Research ArticleClinical Studies

Serum 25-Hydroxyvitamin D and Prevention of Breast Cancer: Pooled Analysis

SHARIF B. MOHR, EDWARD D. GORHAM, JOHN E. ALCARAZ, CHRISTOPHER J. KANE, CAROLINE A. MACERA, J. KELLOGG PARSONS, DEBORAH L. WINGARD and CEDRIC F. GARLAND
Anticancer Research September 2011, 31 (9) 2939-2948;

This article has a correction. Please see:

Abstract

Background: Low serum levels of 25-hydroxyvitamin D [25(OH)D] have been associated with a high risk of breast cancer. Since publication of the most current meta-analysis of 25(OH)D and breast cancer risk, two new nested case–control studies have emerged. Materials and Methods: A PubMed search for all case–control studies on risk of breast cancer by 25(OH)D concentration identified 11 eligible studies. Data from all 11 studies were combined in order to calculate the pooled odds ratio of the highest vs. lowest quantile of 25(OH)D across all studies. Results: The overall Peto odds ratio summarizing the estimated risk in the highest compared to the lowest quantile across all 11 studies was 0.61 (95% confidence interval 0.47, 0.80). Conclusion: This study supports the hypothesis that higher serum 25(OH)D levels reduce the risk of breast cancer. According to the review of observational studies, a serum 25(OH)D level of 47 ng/ml was associated with a 50% lower risk of breast cancer.

Prevention of breast cancer remains one of the greatest public health challenges of our time. In 2010, there were 192,000 cases and 40,000 deaths in the United States from breast cancer, making it the most commonly occurring neoplasm in women, and the second most common cause of death from cancer in women (1). Globally, a wide range of ecological studies have linked low levels of sunlight or ultraviolet B (UVB) irradiance, the main source of circulating vitamin D in humans, with high breast cancer rates (2-7). In another study, women who were regularly exposed to sunlight or consumed above-average amounts of vitamin D were found to have significantly lower incidence rates of breast cancer (8).

Ultraviolet B is needed to make vitamin D, which is synthesized by the skin. Exposure to UVB and supplemental vitamin D intake increase serum 25-hydroxyvitamin D [25(OH)D] levels in a dose-dependent manner (9). Lack of exposure to UVB or everyday use of sunscreens may result in vitamin D deficiency unless there is adequate oral intake (10). A low serum level of 25(OH)D is the main marker of vitamin D deficiency, and has been linked to increased risk of several types of cancer, including cancer of the breast (11).

Normal breast epithelial cells have a vitamin D receptor that is highly sensitive to 1,25(OH)2D the most highly active vitamin D metabolite (12). Numerous laboratory studies have demonstrated the ability of 1,25(OH)2D to promote differentiation and apoptosis in breast cancer cell lines (13, 14).

In 2010, Yin and colleagues extracted data from ten case–control studies on the relationship between serum 25(OH)D levels and risk of breast cancer (15). According to the analysis performed in that study, higher concentrations of serum 25(OH)D were significantly inversely related to breast cancer risk across all studies. The purpose of the present study is to perform an up-to-date pooled analysis of data extracted from all case–control and nested case–control studies of breast cancer and serum 25(OH)D status performed to date.

Materials and Methods

A PUBMED search was conducted by two investigators for observational studies of serum 25(OH)D and risk of breast cancer performed between 1966-2010. The search was performed by using the terms “Vitamin D” or “cholecalciferol” or “calcidiol” or “calcitriol” or “25-hydroxyvitamin D”, and “case–control” or “epidemiology” and “human” as medical subject heading (MeSH) terms and words in the abstract, combined with the subject term “breast neoplasms”. Articles were included if they were published in medical journals, were ordinary case–control, nested case–control, or cohort studies of breast cancer, and included measures of association by quantiles of serum 25(OH)D. Fourteen studies reporting odds ratios for breast cancer by quantiles of serum 25(OH)D in association with breast cancer risk were identified (16-29). One study, a case-control study by Janowsky et al., which analyzed the association between 25(OH)D and breast cancer risk, did not report odds ratios but reported no effect (25).

Two studies that investigated the relationship between 1,25(OH)2D and breast cancer risk also were identified, including the above study by Janowsky et al., (25) and a case-control study by Hiatt et al. (30). Janowsky et al. found a strong inverse association with risk of breast cancer (25), while Hiatt et al. found no association (30).

The study published by Colston et al. (20) did not provide cell frequencies and utilized the same cases and controls that were used by Lowe and colleagues (26). The study by Green et al. (24) was left out of the analysis because it was a subset of the larger study by Bertone-Johnson et al. (18). Therefore, these studies (20, 24, 26) were excluded, leaving a total of 11 case–control or nested case–control studies on 25(OH)D status and breast cancer risk. Data from these eleven studies were independently extracted by two investigators, pooled, and divided into quintiles according to serum 25(OH)D.

Statistical analysis. A pooled odds ratio for all studies was obtained using Peto's Assumption-Free Method for combining odds ratios (31). This method provides a weighted average of the natural logarithms of the odds ratios from each study. The weights were the inverse of the variances of the logarithms of each odds ratio (32).

Sensitivity analysis was performed to assess the effect of inherent differences between case–control and nested case–control studies regarding the point in time of determination of serum 25(OH)D relative to diagnosis of breast cancer in the cases. This analysis was carried out by calculating the pooled odds ratios separately for nested case–control studies of pre-diagnostic serum and ordinary case–control studies. Pooled odds ratios were calculated using random effects models. Since the publication of an earlier meta-analysis by Yin and colleagues (15), two nested case–control studies have been published (22, 29), further necessitating new calculations of pooled odds ratios.

One major study reported finding a statistically significant inverse effect of 25(OH)D in individuals residing at latitudes >37 degrees N (27). Therefore, a further sensitivity analysis was performed in which the pooled odds ratio was calculated for all studies of populations residing at >37 degrees N, regardless of study design.

