Whether hormone therapy and vitamin D (or calcium+vitamin D) has a synergistic relationship on the cardiovascular system in women has gained traction in clinical trials and basic science research.1–3 A recent study suggested that menopausal monkeys taking daily oral estrogen and who had greater percent plasma 25-hydroxycholecalciferol increases over the course of the study had the least severe cardiovascular disease and greater coronary artery remodeling compared with those not taking estrogen with lower plasma 25-hydroxycholecalciferol concentrations.3 Daily use of 1.25 mg conjugated equine estrogen has been shown to increase the biologically active form of vitamin D and vitamin D carrier protein in menopausal women.4 In ovariectomized rats, estrogen upregulates the expression of the vitamin D receptors in the small intestine.5 An analysis from the Women's Health Initiative (WHI) reported a statistically significant reduction (a synergistic effect) in the risk of hip fracture attributed to hormone therapy among participants randomized to calcium plus vitamin D compared with WHI–Hormone Therapy trial participants who were randomized to hormone therapy and placebo.1
Because women in the WHI were receiving both vitamin D+calcium, this trial offers an ideal opportunity to study whether there is a synergistic effect on cardiovascular disease risk factors in menopausal women with calcium+vitamin D as well as hormone therapy. The WHI calcium+vitamin D trials were double-blind, randomized, placebo-controlled studies analyzing multiple health outcomes in menopausal women. In the calcium+vitamin D trial, low-density lipoprotein cholesterol (LDL-C) was significantly reduced for women randomized to calcium+vitamin D,6 and for the WHI–Hormone Therapy trials, both estrogen plus progestin and estrogen alone also significantly reduced LDL-C.7,8 Moreover, both hormone therapy preparations had statistically significant favorable effects on high-density lipoprotein cholesterol, glucose, and waist circumference, but significantly unfavorable effects on triglycerides and systolic blood pressure.7,8
We measured changes in our primary outcome (LDL-C) as well as our secondary outcomes (multiple cardiovascular disease risk factors) in four groups of women randomly assigned to different therapeutic combinations: calcium+vitamin D alone, hormone therapy alone, both hormone therapy and calcium+vitamin D, and neither hormone therapy nor calcium+vitamin D. The study hypothesis is that a statistically significant interaction exists between hormone therapy and calcium+vitamin D in terms of the effect on primary study endpoints, LDL-C as well as secondary outcomes including other cardiovascular risk factors. Conceptually, a significant interaction means that we observed a larger benefit among women randomized to both calcium+vitamin D and hormone therapy than the benefit observed among women randomized to only hormone therapy plus the benefit observed among women randomized to only calcium+vitamin D. In other words, a significant interaction corresponds to a synergistic rather than an additive effect relative to the placebo group.
MATERIALS AND METHODS
The WHI clinical trials were designed to evaluate the risks and benefits of dietary modification, hormone therapy, and supplementation with calcium+vitamin D. The protocol and consent forms were approved by the institutional review boards for all participating institutions (see Acknowledgments in Appendix 1, available online at http://links.lww.com/AOG/A892).
