Interaction of estrogen therapy with calcium and vitamin D supplementation on colorectal cancer risk: Reanalysis of Women's Health Initiative randomized trial
Article first published online: 18 DEC 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 122, Issue 8, pages 1690–1694, 15 April 2008
How to Cite
Ding, E. L., Mehta, S., Fawzi, W. W. and Giovannucci, E. L. (2008), Interaction of estrogen therapy with calcium and vitamin D supplementation on colorectal cancer risk: Reanalysis of Women's Health Initiative randomized trial. Int. J. Cancer, 122: 1690–1694. doi: 10.1002/ijc.23311
- Issue published online: 19 FEB 2008
- Article first published online: 18 DEC 2007
- Manuscript Accepted: 23 OCT 2007
- Manuscript Received: 22 APR 2007
- National Institutes of Health. Grant Number: R25-CA098566
- vitamin D;
- colorectal cancer;
- randomized trial;
Although calcium and vitamin-D intake were consistently shown to be inversely associated with colorectal cancer risk in several large prospective studies and protective against adenoma and cancer in multiple randomized trials, the Women's Health Initiative (WHI) of calcium and low-dose vitamin-D supplementation trial found no overall effects on colorectal cancer. However, the previous report did not recognize an important biologic interaction with estrogen therapy. We investigated the treatment interaction of estrogen with calcium and vitamin-D on risk of colorectal cancer via a reanalysis of primary data results from the WHI calcium and vitamin-D supplementation trial (1,000 mg elemental calcium, 400 IU of vitamin-D3, or placebo), reanalyzing results from women concurrently randomized to estrogen interventions and placebo. Results indicate that concurrent estrogen therapy was a strong effect modifier of calcium and vitamin-D supplementation on colorectal cancer risk. While calcium plus vitamin-D supplementation among women concurrently assigned to estrogen therapies suggested increased risk (Hazard Ratio = 1.50, 95% CI: 0.96–2.33), among women concurrently assigned to placebos arms of the estrogen trials, calcium plus vitamin-D indicated suggestive benefits (HR = 0.71, 95% CI: 0.46–1.09) (p-for-estrogen-interaction = 0.018). Consistent interaction was also found by reported estrogen use (p interaction = 0.037). Results indicate contrasting effects of calcium and vitamin-D by concurrent estrogen therapy on colorectal cancer risk. Although further clinical and mechanistic studies are warranted, the potential clinical implications of the apparent interaction of estrogen therapy with calcium and vitamin-D supplementation should be recognized. Important biological mechanisms related to the key membrane receptor megalin and estrogen-dependent protein calbindin are discussed. © 2007 Wiley-Liss, Inc.
In 1980, vitamin D was first proposed to have benefit against colorectal cancer based on ecological studies of colorectal cancer incidence and low sunlight exposure.1 Vitamin D has been documented to possess a variety of anti-cancer properties,2 including inhibit growth and induce differentiation and apoptosis of colon tumor cells,3 as well as shown to be associated with lower risk of colorectal cancer and adenoma in prospective studies.4–7 In addition to a meta-analysis of observational studies supporting the relation of vitamin D and lower colorectal cancer risk,8 a recent meta-analysis of randomized trials indicates vitamin D supplementation likely decreases risk of all-cause mortality, with benefits for cancer thought to be a major explanatory factor.9 Consistent with a 5 year vitamin D trial that suggested benefits against colorectal cancer mortality,10 a large randomized trial indeed demonstrated that moderate dose 1,000 IU vitamin D and calcium supplementation substantially reduces all-cancer risk after 1–4 years follow-up RR = 0.23 (95% CI: 0.09–0.60),11 where vitamin D significantly contributed benefit beyond calcium supplementation alone.
