Clinical Opinion: The Biologic and Pharmacologic Principles for Age-Adjusted Long-term Estrogen Therapy

Morris Notelovitz, MD, PhD, MB BCh, FACOG, FRCOG

In This Article

Cardiovascular Health and Disease

Cardiovascular disease (CVD) is the leading cause of death in women. Atherosclerosis, the major underlying pathology, is a multifactorial disease that develops from early childhood but only presents clinically in women a decade or more beyond menopause, as atherothrombosis (myocardial infarction or stroke). The majority of women in Western societies have varying degrees of atherosclerosis as they enter menopause. Central to this issue is the timing of ET.

Lipid and Lipoprotein Metabolism. The majority of women (66%) in the Framingham Study who developed incident coronary heart disease (CHD) during a 12-year follow up did not have elevations of their total cholesterol or low-density lipoprotein (LDL)-cholesterol.[24] The most common lipid abnormalities in these women were elevated triglycerides and low high-density lipoprotein (HDL)-cholesterol. Individuals with elevated triglyceride levels have an increased concentration of LDL-cholesterol; the particles are reduced in size and are therefore more atherogenic. Triglyceride values in excess of 200 mg/dL are associated with increased amounts of atherogenic remnant lipoproteins, which may substantially elevate the risk for CVD beyond that predicted on the basis of LDL-cholesterol alone. Hypertriglyceridemia results in decreased HDL-cholesterol and impairment of reverse cholesterol transport, and is also associated with insulin resistance, glucose intolerance, hypertension, and prothrombotic states.[25] High levels of the apolipoprotein Lp(a) have been identified as an independent risk factor for CVD. Concentrations of Lp(a) tend to increase after menopause.[26]

Glucose, Insulin Resistance, and Diabetes. CVD is more prevalent in women with both insulin-dependent and independent diabetes and latent glucose intolerance.[27,28] The association between abnormal carbohydrate metabolism and the pathogenesis of atherosclerosis is unclear: hyperglycemia has an adverse effect on the vessel wall; hyperinsulinemia appears to be an independent risk factor acting through alternative mechanisms.[29] The menopause is associated with a progressive decrease in insulin sensitivity, and insulin resistance is more common in postmenopausal than premenopausal women.[30]

Clinical Message. Hypertriglyceridemia serves as an excellent surrogate marker for dyslipidemia, hypercoagulability, and insulin resistance. Modifying factors related to hypertriglyceridemia may in turn improve insulin sensitivity, fibrinolysis, and, therefore, the potential for CVD.

Blood Pressure and Hypertension. Estrogen vasodilates arterial vessels via a number of mechanisms that have a direct effect on the smooth muscle of the arterial wall. This is also true for estrogen metabolites that may be even more potent than estradiol.[31]

Two mechanistic processes influence arterial wall disease and function: stability of the atheromatous plaque and vasoreactivity of the arterial endothelium.

Atheromatous Plaque. Plaque stability is determined by pro-inflammatory cytokines, such as monocyte chemoattractant protein (MCP-1), and cell adhesion molecules that recruit monocytes from the blood into the intima of the arteries where they subsequently become lipid-laden foam cells.[32,33] C-reactive protein (CRP), an acute-phase inflammatory biomarker, has emerged as one of the most important predictors of CVD.[34,35] Mature plaques consist of a lipid-rich core and a fibrous cap of extracellular matrix proteins, and elevated plasminogen activator inhibitor-1 (PAI-1) levels are associated with thin wall plaques.[36] Fissuring of the weakest cap sites are initiated by proteinases that degrade the cap.[37] One such proteinase is matrix metalloproteinase (eg, MMP-9), which accumulates at the edges of atherosclerotic plaques. Estrogen increases circulating levels of MMP-9 in postmenopausal women.[38] Thus, postmenopausal women who have a systemic and/or local predisposition to rupture of atherosclerotic plaques (ie, have high endogenous PAI-1 levels and/or thin plaques) may be at increased risk for an estrogen-associated cardiac event prior to the initiation of ET.

