G Model SBMB-4010; No. of Pages 8

ARTICLE IN PRESS Journal of Steroid Biochemistry & Molecular Biology xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Journal of Steroid Biochemistry and Molecular Biology journal homepage: www.elsevier.com/locate/jsbmb

Review

Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis夽 Howard N. Hodis a,∗ , Wendy J. Mack b,1 a b

Keck School of Medicine, University of Southern California, 2250 Alcazar Street, CSC 132, Los Angeles, CA 90033, United States Keck School of Medicine, University of Southern California, 2001 Soto Street, SSB 202Y, Los Angeles, CA 90033, United States

a r t i c l e

i n f o

Article history: Received 11 November 2012 Received in revised form 10 June 2013 Accepted 28 June 2013 Keywords: Hormone therapy Estrogen Menopause Women Coronary heart disease Stroke Mortality Clinical trials Observational studies Timing hypothesis

a b s t r a c t Conclusions from randomized controlled trial (RCT) data over the past 10 years has spanned from presumed harm to consistency with observational data that hormone replacement therapy (HRT) decreases the risk for coronary heart disease (CHD) as well as overall mortality in women who are recently postmenopausal. Multiple clinical studies including randomized trials and observational studies converge with animal experimentation to show a consistency that HRT decreases CHD risk and overall mortality in primary prevention when HRT is started at the time of or soon after menopause. The totality of data supports the “timing” hypothesis that posits that HRT effects are dependent on when HRT is started in relation to age and/or time-since-menopause. The totality of data shows that HRT decreases CHD and overall morality when started in women who are less than 60 years old and/or less than 10 years postmenopausal, providing a “window-of-opportunity”. Further evidence shows that women who start HRT when in their 50s and continued for 5–30 years that there is an increase of 1.5 quality-adjusted life-years (QALYs). Additionally, HRT is highly cost-effective at $2438 per QALY gained. The totality of data converges to show a consistency between randomized trials and observational studies that when started in women at or near menopause and continued long-term, HRT decreases CHD and overall mortality compared with women who do not use HRT. This article is part of a Special Issue entitled ‘Menopause’. © 2013 Published by Elsevier Ltd.

1. Early randomized controlled trials Over the past 5 decades, 40–50 postmenopausal observational studies have shown that hormone replacement therapy (HRT) is consistently associated with a 30–50% decrease in coronary heart disease (CHD) risk and total mortality [1–10]. Consistent data from postmenopausal observational studies resulted in development of the hypothesis of “estrogen cardioprotection” [11]. Most recently, the observational study from the Women’s Health Initiative (WHI) showed that users of HRT had a 50% reduction in CHD relative to women who did not use HRT [7]. The first large randomized controlled trial (RCT) of HRT and CHD outcomes was the Heart and Estrogen–progestin Replacement Study (HERS) [12]. HERS included women with established CHD and when randomized women had a mean age of 66.7 years

夽 Funded in part by the National Institutes of Health, National Institute on Aging, R01AG-024154. ∗ Corresponding author. Tel.: +1 323 442 1478; fax: +1 323 442 2685. E-mail addresses: [email protected] (H.N. Hodis), [email protected] (W.J. Mack). 1 Tel.: +1 323 442 1820; fax: +1 323 442 2993.

and were 18 years postmenopausal. Relative to placebo, daily continuous combined conjugated equine estrogen (CEE) plus medroxyprogesterone acetate (MPA) had a null effect on CHD outcomes in HERS (hazard ratio (HR), 0.99; 95% confidence interval (CI), 0.80–1.22) [12]. Consistent with HERS were the results from the Estrogen Replacement and Atherosclerosis (ERA) trial that included women with a mean age of 65.8 years and who were 23 years postmenopausal when randomized as ERA showed that neither unopposed CEE nor CEE + MPA reduced coronary artery atherosclerosis progression [13]. On the other hand, the Estrogen in the Prevention of Atherosclerosis Trial (EPAT) showed that in healthy postmenopausal women without established cardiovascular disease that oral 17␤-estradiol alone reduced progression of subclinical atherosclerosis relative to placebo [14]. Since women randomized to EPAT were younger (mean age 62.2 years) than the HERS and ERA cohorts, and the years postmenopausal at the time of randomization was 5–10 years earlier in EPAT, the divergence in outcomes between EPAT and observational studies versus HERS and ERA was hypothesized to be dependent upon timing of HRT initiation. In particular, the key to preventing CHD appears to be starting HRT at an early stage in the process of atherosclerosis progression at the start of menopause [14]. This hypothesis, further

0960-0760/$ – see front matter © 2013 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.jsbmb.2013.06.011

Please cite this article in press as: H.N. Hodis, W.J. Mack, Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis, J. Steroid Biochem. Mol. Biol. (2013), http://dx.doi.org/10.1016/j.jsbmb.2013.06.011

G Model SBMB-4010; No. of Pages 8 2

ARTICLE IN PRESS H.N. Hodis, W.J. Mack / Journal of Steroid Biochemistry & Molecular Biology xxx (2013) xxx–xxx

supported by EPAT’s sister study, the Women’s Estrogen–progestin Lipid-Lowering Hormone Atherosclerosis Regression Trial (WELLHART) (mean time postmenopausal 18.2 years) and animal studies later became known as the “timing” hypothesis for the reduction of CHD with HRT in postmenopausal women [15]. Over the last decade, a large quantity of data strongly supporting the “timing” hypothesis, including WHI data has accumulated [11].

2. Studies supporting the timing hypothesis Cumulated data from HRT RCTs demonstrate two populations of postmenopausal women who differ in their response to HRT; that is, modification of response based on when HRT is started in relation to age and/or years postmenopausal [11]. Specifically, CHD events and overall mortality are decreased when HRT is started in women less than 60 years old and/or less than 10 years postmenopausal whereas, there is a null effect on these outcomes (and possible adverse effect) when HRT is started in women greater than 60 years old and/or greater than 20 years postmenopausal [11]. The differential effect of HRT on CHD events in relation to age and/or years postmenopausal is summarized by a meta-analysis that included 23 RCTs comprising 191,340 women-years [16]. When analyzed across the entire age range of these randomized trials, CHD events were unaffected by HRT (HR, 0.99; 95% CI, 0.88–1.11) as was the effect of HRT when started in women greater than 60 years old and/or greater than 10 years postmenopausal (HR, 1.03; 95% CI, 0.91–1.16). On the other hand, when started in women who were less than 60 years old and/or less than 10 years postmenopausal, CHD is decreased 32% with HRT compared to placebo (relative risk (RR), 0.68; 95% CI, 0.48–0.96) [16]. The magnitude of CHD reduction for women less than 60 years old and/or less than 10 years postmenopausal at the time of randomization to HRT is comparable to the women in observational studies who started HRT at menopause [1–10]. It should be noted that other factors may be important in explaining the differences between observational studies and randomized trials of the effects of HRT on CHD. One prominent factor is body mass index (BMI). In general, women who used HRT in observational studies were relatively thin (BMI approximately 25 kg/m2 ), whereas women randomized to clinical trials were primarily overweight to obese (BMI approximately 29 kg/m2 ). For example, average BMI in the Women’s Health Initiative (WHI) trials was 28.5 kg/m2 in the CEE + MPA (WHI-E + P) trial and 30.1 kg/m2 in the CEE alone (WHI-E) trial. In addition, 34% of the women in the WHI-E + P trial and 45% of the women in the WHI-E trial were obese with BMI > 30 kg/m2 [17,18]. In HERS, 56% of the women had a BMI >27 kg/m2 [12]. The Cancer Prevention Study II, a 12-year observational study of 290,823 women who were free from cancer and cardiovascular disease at enrollment best demonstrates the effect of HRT on CHD according to BMI [19]. Among HRT users, all cause mortality was reduced relative to never-users (RR, 0.82; 95% CI, 0.78–0.87). Specifically, CHD death was lowest for women using HRT with BMI <22 kg/m2 while there was no association between HRT use and CHD in women with BMI >30 kg/m2 , p-for-interaction = 0.02. For women who used HRT, CHD mortality was reduced 51% (RR, 0.49; 95% CI, 0.37–0.65) among those with BMI <22 kg/m2 , reduced 28% (RR, 0.72; 95% CI, 0.57–0.91) among women with BMI 22 kg/m2 to <25 kg/m2 and reduced 23% (RR, 0.77; 95% CI, 0.59–1.01) among women with BMI 25 kg/m2 to <30 kg/m2 . On the other hand, CHD mortality was increased 45% (RR, 1.45; 95% CI, 1.00–2.11) among women who used HRT with BMI of >30 kg/m2 . Assuming that the Cancer Prevention Study II results extend to both fatal and nonfatal myocardial events, the Cancer Prevention Study II results predict based on the average BMI of women

