Vitamin D intake and breast cancer risk in postmenopausal women: the Iowa Women’s Health Study
- First Online:
- Cite this article as:
- Robien, K., Cutler, G.J. & Lazovich, D. Cancer Causes Control (2007) 18: 775. doi:10.1007/s10552-007-9020-x
- 490 Views
Vitamin D, a prosteroid hormone with anti-proliferative and pro-differentiation activity, is thought to act as a cancer chemopreventive agent. This study evaluated the association between vitamin D intake and breast cancer risk among women in a large prospective cohort study. A total of 34,321 postmenopausal women who had completed a questionnaire that included diet and supplement use were followed for breast cancer incidence from 1986 to 2004. Adjusted relative risks (RR) for breast cancer were calculated for dietary, supplemental, and total vitamin D intake among all women. The adjusted RR of breast cancer for women consuming >800 IU/day versus <400 IU/day total vitamin D was 0.89 (95% CI: 0.77–1.03). RRs were stronger among women with negative than positive ER or PR status. The association of high vitamin D intake with breast cancer was strongest in the first 5 years after baseline dietary assessment (RR = 0.66; 95% CI: 0.46–0.94 compared with lowest-intake group), and diminished over time. Changes in vitamin D intake over time might have contributed to the diminished association observed in later years. Vitamin D intake of >800 IU/day appears to be associated with a small decrease in risk of breast cancer among postmenopausal women. Studies evaluating all sources of vitamin D, especially sun exposure, are needed to fully understand the association between vitamin D and breast cancer risk.
KeywordsVitamin DBreast cancerPostmenopausal women
Breast cancer is the most common cancer among women. In 2006, approximately 213,000 women will be diagnosed and 41,000 women will die from the disease in the US . Vitamin D is a prosteroid hormone with anti-proliferative and pro-differentiation activity , which has led to interest in its possible role in cancer chemoprevention . Cancer incidence and mortality have been positively correlated with latitude by geographic regions in several ecologic studies [4–8], and is thought to be due to diminishing vitamin D synthesis from UV exposure at increasing latitudes.
Vitamin D is produced endogenously in the skin upon exposure to UV radiation (specifically UVB), and is available from dietary intake of animal products (especially fatty fish), fortified foods (especially milk and breakfast cereals), and dietary supplements. Discussion regarding the adequacy of current recommendations for 200–600 IU/day of dietary and/or supplemental vitamin D intake in adults [9–11], especially in light of recommendations to avoid sun exposure and for regular use of sunscreens [12, 13], are contributing to the need for a better understanding of the biological roles of vitamin D in general health and in the development of diseases such as cancer.
To date, three cohort studies have reported modest reductions in breast cancer risk among women with higher levels of dietary or supplemental vitamin D intake [14–16]. The results of case–control studies of vitamin D and breast cancer risk are inconsistent, with some studies reporting an increased risk of breast cancer with increasing vitamin D intake [17, 18], while others report no association between vitamin D intake and breast cancer risk [19, 20]. These inconsistencies may be the result of different methods for selecting cases and controls, and differing dietary intake data collection tools and referent time periods, or may simply indicate no real association.
The goal of this study was to evaluate the association between dietary and supplemental vitamin D intake in relation to currently recommended intake levels and risk of breast cancer, specifically among postmenopausal women. We hypothesized that higher levels of vitamin D intake would be associated with lower risk of breast cancer. We were also able to assess changes in dietary and supplemental vitamin D over the follow up period, and consider how these changes might affect the observed associations between vitamin D intake and breast cancer risk.