The p-value for the overall summary odds ratio was calculated using a z-score, where the numerator was the natural logarithm of the pooled odds ratio and the denominator was the standard error of the natural logarithm of the pooled odds ratio (31). This is the standard method for calculating the p-value using Peto's Assumption-Free Method (31, 33). Odds ratios comparing the highest with the lowest quantiles for each study were displayed in a forest plot (34, 35). Odds ratios (ORs) from the most highly adjusted models were chosen from each study. Confidence intervals (CIs) were computed using the method of Woolf (36, 37). The DerSimonian-Laird statistic was calculated to assess heterogeneity (38). The calculations were performed using Rev Man 5 (Oxford, UK: The Cochrane Collaboration).

Dose–response gradient. A data set was created consisting of one record per participant in each study. The records in this data set identified whether the participant was a case or noncase, the midpoint or median value of the participant's quantile of serum 25(OH)D at baseline, in ng/ml, a study identification number, and a serial number. If the median value for quantiles of 25(OH)D was provided in the study or contributed by the corresponding authors (16, 17, 21-23, 27, 28), it was used. If not (18-20, 24, 26), midpoint values were estimated by computing the arithmetic mean of the upper and lower bounds of the quantiles. If the lower limit of the lowest quantile was not available by inspection or correspondence, the midpoint was calculated using an assumption that the lower bound was zero. If the upper limit of the highest quantile was not available, the lower bound was used as the midpoint value. Data presented in nmol/l were converted to ng/ml using the conversion factor: 1 ng/ml=2.5 nmol/l. The records were put into order by serum 25(OH)D level, then divided into five quintiles using the following cut-off points: 0-10 ng/ml; 11-20 ng/ml; 21-30 ng/ml; 31-40 ng/ml; >40 ng/ml. Odds ratios were then calculated for the association between each quintile of serum 25(OH)D and risk of breast cancer (39). Since raw cell frequencies were used from each study, the odds ratios calculated for the dose–response analysis were unadjusted for potential confounders. The medians of the quintiles were: 10, 15, 25, 35, and 51 ng/ml. The lowest quintile was used as the reference group. Confidence intervals were calculated using the method of Woolf, as for the Forest plots (36, 37).

A dose–response curve was plotted showing the pooled odds ratios for each quintile of the pooled data (36, 37). A least-squares line was drawn to assess the dose–response relationship (40, 41). P-values for trend were calculated using the Mantel-Haenszel chi-square test (42). Serum 25(OH)D concentrations associated with a 50% reduction in breast cancer risk, compared to the lowest quintile of 25(OH)D, were obtained by drawing a vertical line from the point on the dose–response curve corresponding to an odds ratio of 0.5. Computations were performed using SAS, Version 9.2 (SAS Institute, Cary, NC, USA).

Results

To date, 12 epidemiological studies have been performed on the relationship between levels of serum 25(OH)D, the main circulating vitamin D metabolite, and risk of breast cancer (16-19, 21-24, 26-29). Of these, eight found a significant association between higher levels of serum 25(OH)D and a reduced risk of breast cancer (16-18, 21, 22, 24, 26, 28), while four failed to detect a significant association (19, 23, 27, 29), although a significant inverse association was present in the unadjusted data reported by one of these studies (19).

Eleven studies that analyzed risk of breast cancer by quantiles of serum 25(OH)D were identified through the PubMed search. Six were nested case–control studies and five were regular case–control studies. All of these studies were included in the pooled analysis. There was a downward linear gradient in risk of breast cancer with increasing serum 25(OH)D in the pooled analysis (Figure 1). Serum 25(OH)D levels accounted for 76% of the variation in breast cancer risk (R2=0.76, p for trend <0.001). The ORs for the pooled data from lowest to highest quintile were: 1.00, 0.78, 0.71, 0.66, and 0.67 (p for trend <0.001) (Table I). According to the pooled analysis, a serum 25(OH)D concentration of ≥47 ng/ml (110 nmol/l) was associated with a 50% lower risk of breast cancer, compared with <10 ng/ml (Figure 1). This would also correspond to an approximately 10% reduction in risk for every 10 ng/ml increase in 25(OH)D.

Figure 1.
Figure 1.

Pooled analysis of studies of serum 25(OH)D concentration and risk of breast cancer.

The overall Peto OR summarizing the estimated risk in the highest compared to the lowest quantile across all studies was 0.63 (95% CI=0.47-0.80) (Figure 2). In the Forest plot analysis, the 11 studies were heterogeneous (DerSimonian-Laird Chi-square=77.93, df=10, p<0.0001). In a sensitivity analysis, 25(OH)D was significantly, inversely related to breast cancer risk in both nested case–control and ordinary case–control studies. In the nested case–control studies, the overall pooled OR was 0.87 (95% CI=0.77–0.99) (Figure 2). The nested case–control studies were homogenous (DerSimonian-Laird chi-square=4.35, df=5, p=0.50). In the ordinary case–control studies, the overall pooled OR was 0.41 (95% CI=0.31-0.56) (Figure 2). According to the DerSimonian-Laird test, the case–control studies were heterogeneous (chi-square=14.85, df=4, p=0.005). The analyses were repeated using a fixed effects model, and the results were nearly identical.

In a sensitivity analysis of studies performed on individuals residing at a latitude of >37 degrees N, the pooled OR comparing the top with the bottom quintile of 25(OH)D was 0.56 (95% CI=0.50-0.62) (Figure 3). These studies were heterogeneous (Chi-square=58, df=7, p=0.001) (Figure 3). According to the results of a funnel plot analysis (not shown), there was no indication of publication bias.