randomized to clinical trials, no overall significant HT effect on CHD outcome relative to placebo (with the exception of the Danish Osteoporosis Study (DOPS) in which average BMI was 25.2 kg/m2 ) [20]. Specifically, the Cancer Prevention Study II results predict no HRT effect on CHD in women with BMI >25 kg/m2 and a potential increased CHD risk with HRT among women with BMI >30 kg/m2 [19]. In contrast to the clinical trials, women enrolled in observational studies were much leaner (average BMI of approximately 25 kg/m2 ) and those who used HRT had reductions in CHD risk [1–10]. In terms of BMI, women randomized to clinical trials were similar to HRT nonusers in observational studies [1–10]. The CHD outcome according to BMI from the WHI-E trial showed a similar pattern as to the results from the Cancer Prevention Study II as described above [21]. Women with BMI <25 kg/m2 , 25 kg/m2 to <30 kg/m2 and >30 kg/m2 and randomized to CEE therapy relative to placebo in the WHI-E trial showed respectively, a 24% reduction, 13% reduction and 11% increase in CHD events [21]. As reviewed above, comparisons across the cumulative studies clearly show that the women randomized to clinical trials were very different across several parameters from the women studied in observational studies. As such, the hormone cardioprotective hypothesis has yet to be appropriately tested since the population of women studied in observational studies from which the hypothesis was derived has not been studied in randomized trials; the one exception is DOPS in which time from menopause when HT was initiated and BMI of the cohort were similar to women studied in observational studies [20].

3. WHI data supporting the timing hypothesis In response to the growing accumulation of data supportive of the timing hypothesis, WHI post hoc analyses were conducted that are also supportive of the “timing” hypothesis by showing significant trends of an HRT effect on CHD relative to years postmenopausal when HRT is initiated [22]. Compared with placebo, there was a 52% (HR, 0.48; 95% CI, 0.20–1.17) reduced CHD risk in the women who were randomized to CEE and less than 10 years postmenopausal in the WHI-E trial [22]. On the other hand, compared with placebo there was no CHD benefit (HR, 0.96; 95% CI, 0.64–1.44) in the women who were randomized to CEE and 10–19 years postmenopausal or when randomized to CEE and more than 20 years postmenopausal (HR, 1.12; 95% CI, 0.86–1.46) [22]. Compared with placebo, there was a 12% (HR, 0.88; 95% CI, 0.54–1.43) reduced CHD risk in the women who were randomized to CEE + MPA and less than 10 years postmenopausal in the WHI-E + P trial [21]. On the other hand, compared with placebo there was a 23% (HR, 1.23; 95% CI, 0.85–1.77) elevated CHD risk in the women who were randomized to CEE + MPA and 10–19 years postmenopausal and a 66% (HR, 1.66; 95% CI, 1.14–2.41) increased risk in the women who were randomized to CEE + MPA and more than 20 years postmenopausal [22]. In both WHI trials combined, the women randomized to CEE and CEE + MPA and 10 years postmenopausal had a 24% (HR, 0.76; 95% CI, 0.50–1.16) reduced CHD risk compared with placebo [22]. On the other hand, compared with placebo there was no CHD benefit (HR, 1.10; 95% CI, 0.84–1.45) in the women who were randomized to CEE and to CEE + MPA and 10–19 years postmenopausal and a 28% (HR, 1.28; 95% CI, 1.03–1.58) increased risk in the women who were randomized to CEE and CEE + MPA and more than 20 years postmenopausal [22]. Women who were 50–59 years old when randomized to CEE showed significant 34–45% reductions in several categories comprising the CHD composite outcomes of CHD death, myocardial infarction (MI), confirmed angina pectoris and coronary artery revascularization compared with placebo [21]. The 11-year WHI

Please cite this article in press as: H.N. Hodis, W.J. Mack, Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis, J. Steroid Biochem. Mol. Biol. (2013), http://dx.doi.org/10.1016/j.jsbmb.2013.06.011

G Model SBMB-4010; No. of Pages 8

ARTICLE IN PRESS H.N. Hodis, W.J. Mack / Journal of Steroid Biochemistry & Molecular Biology xxx (2013) xxx–xxx

CEE trial follow-up that included randomized treatment for 7 years and post-intervention follow-up for 4 years, showed that the women randomized to CEE and 50–59 years old had significant reductions in CHD (HR, 0.59; 95% CI, 0.38–0.90), total MI (HR, 0.54; 95% CI, 0.34–0.86) and total mortality (HR, 0.73; 95% CI, 0.53–1.00) [23]. Compared with women who were 60–69 and 70–79 years old, the p-for-interaction was statistically significant for CHD (p = 0.05), total MI (p = 0.007) and total mortality (p = 0.04), indicating that the CEE effect on these outcomes differs by age [23]. Importantly, invasive breast cancer was statistically significantly reduced 23% (HR, 0.77; 95% CI, 0.62–0.95) in the women randomized to CEE relative to placebo regardless of age at randomization [23]. Another line of evidence that HRT may reduce CHD events derives from 1064 women who participated in the WHI Coronary Artery Calcium (CAC) sub-study. In the WHI-CAC sub-study, women who were 50–59 years old and randomized to CEE showed less CAC at year 7 of the trial relative to women who received placebo [24]. There was no older age group in this sub-study to evaluate whether HRT influenced CAC when initiated in women more than 60 years old. Even though only one-third of the WHI trial cohorts were comprised of women less than 60 years old and fewer than 5% of the women randomized to WHI were 1 or 2 years postmenopausal, this small subgroup showed a reduction in CHD events and overall mortality with HRT compared to placebo similar to the populations of women in observational studies who typically started HRT at the time of menopause. On the other hand, women randomized to HRT who were more than 60 years old and more than 10 years postmenopausal and not similar to the populations of women in observational studies had no reductions in CHD events or overall mortality with HRT relative to placebo [11].