Study design and population
The Iowa Women’s Health Study (IWHS) is a prospective cohort study focused on examining the relation of dietary exposures and lifestyle factors with cancer occurrence. In 1986, a mailed questionnaire was sent to 99,826 women between the ages of 55 and 69 randomly drawn from the 1985 Iowa Department of Transportation list of licensed drivers. A total of 41,836 women responded to the baseline questionnaire (42% response rate). Subjects have been followed through 5 follow-up questionnaires (1987, 1989, 1992, 1997, 2004), and through annual determination of vital status by linkage to Iowa death records and the National Death Index. For this study, we excluded women who reported any of the following at baseline: any cancer other than non-melanoma skin cancer (n = 3,830), premenopausal status (n = 569), or a mastectomy prior to baseline (n = 1,884). Women who did not report breast cancer at baseline, but who were determined through linkage to the State Health Registry of Iowa to have had a diagnosis of breast cancer before baseline were also excluded (n = 72). Finally, we excluded women who left 30 or more food items blank on the food frequency questionnaire, or who had total energy intake less than 600 kcal/day or more than 5,000 kcal/day (n = 2,712). Exclusion categories were not mutually exclusive. A total of 34,321 women remained in the analysis.
The baseline questionnaire collected information on education, physical activity, individual medical history and family cancer history, anthropometric variables, reproductive health, and smoking history. Subjects also completed a 127-item food frequency questionnaire (FFQ) developed by Willett and colleagues [21, 22]. Participants were asked to report their average frequency of intake over the past year. A common unit or portion size was specified for each food, and responses ranged from “never or less than once per month” to “six or more times per day.” The questionnaire also assessed the use of multivitamins, along with the use of 18 single vitamin and mineral supplements, including vitamin D. Information on current use (yes/no), frequency, and brand of multivitamin supplements was collected with both the 1986 and 2004 questionnaires. If a respondent answered “yes” to taking an individual vitamin or mineral supplement on the 1986 questionnaire, they were then asked to provide a typical daily dose. The 2004 questionnaire did not collect dosage information from respondents; individual vitamin D supplements were assumed to provide 400 IU. Data on duration of supplement use was not collected at either time point. Supplemental vitamin D intake was calculated from reported intake of both multivitamin and individual vitamin D supplement. Daily intakes of nutrients were calculated using the Harvard Nutrient Database. Use of this FFQ has been shown to be reasonably reliable for assessing dietary intake in the IWHS cohort . The Pearson’s correlation coefficient of vitamin D intake from the FFQ with vitamin D intake from five 24-h dietary recalls was 0.51 . The FFQ was repeated in the 2004 follow-up questionnaire.
Data on body measurements were self-reported. Body mass index was used as a measure of relative weight, and was calculated as weight (kg) divided by the square of height (m2). History of screening mammography was assessed as part of the 1989 follow-up questionnaire. Physical activity level was categorized into three levels based on responses to questions assessing moderate and vigorous activity . Vigorous activity two or more times per week or moderate activity more than four times per week was considered “high” activity, vigorous activity once a week or moderate activity 1–4 times per week was considered “medium” activity, and vigorous or moderate activity a few times per month or less was considered “low” activity.
Incident cases of breast cancer were identified between 1986 and 2004 through linkage to the State Health Registry of Iowa, part of the National Cancer Institute’s Surveillance, Epidemiology and End Results program (SEER). Person-years of follow-up were calculated from baseline until the first of the following five outcomes: date of breast cancer diagnosis, date of death (if death occurred in Iowa), midpoint of the interval between the date of last contact and date of death (if death occurred outside of Iowa), date of emigration from Iowa (if known), midpoint of the interval between the last follow-up contact and 31 December, 2004 (if date of emigration from Iowa not known). It was assumed that all other women were alive in Iowa, and contributed follow-up time until 31 December, 2004.
Daily vitamin D intake was examined in three ways: intake from diet, intake from supplements, and intake from diet and supplements combined (total intake). Women were grouped into categories according to the daily vitamin D Adequate Intake (AI) recommendations. The current AI recommendation for daily vitamin D intake is 400 IU for adults aged 50–70. Women were grouped into three categories according to their daily vitamin D intake: did not meet AI recommendation (<400 IU/day), met AI recommendation (400–799 IU/day), or had an intake that was twice the current AI recommendation (≥800 IU/day). To examine vitamin D from supplements, a fourth category was included to account for the large number of women with no supplemental intake of vitamin D (no supplemental vitamin D, up to 400 IU supplemental vitamin D/day, 400–799 IU supplemental vitamin D/day, ≥800 IU supplemental vitamin D/day). Analysis of variance or chi-square tests were conducted to test differences in means or proportions of covariates by category of dietary, supplemental, or total vitamin D intake. Cox proportional hazards regression was used to derive age-adjusted and multivariate-adjusted relative risks for breast cancer among categories of dietary, supplemental, and total vitamin D intake compared to the lowest intake groups.