Similarly, however, calcium may also impede colonic carcinogenesis via a variety of mechanisms,3 as calcium has been inversely associated with colorectal cancer risk in prospective studies,12, 13 and also been demonstrated to be protective against colorectal carcinogenesis and adenomas in multiple randomized trials.14–17 Furthermore, a large pooled analysis of 10 international prospective cohorts found that milk and calcium intakes were both associated with significantly lower risk of colorectal cancer, with the relative risks, comparing highest vs. lowest categories of intake, of 0.85 (95% CI: 0.78, 0.94) for milk and 0.78 (95% CI: 0.69, 0.88) for calcium.12
However, despite the preponderance of prior observational and experimental evidence, the recent Women's Health Initiative (WHI) trial found no benefit of combined calcium and vitamin D supplementation against invasive colorectal cancer.18 As recognized by both the original WHI authors and many expert commentators,19, 20 there is currently a high priority to understand these findings and resolve the seemingly inconsistent results with past randomized trials and observational studies. Methodological limitations ranging from poor treatment compliance, inadequate vitamin D dose and duration have previously been cited in the trial,18–20 and are likely to have contributed to the null findings. However, an emerging body of research from multiple lines of evidence suggests that biological treatment interactions related to the concurrent estrogen interventions may contribute to the unique null findings in the WHI. In this study and review, we highlight a relatively under-recognized but clinically significant interaction between estrogen therapy and calcium + vitamin D for colorectal cancer risk, as well as conduct a reanalysis of the WHI calcium + vitamin D trial results to explore this hypothesis, and discuss mechanisms that may clarify the findings and inconsistencies.
The WHI calcium and vitamin D supplementation trial was a randomized, double-blind, placebo-controlled trial involving 36,282 postmenopausal women from 40 WHI centers in which 18,176 women received 500 mg of elemental calcium as calcium carbonate with 200 IU of vitamin D3 twice daily (1,000 mg of elemental calcium and 400 IU of vitamin D3) and 18,106 received a matching placebo for an average of 7.0 years.18 Utilizing data from the WHI calcium and vitamin D supplementation trial, we conducted a reanalysis of the data in the study report on risk of colorectal cancer (extracting data from publication). We obtained the stratified hazard ratios and variance data for effects of calcium + vitamin D supplementation among various estrogen intervention populations of the multifactorial WHI studies. To examine the interactive effects on estrogen on calcium and vitamin D supplementation, the hazard ratios (HR) of colorectal cancer among women in the calcium and vitamin D supplementation trial who were concurrently randomized to the estrogen–progestin and estrogen-only arms of the factorial WHI trial (n = 8,117) were pooled. This pooled estimate was compared with the pooled estimate among women in the calcium and vitamin D supplementation trial who were concurrently randomized to the 2 respective placebo arms of the factorial estrogen trials (n = 7972). Estimates were pooled via random-effects models.21 In the primary analysis, only these 4 subgroups were included, as only these estrogen trial arms and respective placebo arms can be guaranteed to be balanced for all baseline covariates by randomization. To explore consistency, secondary analysis of interaction by reported estrogen use was also conducted. Treatment interaction of calcium and vitamin D supplementation with concurrent estrogen was assessed via meta-regression.22 All statistical tests were two-sided, unless otherwise noted. All analyses were conducted using STATA 8.2 (College Station, TX).
Among those women concurrently assigned to estrogen treatments in the overlapping WHI estrogen trials, there were a total of 4,039 women randomized to calcium plus vitamin-D and 4,078 women randomized to placebo; a total of 45 and 35 women developed colorectal cancer in these 2 groups, respectively. Among those women concurrently assigned to the comparative placebo groups of the overlapping WHI estrogen trials, there were a total of 4,015 women randomized to calcium plus vitamin-D and 3,957 women randomized to placebo; a total of 36 and 51 women developed colorectal cancer in these 2 groups, respectively. In time-to-event analysis, results indicated that concurrent estrogen therapy was a strong attenuating modifier of calcium and vitamin D supplementation on colorectal cancer risk. The hazard ratio of colorectal cancer for calcium plus vitamin D supplementation among women concurrently assigned to estrogen therapies was 1.50 (95% CI: 0.96–2.33), borderline suggestive of increased risk. Meanwhile the hazard ratio of colorectal cancer for calcium plus vitamin D supplementation among women concurrently assigned to comparative placebos of the estrogen trials was 0.71 (95% CI: 0.46–1.09), indicating possible protective benefit. The p value for interaction by concurrent estrogen assignment was statistically significant (p = 0.018). Results are summarized in Figure 1. Given the strong a priori hypothesis for potential benefits from previous studies, analysis via a 1-sided test yielded HR = 0.71 (90% CI: 0.49–1.01). In further analysis, the significant estrogen interaction was also consistently observed in alternative stratification of reported hormone therapy use, where results again indicated divergent effect by concurrent estrogen use; there was no protective effect among current estrogen users (pooled HR = 1.36, 95% CI: 0.98–1.87), while there is a suggestion of protective benefit among never users of estrogen (HR = 0.81, 95% CI: 0.56–1.16). The p value for interaction by reported current-vs.-never estrogen use was also significant (p = 0.037).