4. Randomized HRT trial in perimenopuasal and recently postmenopausal women Of great importance to our understanding of long-term use of HRT after initiation in close proximity to menopause is comparison of clinical outcomes in healthy young perimenopausal and recently postmenopausal women in the Danish Osteoporosis Prevention Study (DOPS) [20]. DOPS is the only prospective longitudinal randomized trial designed to examine clinical outcomes among women who were specifically a priori randomized to HRT in the perimenopausal/early postmenopausal period. End points were determined with a PROBE (Prospectively, Randomized, Open with Blinded Endpoint evaluation) design. These results included 1006 women who averaged 50 years old and 7 months postmenopausal when randomized for 10 years to oral 17␤-estradiol plus sequential norethisterone acetate or to an untreated group. Hysterectomized women received oral 17␤-estradiol 2 mg daily. Subsequent to randomized treatment, the women were followed for another 6 years for a total follow-up of 16 years. After 10 years of randomized treatment, the pre-specified composite trial end point of overall mortality, MI or heart failure (HF) was significantly reduced by 52% (HR, 0.48; 95% CI, 0.27–0.89), overall mortality was decreased by 43% (HR, 0.57; 95% CI, 0.30–1.08) and stroke was decreased by 23% (HR, 0.77; 95% CI, 0.35–1.70) in the HRT group compared with the control group. In addition, breast cancer was decreased by 42% (HR, 0.58; 95% CI, 0.27–1.27) in the HRT group relative to the control group [20]. After 10 years of randomized treatment, 3 women had confirmed deep vein thrombosis (DVT) (2 in HRT group and 1 in control group; HR, 2.01; 95% CI, 0.18–22.16); one woman in the control group was hospitalized with pulmonary embolism (PE). After 10 years of randomized treatment and 6 years of post-randomization follow-up (16 years of total

3

follow-up), the composite trial end point continued to be significantly reduced by 39% (HR, 0.61; 95% CI, 0.39–0.94), total mortality decreased by 34% (HR, 0.66; 95% CI, 0.41–1.08) and stroke decreased by 11% (HR, 0.89; 95% CI, 0.48–1.65) among women originally randomized to HRT relative to those women randomized to the control group. Breast cancer remained reduced by 10% (HR, 0.90; 95% CI, 0.52–1.57) among the women originally randomized to HRT relative to those women randomized to the control group. After 10 years of randomized treatment and 6 years of post-randomization follow-up, 9 women had confirmed DVT (4 in HRT group and 5 in control group; HR, 0.80; 95% CI, 0.22–2.99); four women were hospitalized with PE, 1 in HRT group and 3 in control group (HR, 0.33; 95% CI, 0.04–3.21) [20]. The results of DOPS are remarkably similar to the 11-year WHI CEE trial follow-up data among women 50–59 years of age and randomized to WHI (41% decrease in CHD and 27% decrease in overall mortality) [23] and to the 32% decrease in CHD [16] and 39% decrease in total mortality [25] shown in meta-analyses of RCTs of women less than 60 years old and/or less than 10 years postmenopausal when randomized to HRT relative to placebo (Table 1).

5. Timing hypothesis supported by additional estrogen receptor-binding ligands In addition to mammalian hormones, agents other than HRT that bind to the estrogen receptor exert similar CHD beneficial effects in younger relative to older postmenopausal women. In the Raloxifene Use for the Heart (RUTH) trial (n = 10,101 women), women randomized to raloxifene (a selective estrogen receptor modulator) and aged 60–70 years (HR, 1.06; 95% CI, 0.88–1.28) and older than 70 years (HR, 0.98; 95% CI, 0.82–1.17) had a null effect relative to placebo on CHD events after an average treatment period of 5.6 years [26]. However, among women randomized to raloxifene and less than 60 years old, CHD was statistically significantly decreased by 41% (HR, 0.59; 95% CI, 0.41–0.83) relative to placebo [26]. The results of RUTH are similar to DOPS [20] as well as similar to WHI among women randomized to CEE and 50–59 years old and followed for 11 years [23] as well as similar to the decrease in CHD shown in meta-analyses of RCTs of women less than 60 years old and/or less than 10 years postmenopausal when randomized to HRT relative to placebo [16] (Table 1). Although age and years-since-menopause was not considered, randomization for 5 years to the SERM lasofoxifene compared with placebo, significantly reduced CHD by 32% (HR, 0.68; 95% CI, 0.50–0.93) in a RCT of 8556 postmenopausal women with osteoporosis who were 59–80 years old [27]. In the Women’s Isoflavone Soy Health (WISH) study, women randomized to high-dose isoflavone soy protein supplementation and less than 5 years postmenopausal had a statistically significant decrease in subclinical atherosclerosis progression relative to placebo whereas isoflavone supplementation among women 5 or more years postmenopausal when randomized had no significant effect [28]. Isoflavones are plant estrogens that preferentially bind to estrogen receptor-beta. The totality of randomized trial data consistently show that HRT decreases CHD incidence and overall mortality in women less than 60 years old who start HRT when less than 10 years postmenopausal [11] (Table 1). These findings are similar to the consistency of CHD and overall mortality reduction reported from observational studies in which most women started HRT within 6 years of menopause [1–10]. Focused in young healthy postmenopausal women on average 50 years old and 7 months postmenopausal when randomized, DOPS provides strong evidence for the long-term efficacy and safety of HT for reducing CHD and overall mortality [20].

Please cite this article in press as: H.N. Hodis, W.J. Mack, Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis, J. Steroid Biochem. Mol. Biol. (2013), http://dx.doi.org/10.1016/j.jsbmb.2013.06.011

G Model SBMB-4010; No. of Pages 8

ARTICLE IN PRESS H.N. Hodis, W.J. Mack / Journal of Steroid Biochemistry & Molecular Biology xxx (2013) xxx–xxx

4

Table 1 Studies reporting coronary heart disease and total mortality in women initiating hormone replacement or selective estrogen receptor modulator (SERM) therapy before age 60 years and/or within 10 years of menopause. Studies

Age (years); time-since-menopause

Coronary heart disease % reduction (risk ratio; 95% confidence interval)

Total mortality % reduction (risk ratio; 95% confidence interval)

DOPS, 10 year [20] DOPS, 16 year [20] WHI-E, 11-year [23] WHI-E [22] WHI-E + P [22] WHI-HRT [22] WHI-E [22] WHI-E + P [22] WHI-HRT [22] Meta-analysis [16] Meta-analysis [25] Bayesian meta-analysis [42] RUTH-SERM [26] Observational studies [1–10]

M = 50; M = 7 months-since-menopause

↓52% (0.48; 0.27–0.89) ↓49% (0.61; 0.39–0.94) ↓41% (0.59; 0.38–0.90) ↓52% (0.48; 0.20–1.17) ↓12% (0.88; 0.54–1.43) ↓24% (0.76; 0.50–1.16) ↓37% (0.63; 0.36–1.09) ↑29% (1.29; 0.79–2.12) ↓7% (0.93; 0.65–1.33) ↓32% (0.68; 0.48–0.96)

↓43% (0.57; 0.30–1.08) ↓34% (0.66; 0.41–1.08) ↓27% (0.73; 0.53–1.00) ↓35% (0.65; 0.33–1.29) ↓19% (0.81; 0.52–1.24) ↓24% (0.76; 0.53–1.09) ↓29% (0.71; 0.46–1.11) ↓31% (0.69; 0.44–1.07) ↓30% (0.70; 0.51–0.96)

<60 <10 YSM <10 YSM <10 YSM <60 <60 <60 <60; <10 YSM M = 54 M = 55 <60 30–55; <5 YSM

↓39% (0.61; 0.39–0.95) ↓27% (0.73; 0.52–0.96) ↓41% (0.59; 0.41–0.83) ↓30–50%

↓30–50%

10 year = 10 years of randomized treatment; estradiol + sequential norethisterone acetate and estradiol therapy. 16 year = 10 years of randomized treatment + 6 years of post-intervention follow-up. 11 year = 7 years of randomized treatment + 4 years of post-intervention follow-up. E = conjugated equine estrogen therapy (CEE). E + P = conjugated equine estrogen + medroxyprogesterone acetate (CEE + MPA), daily continuous combined therapy. HRT = combined CEE and CEE + MPA analysis. SERM = selective estrogen receptor modulator, raloxifene. M = mean. YSM = years-since-menopause.

