Effect of menopausal hormone therapy on components of the metabolic syndrome

Obesity

The increased body mass and adiposity that occur after menopause are difficult to differentiate from the effects of normal aging. Multiple cross-sectional studies have suggested that postmenopausal women exhibit increased total fat mass, increased abdominal fat, and decreased lean body mass (LBM) in comparison to premenopausal women independent of aging [Ley et al. 1992; Svendsen et al. 1995; Panotopoulos et al. 1996; Tremollieres et al. 1996; Guo et al. 1999; Sternfeld et al. 2005; Sowers et al. 2007]. A cross-sectional study in pre-, peri- and postmenopausal obese women showed that both post- and perimenopausal obese women had a higher truncal fat distribution than premenopausal women after adjustment for age and total fat mass [Panotopoulos et al. 1996]. Another cross-sectional study in healthy women using dual-energy X-ray absorptiometry (DEXA) showed an increase in abdominal fat and a decrease in lean tissue mass in perimenopausal years independent of age [Svendsen et al. 1995]. A study in which body fat distribution was measured using abdominal circumference and computerized tomography (CT) showed that visceral-to-subcutaneous abdominal-adipose-tissue-area ratios were significantly higher in postmenopausal women [Zamboni et al. 1992]. Using DEXA to define the effects of menopause on body composition in nonobese postmenopausal women, Ley and colleagues reported that menopause increases android fat (ventral or upper body fat) distribution and reduces gynoid fat (lower-body-segment fat) [Ley et al. 1992]. Lovejoy and colleagues showed that the menopausal transition is associated with an increase in total body fat and visceral abdominal tissue (VAT). The authors also observed that menopause onset was associated with decreased fat oxidation and energy expenditure that could explain the predisposition to obesity [Lovejoy et al. 2008]. Androgens also affect adiposity. A recent study evaluating the relationship between endogenous androgens and body fat distribution in early and late postmenopausal women suggested that the free testosterone in early postmenopausal women and dehydroepiandrosterone levels in late postmenopausal women, respectively, correlated positively, and therefore could influence abdominal fat accumulation [Cao et al. 2013]. The Study of Women’s Health Across the Nation (SWAN) Heart study looked at the relationship between cardiovascular fat (CF) deposits, menopausal status, and endogenous sex hormones [El Khoudary et al. 2015]. This study reported for the first time that late perimenopausal and postmenopausal women have significantly greater volumes of CF independent of age, race, obesity, physical activity, smoking, medications, alcohol, and comorbidity. As fat is a metabolically active organ secondary to accumulation of pro-inflammatory cells, CF could affect the heart vasculature mechanically and functionally and even may contribute to adiposity-related atherosclerosis [Iacobellis et al. 2008; El Khoudary et al. 2015; Wensveen et al. 2015]. Given such close connection between obesity and markers of cardiovascular disease, estrogen effect on both of these processes is of great importance in postmenopausal treatment. In summary, most studies agree that menopause is associated with increased visceral fat, independent of age.

Diabetes

Studies using rodent models conclusively show that that estrogen deficiency alters insulin sensitivity and predisposes to diabetes [Mauvais-Jarvis et al. 2013]. In women, however, the effect of menopause on diabetes risk independent of adiposity and aging is still unclear. In one study of early postmenopausal women, an increase in years since menopause conferred an increased risk of impaired glucose tolerance (IGT) of 6% per year after menopause [Wu et al. 2001]. Another large study including 46,239 adults in China reported a lower prevalence of diabetes in women than in men aged 60 years, whereas the prevalence increased in women compared with men between the ages of 60 and 70 [Yang et al. 2010]. The InterAct study, a prospective case-cohort study with a median follow up of 11 years, reported that women with early-onset menopause were also at higher risk of developing type 2 diabetes (T2D) [Brand et al. 2013].