The results of our analysis of the effects of calcium and vitamin D, by individual-level stratification of concurrent estrogen treatment in the WHI, indicate that estrogen therapy may be a significant modifier of the effect of calcium and vitamin D on colorectal cancer. Although the results of calcium and vitamin D among women randomized to placebo arms of estrogen trials were not quite significant (HR = 0.71, 95% CI: 0.46–1.09), the suggestive protective effects are highly consistent with previous calcium trials against colorectal adenoma, and borderline nonsignificance expected given the reduced sample size (<50%) of the original WHI power calculations assuming no interaction with estrogen therapies. Such an interaction of estrogen with calcium and vitamin D has been under-recognized in the clinical realm; yet it has been mechanistically supported in an emerging body of experimental and observational studies.
One potential mechanism of the observed estrogen interaction is through modulation of intracellular levels of calbindin, a cytosolic calcium-binding protein expressed in various tissues including those of the intestine tract.23 Estrogen has been consistently shown to increase calbindin mRNA and calbindin protein levels,23–25 notably with certain forms of calbindin even recognized to carry an estrogen-response element on its nuclear promoter.23 Calbindin is known to act as an intracellular calcium buffer to prevent calcium cytotoxicity during active calcium absorption, via mechanisms independent of vitamin D receptor (VDR) action.26, 27 However, increase in calbindin consequently decreases intracellular free calcium,28 with such intracellular free calcium decrease demonstrated to be important for subsequent inhibition of calcium-mediated apoptosis in multiple experimental studies.29–31 Furthermore, in addition to studies indicating calbindin gene overexpression in mice colorectal adenomas,32 a body of other evidence also indicates calbindin overexpression may stimulate cell growth and proliferation33 as well as increase insulin levels,34 which may also have cell proliferative effects for colorectal cancer.35–37
However, modulation of vitamin D's effects by estrogen may also be important. Indeed, estrogen analogues have also been demonstrated to modulate the anti-proliferative effects of vitamin D on colon cancer cells.38 Notably, in addition to inhibition of calcium-mediated apoptosis,29–31 increased estrogen-dependent calbindin can also simultaneously inhibit vitamin-D induced apoptosis.28 Furthermore, a synergy of calcium and vitamin D has previously been demonstrated in which higher levels of both are simultaneously necessary for colorectal cancer prevention. For example, some randomized trials and large pooling projects have observed significant beneficial effects of calcium against colorectal adenomas and cancer only among those with high vitamin D levels or intakes.12, 15 Similarly for hip fractures, the protective benefits of vitamin D versus placebo are only observed when vitamin D is combined with calcium.39 This synergy of calcium and vitamin D is important because it is likely that estrogen's modulation and disturbance of either calcium or vitamin D-related mechanisms may attenuate the protective effects of calcium and vitamin D supplementation against colorectal cancer.