6. Animal studies and the timing hypothesis Other lines of evidence, including those from studies in animal models support the timing hypothesis. Examining three independently designed cynomolgus monkey studies of surgical menopause and unopposed estrogen therapy the beneficial anti-atherosclerosis effects of HT appear to depend on timing of initiation of therapy relative to menopause as well as the underlying pathology of the arterial wall. The first set of studies was conducted in monkeys fed normal monkey chow before bilateral oophorectomy and thus had no atherosclerosis at the time of menopause. Immediately following ovariectomy, the monkeys were fed a diet to induce atherosclerosis and started on estrogen therapy. After 2 years of estrogen therapy and while receiving an atherogenic diet, there was 70% less atherosclerosis in this group of monkeys compared with a control group of monkeys who also underwent bilateral oophorectomy and were fed a diet to induce atherosclerosis but given a placebo [29,30]. The second study included monkeys that were fed an atherogenic diet before oophorectomy and thus had atherosclerosis at the time of menopause. Immediately following ovariectomy, the monkeys were started on estrogen therapy and the atherosclerosis-inducing diet was maintained. After 2 years of estrogen therapy and the continued atherogenic diet after menopause, there was 50% less atherosclerosis in this group of monkeys compared with a control group of monkeys that also underwent bilateral oophorectomy and were fed a diet to induce atherosclerosis before and after menopause but given a placebo [31]. The third study included monkeys fed a diet to induce atherosclerosis immediately following bilateral oophorectomy. In this study however, the monkeys were deprived of estrogen therapy for 2 years following menopause. Two years of life for cynomolgus monkeys is roughly the same as 6 years of human life. The atherogenic diet was stopped at the time estrogen therapy was started. Although the atherogenic diet was stopped and the monkeys received estrogen therapy for 2 years, the degree of atherosclerosis in this group of monkeys was similar to that of monkeys given placebo in whom the atherogenic diet was also stopped [32]. Collectively, these monkey studies indicate that a delay in administering estrogen therapy following menopause (2 years in these monkey studies) does not alter the course of atherogenesis. However, when administered immediately following menopause and whether atherosclerosis is present or not at the time of menopause, there is a decrement in the degree

of atherosclerosis relative to a control group of monkeys treated under the same conditions. Similar findings as to the monkey studies have been reported using an apoE-deficient mouse model. In the apoE-deficient mouse study, estrogen therapy did not reduce atherosclerosis in mice with previously established atherosclerosis relative to control mice. However, in the same apoE-deficient mouse study, estrogen therapy prevented the development of atherosclerosis when started concomitantly with the atherogenic diet [33]. This study showed that estrogen therapy was able to prevent the development of atherosclerosis but unable to reduce atherosclerosis once established. Animal studies provide a consistent pathobiological explanation for the discordance between observational studies and randomized trials. Animal studies show that the efficacy of estrogen in preventing atherosclerosis is dependent upon the time when estrogen is initiated in relation to menopause as well as to the underlying extent of preexisting atherosclerosis. Although the underlying mechanisms of these findings remain unknown, animal and human studies provide support for the concept of the healthy endothelium. Together, animal and human studies show that the vascular benefits of estrogen are mediated through estrogen receptors and that these benefits depend upon the integrity of the vascular endothelium and its functional state. Aging and atherosclerosis result in vascular wall damage and loss of estrogen receptors and/or function compromising the functional status of the endothelium [34]. Estrogen receptor function appears to be lost through methylation of the promotor site of the estrogen alpha receptor gene [35]. A decrease in the expression of estrogen receptors and/or loss of estrogen receptor function with aging and atherosclerosis may result in the reduction of the responsiveness of vascular endothelium to estrogen [34,35]. Since estrogen receptor expression is dependent upon circulating estrogen levels, estrogen receptors in vascular endothelial and vascular smooth muscle cells begin disappearing soon after menopause when circulating estrogen levels drop. It appears that maintaining premenopausal circulating estrogen levels is required to preserve both number and function of estrogen receptors after menopause, a requisite for vascular benefits of estradiol therapy. 7. Stroke The preponderance of observational studies indicates that stroke is either reduced or not affected by HRT use [36]. Approximately 8 observational studies indicate reduced stroke risk with

Please cite this article in press as: H.N. Hodis, W.J. Mack, Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis, J. Steroid Biochem. Mol. Biol. (2013), http://dx.doi.org/10.1016/j.jsbmb.2013.06.011

G Model SBMB-4010; No. of Pages 8

ARTICLE IN PRESS H.N. Hodis, W.J. Mack / Journal of Steroid Biochemistry & Molecular Biology xxx (2013) xxx–xxx

HRT and 23 observational studies indicate no relationship between HRT use and stroke. There are 4 observational studies that show an elevated stroke risk with HRT. The most recent observational study, from WHI, showed a 48% reduction in stroke risk in women who use HRT relative to those women who do not use HRT [7]. The Women’s Estrogen for Stroke Trial (WEST) has been the only RCT specifically designed to study the effects of HRT on stroke risk in which stroke was a priori planned as the primary trial outcome [37]. Other trials however, have yielded data concerning the effects of HRT on stroke risk where stroke was considered as an additional outcome. In WEST, 664 postmenopausal women who were ischemic stroke survivors and of median age 71 years and 20 years postmenopausal when randomized, oral 17␤-estradiol 1 mg daily had a null effect on the combined end point nonfatal stroke or overall mortality relative to placebo (RR, 1.1; 95% CI, 0.8–1.4) [37]. Although the women had non-disabling stroke or transient ischemic attack before randomization to WEST (secondary stroke prevention), the effect of HRT on nonfatal and fatal stroke as well as both strokes combined was null in this trial of women at high risk for recurrent stroke [37]. Results from HERS showed that CEE + MPA was associated with a non-significant increase in primary stroke risk (HR, 1.23; 95% CI, 0.89–1.70) in postmenopausal women with established CHD and who were on average 68 years old at randomization [38]. In the CEE + MPA WHI trial, there was nominal statistical significance of 8 additional strokes per 10,000 women per year of CEE + MPA therapy compared with placebo (HR, 1.31; 95% CI, 1.02–1.68) [22]; however, the difference in stroke risk between treatment groups was not statistically significant in the a priori defined outcome adjusting for multiple testing across time and outcomes (HR, 1.31; 95% CI, 0.93–1.84) [39]. In the CEE WHI trial, there was nominal statistical significance of 11 additional strokes per 10,000 women per year of CEE therapy compared with placebo (HR, 1.33; 95% CI, 1.05–1.68) [22]; however, the difference in stroke was not statistically significant in the a priori defined outcome adjusting for multiple testing across time and outcomes (HR, 1.39; 95% CI, 0.97–1.99) [18]. Women randomized to the WHI trials were on average 64 years old and more than 12 years postmenopausal. In a subset of 1403 WHI participants 65–79 years old who underwent magnetic resonance brain imaging on average 8 years following randomization to CEE and CEE + MPA, no significant difference in ischemic lesion volume was found between women randomized to CEE + MPA or CEE relative to those women who received placebo [40]. As such, this neuroradiologic sub-study provides no evidence within or across WHI trials by treatment assignment that HRT exposure increases ischemic lesion volume [40]. On the other hand, ischemic lesion volume was associated with several vascular risk factors including age, smoking, history of cardiovascular disease and hypertension, indicating internal validity of the neuroradiologic outcomes and stroke risk factors in WHI [40]. It is noteworthy that the association of CEE and CEE + MPA therapy with stroke discussed above is predominantly reported amongst older women (average ages 64–68 years when randomized) who were more than 10 years postmenopausal (average 12–20 years) when initiating these therapies. Importantly, cases of stroke are rare among women who start HRT when less than 60 years old [22]. WHI showed that when CEE + MPA is initiated in women less than 60 years of age, there are 5 additional strokes per 10,000 women per year of CEE + MPA therapy (HR, 1.41; 95% CI, 0.75–2.65); when CEE therapy is initiated in women less than 60 years of age, there are 2 fewer strokes per 10,000 women per year of CEE therapy (HR, 0.89; 95% CI, 0.47–1.69) [22]. With both the CEE and CEE + MPA WHI trials combined, there are 2 additional strokes per 10,000 women per year of CEE and CEE + MPA therapy when initiated in women less than 60 years of age (HR, 1.13; 95% CI, 0.73–1.76) [22]. These results are similar to the Nurse’s

