Beneficial effects of estrogen replacement therapy include its ability to reduce low-density lipoprotein (LDL) cholesterol levels and increase high-density lipoprotein (HDL) cholesterol levels [13]. Nonetheless, multiple regression analyses suggest that only 25% to 50% of the reduction in cardiovascular events is attributable to the lipid-lowering effects of estrogen replacement therapy [14], suggesting that other mechanisms contribute to its cardioprotective potential. One such mechanism may involve the effect of estrogen on vascular function. Specifically, estrogen may directly enhance the activity of the endothelium-derived relaxing factor nitric oxide [10, 15-17] and thereby lessen the potential for coronary vasoconstriction and thrombosis [18-21]. When administered to rabbits that had ovariectomy, long-term estrogen replacement improved endothelium-dependent relaxation in vitro [15, 16]. In addition, both short-term and long-term estrogen administration improves in vivo the endothelium-dependent vasodilation in coronary arteries of atherosclerotic cynomolgus monkeys that had ovariectomy [10, 17]. Several recent studies have suggested that short-term estrogen administration improves endothelium-dependent vasodilation of the coronary arteries of postmenopausal women [22, 23].

We hypothesized that long-term estrogen administration would improve vasomotor function in postmenopausal women. Accordingly, we conducted a double-blind, randomized, crossover, placebo-controlled trial to assess the effect of estrogen replacement therapy on endothelium-dependent vasodilation in postmenopausal women. We used high-resolution ultrasonography to serially assess vasomotor function in a peripheral conduit vessel, the brachial artery.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

The patients enrolled in this study were recruited from a larger cohort that was participating in a clinical research trial investigating the effect of estrogen replacement therapy on lipoprotein metabolism. Thirteen postmenopausal women aged 44 to 69 years (average age, 55 ±7 years) participated in this study. Women were eligible if menopause had occurred at least 1 year previously. Menopause was confirmed by measuring serum follicle-stimulating hormone levels. No patient had received hormone replacement therapy for at least 2 months before the study began. A history and physical examination were done to exclude persons with clinical evidence of coronary or peripheral atherosclerosis. Initial evaluation included a Papanicolaou smear (if not done in past year), complete blood count, routine chemistry panel, and lipid profile. Inclusion criteria included mild hypercholesterolemia, defined as a serum cholesterol level of 5.17 mmol/L to 6.20 mmol/L and an LDL cholesterol level of 3.36 mmol/L to 4.13 mmol/L. Exclusion criteria included hypertension, diabetes mellitus, tobacco use, obesity (body weight > 135% of ideal weight), history of breast or uterine cancer, thromboembolism, and liver or renal disease.

Eligible patients were placed on a low-cholesterol diet (American Heart Association phase I) for 6 weeks before randomization. Patients were randomly assigned to one of three treatment groups: placebo, oral 17 ß-estradiol (Estrace, Mead Johnson, Evansville, Indiana.) at a dose of 1 mg/d, or oral 17 ß-estradiol at a dose of 2 mg/d. At the conclusion of each 9-week treatment period, the patients were given progesterone (Provera, Upjohn, Kalamazoo, Michigan) at 10 mg/d for 10 days. Hormone replacement therapy was discontinued for 3 weeks before patients crossed over to the next treatment regimen. All patients received placebo and the two doses of estrogen in random order. Vascular function studies (described below) were done during the eighth or ninth week of each treatment period.

Experimental Protocol

Studies were done in a temperature-controlled vascular research laboratory. All patients were placed in the supine position. We studied vascular reactivity in a conduit vessel, the brachial artery, as previously described. An imaging study of the brachial artery was done using a high-resolution ultrasound machine (Toshiba, Model SSA-270, Otawara-shir, Tochigi-Ken, Japan) that was equipped with a 7.5 MHz linear-array transducer. Baseline images of the brachial artery were obtained proximal to the antecubital fossa. Imaging of the artery was done longitudinally, allowing clear visualization of the posterior wall intima-lumen interface and the anterior wall media-adventitial interface. We assessed endothelium-dependent vasodilation by measuring the change in the caliber of the brachial artery during reactive hyperemia, a maneuver that increases flow through the conduit segment being studied (flow-mediated vasodilation). To create this stimulus, a cuff placed on the upper arm was inflated to suprasystolic pressure for 5 minutes, thereby occluding flow to the forearm. This results in dilatation of downstream forearm resistance vessels. After cuff deflation, reactive hyperemia occurs, as brachial artery blood flow increases to accommodate the dilated resistance vessels. Imaging of the brachial artery was continually done for the 5-minute period after cuff deflation until basal conditions were re-established. Thereafter, sublingual nitroglycerin (at a dose of 0.4 mg) was administered to assess endothelium-independent vasodilation. The artery was studied for an additional 5 minutes. Blood pressure and heart rate were monitored continuously throughout the procedure.

