest quartile of intake.11

Use of HRT could interfere with the previously mentioned biological mechanisms of lignans/enterolactone and consequently the possible benefit of whole grain product intake. Normally, the endogenous level of oestrogens is low in postmenopausal women. A consequence of the use of HRT among postmenopausal women is considerable higher levels of oestrogens compared with women not using HRT. Under such circumstances, the exogenous oestrogens may override the potential preventive effect of whole grains, leading to a higher risk of breast cancer. In contrast, the potential risk-reducing effect of whole grains may be fully expressed in postmenopausal women not using HRT. The present study did, however, not support this hypothesis, but a previous prospective study has been suggestive of an increasing risk of breast cancer with increasing whole grain intake among current oestrogen users, but no association among nonusers and former users.13

Risk factors for breast cancer may be different for subtypes characterized by hormone receptor status or tumour histology; however, few data are available on dietary components and breast cancer subtypes.17–20, 25 In the present study, the association between whole grain products and risk of breast cancer did not differ according to either hormone receptor status or histology. Only one study has previously evaluated whether the effect of whole grains is directed towards breast cancers with a specific hormone receptor status. This study did, however, not find an association between intake of whole grains and risk of ER− breast cancer. The study did not present results for ER+, PR+ or PR− breast cancer.21

In conclusion, in this large prospective cohort study a higher intake of whole grain products was not associated with a lower risk of breast cancer or specific breast cancer subtypes defined by hormone receptor status or histology among postmenopausal women. Further, the whole grain product–breast cancer association did not differ according to HRT use. Because relatively few epidemiological studies on whole grains and breast cancer have been published, and no previous study has considered the importance of both ER and PR receptor status of the tumour and breast cancer histology, more large prospective studies are needed.

Acknowledgements

The authors thank Mrs. Katja Boll (data manager), Mrs. Connie Stripp and Mrs. Jytte Fogh Larsen (project coordinator) for assistance with the data collection and handling.

Ancillary