Discussion

Since the publication of the meta-analysis by Yin et al. (15), two additional nested case–control studies have been reported that found a beneficial association between 25(OH)D levels and breast cancer risk. In the nested case–control study performed by Almquist et al. (29), subjects in the highest category of serum 25(OH)D (37.4 ng/ml) had a 7% lower risk of developing breast cancer (OR 0.93; 95% CI=0.66-1.33) compared to subjects in the lowest category of serum 25(OH)D (18.1 ng/ml) (29), a finding that was not statistically significant. The study performed by Engel and colleagues found that for women <53 years of age, subjects with serum 25(OH)D levels >27 ng/ml had a 40% lower risk of developing breast cancer (OR 0.60; 95% CI=0.37-0.96) than those with a 25(OH)D level <19.8 ng/ml (22). This finding persisted in spite of adjustment for physical activity, a variable that has been shown to have considerable influence on 25(OH)D status (43).

View this table:
Table I.

Case–control and nested case–control studies of serum 25-hydroxyvitamin D metabolites and risk of cancer of the breast, ICD-CM Code 174, according to PubMed search, 1966-2010.

In the present study higher levels of 25(OH)D were significantly inversely associated with breast cancer risk. When comparing the highest vs. lowest quintile of serum 25(OH)D across all studies, individuals in the highest category of 25(OH)D had an overall reduction in risk of breast cancer of approximately 39% compared to those in the lowest category of serum 25(OH)D (Figure 2). Data from all 11 studies provided an estimate of the dose–response relationship between serum 25(OH)D concentration and breast cancer risk. This revealed that a serum level of 44 ng/ml would cut risk of breast cancer by 50%, compared to a median level of 9.7 ng/ml (Figure 1).

Figure 2.
Figure 2.

Pooled odds ratios of breast cancer risk according to serum 25-hydroxyvitamin D, 1966-2010, comparing highest to lowest quintile.

When pooled ORs were calculated separately according to study design (case–control vs. nested case–control) higher serum 25(OH)D levels were associated with a lower risk of breast cancer in both ordinary case–control and nested case–control studies (Figure 1). These findings differed in magnitude, but not direction, from those of Yin et al. (15). However, the current pooled analysis benefited from the pooling of two additional nested case–control studies that had not been published at the time Yin and colleagues (15) performed their meta-analysis.

Pooled ORs were calculated using random effects models, which provide a more conservative estimate of the pooled OR. The random effects model incorporates inter-study variance into the estimate and is the most appropriate method when dealing with heterogeneous studies (31). The nested case–control design is superior to the standard case–control design, since it establishes a temporal sequence.

In a previous meta-analysis performed by Yin and colleagues, serum 25(OH)D was not significantly associated with breast cancer risk in nested case–control studies utilizing pre-diagnostic serum (15). However, 25(OH)D was significantly inversely associated with breast cancer risk in regular case–control studies in which serum 25(OH)D levels were measured shortly after diagnosis. The authors attributed the association in the regular case–control studies to 25(OH)D concentrations being diminished as a result of the disease process itself, or of changes in dietary and other lifestyle habits as a result of the disease, possibly creating a spurious association (15).

However, the majority of incident cases of breast cancer are discovered as a result of self-examination when a lump is found in the breast or armpit, or by mammography (44). Therefore, most incident cases lack the severe symptoms characteristic of a more advanced stage of the disease that might be likely to cause a drastic change in lifestyle habits. For example, in the study performed by Rejnmark et al., 25(OH)D was measured shortly before diagnosis, when changes in lifestyle habits that could influence 25(OH)D were unlikely to have occurred. Women in the Rejnmark et al. study in the highest tertile of serum 25(OH)D had half the risk as those in the lowest (28).

Figure 3.
Figure 3.

Pooled odds ratios of breast cancer risk according to serum 25-hydroxyvitamin D, 1966-2010, comparing highest to lowest quintile, studies based on populations residing at latitude >37 degrees N.

Serum 25(OH)D levels may not be greatly affected by a diagnosis of breast cancer unless the disease itself was somehow responsible for reducing serum 25(OH)D levels, a concept for which there is no known biological basis. One study performed by Goodwin and colleagues found that in women with breast cancer, mortality was significantly reduced in women in the highest tertile of 25(OH)D compared to the lowest (45). The Goodwin et al. analyses of serum 25(OH)D concentrations indicate that, from a biological standpoint, having breast cancer does not necessarily influence the serum 25(OH)D concentration. If the presence of breast cancer itself, rather than behavioral changes as a result of the diagnosis, influenced 25(OH)D concentration, it is highly unlikely that Goodwin et al. would have found an significant inverse association between 25(OH)D status and breast cancer mortality. Two additional studies found that serum 25(OH)D concentration did not markedly change in breast cancer patients after receiving chemotherapy (46, 47). Furthermore, Abbas et al. performed sensitivity analyses excluding cases who underwent chemotherapy before measurement of 25(OH)D or had their 25(OH)D measured greater than 6 months after the diagnosis, with no appreciable effect on the results (17, 27).

All of the ordinary case–control studies on 25(OH)D and breast cancer risk were performed in populations residing in a well-defined geographic area above a latitude of 37 degrees N. All of these studies found a beneficial association between serum 25(OH)D and breast cancer risk (Figure 3). One possible explanation is that measurement of serum 25(OH)D in populations residing at higher latitudes, may be a better indication of lifetime 25(OH)D status since migration from north to south is far more common in the US than migration from south to north (27). The case–control study performed by McCullough et al. was nested in the Cancer Prevention Study (CPS) II cohort, with volunteers from throughout the United States participating. In that study, individuals in the highest category of 25(OH)D had a 65% lower risk than those in the lowest category when the analysis was restricted to women residing at a latitude of >37 degrees N, a finding that was highly statistically significant (27).

In order to explore the effect of latitude, we performed a sub-analysis in which the pooled OR was calculated for studies performed at latitudes >37 degrees North, including the data from the McCullough et al. study (Figure 3). Eight studies were included in this analysis. The data from McCullough and colleagues' sub-analysis of latitude was included as a separate study. All eight studies found significantly lower breast cancer risk in women in the highest quantile of 25(OH)D compared to women in the lowest quantile. The overall pooled OR was 0.56 (95% CI=0.38, 0.65). Interestingly, the nested case–control studies performed in fixed populations residing at a latitude >37 degrees N, found a statistically significant inverse association between 25(OH)D levels and breast cancer risk (22, 27, 28).