5

Health Study, which showed in women less than 60 years old, 2.2 additional strokes per 10,000 women per year of HRT (assuming a relative risk of 1.4 for stroke in current HRT users compared with never-users) [41]. DOPS is consistent with these and the WEST findings as oral 17␤-estradiol plus sequential norethisterone acetate and unopposed 17␤-estradiol 2 mg daily initiated on average 7 months from menopause reduced stroke by 23% (HR, 0.77; 95% CI, 0.35–1.70) compared with the control group accounting for 6 fewer strokes per 10,000 women per year of 17␤-estradiol plus sequential norethisterone acetate and unopposed 17␤-estradiol therapy after 10 years of randomization. After 16 years of total follow-up (10 years of randomized treatment and post-randomization follow-up for 6 years), stroke was reduced by 11% (HR, 0.89; 95% CI, 0.48–1.65) in the women originally randomized to HRT compared with the control group in DOPS [20]. The preponderance of observational studies does not support an association between HRT use and stroke risk. The CEE and CEE + MPA WHI trials support these findings in women less than 60 years old and/or less than 10 years postmenopausal when initiating HRT. Brain imaging studies show no difference in ischemic lesion volume between CEE- or CEE + MPA- and placebo-treated women. Randomized trials show no risk of stroke with oral 17␤-estradiol plus sequential progestin therapy and with oral 17␤-estradiol alone, and DOPS shows no increased risk of stroke after 10 years of randomization to HRT and no risk of stroke with 16 years of total follow-up.

8. Total mortality Although the benefits and risks of HRT continue to be debated, it is important to realize that postmenopausal HRT is the only primary prevention therapy of CHD that reduces total mortality and extends life in women [11]. The positive HRT effect on overall mortality varies by age as shown in a meta-analysis of 30 RCTs with 119,118 women-years [25]. When analyzed across the entire age range in these trials, overall mortality was null (HR, 0.98; 95% CI, 0.87–1.18) as was the effect of HRT when started in women greater than 60 years old (average 66 years old) (HR, 1.03; 95% CI, 0.91–1.16) [25]. On the other hand, when started in women less than 60 years old (average 54 years old), overall mortality is decreased 39% with HRT compared with placebo (RR, 0.61; 95% CI, 0.39–0.95) [25]. The degree of reduction in overall mortality in women less than 60 years old after randomization to HRT is similar to the decrease in overall mortality observed in populations of women who started HRT in close proximity to menopause (typically at or within 2 years of menopause) in observational studies [1–10]. Similar to CHD trends, overall mortality in WHI was reduced 30% with HRT among 50–59 year old women [22]. Compared with placebo, there was a 29% (HR, 0.71; 95% CI, 0.46–1.11) decreased overall mortality risk in the women who were randomized to CEE and less than 60 years old. On the other hand, compared with placebo there was no overall mortality benefit in the women who were randomized to CEE and 60–69 years old (HR, 1.02; 95% CI, 0.64–1.44) or when randomized to CEE and 70–79 years old (HR, 1.20; 95% CI, 0.93–1.56) [22]. Compared with placebo, there was a 31% (HR, 0.69; 95% CI, 0.44–1.07) decreased overall mortality risk in the women who were randomized to CEE + MPA and less than 60 years old. On the other hand, compared with placebo there was no overall mortality benefit in the women who were randomized to CEE + MPA and 60–69 years old (HR, 1.09; 95% CI, 0.83–1.44) or when randomized to CEE + MPA and 70–79 years old (HR, 1.06; 95% CI, 0.80–1.41) [22]. In both trials combined, the women randomized to CEE and CEE + MPA therapies and less than 60 years old had a statistically significant 30% (HR, 0.70; 95% CI, 0.51–0.96) reduced overall mortality risk compared with placebo. On the other hand,

Please cite this article in press as: H.N. Hodis, W.J. Mack, Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis, J. Steroid Biochem. Mol. Biol. (2013), http://dx.doi.org/10.1016/j.jsbmb.2013.06.011

G Model SBMB-4010; No. of Pages 8 6

ARTICLE IN PRESS H.N. Hodis, W.J. Mack / Journal of Steroid Biochemistry & Molecular Biology xxx (2013) xxx–xxx

compared with placebo there was no overall mortality benefit in the women who were randomized to CEE and to CEE + MPA and 60–69 years old (HR, 1.05; 95% CI, 0.87–1.26) or when randomized to CEE and to CEE + MPA and 70–79 years old (HR, 1.14; 95% CI, 0.94–1.37) (22). To address HRT benefits and risks, randomized trials and observational studies that evaluated overall mortality in young postmenopausal women starting HRT at or near menopause were analyzed in a Bayesian meta-analysis [42]. This meta-analysis included 19 randomized trials comprised of 16,283 women who were on average 54.5 years old and studied for 83,043 womenyears over a period of 1–6.8 years (average, 5.1 years). Results showed a 27% (RR, 0.73; 95% Credible Interval (CrI), 0.52–0.96) decrease in overall mortality in the women randomized to HRT compared with placebo [42]. Results from pooling the data from 8 independent prospective observational studies comprised of 212,717 women studied for 2,935,495 women-years over a period of 6–22 years (average, 13.8 years) showed a 22% (RR, 0.78; 95% CrI, 0.69–0.90) decrease in overall mortality in women who used HRT compared with women who did not use HRT [42]. Combining data from the randomized trials and prospective observational studies shows that overall mortality was reduced 28% (RR, 0.72; 95% CrI, 0.62–0.82) [42]. These Bayesian meta-analysis results show that data from several sources converge providing evidence for a beneficial effect of HRT on overall mortality in women who start HRT at or near menopause. Importantly, these Bayesian meta-analysis results strongly show that randomized trial data and observational study data are consistent, each showing an approximate 25% decrease in overall mortality. These Bayesian meta-analysis results are consistent with results from the 10-year randomized trial data from DOPS (43% reduction in overall mortality) and to the 16-year total follow-up of DOPS (34% reduction in overall mortality) [20], similar to the 11-year follow-up of WHI CEE (27% reduction in overall mortality) [23] and consistent with the statistically significant 30% reduction (HR, 0.70; 95% CI, 0.51–0.96) in overall mortality demonstrated in postmenopausal women randomized to HRT relative to placebo and less than 60 years old in the CEE and CEE + MPA WHI trials [22] (Table 1). 9. Postmenopausal hormone therapy cost-effectively extends life A cost-effectiveness analysis indicates that compared with no therapy, HRT started in postmenopausal women in the 50 year old age range and continued for 5–30 years yields a significant increase in 1.5 quality-adjusted life years (QALY) at $2438 per QALY gained [43]. This analysis shows that the data are robust for younger postmenopausal women since HRT remained highly cost effective in all of the sensitivity analyses [43]. Cost effective ratios <$50,000 per QALY gained are considered worthwhile and those <$5000 per QALY gained are considered highly cost effective while a cost effective ratio >$100,000 per QALY gained is considered unattractive [44]. As such, HRT is a highly cost effective therapy for increasing QALYs in young postmenopausal women. In comparison with younger postmenopausal women, starting HRT in postmenopausal women older than 65 years results in 0.11 QALYs costing $27,953 per QALY gained [43]. 10. Testing the “timing” hypothesis In the wake of early trial results showing discordance between RCTs and observational studies, the Early versus Late Intervention Trial with Estradiol (ELITE; clinicaltrials.gov NCT00114517) was funded by the National Institutes of Health; enrollment initiated in 2004. ELITE is specifically designed to study the “timing”