Estrogen has an essential role in mediating arterial vasodilation and blood flow.[39] This is achieved by stimulating endothelial prostacyclin synthesis (a potent vasodilator) and by inhibiting the vasoconstrictive effect of endothelin-I on the arterial subendothelial smooth muscle cells.[40,41] Estrogen also mediates vascular reactivity via nitric oxide synthesis. The respective roles of ERalpha and ERbeta in this regard are unclear, but animal studies suggest that ERbeta may play a significant role. Atherosclerosis decreases the number of ERs (71% in healthy arteries vs 32% in atherosclerotic arteries) and increases the degree of DNA methylation of ERalpha (which alters its expression).[42,43] The reduced ER expression in atherosclerotic lesions may be further compromised by exogenous estrogen plus progestin therapy (HT) - namely, by the co-prescribed progestin, which has been found in animal studies to reverse estrogen's atheroprotective effects. The bioavailability of the co-prescribed estrogen in combined continuous HT products to compensate for this potential adverse event is, therefore, important.

A meta-analysis of 248 studies reviewed the effects of 42 HT regimens on lipid and lipoprotein concentrations.[44] All estrogen-alone regimens raised HDL-cholesterol levels and lowered LDL and total cholesterol levels. The results of this meta-analysis confirmed more recent studies finding that EE 5 mcg increased triglycerides by approximately 39% over baseline[14]; the comparable mean triglyceride elevations for CEE (0.625 mg) and E2 (1 mg) were 35% and 10%, respectively.[45,46]

Oral ET has been shown to have a bimodal effect on insulin sensitivity and glucose tolerance.[47] Standard-dose CEE (0.625 mg) improved insulin sensitivity; higher doses of CEE (1.25 mg) resulted in a deterioration of insulin activity.[47] Other studies have shown that transdermal ET improves or does not adversely influence insulin activity.[48,49,50] Improvement in insulin sensitivity with an estrogen-alone or an estrogen plus progestin regimen has been reported in both nondiabetic postmenopausal women[49,51,52,53,54] and postmenopausal women with type 2 diabetes.[55,56,57] Other studies have found that insulin sensitivity is unaffected or reduced with estrogen-alone or estrogen plus progestin regimens.[49,50,51,58,59,60,61] There is an interindividual variation in response to glucose tolerance among women on ET: subjects with higher basal fasting glucose or insulin levels appear to have an improved insulin sensitivity and clinical response to ET, suggesting that ET may have the ability to pharmacologically restore glucose homeostasis to normal.[62] This is consistent with results from the WHI that showed a reduced risk of non-insulin-dependent diabetes in current hormone therapy users.[63,64]

Although estrogen lowers blood pressure, some women have an idiosyncratic response to ET with elevation in blood pressure. This response occurs most often with oral ET. It is associated with an increased production of renin substrate by the liver and a reversible stimulation of the renin-angiotensin-aldosterone system,[31] and is more likely to occur with CEE due to abnormal hepatic protein synthesis and nonphysiologic angiotensin I/II cleavage products.[65] Oral or transdermal E2 is the preferred estrogen to prescribe under these circumstances.[31] Hypertension should generally not be a contraindication to ET.[31]

The authors of the Estrogen in the Prevention of Atherosclerosis (EPAT) study in healthy postmenopausal women with subclinical atherosclerosis reported that 1 mg of oral E2 slowed the progression of atherosclerosis compared with matched placebo-treated women.[66] This finding was more pronounced in women who were not taking lipid-lowering drugs. This result differs from those of an earlier study, the Estrogen Replacement on Progression of Coronary Artery Atherosclerosis (ERA) trial, in which CEE failed to slow the progression of coronary artery atherosclerosis in women with established disease.[67] The main differences between these studies were younger age and absence of overt disease in the women studied in EPAT. Further RCTs will be needed to determine whether other estrogens have the same positive effect on subclinical atherosclerosis as unopposed E2.