The nested case–control study performed by Bertone-Johnson et al. found an inverse association between 25(OH)D and breast cancer risk, despite the study being carried out in a cohort of individuals who were spread throughout the country. One possible explanation for this is that the participants were nurses. Vitamin D insufficiency is common among healthcare professionals, even those living at very sunny latitudes (48, 49).

In the McCullough et al. study, an association of 25(OH)D with breast cancer risk was not present in women residing at latitude <37 degrees N. The investigators acknowledged that while 95% of those residing in the north were born in the north, 36% of the women living in the south were also born in the north (27). One possible reason that the other nested case–control study performed by Freedman et al. (23) did not find an association between 25(OH)D and breast cancer risk is that it was performed in cohorts where many people may have migrated from north to south. Therefore, the measurement of vitamin D status in these studies may not have been a good indicator of lifetime vitamin D status, or of vitamin D status during the relevant window of time during which vitamin D would exert its protective effect. This non-differential misclassification would most likely result in underestimation of the true relative risk (50).

The present study had several advantages. All known published case–control or nested case–control studies of serum 25(OH)D and risk of breast cancer were included, to the Authors' knowledge. An advantage of serum studies is that they are free from most of the uncertainties of collecting questionnaire data. An advantage of pooled analysis is that by combining the results of many studies, statistical power is increased, making it easier to detect an association between exposure and disease. In addition, the pooling of data from many studies in this analysis allowed for estimation and display of the dose–response gradient.

There were also some important limitations to the present study. One limitation was the inability to study menopausal status as a modifier of the relationship between serum vitamin D levels and risk of breast cancer. Previous studies have found that the relationship between serum 25(OH)D levels and breast cancer risk may be modified by menopausal status (16-19, 21, 22), however, this study was unable to investigate the effect of menopausal status on risk because not all studies provided cell frequencies according to 25(OH)D quantile and menopausal status.

It was also not possible to obtain data on serum 25(OH)D concentrations for each individual for the pooled analysis. Therefore, the median value for quantiles of 25(OH)D, where possible, or the mean of the upper and lower bounds of the quantiles, were used as the serum 25(OH)D value for each of the individual records in the pooled dataset. This may have resulted in a loss of precision for this variable, and possibly, the measures of association based on it.

In addition, the individual studies used different methods of measuring serum 25(OH)D, which may have introduced non-differential misclassification of exposure in both cases and controls. However, non-differential misclassification tends to obscure associations rather than strengthen them (50). Furthermore, although the Authors obtained the most highly adjusted ORs (such as adjusted for BMI, use of HRT, and age at menarche), confounders may have been measured and controlled for differently across studies. Finally, despite conducting a thorough literature search of the PubMed database, we cannot rule out the possibility that this analysis may have excluded a pertinent study.

High concentrations of serum 25(OH)D most likely prevent breast cancer through two key mechanisms. Firstly, 25(OH)D plays an important role in the up-regulation of E-cadherin (51), a glue-like substance that keeps cells bound tightly together, and in a well-differentiated state. Secondly, high serum levels of 25(OH)D provide a greater concentration to serve as substrate for synthesis of 1,25(OH)2D, the most biologically active vitamin D metabolite (52). 1,25(OH)2D is synthesized in a wide range of tissues, including the epithelial tissues of the breast (53).

In the estimation of the dose–response gradient, a serum 25(OH)D level of >47 ng/ml would be associated with 50% lower risk of breast cancer, compared to serum 25(OH)D <9.7 ng/ml (Figure 1). Classical dose–response curves for micronutrients are either linear (54) or have a predominantly linear middle segment (40, 41). This appears to be true for most functions of vitamin D (55, 56). More studies of effects at higher vitamin D intakes are needed.

According to an analysis of 30 studies reporting any adverse effect of high serum 25(OH)D in adults, no reproducible toxicity was reported below 100 ng/ml (57). The median minimum threshold for toxicity in all studies was nearly twice this value, at 197 ng/ml. Therefore, the projected serum 25(OH)D level of approximately 44 ng/ml that would be associated with 50% lower breast cancer risk would be below the threshold for minimal toxicity by a safety factor of 4-5. An upper level of 4,000 IU/day and a no observed adverse health effect level of 10,000 IU/day of vitamin D have been established by the Institute of Medicine (IOM) (58). The IOM also reported that no hypercalcemia from vitamin D intoxication has been described for vitamin D intakes <10,000 IU/day (59). Another review reported that no cases of toxicity were documented at doses <40,000 IU per day (55).

A vitamin D3 intake of 2,000 – 4,000 IU/day, and a target of 45 ng/ml of serum 25(OH)D, are the most practical estimates now available for decision-makers who must weigh the potential benefits of actions that could reduce incidence of breast cancer. The current recommended daily intake of the Institute of Medicine of 600-800 IU/day for mature adults (59) would increase median serum 25(OH)D by only 4-6 ng/ml (60), which would be insufficient to raise the median population serum 25(OH)D levels into the range for cancer prevention.

The results of the NHANES 2001-2004 survey revealed that the mean serum 25(OH)D value for the US population was 24 ng/ml (61). Therefore, an increase of vitamin D intake to 2,000-4,000 IU/day of vitamin D3 would boost serum 25(OH)D by approximately 14-28 ng/ml, raising the estimated median level in the population to 38-52 ng/ml (60). According to the findings of this pooled analysis, the optimal oral vitamin D intake would also be 4,000 IU/day since this would be the dose required to raise median serum 25(OH)D levels from 24 ng/ml to 52 ng/ml (60). A serum 25(OH)D concentration of 52 ng/ml is far below the concentration that would be associated with hypercalcemia or adverse health effects, which ranges in different sources from 195-300 ng/ml (55-57, 59, 62, 63). Increasing serum 25(OH)D to these levels would be most efficiently achieved by intake of vitamin D3 (cholecalciferol) rather than vitamin D2 (ergocalciferol) (64).