Potential confounding factors were included in the model if they had biological relevance, had been previously shown to be a risk factor for breast cancer, or if they notably altered the association between vitamin D intake and breast cancer risk. The final model included age (at baseline, in years), smoking status (current, former, never), age at menarche (in categories: <12, 12, 13, 14, >14), age at reported menopause (<45, 45–52, >52 years), first degree relative with breast cancer (yes/no), estrogen use (ever/never), age at first live birth (<20, 20–24, 25–29, ≥30 years), number of live births (0, 1–2, 3–4, >4), education category (<high school, high school, >high school), BMI category (<25, 25–29.9, ≥30 kg/m2), activity level (low, medium, high), live on a farm (yes/no), mammogram history (ever prior to 1989, yes/no), and daily energy (kcal/day), fat (g/day) and alcohol (g/day) intake. For each categorical variable in the final model, an additional category for missing values was included. Potential effect modification was examined by conducting analyses stratified by BMI category (<25, ≥25–<30, ≥30) and by variables that could serve as surrogates for sun exposure, such as activity level (low, medium, high), and farm residence (yes/no). We also investigated the association between vitamin D intake and specific breast cancer subgroups including cancer stage (in situ, localized, distant/regional/unknown), and ER or PR hormone receptor status (positive/negative).
The association of vitamin D intake with breast cancer risk over time was assessed by stratification on years of follow-up time from the baseline 1986 questionnaire. Each 5-year stratum included the breast cancer cases diagnosed during that 5-year interval and all study participants without a breast cancer diagnosis. To assess changes in total, dietary and supplemental vitamin D intake over the follow-up period, we compared intake data for study participants who completed both the 1986 and 2004 FFQ.
Baseline characteristics of the study population by total vitamin D intake, Iowa Women’s Health Study, 1986*
Total vitamin D
<400 IU/day (n = 20,024)
400–799 IU/day (n = 10,860)
≥800 IU/day (n = 3,437)
Age at menarche
Age at menopause
Age at first birth
Number of live births (%)
Activity level (%)
Smoking status (%)
Ever had a mammograma (%)
Benign breast disease (%)
First-degree relative with breast cancer—yes (%)
Estrogen use—yes (%)
Breast cancer cases
ER status (%)
PR status (%)
In this cohort, 42% of participants reported total daily vitamin D intake levels that met or exceeded the current recommended adequate intake levels of 400 IU/day for women aged 50–70 years . The majority of women (64%) reported that they did not use supplements containing vitamin D. Although only 36% of women reported vitamin D intake from supplements, almost all of the women in the highest group of total vitamin D intake used supplements (97.5%).
The correlation between total or supplemental vitamin D and energy intake was low (r = 0.23 and 0.02, respectively), while dietary vitamin D and energy intake was more closely correlated (r = 0.52). Dietary vitamin D intake was also correlated with total fat intake (r = 0.44). Total and dietary vitamin D intake was highly correlated with total calcium intake (r = 0.55 and 0.58, respectively), while supplemental vitamin D was not as highly correlated with total calcium intake (r = 0.30). Due to the differing correlations between dietary and supplemental vitamin D and energy intake, we chose to adjust for energy intake by including energy intake as a covariate in the multivariate models, rather than using the residual method .