All images were recorded on Super VHS videotape for subsequent analysis. Image analysis was done on a personal computer that was equipped with a video frame grabber. Images recorded on videotape were analyzed by an investigator blinded to treatment assignment. Previous studies have shown that the peak diameter change during reactive hyperemia occurs approximately 1 minute after cuff deflation and 3 minutes after nitroglycerin administration [41]. We used these time points in our study. Images corresponding to the end of the T wave on a simultaneous electrocardiograph were selected and digitized. Image analysis was then done using a proprietary analysis software that searched for the shortest distance between the points on the arterial wall, creating 10 to 20 paired measurements along a 10-mm length. We measured arterial diameter from the intima-lumen interface on the posterior wall to the media-adventitial interface on the anterior wall. We calculated brachial artery diameter by averaging these paired lumen measurements and reported them in millimeters using calibration factors derived from real-time ultrasonography. We used an average of three separate measurements for each condition. In our laboratory, this technique has a variability of only 0.0 ±0.1 mm [24].

To assure that the blood flow stimulus during reactive hyperemia was similar during each treatment phase, forearm blood flow was measured by venous occlusion strain gauge plethysmography using calibrated mercury-in-silastic strain gauges as previously described [25].

Statistical Analysis

The variables compared during the placebo period and during therapy with each dose of estrogen included blood pressure, heart rate, forearm blood flow, basal brachial artery diameter, and the percentage increase in diameter during reactive hyperemia and after patients received sublingual nitroglycerin. Values are expressed as the mean ±SE. For statistical analysis, we used repeated-measure analysis of variance and the Scheffe-F post hoc test [26]. Significance was accepted at P 0.05.

Results

Top Methods Results Discussion Author & Article Info References

Baseline Hemodynamic Measurements

The effect of estrogen treatment on blood pressure, heart rate, forearm blood flow and forearm vascular resistance is presented in Table 1. Estrogen therapy did not affect blood pressure or heart rate. Basal forearm blood flow tended to be higher during both estrogen treatment phases (P = 0.08). Basal forearm vascular resistance was similar during all three treatment phases. The peak forearm blood flow during reactive hyperemia was similar during placebo receipt and each estrogen treatment period. However, peak forearm blood flow was greater when patients received the 1-mg dose of estradiol than when they received the 2-mg dose (P = 0.05).

Flow-mediated Endothelium-dependent Vasodilation

We obtained technically adequate ultrasound images during reactive hyperemia for 12 of the 13 patients. The brachial artery diameter, under basal conditions, measured 3.5 mm, 3.4 mm, and 3.3 mm while patients received placebo, estradiol at 1 mg/d, and estradiol at 2 mg/d, respectively (P > 0.2). The percentage increase in brachial artery diameter during reactive hyperemia for each treatment period is shown in Figure 1. The change in brachial artery diameter was greater when patients received estradiol treatment (13.5% and 11.6% for 1-and 2-mg doses, respectively) than when they received placebo (6.8%, P < 0.001 by analysis of variance; P < 0.05 for each dose by post hoc testing). We observed no difference between the two estrogen treatment periods in the vasodilative response of the brachial artery.

View larger version (34K): [in this window] [in a new window]   Figure 1. Effect of estradiol on flow-mediated, endothelium-dependent vasodilation in 12 postmenopausal women. Oral estradiol at 1 mg/d and at 2 mg/d significantly increased the brachial artery diameter during a flow stimulus induced by reactive hyperemia (P < 0.05 for each dose compared with placebo).