This pooled analysis strongly supports the theory that there is an inverse association between serum 25(OH)D and risk of breast cancer. Although confounding is possible, there are four lines of epidemiological evidence that would support the association being causal: the geographic gradient with latitude and solar UVB irradiance (2-7), observational studies linking deficient serum 25(OH)D levels with increased risk (65, 66), studies linking low oral intake of vitamin D with increased risk, and laboratory studies illuminating the mechanisms in vivo and in vitro (13, 14). Vitamin D receptor polymorphisms that interfere with vitamin D are also associated with increased risk of breast cancer, particularly in combination with low levels of serum 25(OH)D (26). It seems unlikely that a single confounder could account for all these lines of evidence.

This pooled analysis provides the most current epidemiological evidence to investigate the relationship between serum vitamin D levels and breast cancer risk. The results support the hypothesis that higher serum 25(OH)D concentrations reduce the risk of breast cancer. The relationship was present in ordinary case–control studies and to a statistically significant, but lesser degree in nested case–control studies. Researchers may wish to consider the effect of geography on serum 25(OH)D concentrations in future nested case–control studies of serum 25-hydroxy vitamin D levels and breast cancer risk.

Numerous laboratory studies and observational studies have demonstrated the presence of a protective effect of serum 25(OH)D against breast cancer risk. The findings overall are consistent and vary only in degree, not direction. Now is the time to translate the accumulated knowledge from decades of research into public health policy. Vitamin D supplementation with 4000 IU/day of vitamin D3 (cholecalciferol) as a breast cancer primary prevention strategy would be highly effective and safe (59).

Acknowledgements

This research was supported by a Congressional allocation to the Penn State Cancer Institute of the Milton S. Hershey Medical Center, Hershey, PA, through the Department of the Navy, Bureau of Medicine and Surgery, under Work Unit No. 60126. The views expressed in this report are those of the Authors and do not represent an official position of the Department of the Navy, Department of Defense, or the U.S. Government.

Footnotes

  • Note added in proof: While this study was in proof, another nested case–control study was released that studied the association between serum 25(OH)D and risk of breast cancer. It found no association. This finding would not substantially affect the results that were based on the 11 studies that were included here, except possibly for a slight flattening of the dose–response curve. The reference to this study is: Eliassen AH, Spiegelman D, Hollis BW, Horst RL, Willett WC, Hankinson SE: Plasma 25-hydroxyvitamin D and risk of breast cancer in the Nurses' Health Study II. Breast Cancer Res 13: R50, 2011.

  • Received April 5, 2011.
  • Revision received May 31, 2011.
  • Accepted June 2, 2011.