Relative risks of breast cancer by category of daily vitamin D intake at baseline, Iowa Women’s Health Study 1986–2004
Daily vitamin D intake category (IU/day)
Mean vitamin D (IU)
Range vitamin D (IU)
Age adjusted (95% CI)
Full model * (95% CI)
No supplement use
Relative risk of breast cancer by total vitamin D intake stratified by stage at diagnosis and ER/PR status, Iowa Women’s Health Study 1986–2004*
Total vitamin D (IU/d)
RR (95% CI)
RR (95% CI)
RR (95% CI)
RR (95% CI)
RR (95% CI)
RR (95% CI)
Relative risk of breast cancer by total vitamin D intake stratified by years of follow-up before diagnosis*
Total vitamin D (IU/d)
RR (95% CI)
Years of follow-up 0–5
Years of follow-up 5–10
Years of follow-up 10–15
Years of follow-up 15+
A total of 17,794 women provided vitamin D intake data for both the baseline and 2004 FFQ. Mean dietary vitamin D intake decreased slightly (250.6 IU in 1986 versus 240.4 IU in 2004), however, mean total vitamin D intake increased by 119.5 IU/day (410.8 in 1986 versus 530.4 in 2004), indicating that use of supplements containing vitamin D increased over time and with age between the two dietary assessments.
We observed a modestly lower risk of breast cancer among postmenopausal women reporting daily total vitamin D intake of at least 800 IU compared to those reporting less than the current recommended adequate intake of 400 IU/day. The three previous reports from other large cohort studies also indicate that vitamin D intake may be inversely associated with breast cancer risk [14–16]. A report from the Nurses’ Health Study  observed decreasing incidence of breast cancer with increased intake of total or dietary vitamin D, but only among premenopausal women whose total vitamin D intake was more than 500 IU/day compared to those with intakes at 150 IU or less (RR = 0.72, 95% CI: 0.55–0.94). Although limited by a relatively small number of breast cancer cases, the NHANES I Epidemiologic Follow-Up Study (1971–1975 to 1992) reported a non-significantly lower breast cancer risk among both pre- and postmenopausal women consuming greater than 200 IU dietary vitamin D/day or daily use of supplements containing vitamin D compared to women consuming less than 100 IU of dietary vitamin D per day without supplements (RR = 0.86, 95% CI: 0.61–1.20) . Similarly, in the Cancer Prevention Study II (CPS II) Nutrition Cohort of postmenopausal women , a non-significant weak inverse association between dietary vitamin D and breast cancer was observed for postmenopausal women consuming more than 300 IU/day compared to women with intakes less than or equal to 100 IU/day (RR = 0.89, 95% CI: 0.76–1.03). However, for diet plus supplemental vitamin D, no association was found with increasing intake. While we observed inverse associations with breast cancer risk for both dietary and supplemental vitamin D intake, our estimates related to dietary intake alone were based on very small numbers of women in the highest intake category.
In our study, a statistically non-significant inverse association was observed for the highest versus lowest total vitamin D intake groups among women with ER− or PR− tumors. These findings are contrary to those of the CPS II Nutrition Cohort, where the vitamin D associations, in particular, dietary vitamin D, were somewhat stronger for women diagnosed with ER+ tumors. Despite having a smaller study population (n = 34,321 vs. 68,567 for CPS II), our study cohort had a larger number of ER− breast cancers (n = 274 vs. 227) than was reported for the CPS II cohort.
We also observed a statistically significantly lower risk of in situ breast cancer among those in the highest category of vitamin D intake compared to the lowest category, but no difference in risk of localized or distant/regional breast cancer by vitamin D intake category. This finding is possibly indicative of general healthy behaviors, rather than a specific effect of vitamin D exposure. Since our analysis included adjustment for breast cancer screening, physical activity, and alcohol intake, this finding may be related to another unidentified/unmeasured health behavior, or may simply be a chance finding.
We identified a decrease in the strength of the association between vitamin D and breast cancer risk over time, as well as an increase in supplement use containing vitamin D during the follow-up periods. Exposure misclassification of vitamin D intake in the latter time periods by using baseline vitamin D as the exposure may have contributed, in part, to the diminished association with time.
