Author + information
- Received October 28, 1999
- Revision received July 6, 2000
- Accepted August 18, 2000
- Published online December 1, 2000.
- Giuseppe M.C Rosano, MD, FACC∗,
- Carolyn M Webb, PhD†,
- Sergio Chierchia, MD, FACC∗,
- Gian Luigi Morgani, MD∗,
- Michele Gabraele, MD∗,
- Phillip M Sarrel, MD‡,
- Dominique de Ziegler, MD§ and
- Peter Collins, MD, FACC†,*
- ↵*Reprint requests and correspondence: Dr. Peter Collins, Cardiac Medicine, Imperial College School of Medicine at the National Heart and Lung Institute, Dovehouse Street, London SW3 6LY, United Kingdom
We sought to compare the effects of estrogen/transvaginal progesterone gel with estrogen/medroxyprogesterone acetate (MPA) on exercise-induced myocardial ischemia in postmenopausal women with coronary artery disease or previous myocardial infarction, or both.
Estrogen therapy beneficially affects exercise-induced myocardial ischemia in postmenopausal women; however, women with an intact uterus also take progestin to protect against uterine malignancies. The effects of combination estrogen/progestin therapy on myocardial ischemia are unknown.
Eighteen postmenopausal women (mean ± SD age 59 ± 7 years) were given 17-beta-estradiol in single-blinded manner for four weeks (1 mg/day for three weeks then 2 mg/day for one week). Estradiol (2 mg/day) was then continued, and the patients were randomized (double-blind) for 12 days to either transvaginal progesterone gel (90 mg on alternate days) and oral MPA placebo (10 mg/day), or vice versa. After another two weeks on estradiol alone, the patients crossed over to progestin treatment and repeated the protocol on the opposite treatment. Patients underwent treadmill exercise testing after each estradiol phase and at day 10 of each progestin phase.
Exercise time to myocardial ischemia increased after the first estrogen phase as compared with baseline (mean difference with 95% confidence interval [CI]: 72 s [34 to 110], p = 0.001), and was increased by combination estradiol/progesterone therapy as compared with estradiol/MPA therapy (92 s [35 to 149], p = 0.001)). Two patients (11%) were withdrawn while taking estradiol/MPA owing to unstable angina.
Combination estrogen/transvaginal progesterone gel increases exercise time to myocardial ischemia, as compared with estrogen/MPA. These results imply that the choice of progestin in women at higher cardiovascular risk requires careful consideration.
Cardiovascular disease is the major cause of death in postmenopausal women, with a fatality rate that is double that of all forms of cancer. There is a wealth of epidemiologic evidence indicating that postmenopausal estrogen therapy can significantly reduce the risk of death due to cardiovascular disease in postmenopausal women (1). However, women with an intact uterus must also be given progestin to prevent the increased risk of endometrial hyperplasia and cancer seen with estrogen-only therapy. One large epidemiologic study suggests that estrogen combined with progestin results in a significantly decreased risk of death from cardiovascular disease (2); however, this study was not able to distinguish between different progestins. The Heart and Estrogen Replacement Study (HERS) has reported a null effect of long-term combined treatment with conjugated equine estrogen and medroxyprogesterone acetate (MPA) in postmenopausal women at increased risk of coronary artery disease (CAD) events, as well as a potential increase in the risk of myocardial infarction during the first year of therapy (3). One of a number of reasons for the null effect may involve inhibition of the beneficial effect of estrogen by MPA. Investigation of the effect of progestins in women with established CAD is therefore important.
Treatment with short- and long-term estrogen improves exercise time to myocardial ischemia in postmenopausal women with proven CAD (4,5). An anti-ischemic effect of estrogen has been demonstrated using different experimental techniques in postmenopausal women (6,7). A variety of estrogens have anti-atherogenic properties in various models (8,9). Studies have consistently demonstrated that MPA antagonizes the inhibitory effects of conjugated equine estrogens on coronary artery atherosclerosis in cynomolgus monkeys (10,11). Estrogen can protect against coronary vasoconstriction to acetylcholine (12,13); however, recently it has been shown that MPA can interfere with this protection against coronary vasospasm (14). There is evidence that progesterone can result in endothelium-independent relaxation of coronary arteries at high concentrations (15), and that natural progesterone, unlike MPA, does not antagonize the effect of estrogen on coronary atheroma (10) or coronary vasospasm (14). The choice of progestin, particularly in patients with established CAD, may be very important. We therefore undertook a study to compare the effects of adding oral MPA or transvaginal progesterone gel to estrogen therapy on exercise-induced myocardial ischemia in postmenopausal women with proven CAD.