hypothesis. Stratified by time-since-menopause, 643 postmenopausal women were randomized to a single-center, placebo-controlled, double-blind trial. In a 2 × 2 design, women without a history or symptoms of cardiovascular disease were randomized within two strata, <6 years- and >10 postmenopausal to oral estradiol 1 mg per day or placebo with progesterone vaginal gel or placebo for 10 days each month. Subclinical atherosclerosis progression determined every 6 months as carotid artery intimamedia thickness is the clinical trial primary outcome. Cognitive decline determined across 7 domains is the clinical trial secondary outcome. Because of the extensive accumulation of evidence since 2003 supporting the original ELITE application to the NIH for the study of the “timing” hypothesis (111,415), the NIH awarded a 3year ELITE extension. The ELITE extension has three specific aims: (1) to increase the randomized treatment period; (2) addition of another vascular imaging outcome consisting of non-contrast and contrast cardiac computed tomography to non-invasively measure coronary artery calcium and coronary artery lesions; and, (3) to extend measurement of cognitive performance with the addition of a third cognitive end point. ELITE trial results are planned for 2013.

11. Discussion Postmenopausal HRT provides the most consistent data for reducing CHD as well as overall mortality in women (Table 1). Cumulated data consistently show that starting HRT in women less than 60 years old and/or less than 10 years postmenopausal reduces CHD events and overall mortality, is associated with rare risk [45], extends life and at approximately $2400 per QALY gained, is extremely cost-effective [43]. The evidence-based data are large and consistent across 40–50 observational studies and meta-analyses encompassing 20–30 RCTs. In addition, the risks and their magnitude caused by standard primary CHD prevention therapies (statins and aspirin) are similar to those associated with HRT [45]. Importantly, the risks associated with HRT are predominantly rare and more rare when HRT is started in women less than 60 years old and/or less than 10 years postmenopausal [11,45]. For example, consistent with the cumulated data that predominantly show a lack of association of HRT with stroke, the additional absolute risk associated with HRT is rare when considered across all ages and more rare when started in women less than 60 years old and/or less than 10 years postmenopausal. The risk in postmenopausal women less than 60 years old who start HRT within 10 years of menopause approximates 2 additional stroke per 1000 women over 10 years of CEE or CEE + MPA therapy. Randomized trials show no stroke risk with oral 17␤-estradiol therapy with or without an added progestin for up to a total follow-up of 16 years. Benefits and risks of HRT vary by dosage, regimen and timing of initiation. As such, broad sweeping conclusions concerning HRT benefits and risks are not possible and attempts to generalize risk as comparable to daily continuous combined CEE + MPA (considered the “worse-case-scenario” for risks although risks associated with CEE + MPA are rare) have resulted in misleading and inaccurate information concerning HRT. Although the benefits and risks of HRT continue to be debated, one evidence-based fact is indisputable, HRT reduces overall mortality 30% when started in young postmenopausal women and continued long-term. Data show that when initiated in women in their 50s and continued for 5–30 years, HRT reduces overall mortality and cost-effectively increases QALYs. Timing of initiation of primary prevention therapy appears to have significant biological and clinical consequences for women. The timing of initiation of primary prevention therapies provides opportunity for CHD event reduction as well as reduction in overall mortality throughout the postmenopausal period and forges a new

Please cite this article in press as: H.N. Hodis, W.J. Mack, Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis, J. Steroid Biochem. Mol. Biol. (2013), http://dx.doi.org/10.1016/j.jsbmb.2013.06.011

G Model SBMB-4010; No. of Pages 8

ARTICLE IN PRESS H.N. Hodis, W.J. Mack / Journal of Steroid Biochemistry & Molecular Biology xxx (2013) xxx–xxx

paradigm for CHD primary prevention in women. It is important to rethink the appropriate clinical application of the evidence-based data; that is, reduction of CHD and overall mortality with rare risks in women who initiate HRT when less than 60 years old and/or less than 10 years postmenopausal. This is especially important since evidence-based data shows that neither statins nor aspirin reduce CHD or overall mortality in women when employed for CHD primary prevention [11]. Women older than 60 years and/or more than 10 years postmenopausal probably have missed the window of opportunity for benefit from HRT and should probably not initiate HRT for CHD primary prevention. DOPS, the only prospective longitudinal randomized trial conducted specifically in women less than 60 years of age (average age = 50 years) and less than 10 years postmenopausal (average = 7 months) with cardiovascular disease and breast cancer outcomes, provides direct and compelling evidence that the benefits of prevention of chronic diseases outweigh the risks, including long-term HRT for >15 years. In fact, the women randomized to DOPS specifically represent the population of women included in observational studies and hence, DOPS is the only randomized trial to date to appropriately test the estrogen cardioprotective hypothesis among the very population of women from which the cardioprotective hypothesis was generated. DOPS also represents the clinical situation in which women are started on HRT. In conclusion, a large body of randomized trial data including WHI data converges with results from observational studies, animal studies and basic science supporting that HRT health outcomes vary by time-since-menopause. DOPS provides direct randomized trial evidence for the beneficial consequences of initiating HRT at or near menopause and continuing HRT long-term with low risk. Inconsistencies in presentation and interpretation of HRT data has created great confusion for health-care providers and patients alike, culminating in the call for an independent commission to evaluate the interpretation and dissemination of the evidence-based data in relation to public health recommendations [46]. Health-care providers and patients can use the cumulated data in making clinical decisions concerning chronic disease prevention including osteoporosis and downstream morbidity and mortality from bone fractures [47] keeping in mind that any prevention strategy must be personalized. The data not only provide strong consistent evidence for the beneficial effects of HRT when initiated in close proximity to menopause but also reassurance of the long-term safety for HRT.