Effect of Estrogen Replacement Therapy on Brachial Artery Responses to Sublingual Nitroglycerin

Seven patients received nitroglycerin during all three treatment phases. Four of the patients received nitroglycerin at least once but refused further doses during subsequent visits because of an adverse reaction (symptomatic hypotension, severe headache) at the previous visit. Two patients refused nitroglycerin at all three visits. Estrogen administration had no significant effect on endothelium-independent vasodilation. Nitroglycerin increased brachial artery diameter 12.3% when patients received placebo, 14.2% when they received estradiol at 1 mg/d, and 16.5% when they received estradiol at 2 mg/d (P > 0.2).

Effect of Estrogen Replacement Therapy on Total Cholesterol Levels

The patients we studied are part of a larger cohort of postmenopausal women who are participating in a study to assess the effect of estrogen replacement therapy on lipoprotein metabolism. The results of that trial are being reported separately. Among the patients we describe in this report, estrogen did not significantly affect the total cholesterol level, which was 5.66 mmol/L when patients received placebo, 5.40 mmol/L when they received estradiol at 1 mg/d, and 5.23 mmol/L when they received estradiol at 2 mg/d, respectively (P = 0.12).

Discussion

Top Methods Results Discussion Author & Article Info References

The new finding in our study is that sustained estrogen replacement therapy improves endothelium-dependent vasodilation in postmenopausal women. Daily administration of estradiol at a dose of either 1 mg or 2 mg increased flow-mediated vasodilation in the brachial artery after 9 weeks of therapy. These observations were made on a peripheral artery, were unaffected by atherosclerosis, and have important implications on the effect of estrogen on endothelial function in other vessels.

Endothelium-dependent Vasodilation

Laboratory experiments using arterial strips have shown that a normal endothelium produces a local short-lived substance, endothelium-derived relaxing factor, which is either nitric oxide or a related nitrosothiol [27-29]. Endothelium-derived relaxing factor is released in response to various chemical stimuli (for example, acetylcholine, serotonin, thrombin, and substance P) and physical factors (for example, blood flow, shear stress, and pulsatile pressure) [30]. This vasodilator mechanism opposes the constrictor actions of hormones, neurotransmitters, and platelet-release products.

Endothelium-dependent relaxation is abnormal in conduit and resistance vessels of humans with atherosclerosis [18-20, 31]. Reduced levels of endothelium-derived nitric oxide may contribute to the pathogenesis of myocardial ischemia and infarction by facilitating coronary vasoconstriction and thrombosis. Reduced activity of endothelium-derived relaxing factor may also be involved in atherogenesis because nitric oxide normally inhibits platelet adhesion and aggregation, leukocyte adhesion to the endothelial surface, and vascular smooth-muscle proliferation [21, 32-35]. Indeed, persons with risk factors for atherosclerosis show abnormalities of endothelium-dependent vasodilation, even before the development of atherosclerosis and in vessels devoid of atherosclerosis. We and others have reported previously that endothelial function is abnormal in forearm resistance vessels of patients with hypercholesterolemia, diabetes mellitus, and hypertension [25, 36-38]. Using ultrasonography, Celermajer and colleagues [39, 40] reported that flow-mediated vasodilation of the brachial artery is abnormal in persons with hypercholesterolemia, in those who smoke cigarettes, and in those with coronary artery disease. We have found that reduced brachial artery flow-mediated vasodilation correlates closely with functional abnormalities of the coronary endothelium [41]. Preliminary studies in our laboratory have shown that flow-mediated vasodilation in the brachial artery is mediated by endothelium-derived nitric oxide because it is inhibited by the nitric oxide synthase antagonist, n^G-monomethyl-l-arginine. High-resolution ultrasonography is a reproducible technique for measuring brachial artery diameter [24]; we were thus able to use this noninvasive method to study the effect of estrogen replacement on endothelium-dependent vasomotion in postmenopausal women.