Regarding the direct effect of menopause predisposing to diabetes, studies using the intravenous glucose tolerance test (IVGTT) followed by mathematical modeling of insulin sensitivity found that menopause was associated with IR or decreased noninsulin-dependent glucose disposal [Walton et al. 1993; Lindheim et al. 1994]. In contrast, another study using the hyperinsulinemic-euglycemic clamp technique observed no difference in IR in postmenopausal women [Toth et al. 2000]. A secondary data analysis of women who were not taking estrogen in The Postmenopausal Estrogen/Progestin Interventions (PEPI) study showed association between increased body mass index (BMI) and waist-to-hip ratio (WHR) with increased fasting glucose and insulin; they concluded that postmenopausal visceral fat accumulation accounts for the observed alteration in glucose homeostasis [Barrett-Connor et al. 1996]. In summary, although studies in rodent models clearly show that estrogen deficiency alters insulin sensitivity and predisposes to glucose intolerance independently of fat accumulation, researchers have not yet reached a consensus to determine how menopause alters glucose homeostasis.

Lipids

The mechanism behind changes in lipid metabolism during menopause is not clear. Similar to MetS, during the perimenopausal transition, there are alterations in lipid metabolism towards a more atherogenic profile with increased low-density lipoprotein (LDL) cholesterol and triglycerides (TGs), and decreased high-density lipoprotein (HDL) cholesterol. During the transition from premenopause to first year postmenopause, the changes in LDL cholesterol, TG, and BMI were larger than those between years 1 and 5 postmenopause [Matthews et al. 2001]. Multiple studies have shown that postmenopausal women have higher total cholesterol, LDL, and TG, and lower HDL [Campos et al. 1988; Jensen et al. 1990; Stevenson et al. 1993; Fukami et al. 1995; Li et al. 1996; Matthews et al. 2001; Anagnostis et al. 2015] as compared with premenopausal women. VAT excess is strongly associated with metabolic disorders such as IR and dyslipidemia [Van Pelt et al. 2002; Goodpaster et al. 2005], which may explain why these features emerge after menopause.

Hypertension (HTN)

Removal of circulating estrogen via ovariectomy significantly increases BP in multiple rodent models of hypertension [Sandberg and Ji, 2012]. The most striking effect is observed in the mRen2 congenic rat, where ovariectomy increases systolic blood pressure (SBP) by over 50 mmHg [Chappell et al. 2003]. Moreover, chronic estradiol (E2) treatment decreases BP in the ovariectomized (OVX) mRen2 rat to a similar extent as antihypertensive therapy. Ovariectomy exacerbates salt-sensitive hypertension in both the Dahl rat [Hinojosa-Laborde et al. 2004] and the spontaneously hypertensive rat [Fang et al. 2001]. In women, surgical menopause induced by total hysterectomy is similarly associated with a higher prevalence of hypertension [Howard et al. 2005] and increased salt-sensitivity of BP [Schulman et al. 2006]. Natural menopause is also associated with elevated BP, most likely a result of decreased circulating estradiol in combination with vascular stiffening [Staessen et al. 1997]. The incidence of hypertension continually increases in aging women and eventually exceeds that of age-matched men [Mozaffarian et al. 2015].

In summary, menopause is associated with the onset of all components of the MetS including VAT accumulation, atherogenic dyslipidemia, IR, and increased BP. However, in a prospective study of the menopausal transition, Matthews and colleagues reported that only LDL exhibited a postmenopausal increase. In contrast, blood glucose, insulin and BP exhibited a linear increase, indicative of chronological aging [Matthews et al. 2009].

Adiposity

Studies have reported beneficial effects of MHT on abdominal fat, WHR, LBM and weight. In recent years, CT and DEXA have made it possible to assess changes in body composition and fat distribution with more accuracy than the clinical measurement of WHR. Earlier studies using DEXA scans for measurement of body composition changes showed positive effects of MHT. A small, prospective, randomized, placebo-controlled study with 62 early-postmenopausal women followed up for 2 years showed that treatment with combined estrogen-progestogen prevented the increase in abdominal fat after menopause compared with placebo [Haarbo et al. 1991]. Another cross-sectional study in 712 postmenopausal female twins showed that MHT users had lower central fat [Samaras et al. 1998]. The Danish Osteoporosis Prevention Study (DOPS) (n = 2016), a 5-year randomized controlled clinical trial in early postmenopausal women aged 45–58 years, showed that women on MHT (E2, 2 mg and norethisterone acetate, 1 mg) gained less fat than controls [not on hormone replacement therapy (HRT)] [Jensen et al. 2003]. Similar findings of decreased intra-abdominal fat in women treated with MHT were seen in an observational, longitudinal study over 2 years [Gower et al. 2006]. It is difficult to make generalized statements regarding the effect of MHT on weight distribution and weight changes, as all of the studies mentioned above enrolled only subjects with a BMI less than 30 kg/m². However, one study in women with a mean BMI of 34 revealed that MHT users exhibited 10.4% lower body mass, 10.1% lower BMI, 13.2% lower fat mass, and 25.6% less VAT as compared with nonusers [Sites et al. 2001].