References [1] F. Grodstein, M. Stampfer, The epidemiology of coronary heart disease and estrogen replacement in postmenopausal women, Progress in Cardiovascular Diseases 38 (1995) 199–210. [2] F. Grodstein, M. Stampfer, Estrogen for women at varying risk of coronary disease, Maturitas 30 (1998) 19–26. [3] S.G. Thompson, T.W. Meade, G. Greenberg, The use of hormonal replacement therapy and the risk of stroke and myocardial infarction in women, Journal of Epidemiology & Community Health 43 (1989) 173–178. [4] M. Falkeborn, I. Persson, H.O. Adami, The risk of acute myocardial infarction after oestrogen and oestrogen–progestogen replacement, An International Journal of Obstetrics & Gynaecology 99 (1992) 821–828. [5] B.M. Psaty, S.R. Heckbert, D. Atkins, R.N. Lemaitre, T.D. Koepsell, P.W. Wahl, D.S. Siscovick, E.H. Wagneret, The risk of myocardial infarction associated with the combined use of estrogens and progestins in postmenopausal women, Archives of Internal Medicine 154 (1994) 1333–1339. [6] F. Grodstein, M.J. Stampfer, J.E. Manson, G.A. Colditz, W.C. Willett, B. Rosner, F.E. Speizer, C.H. Hennekens, Postmenopausal estrogen and progestin use and the risk of cardiovascular disease, The New England Journal of Medicine 335 (1996) 453–461. [7] R.L. Prentice, R. Langer, M.L. Stefanick, B.V. Howard, M.L. Pettinger, G.L. Anderson, D. Barad, J.D. Curb, J. Kotchen, L. Kuller, M. Limacher, J. Wactawski-Wende, Combined postmenopausal hormone therapy and cardiovascular disease: toward resolving the discrepancy between observational studies and the Women’s Health Initiative clinical trial, American Journal of Epidemiology 162 (2005) 404–414.

7

[8] R.L. Prentice, R. Langer, M.L. Stefanick, B.V. Howard, M. Pettinger, G.L. Anderson, D. Barad, J.D. Curb, J. Kotchen, L. Kuller, M. Limacher, J. Wactawski-Wende, Combined analysis of Women’s Health Initiative observational and clinical trial data on postmenopausal hormone therapy and cardiovascular disease, American Journal of Epidemiology 163 (2006) 589–599. [9] B.D. Henderson, A. Paganini-Hill, R.K. Ross, Decreased mortality in users of estrogen replacement therapy, Archives of Internal Medicine 151 (1991) 75–78. [10] F. Grodstein, J.E. Manson, G.A. Colditz, W.C. Willett, F.E. Speizer, M.J. Stampfer, A prospective, observational study of postmenopausal hormone therapy and primary prevention of cardiovascular disease, Annals of Internal Medicine 133 (2000) 933–941. [11] H.N. Hodis, P. Collins, W.J. Mack, L. Schierbeck, The timing hypothesis for coronary heart disease prevention with hormone therapy: past, present and future in perspective, Climacteric 15 (2012) 217–228. [12] S. Hulley, D. Grady, T. Bush, C. Furberg, D. Herrington, B. Riggs, E. Vittinghoff, Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women, JAMA 280 (1998) 605–613. [13] D.M. Herrington, D.M. Reboussin, K.B. Brosnihan, P.C. Sharp, S.A. Shumaker, T.E. Snyder, C.D. Furberg, G.J. Kowalchuk, T.D. Stuckey, W.J. Rogers, D.H. Givens, D. Waters, Effects of estrogen replacement on the progression of coronary artery atherosclerosis, The New England Journal of Medicine 343 (2000) 522–529. [14] H.N. Hodis, W.J. Mack, R.A. Lobo, D. Shoupe, A. Sevanian, P.R. Mahrer, R.H. Selzer, C.R. Liu, C.H. Liu, S.P. Azen, Estrogen in the prevention of atherosclerosis: a randomized, double-blind, placebo-controlled trial, Annals of Internal Medicine 135 (2001) 939–953. [15] H.N. Hodis, W.J. Mack, S.P. Azen, R.A. Lobo, D. Shoupe, P.R. Mahrer, D.P. Faxon, L. Cashin-Hemphill, M.E. Sanmarco, W.J. French, T.L. Shook, T.D. Gaarder, A.O. Mehra, R. Rabbani, A. Sevanian, A.B. Shil, M. Torres, K.H. Vogelbach, R.H. Selzer, Hormone therapy and the progression of coronary artery atherosclerosis in postmenopausal women, The New England Journal of Medicine 349 (2003) 535–545. [16] S.R. Salpeter, J.M.E. Walsh, E. Greyber, E.E. Salpeter, Coronary heart disease events associated with hormone therapy in younger and older women: a metaanalysis, Journal of General Internal Medicine 21 (2006) 363–366. [17] Writing Group for the Women’s Health Initiative Investigators, Risks and benefits of estrogen plus progestin in healthy menopausal women: principal results from the Women’s Health Initiative randomized controlled trial, JAMA 288 (2002) 321–323. [18] The Women’s Health Initiative Steering Committee, Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the Women’s Health Initiative randomized controlled trial, JAMA 291 (2004) 1701–1712. [19] C. Rodriguez, E.E. Calle, A.V. Patel, L.M. Tatham, E.J. Jacobs, M.J. Thun, Effect of body mass on the association between estrogen replacement therapy and mortality among elderly US women, American Journal of Epidemiology 153 (2001) 145–152. [20] L.L. Schierbeck, L. Rejnmark, C.L. Tofteng, L. Stilgren, P. Eiken, L. Mosekilde, L. Kober, J.E.B. Jensen, Effect of hormone replacement treatment on cardiovascular events in recently postmenopausal women: randomized trial, BMJ 345 (2012) e6409. [21] J. Hsia, R.D. Langer, J.E. Manson, L. Kuller, K.C. Johnson, S.L. Hendrix, M. Pettinger, S.R. Heckbert, N. Greep, S. Crawford, C.B. Eaton, J.B. Kostis, P. Caralis, R. Prentice, For the women’s health initiative investigators. Conjugated equine estrogens and coronary heart disease: the Women’s Health Initiative, Archives of Internal Medicine 166 (2006) 357–365. [22] J.E. Rossouw, R.L. Prentice, J.E. Manson, L.L. Wu, D. Barad, V.M. Barnabei, M. Ko, A.Z. LaCroix, K.L. Margolis, M.L. Stefanick, Postmenopausal hormone therapy and risk of cardiovascular disease by age and years since menopause, JAMA 297 (2007) 1465–1477. [23] A.Z. LaCroix, R.T. Chlebowski, J.E. Manson, A.K. Aragaki, K.C. Johnson, L. Martin, K.L. Margolis, M.L. Stefanick, R. Brzyski, J.D. Curb, B.V. Howard, C.E. Lewis, J. Wactawski-Wende, For the WHI investigators. Health outcomes after stopping conjugated equine estrogens among postmenopausal women with hysterectomy: a randomized controlled trial, JAMA 305 (2011) 1305–1314. [24] J.E. Manson, M.A. Allison, J.E. Rossouw, J.J. Carr, R.D. Langer, J. Hsia, L.H. Kuller, B.B. Cochrane, J.R. Hunt, S.E. Ludlam, M.B. Pettinger, M. Gass, K.L. Margolis, L. Nathan, J.K. Ockene, R.L. Prentice, J. Robbins, M.L. Stefanick, For the WHI and WHI-CACS investigators. Estrogen therapy and coronary-artery calcification, The New England Journal of Medicine 356 (2007) 2591–2602. [25] S.R. Salpeter, J.M.E. Walsh, E. Greybe, T.M. Ormiston, E.E. Salpeter, Mortality associated with hormone replacement therapy in younger and older women: a meta-analysis, Journal of General Internal Medicine 19 (2004) 791–804. [26] P. Collins, L. Mosca, M.J. Geiger, D. Grady, M. Kornitzer, M.G. Amewou-Atisso, M.B. Effron, S.A. Dowsett, E. Barrett-Connor, N.K. Wenger, Effects of the selective estrogen receptor modulator raloxifene on coronary outcomes in the raloxifene use for the heart trial: results of subgroup analyses by age and other factors, Circulation 119 (2009) 922–930. [27] S.R. Cummings, K. Ensrud, P.D. Delmas, A.Z. LaCroix, S. Vukicevic, D.M. Reid, S. Goldstein, U. Sriram, A. Lee, J. Thompson, R.A. Armstrong, D.D. Thompson, T. Powles, J. Zanchetta, D. Kendler, P. Neven, R. Eastell, For the PEARL study investigators. Lasofoxifene in postmenopausal women with osteoporosis, The New England Journal of Medicine 362 (2010) 686–696. [28] H.N. Hodis, W.J. Mack, N. Kono, S.P. Azen, D. Shoupe, J. Hwang-Levine, D. Petitti, L. Whitfield, M. Yan, A.A. Franke, R.H. Selzer, For the WISH research group. Isoflavone soy protein supplementation and progression of subclinical atherosclerosis in healthy postmenopausal women: a randomized controlled trial, Stroke 42 (2011) 3168–3175.

Please cite this article in press as: H.N. Hodis, W.J. Mack, Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis, J. Steroid Biochem. Mol. Biol. (2013), http://dx.doi.org/10.1016/j.jsbmb.2013.06.011

G Model SBMB-4010; No. of Pages 8 8

ARTICLE IN PRESS H.N. Hodis, W.J. Mack / Journal of Steroid Biochemistry & Molecular Biology xxx (2013) xxx–xxx

[29] M.R. Adams, T.C. Register, D.L. Golden, J.D. Wagner, J.K. Williams, Medroxyprogesterone acetate antagonizes inhibitory effects of conjugated equine estrogens on coronary artery atherosclerosis, Arteriosclerosis, Thrombosis, and Vascular Biology 17 (1997) 217–221. [30] T.B. Clarkson, M.S. Anthony, C.P. Jerome, Lack of effect of raloxifene on coronary artery atherosclerosis of postmenopausal monkeys, The Journal of Clinical Endocrinology and Metabolism 83 (1988) 721–726. [31] T.B. Clarkson, M.S. Anthony, T.M. Morgan, Inhibition of postmenopausal atherosclerosis progression: a comparison of the effects of conjugated equine estrogens and soy phytoestrogens, The Journal of Clinical Endocrinology and Metabolism 86 (2001) 41–47. [32] J.K. Williams, M.S. Anthony, E.K. Honore, D.M. Herrington, T.M. Morgan, T.C. Register, T.B. Clarkson, Regression of atherosclerosis in female monkeys, Arteriosclerosis, Thrombosis, and Vascular Biology 15 (1995) 827–836. [33] M.E. Rosenfeld, K. Kauser, B. Martin-McNulty, P. Polinsky, S.M. Schwartz, G.M. Rubanyi, Estrogen inhibits the initiation of fatty streaks throughout the vasculature but does not inhibit intra-plaque hemorrhage and the progression of established lesions in apolipoprotein E deficient mice, Atherosclerosis 164 (2002) 251–259. [34] D.W. Losordo, M. Kearney, E.A. Kim, J. Jekanowski, J.M. Isner, Variable expression of the estrogen receptor in normal and atherosclerotic coronary arteries of premenopausal women, Circulation 89 (1994) 1501–1510. [35] W.S. Post, P.J. Goldschmidt-Clemont, C.C. Wilhide, A.W. Heldman, M.S. Sussman, P. Ouyang, E.E. Milliken, J.P.J. Issa, Methylation of the estrogen receptor gene is associated with aging and atherosclerosis in the cardiovascular system, Cardiovascular Research 43 (1999) 985–991. [36] A. Paganini-Hill, Hormone replacement therapy and stroke: risk, protection or no effect, Maturitus 38 (2002) 243–261. [37] C.M. Viscoli, L.M. Brass, W.N. Kernan, P.M. Sarrel, S. Suissa, R.I. Horwitz, A clinical trial of estrogen replacement therapy after ischemic stroke, The New England Journal of Medicine 345 (2001) 1243–1249.

[38] J.A. Simon, J. Hsia, J.A. Cauley, C. Richards, F. Harris, J. Fong, E. Barrett-Connor, S.B. Hulley, For the HERS Research Group. Postmenopausal hormone therapy and risk of stroke: the heart and estrogen-progestin replacement study (HERS), Circulation 103 (2001) 638–642. [39] S. Wassertheil-Smoller, S.L. Hendrix, M. Limacher, G. Heiss, C. Kooperberg, A. Baird, T. Kotchen, J.D. Curb, H. Black, J.E. Rossouw, A. Aragaki, M. Safford, E. Stein, S. Laowattana, W.J. Mysiw, For the WHI Investigators. Effect of estrogen plus progestin on stroke in postmenopausal women: the Women’s Health Initiative: a randomized trial, JAMA 289 (2003) 2673–2684. [40] L.H. Coker, P.E. Hogan, N.R. Bryan, L.H. Kuller, K.L. Margolis, K. Bettermann, R.B. Wallace, Z. Lao, R. Freeman, M.L. Stefanick, S.A. Shumaker, For the women’s health initiative memory study. Postmenopausal hormone therapy and subclinical cerebrovascular disease: the WHIMS-MRI study, Neurology 72 (2009) 125–134. [41] F. Grodstein, J.E. Manson, M.J. Stampfer, K. Rexrode, Postmenopausal hormone therapy and stroke: role of time since menopause and age at initiation of hormone therapy, Archives of Internal Medicine 168 (2008) 861–866. [42] S.R. Salpeter, J. Cheng, L. Thabane, N.S. Buckley, E.E. Salpeter, Bayesian metaanalysis of hormone therapy and mortality in younger postmenopausal women, American Journal of Medicine 12 (2009) 1016–1022. [43] S.R. Salpeter, N.S. Buckley, H. Liu, E.E. Salpeter, The cost-effectiveness of hormone therapy in younger and older postmenopausal women, American Journal of Medicine 122 (2009) 42–52. [44] D.J. Cohen, M.R. Reynolds, Interpreting the results of cost-effectiveness studies, Journal of the American College of Cardiology 52 (2008) 2119–2126. [45] H.N. Hodis, W.J. Mack, Postmenopausal hormone therapy in clinical perspective, Menopause 14 (2007) 944–957. [46] W.H. Utian, A decade post WHI, menopausal hormone therapy comes full circle–need for independent commission, Climacteric 15 (2012) 320–325. [47] E. Davies, N.P. Mangongi, C.L. Carter, Is timing everything? A meeting report of the society for women’s health research roundtable on menopausal hormone therapy, Journal of Women’s Health 22 (2013) 303–311.

Please cite this article in press as: H.N. Hodis, W.J. Mack, Hormone replacement therapy and the association with coronary heart disease and overall mortality: Clinical application of the timing hypothesis, J. Steroid Biochem. Mol. Biol. (2013), http://dx.doi.org/10.1016/j.jsbmb.2013.06.011