Postmenopausal women (17-beta-estradiol <100 pmol/liter and follicular stimulating hormone [FSH] >40 IU) with stable exertional angina (16), reproducible positive exercise tests for myocardial ischemia, angiographically proven CAD (defined as at least 70% narrowing of the lumen diameter in one or more of the major coronary arteries or their major branches) and/or previous myocardial infarction were enrolled. No patient had a myocardial infarction in the previous six months. Patients who had taken hormone therapy in the previous three months, were taking digoxin, had abnormal left ventricular function or had resting electrocardiographic (ECG) conduction abnormalities were excluded. The study was undertaken at two sites; therefore, the protocol was approved by the respective institutional Ethics Committee. All patients gave written, informed consent.
One week before commencing the study, all anti-anginal medications other than sublingual nitroglycerin were withdrawn for the duration of the study. Sublingual nitroglycerin and caffeine-containing foods were prohibited for 3 h before each exercise test.
Patients performed a screening treadmill exercise test according to the modified Bruce protocol. Patients with a positive test for myocardial ischemia (ST segment depression ≥1 mm) then commenced a single-blind phase of 17-beta-estradiol (Estrace, Bristol Myers Squibb, Princeton, New Jersey) for four weeks (1 mg/day for 18 days followed by 2 mg/day for 10 days). Estradiol (2 mg/day) was then continued for the duration of the study. At the end of the fourth week, patients were randomized, in double-blinded manner, for 12 days to either oral MPA (10 mg/day) and transvaginal progesterone gel placebo or MPA placebo and active transvaginal progesterone gel (90 mg every other day; Crinone, Columbia Laboratories; Paris, France). After another 16 days on estradiol alone, the patients crossed over to progestin treatment and repeated the protocol on the opposite treatment regimen. Exercise testing was performed after each estradiol phase and on day 10 of each progestin phase (Fig. 1). Study drug compliance was checked by a drug count at the end of each treatment phase.
A venous blood sample for measurement of plasma levels of 17-beta-estradiol, progesterone and FSH was taken immediately before the first screening exercise test and before each test of the single-blind and double-blind phases. Plasma was frozen and analyzed by chemiluminescence (Chiron Diagnostics) at a core laboratory to avoid knowledge of the study treatment.
Exercise tests were performed using a Marquette CASE-15 computer-assisted recorder system. An averaged 12-lead ECG was obtained at rest, every minute during exercise, at the onset of anginal symptoms, at the onset of 1-mm ST segment depression, at peak exercise and every minute during recovery. Leads II, V2 and V5 were continuously monitored. To accurately capture the time to the onset of 1-mm ST segment depression, the investigator initiated the printing of ECGs as the change from baseline ST-segment to 1-mm ST segment depression approached, and continued recording until the ST segment depression decreased beyond 1 mm. Systolic and diastolic blood pressures were measured at rest, at the end of each stage during exercise and every 2 min during recovery.
A positive ECG response was defined as horizontal or downsloping ST segment depression ≥1 mm at 60 ms after the J point (17) occurring in at least two contiguous leads and at least six consecutive complexes. Exercise tests were concluded at the point of physical exhaustion, ST segment depression ≥3 mm, severe angina, severe dyspnea or a decline in systolic blood pressure >20 mm Hg. Total exercise time, time to 1-mm ST segment depression, heart rate and blood pressure at the onset of 1-mm ST segment depression and peak exercise, maximal ST segment depression and time to development of angina during exercise were recorded. The rate–pressure product was calculated as the product of heart rate and systolic blood pressure.
Exercise ECGs were analyzed by an experienced independent investigator who was unaware of the clinical data. The ECG lead showing the greatest ST segment depression in the screening exercise tests was selected for analysis.
Analysis of variance was used for the main analysis, with treatment, period and dose as fixed variables, and patient and period within the patient as random variables. The prestudy baseline values were considered as a covariate and were included in the analysis of variance as such. A significant treatment–period interaction was indicative of a carry-over effect. Validity checks were made using the Anderson-Darling statistic, Cramer von Mises statistic, Watson statistic and Bartlett’s test. Data were log-transformed when the assumption of equal variances was not met (baseline and peak blood pressures and hormone profile data only). These data are expressed as the geometric mean value (95% confidence interval [CI]). All other exercise data are expressed as the mean value (95% CI). Demographic data are expressed as the mean value ± SD. For the hormone profiles, each treatment mean value was compared with the baseline mean value, and the Bonferroni procedure was used to adjust the p values for multiple comparisons. p values ≤0.05 were considered to be statistically significant.
Twenty-six patients were enrolled. Two patients were withdrawn due to unstable angina shortly after commencing MPA treatment in combination with estrogen. One patient withdrew while taking combination estrogen and transvaginal progesterone gel due to heavy vaginal bleeding, which ceased two days after stopping treatment, and one patient withdrew because of severe breast tenderness after commencing the estrogen-only phase. One patient who was above the upper age limit was inadvertently randomized but did not start treatment, one patient withdrew after a car accident and two patients dropped out for unspecified personal reasons. Therefore, 18 women completed the study (Table 1). Eight patients had a previous myocardial infarction and one had previous coronary artery bypass graft surgery. Four patients had a history of hyperlipidemia (range for all patients: 5.4 to 9.2 mmol/liter) and three of these patients were taking lipid-lowering medication. All patients had clinical and biochemical evidence of postmenopausal status (Table 2).
Baseline hormonal levels confirmed the menopausal status of the participating subjects (Table 2). Estrogen treatment raised plasma 17-beta-estradiol (p < 0.001) and lowered FSH (p < 0.001) (Table 2). Cyclical treatment with MPA or vaginal progesterone further lowered plasma FSH (both p < 0.001) (Table 2). Plasma progesterone increased compared with baseline when women received progesterone gel (p < 0.001) (Table 2). The amplitude of this elevation is concordant with the subphysiologic levels of progesterone achieved with the vaginal gel (18). Plasma progesterone levels, however, were not different from baseline levels when receiving MPA (a molecule not readily measurable). This confirms the lack of cross-reactivity of the progesterone antibody with MPA. The decrease in plasma FSH after MPA treatment provides an indirect reflection of treatment compliance.
Estradiol alone increased the time to 1-mm ST segment depression, as compared with baseline (p < 0.001) (Table 3 and Fig. 2). There was a significant increase in exercise time to the onset of 1-mm ST segment depression by combination estradiol/progesterone as compared with combination estradiol/MPA (Table 3 and Fig. 2), with a mean difference of 92 s (95% CI 35 to 149) (p < 0.001). Nine patients demonstrated at least 1-mm ST segment depression after taking progesterone, and 11 patients demonstrated at least 1-mm ST segment depression after taking MPA. Five patients had a negative exercise test for myocardial ischemia after taking both combination progesterone and combination MPA. For analysis, total exercise time was substituted for time to 1-mm ST segment depression in patients who did not reach 1-mm ST segment depression with treatment. There were no differences between the measured exercise variables after the first and second estrogen-alone phases.
There was a trend toward an increase in total exercise time by combination progesterone therapy as compared with combination MPA therapy (mean difference 31 s (95% CI 7 to 68) (p = 0.083) (Table 3). The rate–pressure product at rest, but not at 1-mm ST segment depression or peak exercise, was significantly decreased after combination progesterone therapy, as compared with after combination MPA therapy (mean difference 0.926 [95% CI 0.859 to 0.998]) (p = 0.038). Treatment with combination MPA tended to increase the rest heart rate as compared with combination progesterone therapy (mean difference 5 beats/min [95% CI 0.31 to 11]) (p = 0.058), but there was no difference in the heart rate at 1-mm ST segment depression or peak exercise. There was no difference in systolic blood pressure between the treatment regimens. Treatment–period interactions did not occur, indicating no carry-over effect for any variable, and no training effect was shown.
Findings and comparison with previous studies
Our results demonstrate that natural progesterone gel, when added to estradiol treatment, results in an increase in exercise time to myocardial ischemia compared with estrogen alone. In contrast, oral MPA has no effect on exercise time to myocardial ischemia in patients already taking estrogen. It has previously been demonstrated that both short-term (4,6) and long-term (5) estrogen therapy can delay the onset of signs of myocardial ischemia on the ECG and increase exercise tolerance. The results of the present study indicate a synergistic effect of estrogen and progesterone, but not estrogen and MPA, on exercise time to myocardial ischemia. This is a novel finding in the setting of CAD.
Similar to the findings of the long-term estrogen study by Webb et al. (5), we found no significant effect of either combined hormone regimen on total exercise time in the present study. However, there was a trend toward an increase in total exercise time by combination progesterone therapy as compared with combination MPA therapy (p = 0.083). Studies of conventional anti-anginal therapy, such as verapamil, have also shown a significant effect on the time to onset of signs of myocardial ischemia, but no effect on total exercise time (19).
Two patients (11%) developed unstable angina shortly after commencing the MPA treatment phase (two and five days) and were withdrawn from the study. These patients were successfully managed medically, then one went on to have percutaneous transluminal coronary angioplasty. Unstable angina is the result of complex pathophysiology that includes vasoconstriction, atherosclerotic plaque rupture and platelet-induced thrombosis (20). An MPA-induced tendency toward coronary vasoconstriction at the site of atheroma is one possible explanation for the development of unstable angina in these patients (14). The other possibility is that MPA results in an increased thrombotic tendency, although animal data suggest that this is unlikely (10).
The observed significant increase in time to myocardial ischemia after combination estrogen plus progesterone gel could involve a number of mechanisms. Progesterone may have a synergistic effect with estrogen, however arguing against this is the very small measured increase in plasma concentration of progesterone with transvaginal progesterone gel. Another hypothesized mechanism may be that because the plasma levels of progesterone were so low, our observed increase in exercise time was purely due to an unopposed effect of a further 12 days of estrogen therapy. As MPA has been shown in animal models to oppose the beneficial vascular effects of estrogen, in the present study, MPA may be nullifying a potential further beneficial effect of more prolonged estrogen exposure on myocardial ischemia. For example, we have shown a further increase in the time to onset of myocardial ischemia after eight weeks of estrogen therapy as compared with four weeks (5). However, we showed no difference in measured exercise variables between the first and second estrogen-alone phases.
Effects of combination hormone therapy on vascular reactivity
Epidemiologic data suggesting a beneficial effect of postmenopausal hormone therapy on CAD risk are largely based on studies using unopposed estrogen; however, combination estrogen/progestin therapy is required to decrease the risk of endometrial hyperplasia and subsequent uterine carcinoma. One of the major unanswered questions in hormone replacement therapy for hypoestrogenic women is how to balance estrogen with progestin. This becomes more important in patients with CAD in whom changes in vasomotion may be detrimental. It is also particularly pertinent because the recent HERS study reported that continuous, combined, oral, conjugated equine estrogen plus oral MPA did not reduce the overall rate of CAD events in postmenopausal women with established CAD (3). In the early phase of HERS, there was an increase in CAD events in patients taking hormone therapy, specifically an increase in nonfatal myocardial infarction during the first year (first few months) of starting therapy. The survival curves then converged after one year and diverged during the second year to show a protective effect of hormone therapy in the latter years of the study. Although the patients in the present study and HERS were similar—namely, postmenopausal women with existing coronary atherosclerosis—exercise-induced myocardial ischemia is a measure of fixed disease in our study, as compared with the end point of cardiovascular events in HERS. Also, different estrogens were used: HERS used a continuous, combined preparation, whereas a cyclical regimen was used in the present study.
It is possible that MPA may oppose the vasodilator actions of estrogen, as has been demonstrated in animal models (21). Miyagawa et al. (14) studied the effect of physiologic levels of 17-beta-estradiol combined with MPA or progesterone on coronary vasomotion in ovariectomized rhesus monkeys. Coronary vasospasm was demonstrated in response to physiologic stimulation without injury when the animals were treated with MPA plus estradiol, but not when treated with progesterone plus estradiol. This latter combination protected against acetylcholine-induced vasospasm. These investigators concluded that MPA, in contrast to progesterone, increases the risk of coronary vasospasm. This finding concurs with our findings in that two patients were withdrawn from the study because of unstable angina after the initiation of the MPA phase of the study.
Effects of combination hormone therapy on atherosclerosis
In addition to its effect on vasomotor tone, MPA antagonizes the inhibitory effects of conjugated equine estrogens on coronary atherosclerosis in cynomolgus monkeys (10) and of 17-beta-estradiol in the rat carotid artery. This may be at least partially due to an inhibitory effect of MPA on the beneficial lipid effects of estrogen. In the Postmenopausal Estrogen-Progestin Interventions (PEPI) trial, MPA partially inhibited the estrogen-associated increase in HDL cholesterol more than micronized progesterone (22). However observational studies have shown a similar reduction in coronary heart disease risk in women using MPA plus estrogen as in women taking unopposed estrogen (23). There are a number of reports that clearly show a detrimental effect of MPA on the beneficial effects of conjugated equine estrogens with regard to atheroma development (10,11) and vascular reactivity (14,21). There are also studies showing that progesterone does not appear to have this inhibitory effect either on atheroma development (8,24) or vascular reactivity in animal models (14,15), nor on vascular reactivity in humans (25). Progesterone has an anti-smooth muscle proliferative effect in vitro (26), a feature that may be important in that atherosclerotic plaques of women who have suffered acute sudden death have more smooth muscle cell proliferation than male subjects (27).
The present study was designed to investigate whether the addition of progestins in women taking estrogen affects the anti-ischemic properties of estrogen. The design might have included a progestin placebo-placebo arm or might have given the estradiol in a double-blind, placebo-controlled fashion. However, this would have added to the already complex design of the study and may not have contributed any further to the current knowledge in this field. Indeed, previous randomized, placebo-controlled studies have reported effects of estrogen alone on myocardial ischemia (4–6).
Both the present study and HERS argue that the choice of progestin may be important in higher risk women with established CAD. The choice of natural progesterone in this group of patients would seem more appropriate, as natural progesterone gel enhances the effects of estrogen on exercise-induced myocardial ischemia. Further randomized studies investigating different estrogens and progestins, as well as doses and routes of administration in different patient groups, will be required before definitive conclusions can be made about the use of hormone therapy in postmenopausal women who have, or are at increased risk of having, cardiovascular disease.
☆ This study was supported by Columbia Laboratories, Paris, France.
- coronary artery disease
- confidence interval
- follicular stimulating hormone
- Heart and Estrogen Replacement Study
- medroxyprogesterone acetate
- Received October 28, 1999.
- Revision received July 6, 2000.
- Accepted August 18, 2000.
- American College of Cardiology
- Alpaslan M,
- Shimokawa H,
- Kuroiwa-Matsumoto M,
- Harasawa Y,
- Takeshita A
- Rosano G.M.C,
- Caixeta A.M,
- Chierchia S.L,
- et al.
- Adams M.R,
- Kaplan J.R,
- Manuck S.B,
- et al.
- Bellinger D.A,
- Williams J.K,
- Adams M.R,
- Honore E.K,
- Bender D.E
- Levine R.L,
- Chen S.J,
- Durand J,
- Chen Y.F,
- Oparil S
- Williams J.K,
- Adams M.R,
- Herrington D.M,
- Clarkson T.B
- Williams J.K,
- Adams M.R,
- Klopfenstein H.S
- Gibbons R.J,
- Chatterjee K,
- Daley J,
- et al.
- Livesley B,
- Catley P.F,
- Campbell R.C,
- Oram S
- Davies M.J,
- Thomas A.C
- Williams J.K,
- Honore E.K,
- Washburn S.A,
- Clarkson T.B
- Register T.C,
- Adams M.R,
- Golden D.L,
- Clarkson T.B
- Gerhard M,
- Walsh B.W,
- Tawakol A,
- et al.