References

  1. ↵
    1. Jemal A,
    2. Siegel R,
    3. Xu J,
    4. Ward E
    : Cancer statistics, 2010. CA Cancer J Clin 60: 277-300, 2010.
    OpenUrlCrossRefPubMed
  2. ↵
    1. Gorham E,
    2. Garland C,
    3. Garland F
    : Acid haze air pollution and breast and colon cancer in 20 Canadian cities. Can J Publ Health 80: 96-100, 1989.
    OpenUrl
    1. Garland F,
    2. Garland C,
    3. Gorham E,
    4. Young J Jr..
    : Geographic variation in breast cancer mortality in the United States: a hypothesis involving exposure to solar radiation. Prev Med 19: 614-622, 1990.
    OpenUrlCrossRefPubMed
    1. Gorham E,
    2. Garland F,
    3. Garland C
    : Sunlight and breast cancer incidence in the USSR. Int J Epidemiol 19: 820-824, 1990.
    OpenUrlAbstract/FREE Full Text
    1. Garland C,
    2. Garland F,
    3. Gorham E
    : Sunlight, sulfur dioxide and breast and colon cancer in Italy (abstract). Annu Meet Amer Assoc Advance Science, 1992.
    1. Grant WB
    : An ecologic study of dietary and solar ultraviolet-B links to breast carcinoma mortality rates. Cancer 94: 272-281, 2002.
    OpenUrlCrossRefPubMed
  7. ↵
    1. Freedman D,
    2. Dosemeci M,
    3. McGlynn K
    : Sunlight and mortality from breast, ovarian, colon, prostate, and non-melanoma skin cancer: a composite death certificate based case-control study. Occup Environ Med 59: 257-262, 2002.
    OpenUrlAbstract/FREE Full Text
  8. ↵
    1. John E,
    2. Schwartz G,
    3. Dreon D,
    4. Koo J
    : Vitamin D and breast cancer risk: The NHANES I epidemiologic follow-up study, 1971-1975 to 1992. Cancer Epidemiol Biomark Prev 8: 399-406, 1999.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Looker AC,
    2. Dawson-Hughes B,
    3. Calvo MS,
    4. Gunter EW,
    5. Sahyoun NR
    : Serum 25-hydroxyvitamin D status of adolescents and adults in two seasonal subpopulations from NHANES III. Bone 30: 771-777, 2002.
    OpenUrlCrossRefPubMed
  10. ↵
    1. Matsuoka L,
    2. Wortsman J,
    3. Holick M
    : Chronic sunscreen use decreases the concentration of 25-hydroxyvitamin D: a preliminary study. Arch Dermatol 124: 1802-1804, 1988.
    OpenUrlCrossRefPubMed
  11. ↵
    1. Garland C,
    2. Garland F,
    3. Gorham E,
    4. Lipkin M,
    5. Newmark H,
    6. Mohr S,
    7. Holick M
    : The role of vitamin D in cancer prevention. Am J Public Health 96: 252-261, 2006.
    OpenUrlCrossRefPubMed
  12. ↵
    1. Eisman J,
    2. Martin T,
    3. MacIntyre I,
    4. Mosely J
    : 1,25-dihydroxyvitamin D3 receptor in breast cancer cells. Lancet 2: 1335-1336, 1979.
    OpenUrlCrossRefPubMed
  13. ↵
    1. Pendas-Franco N,
    2. Gonzalez-Sancho JM,
    3. Suarez Y,
    4. Aguilera O,
    5. Steinmeyer A,
    6. Gamallo C,
    7. Berciano MT,
    8. Lafarga M,
    9. Munoz A
    : Vitamin D regulates the phenotype of human breast cancer cells. Differentiation 75: 193-207, 2007.
    OpenUrlCrossRefPubMed
  14. ↵
    1. Wang Q,
    2. Lee D,
    3. Sysounthone V,
    4. Chandraratna RAS,
    5. Christakos S,
    6. Korah R,
    7. Wieder R
    : 1,25-Dihydroxyvitamin D3 and retinoic acid analogues induce differentiation in breast cancer cells with function- and cell-specific additive effects. Breast Cancer Res Treat 67: 157-168, 2001.
    OpenUrlCrossRefPubMed
  15. ↵
    1. Yin L,
    2. Grandi N,
    3. Raum E,
    4. Haug U,
    5. Arndt V,
    6. Brenner H
    : Meta-analysis: serum vitamin D and breast cancer risk. Eur J Cancer 46: 2196-2205, 2010.
    OpenUrlCrossRefPubMed
  16. ↵
    1. Abbas S,
    2. Chang-Claude J,
    3. Linseisen J
    : Plasma 25-hydroxyvitamin D and premenopausal breast cancer risk in a German case-control study. Int J Cancer 124: 250-255, 2009.
    OpenUrlCrossRefPubMed
  17. ↵
    1. Abbas S,
    2. Linseisen J,
    3. Slanger T,
    4. Kropp S,
    5. Mutschelknauss EJ,
    6. Flesch-Janys D,
    7. Chang-Claude J
    : Serum 25-hydroxyvitamin D and risk of post-menopausal breast cancer – results of a large case-control study. Carcinogenesis 29: 93-99, 2008.
    OpenUrlAbstract/FREE Full Text
  18. ↵
    1. Bertone-Johnson ER,
    2. Chen WY,
    3. Holick MF,
    4. Hollis BW,
    5. Colditz GA,
    6. Willett WC,
    7. Hankinson SE
    : Plasma 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D and risk of breast cancer. Cancer Epidemiol Biomarkers Prev 14: 1991-1997, 2005.
    OpenUrlAbstract/FREE Full Text
  19. ↵
    1. Chlebowski R,
    2. Johnson K,
    3. Kooperberg C,
    4. Pettinger M,
    5. Wactawski-Wende J,
    6. Rohan T,
    7. Rossouw J,
    8. Lane D,
    9. O'Sullivan M,
    10. Yasmeen S,
    11. Hiatt R,
    12. Shikany J,
    13. Vitolins M,
    14. Khandekar J,
    15. Hubbell F,
    16. Women's Health Initiative Investigators
    : Calcium plus vitamin D supplementation and the risk of breast cancer. J Natl Cancer Inst 100: 1581-1591, 2008.
    OpenUrlAbstract/FREE Full Text
  20. ↵
    1. Colston KW,
    2. Lowe LC,
    3. Mansi JL,
    4. Campbell MJ
    : Vitamin D status and breast cancer risk. Anticancer Res 26: 2573-2580, 2006.
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. Crew KD,
    2. Gammon MD,
    3. Steck SE,
    4. Hershman DL,
    5. Cremers S,
    6. Dworakowski E,
    7. Shane E,
    8. Terry MB,
    9. Desai M,
    10. Teitelbaum SL,
    11. Neugut AI,
    12. Santella RM
    : Association between plasma 25-hydroxyvitamin D and breast cancer risk. Cancer Prev Res (Phila) 2: 598-604, 2009.
    OpenUrlAbstract/FREE Full Text
  22. ↵
    1. Engel P,
    2. Fagherazzi G,