E2 and CEE improve brachial artery flow-mediated dilation, a surrogate biomarker of coronary artery vascular reactivity.[41] Treatment of postmenopausal women with a variety of ETs reduces the pro-inflammatory cytokine MCP-I concentration to premenopausal levels. The same was found for cell adhesion molecules, all of which were reduced with ET.[68] CEE has been reported in to increase CRP, but the observed associated increase in cardiac events was related to baseline pretreatment CRP concentrations.[69] However, a recent analysis has found that it was progestin in combination with CEE that potentiated the IL-6 (inflammatory)-mediated stimulation of CRP, and not CEE alone.[70] Transdermal E2 is less likely to stimulate an increase in CRP, and a recent review found that whereas oral hormone therapy was generally associated with elevated CRP transdermal E2, it was not associated with elevated levels of IL-6 or CRP alone or with the addition of progestins.[71]

The use of CEE alone therapy in the WHI study was not associated with an increase in the incidence of CHD when the data for the total trial group were analyzed -- hazard ratio (HR) = 0.91 (0.73-1.14).[16] The HR for the younger subjects (aged 50 to 59 years), however, was reduced (HR 0.56 [0.3-1.03]), although the confidence limits in this category just exceeded statistical significance. This result is nevertheless consistent with the 50% reduction in CVD noted in the majority of observational studies. Younger age and earlier initiation of ET may be reflective of a healthy user effect that is relevant to the majority of women who seek treatment for symptoms during their early menopause. The effect of lower-dose ET (CEE 0.45 mg and equivalent doses with other estrogens) on cardiovascular health is not known. Lower-dose ET seems to have a similar protective effect as standard-dose ET on metabolic risk factors associated with CVD. [72,73]

Animal experiments provide validated and biologically sound data documenting the ability to prevent cardiovascular disease with early ET. However, application of these data to clinical practice is not advocated due primarily to the absence of comparable data from long-term RCTs in younger healthy women. Subanalysis of data from studies reporting adverse cardiovascular events frequently note that a significant percentage of these subjects had an underlying latent abnormality that would have precluded ET. In short, it is often the inappropriate selection of a patient for ET that is responsible for undesired side effects, rather than the ET per se.

Screening of patients before initiating treatment for any estrogen-related indication should include a metabolic profile (eg, glucose, liver enzymes, LDL and HDL cholesterol, triglycerides, and CRP). The choice of the type of estrogen and the route of ET will be governed by the result. The lowest dose of ET consistent with the indication should be prescribed, and the patient needs to be monitored regularly to assess the efficacy of treatment. All available ETs are safe for the majority of eligible menopausal women. The presence of a metabolic abnormality (eg, dyslipidemia, glucose intolerance, or hypertension) is not a contraindication for ET. The type, dose, and sometimes the route of ET may need to be adjusted according to the pharmacologic profile of the ET. Specific cardiovascular diseases may need treatment with disease-specific drugs.

Venous thrombosis (deep vein thrombosis [DVT] and pulmonary embolism [PE]) is a relatively uncommon condition.[74,75] The pathogenesis of venous thrombosis differs from that of arterial thrombosis. Venous thrombosis is associated with a strong genetic predisposition to hypercoagulation and a contributory decrease in venous blood flow. Arterial thrombosis results from the stimulation of platelets and the initiation of the coagulation cascade, secondary to a damaged vascular endothelium. The commonality between venous and arterial thrombosis is formation of a thrombus. Risk factors and the clinical profile for VTE and arterial disease differ.[74]

Biology: Coagulation. Hemostasis is a delicately balanced process involving a coagulation cascade that has an intrinsic and an extrinsic pathway.[74] A number of inactive zymogens (factors) are converted to proteases (activated factors) that eventually result in the conversion of prothrombin to thrombin. Thrombin catalyzes fibrinogen to fibrin and leads to the formation of an insoluble blood clot. The coagulation cascade is modulated by an anticoagulation system involving primarily antithrombin III (a factor responsible for 50% of the fluidity of blood), protein C, and its cofactor protein S. Fibrin formation is controlled by the fibrinolytic activity of plasmin, which in turn is derived from the activation of the zymogen, plasminogen. This system is controlled by activators (plasminogen activator [t-PA]) and inhibitors, PA1-1.