Vitamin D is unusual as a nutrient in that the primary naturally occurring source of the nutrient is UV exposure rather than dietary intake. Only one of the three previous cohort studies of dietary and supplemental vitamin D in relation to breast cancer risk, the NHANES I Epidemiologic Follow-Up Study report, included self-reported sun exposure in the analysis . A 30% lower breast cancer incidence was observed among women with the highest compared to the lowest quartile of total vitamin D exposure (UV and diet). While our study did not assess participants’ sun exposure, physical activity level and farm residence were considered as potential surrogates for sun exposure. We did not observe significant differences in the associations between vitamin D intake and breast cancer risk upon stratification by these measures, however, it is likely that these measures do not adequately capture sun exposure information. Similarly, sun exposure history was not assessed in the CPS II Nutrition Cohort study, however the inverse association observed between vitamin D and postmenopausal breast cancer risk was reported to be somewhat stronger for women residing in states with lower values of the UV index . Adjusting for area of residence did not affect the associations of total and dietary vitamin D with premenopausal breast cancer risk in the Nurses Health Study .
Studies using circulating vitamin D levels (as a biomarker of both dietary and sun exposure sources of vitamin D) to evaluate the association with breast cancer risk have so far been inconclusive. No differences in serum 1,25-dihydroxyvitamin D (1,25(OH)2D) levels, the metabolically active form of vitamin D, were reported between cases and controls in a nested case–control study , while another case–control study found a decreased risk of breast cancer among white women, but not black women, with higher levels of plasma 1,25(OH)2D . In a nested case–control study from the Nurses Health Study, the relationship between plasma 1,25(OH)2D and breast cancer risk was inconsistent . These last two studies reported conflicting findings for 25-hydroxyvitamin D (the primary circulating form of vitamin D and precursor to 1,25(OH)2D) and breast cancer risk, with a null finding in one  and an inverse relationship in the other .
In addition to the lack of data on sun exposure and circulating vitamin D levels, a limitation of this study, along with the Nurses Health Study and the CPS II Nutrition Cohort study, is that FFQs were used to assess dietary vitamin D intake, whereas the National Health and Nutrition Examination Study I dietary intake data was obtained using a single day 24-h recall. While the 24-h recalls are generally considered more accurate for the period covered than FFQ data , a single day of data does not accurately represent usual dietary intake. The databases used to calculate vitamin D intake from food and supplemental sources, including the database used for our study, do not distinguish between the cholecalciferol (D3) or ergocalciferol (D2) forms of vitamin D. Current research indicates that the cholecalciferol (D3) form of vitamin D is better utilized and has a longer shelf life than ergocalciferol (D2) , although both forms have been used in supplements. Thus, the inability to distinguish between the form of vitamin D used in supplements may have limited the accuracy and precision of these analyses, especially given that the vast majority of the women in the highest intake group in our study were supplement users.
A strength of our study included the study participants’ relatively high dietary and supplemental vitamin D intake, allowing assessment of the association of higher levels of vitamin D intake with breast cancer risk than had previously been possible. The mean dietary vitamin D intake for our study cohort was 251.5 IU/day at baseline (1986), which appears to be higher than the mean for the other three cohort studies, as well as being higher than population averages for dietary vitamin D among white women from NHANES data for approximately the same time period as our baseline questionnaire (173 IU/day, ). As with any epidemiologic study of diet, measurement error in assessing nutrient intake is a concern with our study, and our category cut-points of 400 and 800 IU of vitamin D are likely not precise.
Current adequate intake recommendations for vitamin D, as determined by the Food and Nutrition Board of the Institute of Medicine , were established to prevent deficiency diseases such as rickets and osteoporosis. However, recent scientific advancements in the understanding of the hormonal activities of vitamin D metabolites indicate that the definition of adequacy of vitamin D intake needs to be expanded to include levels sufficient to support vitamin D’s role in cancer chemoprevention and other associated activities. Further studies using more detailed vitamin D exposure measurement, especially sun exposure, are needed to clarify optimal dietary/supplemental vitamin D intake levels to reduce cancer risk, especially given current recommendations to limit UV sun exposure to decrease risk of skin cancers.
The authors thank Aaron R. Folsom, MD, MPH for providing the Iowa Women’s Health Study data, and Ching Ping Hong, MS, and Lisa Harnack, DrPH, RD for consultation and assistance in data preparation prior to analysis. Funding: Funding for the Iowa Women’s Health Study was provided by the National Cancer Institute (R01 CA039742).