Another potential mechanism for the interaction may be related to estrogen's well-recognized importance in bone formation.40Via activation of osteoblasts, estrogen induces substantially increased mineralization of calcium into bone. This established effect was also observed in the parallel WHI study of fractures which found stronger protective effects of calcium + vitamin D supplementation against fractures among those simultaneously randomized to estrogen.41 Furthermore, another clinical trial found that even low-dose, short-term estrogen therapy can significantly modulate and attenuate the resorbing effects of PTH, thereby increasing bone formation and calcium sequestration.42 Such a parallel action of estrogen in diverting calcium away from systemic circulation and sequestering calcium into bone mineralization reduces the overall effective bioavailability of calcium for carcinogenesis prevention.
Finally, and most importantly, very recent experimental evidence indicates that the binding of vitamin D to its plasma membrane receptor megalin43–45 may be competitive, as megalin is also now an established nonrecycled endocytosis receptor of estrogen bound to sex hormone-binding globulin (SHBG),44 replacing the outdated free steroid hormone diffusion hypothesis.46 Indeed, an emerging body of evidence indicate that SHBG-bound estrogen and other sex hormones may be biologically active via receptor-mediated endocytosis,44, 47, 48 specifically mediated by the recently identified megalin receptor,44 a 600-kDa transmembrane protein belonging to the LDL-receptor family. Megalin expressed in a variety of cell tissues including the GI,49 not only binds calcium,50 but more importantly, megalin furthermore serves as a key endocytosis cell surface receptor for several vitamins and hormonal ligands, including: retinol-binding protein (RBP)-bound vitamin A,45, 51 vitamin D,43, 45 parathyroid hormone,52 and its recently identified ligands—estrogen and testosterone bound to SHBG,44 facilitating the endocytosis and signaling actions of these key sex hormones. Indeed, this is clinically relevant to the WHI trial findings, as megalin gene knockout has been demonstrated to strongly induce both estrogen deficiency and vitamin D deficiency43, 44—indicating the interplay of estrogen and vitamin D via their shared receptor. However, unlike the LDL receptor which is recycled back to the cell surface after intracellular delivery via endocytosis, a portion of megalin receptors each cycle is immediately degraded in the lysosome,44, 49 resulting in competition of estrogen and vitamin D metabolites for available receptors and cellular uptake. Such a competitive mechanism of estrogen and vitamin D uptake may indeed explain the treatment interaction of concurrent estrogen therapy with calcium and vitamin D. Furthermore, the ligand capability of PTH52 for the megalin receptor also suggests a potential triple interactive interplay with calcium, for which further studies are warranted. A listing of potential mechanisms for the observed interaction is summarized in Table I.
|1. Estrogen increases calbindin (calcium binding protein) gene expression and intracellular calbindin levels.|
|a. Calbindin modulates intracellular free calcium levels.|
|b. Increased calbindin inhibits calcium-mediated and vitamin D-mediated apoptosis, and promotes cell proliferation and tumor progression.|
|2. Estrogen and vitamin D complexes undergo competitive binding for their common cellular uptake membrane receptor Megalin.|
|a. Megalin receptors undergo endocytosis after competitive saturation.|
|b. Megalin receptors are nonrecyclable back to plasma membrane, undergoing lysosomal degradation after intracellular steroid delivery.|
|3. Estrogen-induced osteoblast formation increases bone mineralization of calcium, sequestering calcium from systemic bioavailability.|
|4. Estrogen attenuation of either calcium or vitamin D bioavailability (as described earlier) may disrupt the necessary synergistic levels of both high calcium and high vitamin D levels for colorectal cancer prevention.|
Along with the above described experimental evidence and biological mechanisms, our reanalysis of results from the WHI trial indicating a significant interaction of calcium and vitamin D supplementation treatment efficacy by concurrent estrogen (p = 0.018) is further supported by other epidemiology and clinical research. Notably for the estrogen-regulated calbindin, a clinical follow-up study found that calbindin gene-expression in tumors was strongly associated with increased rate of medulloblastoma recurrence and cancer death.53 Furthermore, in separate studies of both dietary vitamin D and serum 25-OH-vitamin-D, the association between vitamin D intake and lower colorectal cancer and adenoma risk in postmenopausal women were consistently somewhat stronger in HRT nonusers than current users.5, 7 Similarly, in our analysis of the WHI trial data with individual-level stratification of HRT treatment assignment, calcium and vitamin D supplementation was associated with borderline significant lower risk of colorectal cancer in placebo groups of the WHI estrogen trials, while no benefit, and perhaps even increased risk, was seen among women assigned to estrogen treatment.
Interaction by estrogen was also suggested in a pooled analysis of 2 large prospective studies of plasma 25-OH vitamin D. In these 2 studies of vitamin D and colorectal cancer in men and women,55 results indicated substantially stronger beneficial effects of vitamin D against colorectal cancer among lean individuals, but not among overweight individuals, who have higher estrogen levels. Similarly consistent with the estrogen interaction, a stronger inverse association was also observed for vitamin D and colorectal cancer risk among those with high physical activity, who naturally have decreased estrogen levels, while a nonsignificant association was observed among those with lower physical activity. Thus, multiple lines of biological and clinical evidence support an estrogen interaction, as observed in the WHI trial.
Nevertheless, estrogen itself is still likely protective against colorectal cancer via other mechanisms, as the findings in this study pertains only to the effect of calcium and vitamin D supplementation among concurrent estrogen users and nonusers. Moreover, estrogen's overall effects on colorectal cancer in the WHI may also be a contributing mechanism to the observed interaction with calcium and vitamin D. Notably, the previous WHI trial of estrogen plus progestin found decreased risk of colorectal cancer after early trial termination with 6.8 years average followup.56 However, in previous prospective studies, the beneficial effects of both calcium and vitamin D became significantly apparent only after more than 10 years of supplement use.6 Thus, because estrogen's innate beneficial effects act earlier in the prevention of progression to carcinogenesis—the concurrent use of estrogen may somewhat obviate or attenuate the effects of vitamin D that acts in the later phases of progression to invasive colorectal cancer. This may be a further potential explanation for the significant interaction.
In conclusion, multiple lines of evidence from a wide range of experimental and observational studies support an interaction of estrogen with calcium and vitamin D for colorectal cancer risk, which may help resolve the findings from the WHI trial. Consistent with the confluence of previous research, these findings suggest that greater intake of vitamin D and calcium is associated with lower incidence of colorectal cancer in most of the population, including men and the growing majority of women who do not take hormone therapy.
In addition to recognized limitations of the WHI trial, the interaction of concurrent estrogen therapy and resulting reduced power may explain why no readily apparent protective effects of calcium and vitamin D were observed in the WHI trial, particularly among those simultaneously randomized to estrogen therapy. Given that the suggestive beneficial effects of calcium and vitamin D were found in a smaller sample size that was underpowered to explore the effects, future studies of calcium and vitamin D supplementation targeted among estrogen nonusers are warranted. Although further clinical and mechanistic studies are warranted, the potential clinical implications of the interaction of concurrent estrogen therapy with calcium and vitamin D supplementation should be considered.
We would like to thank Dr. Walter Willett (Harvard School of Public Health) for his helpful comments. E. Ding takes responsibility for the integrity of the data and the accuracy of the data analysis. E. Ding was supported by a training grant from the National Institutes of Health (R25-CA098566). The funding source had no role in the study conduct and analysis.
- 19Calcium plus vitamin D and the risk of colorectal cancer. N Engl J Med 2006; 354: 2287–8.http://www.ncbi.nlm.nih.gov/sites/entrez?Db=pubmed&Cmd=ShowDetailView&TermToSearch=16729368, , , , .
- 22Systematic reviews in health care: meta-analysis in context, 2nd ed. London, UK: BMJ Books, 2001., , .
- 39Need for additional calcium to reduce the risk of hip fracture with vitamin D supplementation: evidence from a comparative meta-analysis of randomized controlled trials. J Clin Endocrinol Metab http://jcem.endojournals.org/cgi/content/abstract/92/4/1415, , , , , .
- 42Low dose estrogen replacement therapy (ERT) for postmenopausal hemodialysis (HD) patients. Clin Calcium 2005; 15 ( Suppl 1): 161–6., , , , , , .