    3. Boutten A,
    4. Dupre T,
    5. Mesrine S,
    6. Boutron-Ruault MC,
    7. Clavel-Chapelon F
    : Serum 25(OH) vitamin D and risk of breast cancer: a nested case–control study from the French E3N cohort. Cancer Epidemiol Biomarkers Prev 19: 2341-2350, 2010.
    OpenUrlAbstract/FREE Full Text
  23. ↵
    1. Freedman DM,
    2. Chang SC,
    3. Falk RT,
    4. Purdue MP,
    5. Huang WY,
    6. McCarty CA,
    7. Hollis BW,
    8. Graubard BI,
    9. Berg CD,
    10. Ziegler RG
    : Serum levels of vitamin D metabolites and breast cancer risk in the prostate, lung, colorectal, and ovarian cancer screening trial. Cancer Epidemiol Biomarkers Prev 17: 889-894, 2008.
    OpenUrlAbstract/FREE Full Text
  24. ↵
    1. Green AK,
    2. Hankinson SE,
    3. Bertone-Johnson ER,
    4. Tamimi RM
    : Mammographic density, plasma vitamin D levels and risk of breast cancer in postmenopausal women. Int J Cancer 127: 667-674,
  25. ↵
    1. Janowsky EC,
    2. Lester GE,
    3. Weinberg CR,
    4. Millikan RC,
    5. Schildkraut JM,
    6. Garrett PA,
    7. Hulka BS
    : Association between low levels of 1,25-dihydroxyvitamin D and breast cancer risk. Public Health Nutr 2: 283-291, 1999.
    OpenUrlPubMed
  26. ↵
    1. Lowe LC,
    2. Guy M,
    3. Mansi JL,
    4. Peckitt C,
    5. Bliss J,
    6. Wilson RG,
    7. Colston KW
    : Plasma 25-hydroxy vitamin D concentrations, vitamin D receptor genotype and breast cancer risk in a UK Caucasian population. Eur J Cancer 41: 1164-1169, 2005.
    OpenUrlCrossRefPubMed
  27. ↵
    1. McCullough ML,
    2. Stevens VL,
    3. Patel R,
    4. Jacobs EJ,
    5. Bain EB,
    6. Horst RL,
    7. Gapstur SM,
    8. Thun MJ,
    9. Calle EE
    : Serum 25-hydroxyvitamin D concentrations and postmenopausal breast cancer risk: a nested case–control study in the Cancer Prevention Study-II Nutrition Cohort. Breast Cancer Res 11: R64, 2009.
    OpenUrlCrossRefPubMed
  28. ↵
    1. Rejnmark L,
    2. Tietze A,
    3. Vestergaard P,
    4. Buhl L,
    5. Lehbrink M,
    6. Heickendorff L,
    7. Mosekilde L
    : Reduced prediagnostic 25-hydroxyvitamin D levels in women with breast cancer: a nested case–control study. Cancer Epidemiol Biomarkers Prev 18: 2655-2660, 2009.
    OpenUrlAbstract/FREE Full Text
  29. ↵
    1. Almquist M,
    2. Bondeson AG,
    3. Bondeson L,
    4. Malm J,
    5. Manjer J
    : Serum levels of vitamin D, PTH and calcium and breast cancer risk-a prospective nested case–control study. Int J Cancer 127: 2159-2168, 2010.
    OpenUrlCrossRefPubMed
  30. ↵
    1. Hiatt R,
    2. Krieger N,
    3. Lobaugh B,
    4. Drezner M,
    5. Vogelman J,
    6. Orentreich N
    : Prediagnostic serum vitamin D and breast cancer. J Natl Cancer Inst 90: 461-463, 1998.
    OpenUrlFREE Full Text
  31. ↵
    1. Egger M,
    2. Davey Smith G,
    3. Altman D
    1. Deeks J,
    2. Altman D,
    3. Bradburn M
    : Statistical methods for examining heterogeneity and combining results from several studies in a meta-analysis. In: Systematic Reviews and Health Care: Meta-Analysis in Context. Egger M, Davey Smith G, Altman D (eds.). London: BMJ Publications pp. 285-312, 2002.
  32. ↵
    1. Breslow N,
    2. Day N
    : Statistical Methods in Cancer Research, Vol 1–The Analysis of Case-Control Studies. Lyon: IARC Scientific Publication 1, 1980.
  33. ↵
    1. Deeks J,
    2. Cochrane Collaboration Statistical Methods Working Group
    : Statistical Methods Programmed in MetaView. MetaView Manual, Version 4, 1999.
  34. ↵
    1. Lewis S,
    2. Clarke M
    : Forest plots: trying to see the wood and the trees. BMJ 322: 1479-1480, 2001.
    OpenUrlFREE Full Text
  35. ↵
    1. Yeh J,
    2. D'Amico F
    : Forest plots: data summaries at a glance. J Fam Pract 53: 1007, 2004.
    OpenUrlPubMed
  36. ↵
    1. Woolf B
    : On estimating the relation between blood group and disease. Ann Hum Genet 19: 251-253, 1955.
    OpenUrlCrossRefPubMed
  37. ↵
    1. Fleiss J
    : Statistical Methods for Rates and Proportions, second ed. New York: Wiley, 1981.
  38. ↵
    1. DerSimonian R,
    2. Laird N
    : Meta-analysis in clinical trials. Control Clin Trials 7: 177-188, 1986.
    OpenUrlCrossRefPubMed
  39. ↵
    1. Lilienfeld D,
    2. Stolley P
    : Foundations of Epidemiology, third ed. New York: Oxford, 1994.
  40. ↵
    1. Carpenter J
    : A method for presenting and comparing dose–response curves. J Pharmacol Methods 15: 283-303, 1986.
    OpenUrlCrossRefPubMed
  41. ↵
    1. Steenland K,
    2. Deddens J
    : A practical guide to dose-response analyses and risk assessment in occupational epidemiology. Epidemiology 15: 63-70, 2004.
    OpenUrlCrossRefPubMed
  42. ↵
    1. Mantel N
    : Chi-square tests with one degree of freedom; extensions of the Mantel-Haenszel procedure. J Am Stat Assoc 58: 690-700, 1963.
    OpenUrlCrossRef
  43. ↵
    1. Brock K,
    2. Huang WY,
    3. Fraser DR,
    4. Ke L,
    5. Tseng M,
    6. Stolzenberg-Solomon R,
    7. Peters U,
    8. Ahn J,
    9. Purdue M,
    10. Mason RS,
    11. McCarty C,
    12. Ziegler RG,
    13. Graubard B
    : Low vitamin D status is associated with physical inactivity, obesity and low vitamin D intake in a large US sample of healthy middle-aged men and women. J Steroid Biochem Mol Biol 121: 462-466, 2010.
    OpenUrlCrossRefPubMed
  44. ↵
    1. Baines C,
    2. Miller A
    : The role of clinical breast examination and breast self-examination. Prev Med, 2011.
  45. ↵
    1. Goodwin PJ,
    2. Ennis M,
    3. Pritchard KI,
    4. Koo J,