Like previously published secondary analyses,9 the WHI–Hormone Therapy trials data were combined to improve statistical power and further justified because both WHI arms had the same qualitative effects on the measured cardiovascular disease risk factors. These results can apply to a population similar to those enrolled in the WHI–Hormone Therapy trials: 40% without a uterus taking estrogen therapy or placebo and 60% with an intact uterus taking estrogen and progestogen therapy or placebo. The WHI is the largest cohort (n=16,089) randomized to both hormone therapies (active or placebo) and calcium+vitamin D (active or placebo)1 from whom blood data were collected on 1,521 participants. Because we are using pre-existing data, power calculations were not performed.10,11
A total of 68,132 women aged 50–79 years were recruited between September 1993 and October 1998 and were randomly assigned into the WHI–Dietary Modification trial, WHI–Hormone Therapy trials, or both. A total of 27,347 women in the two parallel WHI–Hormone Therapy trials were randomized to 0.625 mg conjugated equine estrogen alone or placebo among women who had a hysterectomy or 0.625 mg conjugated equine estrogen plus 2.5 mg medroxyprogesterone acetate or placebo taken daily among women who had not had a hysterectomy. A total of 48,835 women in the WHI–Dietary Modification trial were randomized to a dietary modification intervention (dietary modification to lower total fat intake; n=19,541) or comparison (usual diet; n=29,294) group. At the first or second annual visit, 36,282 eligible women from WHI–Hormone Therapy (16,089) and WHI–Dietary Modification (n=25,210) trials were randomized further to calcium+vitamin D (1,000 mg elemental calcium [carbonate] plus 400 international units vitamin D3 daily supplementation [n=18,176]) or placebo (n=18,106) with 14% (n=5,017) of participants in both the Dietary Modification and Hormone Therapy trials. The eligibility criteria to be enrolled in the calcium+vitamin D trial included many safety parameters (eg, no previous hypercalcemia or renal calculi) and no competing risk indicators (eg, no medical condition associated with survival of less than 3 years). Eligibility for the WHI–Hormone Therapy trial included postmenopausal (as defined previously12) who were between 50 and 79 years of age at initial screening. Analysis included women who participated in both the calcium+vitamin D trial and the WHI–Hormone Therapy trial (either estrogen and progestogen or estrogen alone) and were also part of the 6% blood subsample (Fig. 1) (n=1,521). Because the calcium+vitamin D trial was initiated after 1 year of the WHI–Hormone Therapy, year 1 of the WHI–Hormone Therapy was considered as baseline for the calcium+vitamin D trial. Lipids along with other cardiovascular risk factors (blood pressure, weight, waist circumference, waist-to-hip ratio, glucose, insulin) were measured at baseline and years 1, 3, and 6 after randomization into the hormone therapy trials. The cardiovascular disease risk factors, which were measured after the estrogen plus progestin trial and the conjugated equine estrogen-alone trial stopped on July 7, 2002, and February 29, 2004, respectively, were censored. Details of biomarker analysis and laboratory methods have been published previously.13
Repeated-measures models with an unstructured variance–covariance matrix were used to model longitudinal means. Per the randomized partial-factorial design, means were assumed to be equal at baseline for all four hormone therapy+randomization groups and equal at year 1 from hormone therapy randomization for calcium+vitamin D randomization groups; the latter is the calcium+vitamin D “baseline” measure. Constraining the “baseline” means to be equal before randomization is the most efficient use of the data.14,15 To allow for parsimonious 1-degree-of-freedom estimates of treatment effects (treatment group minus placebo) and interactions (calcium+vitamin D+hormone therapy), the postrandomization means at years 3 and 6 are averaged.
For the subgroup analysis, we analyzed whether particular subgroups may modify the calcium+vitamin D+hormone therapy interaction on our main outcome variable, LDL-C (ie, whether a synergistic effect of calcium+vitamin D+hormone therapy might occur for particular subgroups). Statistical significance was based on a test of interaction. We looked at a total of 13 prespecified baseline subgroups (Table 1). No adjustment for multiple testing was made; at most, one interaction was expected to be significant by chance alone. The cutpoints for total vitamin D intake and total calcium were also chosen a priori. The lower cutpoint was suggested previously1 where the effect of hormone therapy appeared to be null for values of total vitamin D below 200 international units and calcium below 800 mg.
To address the skewed distributions of triglycerides, glucose, insulin, and waist-to-hip ratio, log-transformation was used, and geometric means are reported. Statistical significance of a synergistic effect was based on tests of interaction. A significant interaction corresponds to a synergistic rather than an additive effect relative to the placebo group. To graphically show the presence of an additive or synergistic effect between calcium+vitamin D and hormone therapy across all cardiovascular disease biomarkers, Z-scores (treatment effect divided by standard error) are shown. All analyses were done with SAS 9.4 and figures were drawn with R 3.1. All P values are two-sided and P values ≤.05 were regarded as significant.
Baseline characteristics were similarly distributed by treatment groups (Table 2). The effects of calcium+vitamin D+hormone therapy, on all of the cardiovascular disease risk factors except insulin, were larger in magnitude and in the same direction as the effects of hormone therapy alone, regardless of the size or direction of the calcium+vitamin D effect. In other words, the addition of calcium+vitamin D enhanced the effects of hormone therapy either in a positive or a negative direction. However, none of the hormone therapy+calcium+vitamin D interactions were statistically significant (Fig. 2) and therefore did not provide statistical evidence for the synergistic effects of hormone therapy+calcium+vitamin D. For example, although hormone therapy+calcium+vitamin D had a stronger effect on LDL-C compared with either hormone therapy alone or calcium+vitamin D alone, the observed effects were additive (P value for interaction=.26). Estimates for the primary analysis were precise; the 95% confidence interval (CI) for mean LDL-C in all four treatment groups was ±3 mg/dL. The effects on LDL-C (active minus placebo) were −1.6 (−5.5 to 2.2) mg/dL for calcium+vitamin D alone, −9.0 (−13.0 to −5.1) mg/dL for hormone therapy alone, and −13.8 (−17.8 to −9.8) mg/dL for calcium+vitamin D+hormone therapy (Fig. 3). Appendix 2, available online at http://links.lww.com/AOG/A892, displays the profile means for our primary endpoint, LDL-C, by randomization groups during the study. To investigate the influence of temporal trends, we limited postrandomization follow-up to year 3 and observed a similar pattern (P value for interaction=.44); the treatment effects on LDL-C for this sensitivity analysis were −1.5 (−5.1 to 2.2) mg/dL for calcium+vitamin D alone, −13.9 (−17.8 to −10.1) mg/dL for hormone therapy alone, and −17.4 (−21.2 to −13.6) mg/dL for calcium+vitamin D+hormone therapy. Lastly, a sensitivity analysis was conducted to account for compliance to study pills. Specifically, LDL-C measurements that occurred after a participant became nonadherent (took less than 80% of study pills) were censored. Resulting model estimates produced a similar additive pattern without any evidence for a synergistic effect (P value for interaction=.66).
A sensitivity analysis was conducted to determine whether the calcium+vitamin D+hormone therapy interaction depended on hormone therapy preparation by testing the three-way interaction calcium+vitamin D+hormone therapy×cohort (estrogen plus progestogen compared with estrogen therapy alone). As expected, none of the three-way interactions provided statistical evidence against pooling the hormone therapy trials for any of the cardiovascular disease risk factors (all P values >.30). The effect of hormone therapy (active compared with placebo) on LDL-C from baseline to year 1 was −16.8 (−20.4 to −13.2) and −18.4 (−21.5 to −15.3) mg/dL among estrogen therapy and estrogen+progestogen therapy participants, respectively.
Calcium and vitamin D+hormone therapy had a synergistic effect on LDL-C at low total intakes (dietary and supplements) of vitamin D (P value for interaction=.03). In addition, the effect of hormone therapy alone was more attenuated at lower levels of vitamin D intake. Calcium and vitamin D+hormone therapy had an additive (P value for interaction=.06) effect at low intakes of calcium (Fig. 3) and calcium+vitamin D+hormone therapy had a synergistic effect (P value for interaction=.007) among hypertensive women (self-reported of treatment for hypertension or recorded blood pressure 140/90 mm Hg or greater). The effect of calcium+vitamin D+hormone therapy did not vary with age (P=.59). An analysis of the two×two factorial for the main effects of calcium+vitamin D and hormone therapy, without a calcium+vitamin D+hormone therapy interaction term, is presented in Appendix 3 (available online at http://links.lww.com/AOG/A892). As previously shown in the full cohort,6 calcium+vitamin D has a favorable effect on LDL-C with a mean decrease of 3.2 mg/dL (CI −5.9 to −0.5). In addition, we demonstrated a favorable effect on total cholesterol with a mean decrease of 3.2 mg/dL (CI −6.2 to −0.3).