Platelets have a central role in arterial thrombosis. They adhere to various adhesion molecules and to the exposed subendothelium of the damaged arterial wall. Platelet activation mediates coagulation via the intrinsic pathway, to factor X and thrombin formation. Venous thrombosis is initiated via the extrinsic pathway.

The type, dose, and route of ET are important in terms of venous thrombosis. Oral esterified estrogen has been shown not to be associated with an increase in VTE risk (odds ratio [OR] 0.92 [0.69-1.22]) compared with an equivalent dose of CEE (OR 1.65 [1.24-2.19]).[76] A range (0.9 to 2.5 mg) of CEE doses were studied, and there was greater difference with higher daily doses of CEE. The OR was 1.21 [0.46-3.17] with low-dose CEE (mean, 0.3 mg) and 3.80 [1.90-7.61] with high-dose (mean 1.16 mg) CEE.[76]

Since the procoagulant factors are synthesized in the liver, avoidance of the entero-hepatic first-pass effect is associated with a more limited change in the hemostatic profile.[77] Whether this is clinically important needs to be investigated. Most of the changes in the biomarkers of coagulation, anticoagulation, and fibrinolysis result from oral estrogen usage (CEE, E2, or EE). These include an increase in the levels of tissue activator fibrinolysis inhibitor antigen, protein C, D-dimer, and factors VII, IX, and X (all indicators of increased coagulation), and a decrease in levels of protein S, antithrombin III (anticoagulation), and tissue plasminogen activators (fibrinolysis).[78] The degree to which ET increases the risk of VTE will vary primarily with the dose of estrogen and possibly the route of ET. In one study, the OR for VTE in current users of oral and transdermal ET compared with nonusers was 3.5 (1.8-6.8) and 0.9 (0.5-1.6), respectively.[77] The estimated risk for VTE in current users of oral ET compared with transdermal ET users was 4.0 (1.9-8.3).

The risk of VTE (including DVT and PE) in the estrogen-alone (CEE 0.625 mg) arm of the WHI was 28 (treatment group) vs 21 (placebo group) per 10,000 person-years,[16] a finding consistent with earlier data from observational studies. These results must be judged in the context of (1) the relative infrequency of VTE in adults (1 or 2 per 1000) and the 2- to 4-fold increase associated with ET; (2) age; the risk is significantly lower in younger women; (3) weight; a similar increase in the risk of VTE is related to body mass index (BMI), and (4) genetic predisposition; women on ET with heterogeneous and homogenous Factor V Leiden disease polymorphism, for example, have a significant 7.5- to 8-fold increased risk for VTE.[79]

In 2 recent studies in healthy postmenopausal women, CEE (0.625 mg) significantly decreased levels of anticoagulation factors (fibrinogen, antithrombin III, total and free protein S, and t-PA and PAI-1) and increased coagulation factors (VII and D-dimer).[80,81] However, a study that looked at lifestyle factors found that BMI, high levels of total cholesterol and tricglycerides, and low HDL levels had more of an impact on hemostatic variables in healthy postmenopausal women than hormone therapy or menopausal status.[82]

Venous thrombosis is primarily a balance between coagulation and anticoagulation. Interpretation of ET-related changes in the relevant factors must be applied with due caution in practice for the following reasons. Most of the factors assayed are in their inactive form (zymogens), and the system has a built-in redundancy. For example, antithrombin III activity has to be reduced by 50% before the disease is clinically relevant. In addition, significant changes in coagulation may be balanced by an equivalent alteration in anticoagulation.[78] There are no tests appropriate for routine clinical use that assay platelet activity.[77,78,83] On the basis of available data and clinical experience, ET has no clinically significant adverse effects on coagulation in healthy postmenopausal women.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: