Menopause Is Associated With Endothelial Dysfunction in Women

It is well documented that menopause and consequent estrogen deprivation increase the risk of cardiovascular disease in women.¹² However, the mechanisms responsible for the cardioprotective effect of estrogen in women are not clear. Although estrogen can favorably affect lipid profile,³⁴⁵ increasing evidence supports the possibility that estrogens can act directly on the arterial wall.⁶⁷⁸

Endothelium plays a primary role in the local regulation of vascular activity by synthesizing vasodilating and vasoconstricting substances.⁹ Many agonists (including acetylcholine, bradykinin, substance P, etc) or physical stimuli (shear stress) can trigger endothelium-dependent relaxations by releasing an endothelium-derived relaxing factor,^10111213 identified with nitric oxide.¹⁴ Endothelial cells can also produce cyclooxygenase-dependent endothelium-derived contracting factor(s),¹⁵¹⁶ which are very likely to be endoperoxides.¹⁷¹⁸ In humans, impaired endothelium-dependent vasodilation is associated with diseases such as essential or secondary hypertension,¹⁹²⁰ hypercholesterolemia,²¹²² and atherosclerosis²² or physiological conditions such as advancing age,²³ all of which represent well-documented cardiovascular risk factors. Thus, the possibility exists that the augmented prevalence of cardiovascular disease in menopausal women could be related to endothelial dysfunction caused by endogenous estrogen deficiency.

To address this issue, in the present study we examined whether the onset of menopause is associated with detectable impairment of endothelium-dependent vasodilation. In addition, to better characterize the possible role of endogenous estrogen on endothelial function, we evaluated both normotensive subjects and patients with essential hypertension, a clinical condition characterized by impaired endothelium-dependent vasodilation.¹⁹²⁰ In particular, we looked at the relationship between endothelium-dependent vasodilation and advancing age in both men and women to ascertain whether women are characterized by an attenuation of age-related endothelial dysfunction when still menstruating and by a sharp worsening after menopause compared with matched men.

Methods

Study Population

The study population included 73 normotensive control subjects and 73 essential hypertensive patients. Some members of the present study population (53 of the normotensive subjects and 57 of the essential hypertensive patients) participated in another study,²³ and the remaining individuals were specifically recruited for this study. Subjects with hypercholesterolemia (total cholesterol >5.2 mmol/L), diabetes mellitus, cardiac and/or cerebral ischemic vascular disease, impaired renal function, and other major pathologies were excluded from the study. Moreover, individuals who smoked more than five cigarettes per day were excluded from the study. Any pharmacological treatment was discontinued for at least 2 weeks before the study.

Normotensive subjects, defined as such by the absence of a family history of essential hypertension and blood pressure values below 140 to 90 mm Hg, were characterized by a mean age of 44.3±15.1 years (range, 18 to 76 years), blood pressure values of 118.1±6.1/77.6±3.0 mm Hg, and total plasma cholesterol of 4.52±0.56 mmol/L. Essential hypertensive patients were recruited from among the newly diagnosed patients in our outpatient clinic. In all cases, patients reported the presence of a positive family history of essential hypertension, and supine arterial blood pressure (after 10 minutes of rest), measured by mercury sphygmomanometer three times at 1-week intervals, was consistently found to be greater than 140/90 mm Hg. Secondary forms of hypertension were excluded. Mean age was 47.4±15.2 years (range, 20 to 76 years), blood pressure values were 162.3±8.3/111.3±5.2 mm Hg, and plasma total cholesterol was 4.82±0.40 mmol/L.

Both normotensive and hypertensive participants were selected to have comparable hemodynamic and humoral variables along all aging profiles. All normotensive and hypertensive women younger than 45 years had a normal menstrual cycle, and those older than 45 years had had their last menses at least 1 year before enrollment. Neither normally menstruating or menopausal women had ever had hormone replacement therapy. According to institutional guidelines, all participants were aware of the investigational nature of the study and gave written consent to it.

Experimental Procedure

All studies were performed at 8 am, after participants had fasted overnight, with subjects lying supine in a quiet air-conditioned room (22° to 24°C). A 21-gauge polyethylene cannula (Abbot) was inserted into the brachial artery under local anesthesia (2% lidocaine) and connected through stopcocks to a pressure transducer (model MS20, Electromedics) for monitoring of systemic mean blood pressure (1/3 Pulse Pressure+Diastolic Pressure) and heart rate (model VSM1, Physiocontrol) and for intra-arterial infusions. Forearm blood flow (FBF) was measured in both forearms (experimental and contralateral) by strain-gauge venous plethysmography (LOOSCO, GL LOOS).²⁴ Circulation to the hand was excluded 1 minute before each sampling or FBF measurement by inflation of a pediatric cuff around the wrist at suprasystolic pressure. Details concerning the sensitivity and reproducibility of the method as performed in our laboratory have already been published.²⁵

Drug infusion rates were normalized for 1 dL tissue by adjustment of the infusion speed to desired infusion rates. Drugs were infused at systemically ineffective rates through separate ports via three-way stopcocks.

Experimental Design

Endothelium-dependent vasodilation was estimated by a dose-response curve to intra-arterial acetylcholine (cumulative increase of the infusion rates: 0.15, 0.45, 1.5, 4.5, and 15 μg/100 mL forearm tissue per minute for 5 minutes at each dose), an endothelium-dependent vasodilator.²⁶²⁷ Endothelium-independent vasodilation was assessed by a dose-response curve to intra-arterial sodium nitroprusside (cumulative increase by 1, 2, and 4 μg/100 mL forearm tissue per minute for 5 minutes at each dose), a direct smooth muscle cell relaxant compound.²⁸ These rates were selected to induce vasodilation comparable to that obtained with acetylcholine. The infusion sequence of the two drugs was randomized, and 45 minutes of recovery were allowed between the two experimental steps.

Data Analysis

Since arterial pressure did not significantly change during the study, all data were analyzed in terms of FBF; FBF increments were taken as evidence of local vasodilation. Raw data were analyzed by ANOVA for repeated measures, and Duncan's test was applied for multiple comparison testing. Wilcoxon's test was used for checking the statistical significance of differences between nonparametric values. The FBF response to the dose-response curve to acetylcholine and sodium nitroprusside was also analyzed both as the maximal vasodilation to the highest concentration of the agonist and as the slope of the percent increase in FBF above basal induced by both agonists. The slope allows one to summarize the vascular effect of the whole dose-response curve and is calculated with a linear regression analysis (y axis, percent increase in FBF; x axis, log[dose (μg/min) of acetylcholine or sodium nitroprusside]). The correlation coefficient was .92±.08 (range, .77 to .99). Interactions between age and forearm vasodilation in response to acetylcholine and sodium nitroprusside (considered in terms of either maximal effect or slope) were calculated by a multivariate analysis, using a multiple stepwise regression, to exclude the effects of blood pressure and plasma cholesterol.

To evaluate the effect of sex on vasodilation in response to acetylcholine and sodium nitroprusside, we carried out a comparison between the response to each compound in men and women according to a multiphase model²⁹ that determines a change point, α, that identifies the “most likely” age when there is a change in the linear relationship between aging and FBF modifications. Thus, the mathematical model is the linear-linear, which refers to a model made up of two linear equations, each active over a different range of X (age). The estimation equation is Y=A+BX+C(X−D)SGN(X−D), where Y is FBF. The SGN function lets us define a sharp transition from one curve to the next. Mathematically, this function is defined as follows:

For age below the change point α, the equation is Y=a₁+b₁X, whereas for age above or equal to the change point α, the equation is Y=a₂+b₂X. The gradient of FBF modifications with age was calculated as the percentage of the decrease of each slope per year. Men and women were then compared with respect to the gradient of FBF modifications with age, both before and after their estimated change point, with a χ² test.

Results are expressed as mean±SE.

Drugs

Acetylcholine HCl (Farmigea SpA) and sodium nitroprusside (Malesci) were obtained from commercially available sources and diluted freshly to the desired concentration by addition of normal saline. Sodium nitroprusside was dissolved in glucosate solution and protected from light by aluminum foil.

Results

Baseline systemic demographic, hemodynamic, and humoral characteristics for the normotensive and hypertensive participants are summarized in Table 1. Plasma cholesterol, glycemia, body mass index, and FBF did not differ significantly between normotensive subjects and hypertensive patients. In accordance with the enrollment criteria, blood pressure was significantly higher in patients with essential hypertension. However, in both normotensive and hypertensive participants, blood pressure values were not statistically different between younger and older people.

Response to Intrabrachial Acetylcholine and Sodium Nitroprusside in the Overall Population

In both normotensive subjects and essential hypertensive patients, acetylcholine and sodium nitroprusside caused dose-dependent vasodilation. Responses to acetylcholine were significantly (P<.001) blunted in patients with essential hypertension (from 3.4±0.4 to 16.6±2.1 mL/100 mL forearm tissue per minute) compared with responses in normotensive subjects (from 3.5±0.3 to 23.1±2.8 mL/100 mL forearm tissue per minute), whereas the vasodilating effect of sodium nitroprusside was similar between the two groups (normotensive subjects: 3.4±0.3 to 22.4±2.2 mL/100 mL forearm tissue per minute; hypertensive patients: 3.5±0.3 to 21.3±2.4) (Fig 1). In addition, in control subjects and hypertensive patients, responses to acetylcholine showed a strong significant negative correlation with increasing age in terms of both maximal FBF increase (normotensive subjects: r=−.88, P<.001; hypertensive patients: r=−.87, P<.001) and slope of the response to acetylcholine (normotensive subjects: r=−.82, P<.001; hypertensive patients: r=−.85, P<.001). In contrast, the relationship between the response to sodium nitroprusside and aging was less evident in terms of both maximal FBF increase (normotensive subjects: r=−.46, P<.01; hypertensive patients: r=−.48, P<.01) and slope (normotensive subjects: r=−.36, P<.05; hypertensive patients: r=−.28, P<.05).

Effect of Sex on the Response to Intrabrachial Acetylcholine and Sodium Nitroprusside

In normotensive men, change-point analysis identified a constant age-related decline in maximal vasodilation to acetylcholine by 1.8% per year (Fig 2). In contrast, women were found to be characterized by a slight decrease (0.5% per year) in vasodilation to acetylcholine up to the age of around 49 years (P<.001 versus men), after which the vascular response to the compound declined more quickly (2.1% per year) compared with men (P<.01) or women before menopause (P<.001) (Fig 2). It is worth noting that in men and women older than 60 years (11 men and 8 women), acetylcholine-mediated vasodilation was not statistically different (3.4±0.3 to 11.3±1.2 mL/100 mL forearm tissue per minute and 3.5±0.4 to 14.4±1.5, respectively; P<.12).

When data were analyzed in terms of the slope of the response to acetylcholine, the results were very similar to those obtained by examination of the maximal response to the endothelium-dependent vasodilator (Table 2).

Finally, there were no sex-related differences in sodium nitroprusside–induced vasodilation (Fig 3).

In male essential hypertensive patients, vasodilation to acetylcholine started to decline rapidly as early as at 20 years, with a decrement of 2.5% per year (Fig 4). It is worth noting that change-point analysis indicated that at around 40 years, the age-related decrement in the response to acetylcholine slowed to 1.5% per year (P<.02 versus hypertensive patients younger than 40 years) (Fig 4). In essential hypertensive women, there was also an early age-related decrement in the response to acetylcholine, which was found to be less marked (1% per year) than in age-matched men (P<.01) (Fig 4). In the middle of the fifth decade (most likely a change point at 44 years), we observed a steeper, age-related decline in acetylcholine-induced vasodilation (1.98% per year) (P<.01 versus younger women and P<.02 versus men of the same age) (Fig 4). In patients older than 60 years (12 men and 9 women), the response to acetylcholine was similar in men and women (3.3±0.3 to 8.2±1.1 mL/100 mL forearm tissue per minute and 3.2±0.2 to 9.1±1.1, respectively).

When the slope of the response to acetylcholine was analyzed, results showed a similar behavior (Table 2).

Finally, there were no sex-related differences in sodium nitroprusside–induced vasodilation (Fig 5).

Discussion

We designed the present study to evaluate the influence of menopause and consequent estrogen deficiency on endothelium-dependent vasodilation in humans. We studied normotensive subjects and patients with essential hypertension, a clinical condition characterized by endothelial dysfunction and a higher risk of cardiovascular disease. In our study population, we confirmed previous evidence that the response to the endothelium-dependent vasodilator acetylcholine but not to sodium nitroprusside, a smooth muscle cell relaxant compound, is blunted in essential hypertensive patients compared with control subjects, further supporting the hypothesis of endothelial dysfunction in essential hypertension.¹⁹²⁰²³ In addition, in both normotensive and hypertensive participants, a highly significant inverse correlation was confirmed between increasing age and the response to acetylcholine, whereas only a weaker negative correlation was observed between the vasodilating effect of sodium nitroprusside and aging. Taken together, these data reinforce the possibility that aging is associated with endothelial dysfunction.²²²³

However, the main finding in the present study is that menopause affects endothelium-dependent but not endothelium-independent vasodilation in both normotensive and essential hypertensive women. In normotensive men, endothelial responses started to decline in the third decade, and the derangement was gradual and progressive until old age. In contrast, in women up to the end of the fifth decade, advancing age only slightly affected endothelium-dependent vasodilation. After the age of 49 years (and after menopause in this female study population), endothelium-dependent vasodilation showed a decline that was steeper than that in men. It is important to observe that the sex-related difference in endothelium-dependent vasodilation was no longer statistically evident after 60 years despite a vascular response to acetylcholine that was still greater (around 20%) in women compared with men. Thus, although this finding could suggest that in old age the sex difference in endothelial function disappears, an inadequate power of the study to detect a statistical significance in this study subgroup cannot be excluded.

In the men with essential hypertension, advancing age was associated with a decrease in endothelium-dependent vasodilation to acetylcholine, which was more evident up to 40 years, when the rate of the decline started to slow. The more rapid decline in endothelium-dependent vasodilation in younger hypertensive men compared with those older than 40 years could suggest that hypertension anticipates age-related endothelial dysfunction. Thus, in young essential hypertensive patients, age and hypertension seem to act simultaneously to impair endothelial function. Moreover, the fact that after a certain age (around 40 years) the decline seems to slow is probably because the response to acetylcholine cannot be depressed over a certain degree. This possibility is in agreement with our previous observation that the age-related decline in endothelium-dependent vasodilation seems to be arrested after 60 years in essential hypertensive patients not divided according to sex.²³ At variance with results observed in hypertensive men, the pattern of endothelial dysfunction seems to be the opposite in hypertensive women, in whom the rate of age-related dysfunction in vasodilation induced by acetylcholine was evident but slower up to 44 years, with a steeper decline beginning at around the age of menopause. In older groups (>60 years), no sex difference was observed in the age-related decrease in endothelial function. In both normotensive and essential hypertensive participants, we observed no sex difference in the response to sodium nitroprusside.

Thus, taken together, these results seem to indicate that although in normotensive men, endothelial dysfunction associated with increasing age is a constant and homogeneous event, in normotensive women, menopause is a crucial time characterized by a remarkable deterioration of endothelial function. It is worth noting that in both male and female essential hypertensive patients, an impairment in endothelium-dependent vasodilation is already detectable in an early age. However, in the early decades, age-related endothelial dysfunction is more accentuated in male than in female hypertensive patients, whereas after the fifth decade, the extent of impairment of endothelium-dependent vasodilation is lower in men than in postmenopausal women. This finding provides further confirmation that in agreement with the situation in normotensive women, a mechanism protecting endothelial function is also present in hypertensive premenopausal women. In line with this possibility, in both postmenopausal normotensive and hypertensive women older than 60 years, aging-related endothelial dysfunction is similar to that in men of the same age.

The present results in normotensive subjects are in agreement with previous observations of Celermajer and colleagues,³⁰ who demonstrated that a decline in endothelium-dependent flow-mediated vasodilation of the brachial artery, but not in the response to glyceryl trinitrate, is associated with aging in normotensive men a decade earlier than in normotensive women, who were shown to be affected by endothelial dysfunction only after the onset of menopause. However, these results refer to large systemic arteries of normotensive subjects. Thus, our present findings add complementary information concerning resistance arterioles of normotensive control subjects and essential hypertensive patients. Therefore, taken together, these results indicate that menstruation in women presents a protective mechanism against age-related endothelial dysfunction, whereas the postmenopausal period is characterized by loss of such protection.

Both experimental and clinical data suggest that this protective mechanism could be related to the presence of endogenous estrogens. It is well documented that estrogen has a favorable action on endothelial function.⁷⁸ This effect could be at least partially mediated by the beneficial action of estrogen on lipid profile, since it is documented that premenopausal women have reduced low-density lipoprotein and elevated high-density lipoprotein cholesterol levels compared with men,³¹ whereas after menopause, low-density lipoprotein cholesterol levels rise.³² However, this possibility is not supported by the present results because women recruited in our study population were characterized by matched plasma cholesterol values; therefore, lipid profile was not different in premenopausal compared with postmenopausal women. Thus, it is possible that the protective effect of endogenous estrogen on endothelium-dependent vasodilation in women before menopause could result from a direct effect of the hormone on endothelial cells. In vitro and in vivo animal studies have demonstrated that estrogen potentiates endothelium-dependent vasodilator responses to acetylcholine.⁷ Moreover, estrogen replacement after ovariectomy enhances endothelium-dependent vasodilation in normocholesterolemic animals as well as in monkeys with dietary atherosclerosis.⁷³³³⁴ In humans, estrogen replacement therapy corrects coronary endothelium-dependent vasodilation in postmenopausal women with atherosclerosis. Thus, acute administration of 17β-estradiol, conjugated estrogens, and ethinyl estradiol increases coronary responses to acetylcholine.³⁵³⁶³⁷ In addition, long-term estrogen replacement can increase both endothelium-dependent flow-mediated vasodilation in hypercholesterolemic, postmenopausal women³⁸ and serum nitrate and nitrite levels, an index of enhanced nitric oxide production, in postmenopausal women.³⁹

The mechanism by which endogenous estrogens protect the endothelium cannot be determined from the present study. One possibility is suggested by the demonstration that long-term administration of estrogen upregulates the transcription of nitric oxide synthase, the enzyme that regulates nitric oxide production from the catabolism of l-arginine into citrulline and enhances its activity in nonvascular tissue.⁴⁰ However, whether such upregulation is present in vascular tissue is still to be determined. In addition, estrogen can increase endothelium-dependent vasodilation by an acute mechanism, as demonstrated by the above-reported studies in which estrogen effects were observed after 15 to 20 minutes of administration.³⁵³⁶³⁷ A further possibility is represented by the finding that estrogen can act as an antioxidant.⁴¹ If this is the case, aging-related impairment in endothelium-dependent vasodilation could be at least partially related to the production of oxygen-derived free radicals, which inactivate nitric oxide.

Endothelial dysfunction associated with menopause could be one of the possible explanations for increased cardiovascular morbidity and mortality in postmenopausal women. This hypothesis is in agreement with the findings that after menopause the extent of derangement of endothelium-dependent vasodilation becomes progressively similar between men and women; furthermore, the incidence of cardiovascular disease, which is three times lower in women than in men before menopause, is approximately equal in men and women after 75 years. In addition, estrogen replacement therapy, which as described above improves endothelial function, can also reduce adverse coronary events up to 50%. The suggestion that the beneficial effect of estrogen on cardiovascular risk could be exerted through a mechanism involving endothelial cells is based on the finding that nitric oxide is not only a vasodilator substance but can also inhibit monocyte adhesion,⁴²⁴³ smooth muscle cell proliferation,⁴⁴ and platelet aggregation.⁴⁵⁴⁶ All of these effects could play an important role in inhibiting or at least retarding the development of atherosclerosis.

In conclusion, the present study indicates that in both normotensive and essential hypertensive women, menopause is associated with the onset or worsening of aging-associated endothelial dysfunction, respectively, a finding that could indicate a possible mechanism responsible for the acceleration of cardiovascular risk in postmenopausal women.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Footnotes