Although most randomized control trials (RCTs) did not specifically describe the effect of MHT-induced changes in visceral adiposity, they did review weight, BMI, WHR, and waist circumference (WC). The Women’s Health Initiative (WHI) [conjugated estrogen (CE), 0.625 mg/ medroxyprogesterone acetate (MPA), 2.5 mg] trial showed a small but statistically significant decrease in BMI (−0.19 ± 0.04, p < 0.01) and WC (0.77 ± 0.10 cm, p < 0.01) during the first year of treatment in women assigned to MHT [Margolis et al. 2004]. Similar results were found in the Heart and Estrogen/Progestin Replacement Study (HERS) (CE, 0.625 mg/MPA, 2.5 mg) where women assigned to MHT experienced slight but significant weight loss (–0.5 kg), decreased BMI (–0.2 kg/m²), and decreased abdominal obesity (WHR: –0.01; WC: –0.8 cm) during follow up compared with placebo [Kanaya et al. 2003]. All women in the PEPI trial [CE, 0.625 mg/d ± MPA, 2.5 mg or micronized progesterone (MP), 200 mg] gained weight [Miller et al. 1995]. However, the mean increase from baseline was smaller among women assigned to unopposed CE (mean 0.7 kg at 36 months) [Miller et al. 1995]. Of note, by the third year of the WHI trial, glucose and BMI differences were smaller or no longer statistically significant, likely due to smaller sample size and nonadherence to study medication [Margolis et al. 2004].

Post hoc analysis of SMART trials showed no significant increase in body mass or BMI in postmenopausal women receiving CE, 0.45  mg/ bazedoxifene (BZA), 20  mg or CE, 0.625  mg/BZA, 20  mg for up to 2 years in SMART trials; however greater increases in weight (p  =  0.015) and BMI (p  =  0.014) were seen with placebo versus CE, 0.625  mg/BZA, 20  mg at month 12 [Black et al. 2016].

While the above data demonstrate that HRT results in decreased BMI, weight, and fat, others have found no change [Walker et al. 2001; Ryan et al. 2002].

In conclusion, the most significant studies agree that MHT decreases abdominal fat (see Table 1 for summary of findings from RCTs).

Insulin resistance and diabetes

The Nurses’ Health Study (n = 21,028) was the first large prospective study to report that MHT users (estrogen alone, progesterone alone, or combination) had a 20% decreased risk of diabetes as compared with nonusers, after adjustment for age and BMI [Manson et al. 1992]. In the PEPI trial, fasting insulin (and glucose) levels were also decreased in women assigned to active treatment [Miller et al. 1995]. In a 1-year metabolic substudy of the WHI, osteoporosis, progestin and estrogen (Women’s HOPE, n = 749), a significant decrease in fasting glucose was also seen in women taking MHT [CE, 0.625 mg/MPA, 2.5 mg] [Lobo et al. 2001].

Impaired insulin sensitivity is a major factor predisposing to T2D. A meta-analysis of 107 RCTs reported that women with diabetes assigned to HRT had a reduction of IR [homeostatic model assessment (HOMA)-IR] of 35.8% compared with placebo or no treatment [Salpeter et al. 2006]. Women without diabetes assigned to HRT had smaller a reduction in HOMA-IR of 12.9% compared with controls. Furthermore, MHT reduced fasting glucose by 11.5% and fasting insulin by 20.2%. The relative ratio for developing diabetes was 0.7 (95% confidence interval, 0.6–0.9) indicating a 30% reduction in new-onset diabetes for those receiving MHT [Salpeter et al. 2006]. Further, MHT resulted in decreased incidence of diabetes in the two largest RCTs. In the HERS study, women randomly assigned to hormone therapy (CE, 0.625 mg/MPA, 2.5 mg) had a 35% lower risk for diabetes than those assigned to placebo [Kanaya et al. 2003]. In the WHI trial, treatment with MHT (0.625 mg CE plus 2.5 mg MPA) resulted in a 20% lower incidence of self-reported diabetes in predominantly healthy women [Margolis et al. 2004]. Consistent with the antidiabetic effect of MHT, in the postintervention follow up of the WHI trials, the diabetes risk reductions disappeared in the years after discontinuation of MHT [Manson et al. 2013]. Therefore, the most significant trials agree that MHT decreases the risk of developing diabetes. One should note, however, that although most trials reported MHT improves fasting glucose, postchallenge glucose was also slightly altered [Miller et al. 1995] (see Table 1 for summary of findings from RCTs).