Although there are data to suggest calcium+vitamin D has some beneficial effects on cardiovascular disease risk factors,11,14,15 this has not been well established, and there is a paucity of prospective data regarding the effect of calcium+vitamin D on cardiovascular disease outcomes.16,17 More recent data have suggested that estrogen therapy alone in younger women closer to the time of menopause (the timing hypothesis) could have beneficial cardiovascular disease outcomes, particularly lower rates of myocardial infarction,12,18–21 but no significant reduction was observed with younger women randomized to estrogen plus progestogen therapy.9 Although highly controversial, some observational data have raised questions about the safety of high doses of calcium supplements and potential cardiovascular disease risks,22 although the WHI calcium+vitamin D trial did not.6 Our findings suggest that calcium combined with vitamin D is not detrimental, at least in regard to most cardiovascular disease risk factors.
The well-decomunted7,23 beneficial effect that hormone therapy has on cholesterol parameters aside from triglyceride is felt to be moderate compared with other cholesterol-lowering therapies. Hormone therapy has been shown to have beneficial effects on other cardiovascular disease risks as well, like glucose24 and weight distribution,25 but has increased mean systolic blood pressure in both WHI–Hormone Therapy trials26 and in other randomized clinical trials.27 It would be helpful to know whether the effects of hormone therapy or other interventions with a moderate effect on cardiovascular disease risk would be additive or synergistic with calcium+vitamin D. The data we present suggest an additive relationship with hormone therapy, which is modestly beneficial for some cardiovascular disease risk factors (eg, LDL-C, high-density lipoprotein cholesterol, total cholesterol, glucose, insulin, waist circumference, and the waist-to-hip ratio), but modestly harmful for others (eg, systolic and diastolic blood pressure, triglycerides; Fig. 2).
In a similar study from this population, the effect of calcium+vitamin D and hormone therapy on bone density was also additive.1 A synergistic effect was identified, however, when the effect of hormone therapy and calcium+vitamin D was studied on the primary outcome, fracture.1 Hence, it is possible that calcium+vitamin D and hormone therapy may have a synergistic effect for cardiovascular disease as a primary outcome. Calcium and vitamin D+hormone therapy has a greater effect on cardiovascular disease risk categories when compared with all other combinations. Moreover, for all endpoints except insulin, the effect of calcium+vitamin D+hormone therapy and hormone therapy alone were in the same direction, but the magnitude of calcium+vitamin D+hormone therapy was greater. Therefore, results suggest that the addition of calcium+vitamin D supplementation to a hormone therapy regimen could enhance the effects of hormones. In contrast, for more than half of the endpoints, hormone therapy+calcium+vitamin D and calcium+vitamin D alone went in opposite directions, so the addition of hormone therapy may swamp the effect of calcium and vitamin D supplementation (Fig. 2).
In the subgroup analysis of total vitamin D intake, the effect of hormone therapy alone had an impressive decreasing effect on LDL-C as the intake of total vitamin D increased (CI −3.2 to −7.3, −12.2 to −22.8; P=.03), implying a synergistic relationship (Fig. 3, hormone therapy-alone column for vitamin D effect). Looking at the effect of hormone therapy+calcium+vitamin D on LDL-C, the effect (CI −10.7 to −15.5, −20.1 to −14.9; P=.03) seems to progressively increase until the total vitamin D intake exceeds 600 international units (Fig. 3, hormone therapy+calcium+vitamin D column for vitamin D effect). This implies a threshold phenomenon in which total vitamin D intake is more beneficial to hormone therapy and calcium+vitamin D up to a certain point (or threshold). Based on these findings, for women on estrogen and who have low intake of vitamin D, one should consider supplementation to lower LDL-C that may decrease the risk of heart disease.