Localization of estrogen receptors on vascular endothelium and smooth-muscle cells of several mammalian species suggests that these hormones may influence vascular function [42-44]. Short-term administration of estrogen improved endothelium-dependent vasodilation of atherosclerotic coronary arteries in monkeys that had ovariectomy and in postmenopausal women when studied in vivo [17, 22, 23]. Long-term estrogen treatment improved in vivo the endothelium-dependent relaxation of atherosclerotic femoral arteries excised from rabbits that had ovariectomy and of atherosclerotic coronary arteries in monkeys that had ovariectomy [10, 15, 16]. High levels of total serum cholesterol and LDL cholesterol and low levels of HDL cholesterol are associated with reduced endothelium-dependent vasodilation [24, 45, 46]. Estrogen increases HDL cholesterol and lowers LDL cholesterol levels and thereby might be expected to improve endothelium-dependent relaxation in patients with lipid disorders [13]. The short-term effects of estrogen on vascular function (that is, in the absence of lipid changes) suggest, however, that additional mechanisms must be involved. In our patients, we observed no significant effect of estrogen treatment on total cholesterol level. Our study, however, did not have the power to address the possibility that estrogen altered one or more lipid fractions that are important in atherogenesis. Our participants are part of a larger cohort of persons in whom the effect of estrogen on lipid metabolism is being studied. Therefore, we cannot discount the possibility that a modest decrease in cholesterol levels contributed to the favorable effect of estrogens on endothelial function that we observed in our study.

Mechanisms of Endothelium-dependent Vasodilation

Conceptually, estrogen could improve endothelium-dependent vasodilation by increasing synthesis and release of nitric oxide, by inhibiting its degradation, or by altering the balance of vasodilator and vasoconstrictor prostaglandins. For example, estrogen may increase the release of nitric oxide through up-regulation of muscarinic receptors [47]. Also, the antioxidant properties of estrogen might stabilize nitric oxide and prevent its inactivation, explaining how vasodilative effects can be observed within minutes of the administration [48]. Recently, Keaney and colleagues [49] reported that estrogen replacement increased resistance to LDL oxidation and improved endothelium-dependent vasodilation in swine that had ovariectomy.

Estrogen may also influence vascular tone through effects on other endothelium-derived vasoactive substances. Several but not all studies have also shown that estrogens stimulate the production of vasodilator prostaglandins and reduce the effects of vasoconstricting prostanoids [15, 50, 51]. Estrogen also attenuates vasoconstriction to endothelin-1 [52]. In isolated canine coronary arteries, estrogen administration eliminates induced-action potentials and causes cell membrane hyperpolarization.

Implications for Estrogen Dosing

Our findings support the use of estrogen replacement therapy to improve endothelial function but do not enable us to recommend a specific dose of estrogen. The similar responses seen with the 1- and 2-mg dosages of estradiol suggest that the observations were made at the plateau region of the dose-response curve. We are hopeful that lower doses will be effective because clinical use of higher doses of estrogen may increase the risk for endometrial cancer [53, 54]. In addition, we cannot comment on whether the addition of progesterone will alter the favorable effects of estrogen on endothelium-dependent vasodilation.

Study Limitations

Our study design has several limitations. First, all the participants had mild hypercholesterolemia because they were recruited from a larger study examining the effect of estrogen on lipid metabolism. Thus, the results cannot necessarily be extrapolated to postmenopausal women whose cholesterol levels are entirely within normal limits. Furthermore, we cannot be certain that the effect of estrogen on endothelium-dependent vasodilation occurred as a result of its lipid-lowering effects and not as a direct effect of estrogen on the release or activity of endothelium-derived nitric oxide. Second, we do not discount the likelihood that these persons had occult atherosclerosis. To avoid the confounding effects of atherosclerosis, we chose women who had no historical or clinical evidence of coronary artery disease, peripheral arterial occlusive disease, or cerebrovascular disease. Furthermore, we studied a conduit vessel from an extremity that is infrequently affected by atherosclerosis and used a technique that would have enabled us to detect intimal thickening or plaque formation if it were present. Third, we cannot exclude a carry-over effect of estradiol in this crossover trial. If this were the case, our findings are even more impressive because a carry-over effect would have the potential of enhancing flow-mediated vasodilation during placebo receipt and therefore lessen the distinction between placebo receipt and estrogen treatment periods. Finally, one might argue that too few patients received nitroglycerin to enable us to distinguish an effect of estrogen on endothelium-independent vasodilation. Nonetheless, the average difference between treatment groups observed after nitroglycerin treatment was considerably less than that observed during reactive hyperemia, indicating that any effect of estrogen on endothelium-independent vasodilation by nitroglycerin would be small.

Conclusions

Our study shows that estrogen replacement therapy improves flow-mediated endothelium-dependent vasomotion in postmenopausal women with mild hypercholesterolemia. Further investigation is warranted to enhance our understanding of the mechanisms of the effect of estrogen on vascular function and to determine whether its effect on endothelial function contributes to reduction in cardiovascular-related morbidity and mortality.