Lipids

All major MHT trials using CE, including HERS, PEPI, and WHI confirmed that MHT produced a reduction in LDL, an increase in HDL, and an increase in TGs relative to placebo [Miller et al. 1995; Hulley et al. 1998; Rossouw et al. 2002]. Thus, in the HERS study (CE, 0.625 mg/MPA, 2.5 mg), when comparing the hormone group with placebo, mean LDL decreased by 11%, mean HDL increased by 6% and mean TG levels increased by 7% [Hulley et al. 1998]. The PEPI trial (CE, 0.625 mg ± MPA, 2.5 mg or MP, 200 mg) showed that all hormone treatments improved lipids compared with placebo [Miller et al. 1995]. HDL increased by an average 0.06 mmol/l in women treated with the least effective of the active regimens compared with placebo, and LDL decreased by an average 0.26 mmol/l in women with the least effective of the active regimens relative to placebo. In the WHI (CE, 0.625 mg/MPA, 2.5 mg), subsamples of fasting blood specimen were assessed in 8.6% of study participants that revealed a reduction in LDL by 12.7%, an increase in HDL by 7.3% and an increase in TGs by 6.9% with MHT, relative to placebo [Rossouw et al. 2002]. In a 1-year metabolic substudy of the Women’s HOPE Study, all hormone regimens (0.625 mg CE; 0.625 mg CE/2.5 mg MPA; 0.45 mg CE; 0.45 mg CE/2.5 mg MPA; 0.45 mg CE/1.5 mg MPA; 0.3 mg CE; 0.3 mg CE/1.5 mg MPA) increased HDL by 5–18% in different MHT groups, and decreased LDL by 1.8–10.9% [Lobo et al. 2001]. Finally, in a meta-analysis of 107 RCTs, overall MHT increased HDL by 5.1%, while decreasing LDL by 11% compared with placebo or no treatment [Salpeter et al. 2006]. In a subgroup analysis, these effects were more pronounced with oral agents compared with transdermal E2 and were dose dependent. Overall, oral agents increased TG levels by 6.0%, while transdermal E2 had no effect (see Table 1 for summary of findings from RCTs).

Hypertension

The effect of MHT on BP is controversial. Sublingual estradiol decreases peripheral resistance in menopausal women within minutes, to a greater extent in hypertensive versus normotensive women [Pines et al. 1998]. In a small placebo-controlled, randomized crossover study in which women were randomized to 17β-estradiol plus cyclic norethisterone acetate (NETA) or placebo, E2 had a modest BP-lowering effect that was amplified by NETA [Sorensen et al. 2000]. In a meta-analysis of more than 100 trials, overall MHT produced a small 1.7% reduction in mean BP [Salpeter et al. 2006]. In subgroup analysis, only oral CE reduced BP, while transdermal agents and oral esterified estrogens did not have any significant effects. In the WHI, after adjusting for age, BMI, and other risk factors, there was a greater likelihood for hypertension in current MHT users [Wassertheil-Smoller et al. 2000]. The PEPI trial did not show any significant effect of MHT on BP [Miller et al. 1995]. In contrast, the Baltimore Longitudinal Study on Aging found that over the 6 years of the trial, women taking MHT (oral or transdermal estrogen and progestin) had a smaller increase in SBP (1.6 mmHg) in comparison with nonusers (8.9 mmHg) [Scuteri et al. 2001]. The Estrogen in the Prevention of Atherosclerosis Trial (oral E2, 1mg) showed a neutral effect of MHT on BP, with no difference in BP changes with placebo, or treatment with E2, daily for 2 years [Steiner et al. 2005]. Therefore, the impact of MHT on BP is complicated, perhaps because of individual variability in underlying cardiovascular disease before the start of treatment (see Table 1 for summary of findings from RCTs).