A major strength of the study is the double-blind, randomized, placebo-controlled design in a well-characterized population. Given the numbers and demographic diversity of this cohort, the findings should be generalizable to the U.S. population. This is a large study in which women were randomized to calcium+vitamin D, hormone therapy, or both with nearly 400 women in each arm. Several studies have suggested that vitamin D may have a therapeutic window phenomenon with detriment at the extremes and benefit at midlevels.28,29 This may explain why hormone therapy and calcium+vitamin D seemed to be synergistic at lower calcium+vitamin D intakes. Limitations, therefore, include the 400 international units vitamin D, which is typically used to prevent rickets, but may be inadequate to lower LDL-C. Women were allowed to continue their own calcium supplements because it would have been unethical to prohibit concurrent calcium use in a long-term, placebo-controlled trial. Also, the supplement trial used a combination of calcium+vitamin D so that the effects of either nutrient alone cannot be ascertained. We were not able to further explore the observations that calcium+vitamin D+hormone therapy had a synergistic effect on LDL-C at low total intakes (dietary and supplements) of vitamin D and calcium by correlating blood concentration of vitamin D and calcium with total intake, because only a small percentage of women had these serum markers measured.
In summary, with the exception of insulin, the absolute effect of calcium+vitamin D and hormone therapy on cardiovascular disease risk factors was larger compared with hormone therapy alone or calcium+vitamin D alone, including LDL-C, our primary endpoint. For clinicians and most patients deciding to begin calcium+vitamin D supplementation, current use of hormone therapy should not influence that decision. However, based on these findings, for women on estrogen and who have low intake of vitamin D, one should consider calcium+vitamin D supplementation to lower LDL-C that may decrease the risk of heart disease.
1. Robbins JA, Aragaki A, Crandall CJ, Manson JE, Carbone L, Jackson R, et al. Women's health initiative clinical trials: interaction of calcium and vitamin D with hormone therapy. Menopause 2014;21:116–23.
2. Schnatz PF, Marakovits KA, O'Sullivan DM, Ethun K, Clarkson TB, Appt SE. Response to an adequate dietary intake of vitamin D3 modulates the effect of estrogen therapy on bone density. J Womens Health (Larchmt) 2012;21:858–64.
3. McCurdy RJ, Jiang X, Clark M, Nudy M, Schnatz PF. Vitamin D and conjugated equine estrogen: the association with coronary artery atherosclerosis in cynomolgus monkeys. Menopause 2016;23:481–7.
4. Cheema C, Grant BF, Marcus R. Effects of estrogen on circulating “free” and total 1,25-dihydroxyvitamin D and on the parathyroid-vitamin D axis in postmenopausal women. J Clin Invest 1989;83:537–42.
5. Liel Y, Shany S, Smirnoff P, Schwartz B. Estrogen increases 1,25-dihydroxyvitamin D receptors expression and bioresponse in the rat duodenal mucosa. Endocrinology 1999;140:280–5.
6. Hsia J, Heiss G, Hong Ren H, Allison M, Dolan NC, Greenland P, et al. Calcium/vitamin D supplementation and cardiovascular events. Circulation 2007;115:846–54.
7. Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, Kooperberg C, Stefanick ML, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women's Health Initiative randomized controlled trial. JAMA 2002;288:321–33.
8. Anderson GL, Limacher M, Assaf AR, Bassford T, Beresford SA, Black H, et al. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the Women's Health Initiative randomized controlled trial. JAMA 2004;291:1701–12.
9. Rossouw JE, Prentice RL, Manson JE, Wu L, Barad D, Barnabei VM, et al. Postmenopausal hormone therapy and risk of cardiovascular disease by age and years since menopause. JAMA 2007;297:1465–77.
10. Hoenig JM, Heisey DM. The abuse of power: the pervasive fallacy of power calculations for data analysis. Am Statistician 2001;55:19–24.