    5. Hood N
    : Prognostic effects of 25-hydroxyvitamin D levels in early breast cancer. J Clin Oncol 27: 3757-3763, 2009.
    OpenUrlAbstract/FREE Full Text
  46. ↵
    1. Kailajarvi ME,
    2. Salminen EK,
    3. Paija OM,
    4. Virtanent AM,
    5. Leino AE,
    6. Irjala KA
    : Serum bone markers in breast cancer patients during 5-fluorouracil, epirubicin and cyclophosphamide (FEC) therapy. Anticancer Res 24: 1271-1274, 2004.
    OpenUrlAbstract/FREE Full Text
  47. ↵
    1. Gao Y,
    2. Shimizu M,
    3. Yamada S,
    4. Ozaki Y,
    5. Aso T
    : The effects of chemotherapy including cisplatin on vitamin D metabolism. Endocr J 40: 737-742, 1993.
    OpenUrlCrossRefPubMed
  48. ↵
    1. Multani SK,
    2. Sarathi V,
    3. Shivane V,
    4. Bandgar TR,
    5. Menon PS,
    6. Shah NS
    : Study of bone mineral density in resident doctors working at a teaching hospital. J Postgrad Med 56: 65-70, 2010.
    OpenUrlCrossRefPubMed
  49. ↵
    1. Mahdy S,
    2. Al-Emadi SA,
    3. Khanjar IA,
    4. Hammoudeh MM,
    5. Sarakbi HA,
    6. Siam AM,
    7. Abdelrahman MO
    : Vitamin D status in health care professionals in Qatar. Saudi Med J 31: 74-77, 2010.
    OpenUrlPubMed
  50. ↵
    1. Szklo M,
    2. Nieto J
    : Epidemiology: Beyond the Basics. Gaithersburg, MD: Aspen Publishers, pp. 141-145, 2000.
  51. ↵
    1. Palmer HG,
    2. Gonzalez-Sancho JM,
    3. Espada J,
    4. Berciano MT,
    5. Puig I,
    6. Baulida J,
    7. Quintanilla M,
    8. Cano A,
    9. de Herreros AG,
    10. Lafarga M,
    11. Munoz A
    : Vitamin D(3) promotes the differentiation of colon carcinoma cells by the induction of E-cadherin and the inhibition of beta-catenin signaling. J Cell Biol 154: 369-387, 2001.
    OpenUrlAbstract/FREE Full Text
  52. ↵
    1. Newmark H,
    2. Heaney R,
    3. Lachance P
    : Should calcium and vitamin D be added to the current enrichment program for cereal-grain products? Am J Clin Nutr 80: 264-270, 2004.
    OpenUrlAbstract/FREE Full Text
  53. ↵
    1. Welsh J
    : Vitamin D and breast cancer: insights from animal models. Am J Clin Nutr 80: 1721S-1724S, 2004.
    OpenUrlAbstract/FREE Full Text
  54. ↵
    1. Moore L,
    2. Loring-Bradlee M,
    3. Singer M,
    4. Rothman K,
    5. Milunsky A
    : Folate intake and the risk of neural tube defects: an estimation of dose–response. Epidemiology 14: 200-205, 2003.
    OpenUrlCrossRefPubMed
  55. ↵
    1. Vieth R,
    2. Chan PC,
    3. MacFarlane GD
    : Efficacy and safety of vitamin D3 intake exceeding the lowest observed adverse effect level. Am J Clin Nutr 73: 288-294, 2001.
    OpenUrlAbstract/FREE Full Text
  56. ↵
    1. Vieth R
    : Why the optimal requirement for vitamin D3 is probably much higher than what is officially recommended for adults. J Steroid Biochem Mol Biol 89-90: 575-579, 2004.
    OpenUrlCrossRefPubMed
  57. ↵
    1. Vieth R
    : Vitamin D supplementation, 25-hydroxyvitamin D concentrations, and safety. Am J Clin Nutr 69: 842-856, 1999.
    OpenUrlAbstract/FREE Full Text
  58. ↵
    1. National Academy of Sciences
    . Commitee to Review Dietary Reference Intakes for Vitamin D and Calcium. Washington DC: National Academies Press, 2010.
  59. ↵
    1. National Academy of Sciences–Institute of Medicine–Food and Nutrition Board
    . Dietary Reference Intakes for Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride. Washington DC: National Academy Press, 1997.
  60. ↵
    1. Heaney RP,
    2. Davies KM,
    3. Chen TC,
    4. Holick MF,
    5. Barger-Lux MJ
    : Human serum 25-hydroxycholecalciferol response to extended oral dosing with cholecalciferol. Am J Clin Nutr 77: 204-210, 2003.
    OpenUrlAbstract/FREE Full Text
  61. ↵
    1. Ginde AA,
    2. Liu MC,
    3. Camargo CA Jr..
    : Demographic differences and trends of vitamin D insufficiency in the US population, 1988-2004. Arch Intern Med 169: 626-632, 2009.
    OpenUrlCrossRefPubMed
  62. ↵
    1. Honkanen R,
    2. Alhava E,
    3. Parviainen M,
    4. Talasniemi S,
    5. Monkkonen R
    : The necessity and safety of calcium and vitamin D in the elderly. J Am Geriatr Soc 38: 862-866, 1990.
    OpenUrlPubMed
  63. ↵
    1. Vieth R
    : Vitamin D nutrition and its potential health benefits for bone, cancer and other conditions. J Nutr Environ Med 11: 275-291, 2001.
    OpenUrlCrossRef
  64. ↵
    1. Trang HM,
    2. Cole DE,
    3. Rubin LA,
    4. Pierratos A,
    5. Siu S,
    6. Vieth R
    : Evidence that vitamin D3 increases serum 25-hydroxyvitamin D more efficiently than does vitamin D2. Am J Clin Nutr 68: 854-858, 1998.
    OpenUrlAbstract
  65. ↵
    1. Shin MH,
    2. Holmes MD,
    3. Hankinson SE,
    4. Wu K,
    5. Colditz GA,
    6. Willett WC
    : Intake of dairy products, calcium, and vitamin D and risk of breast cancer. J Natl Cancer Inst 94: 1301-1311, 2002.
    OpenUrlAbstract/FREE Full Text
  66. ↵
    1. Knight JA,
    2. Lesosky M,
    3. Barnett H,
    4. Raboud JM,
    5. Vieth R
    : Vitamin D and reduced risk of breast cancer: a population-based case–control study. Cancer Epidemiol Biomarkers Prev 16: 422-429, 2007.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Reprints and Permissions
Serum 25-Hydroxyvitamin D and Prevention of Breast Cancer: Pooled Analysis
SHARIF B. MOHR, EDWARD D. GORHAM, JOHN E. ALCARAZ, CHRISTOPHER J. KANE, CAROLINE A. MACERA, J. KELLOGG PARSONS, DEBORAH L. WINGARD, CEDRIC F. GARLAND
Anticancer Research Sep 2011, 31 (9) 2939-2948;
Twitter logo Facebook logo Mendeley logo

Jump to section