11. Bacchetti P. Peer review of statistics in medical research: the other problem. BMJ 2002;324:1271–3.
12. LaCroix AZ, Chlebowski RT, Manson JE, Aragaki AK, Johnson KC, Martin L, et al. Health outcomes after stopping conjugated equine estrogens among postmenopausal women with prior hysterectomy: a randomized controlled trial. JAMA 2011;305:1305–14.
13. Schnatz PF, Jiang X, Vila-Wright S, Aragaki AK, Nudy M, O'Sullivan DM, et al. 25(OH) calcium/vitamin D (calcium + vitamin D) supplementation, serum 25-hydroxyvitamin D concentrations, and cholesterol profiles in the Women's Health Initiative calcium/vitamin D randomized trial. Menopause 2014;21:823–33.
14. Liu GF, Lu K, Mogg R, Mallick M, Mehrotra DV. Should baseline be a covariate or dependent variable in analyses of change from baseline in clinical trials? Stat Med 2009;28:2509–30.
15. Fitzmaurice GM, Laird NM, Ware JH. Applied longitudinal analysis. New York (NY): John Wiley & Sons; 2012.
16. Schnatz PF, Manson JE. Vitamin D and cardiovascular disease: an appraisal of the evidence. Clin Chem 2014;60:600–9.
17. Schnatz PF, Nudy M, Jiang X, Demko JE, Appt SE. Vitamin D deficiency and cardiovascular disease in post-menopausal women: contributions from human and non-human primate studies. Menopause 2015;22:554–63.
18. Manson JE, Chlebowski RT, Stefanick ML, Aragaki AK, Rossouw JE, Prentice RL, et al. Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women's health initiative randomized trials. JAMA 2013;310:1353–68.
19. Schierbeck LL, Rejnmark L, Tofteng CL, Stilgren L, Eiken P, Mosekilde L, et al. Effect of hormone replacement therapy on cardiovascular events in recently postmenopausal women: randomized trial. BMJ 2012;345:e6409.
20. Schnatz PF. Hormonal therapy: does it increase or decrease cardiovascular risk? Obstet Gynecol Surv 2006;61:673–81.
21. Clarkson TB, Meléndez GC, Appt SE. Timing hypothesis for postmenopausal hormone therapy: its origin, current status, and future. Menopause 2013;20:342–53.
22. Wang L, Manson JE, Sesso HD. Calcium intake and risk of cardiovascular disease: a review of prospective studies and randomized clinical trials. Am J Cardiovasc Drugs 2012;12:105–16.
23. Effects of estrogen or estrogen/progestin regimens on heart disease risk factors in postmenopausal women. The Postmenopausal Estrogen/Progestin Interventions (PEPI) Trial. The Writing Group for the PEPI Trial. JAMA 1995;273:199–208.
24. Kim JH, Cho HT, Kim YJ. The role of estrogen in adipose tissue metabolism: insights into glucose homeostasis regulation. Endocr J 2014;61:1055–67.
25. Brown LM, Clegg DJ. Central effects of estradiol in the regulation of food intake, body weight, and adiposity. J Steroid Biochem Mol Biol 2010;122:65–73.
26. Shimbo D, Wang L, Lamonte MJ, Allison M, Wellenius GA, Bavry AA, et al. The effect of hormone therapy on mean blood pressure and visit-to-visit blood pressure variability in postmenopausal women: results from the Women's Health Initiative randomized controlled trials. J Hypertens 2014;32:2071–81.
27. Ashraf MS, Vongpatanasin W. Estrogen and hypertension. Curr Hypertens Rep 2006;8:368–76.
28. Sato M, Lu J, Iturria S, Stayrook KR, Burris LL, Zeng QQ, et al. A nonsecosteroidal vitamin D receptor ligand with improved therapeutic window of bone efficacy over hypercalcemia. J Bone Miner Res 2010;25:1326–36.
29. Querfeld U, Mak RH. Vitamin D deficiency and toxicity in chronic kidney disease: in search of the therapeutic window. Pediatr Nephrol 2010;25:2413–30.