|
|
Background In postmenopausal women with coronary artery disease, conjugated equine estrogen with or without continuous administration of medroxyprogesterone acetate has failed to slow the progression of atherosclerosis. Whether 17
Methods We conducted a double-blind, placebo-controlled trial in 226 postmenopausal women (mean age, 63.5 years) who had at least one coronary-artery lesion. Participants were randomly assigned to usual care (control group), estrogen therapy with micronized 17
Results After a median of 3.3 years of follow-up, the mean (±SE) change in the percent stenosis in the 169 participants who had a pair of matched angiograms was 1.89±0.78 percentage points in the control group, 2.18±0.76 in the estrogen group, and 1.24±0.80 in the estrogen–progestin group (P=0.66 for the comparison among the three groups). The mean difference in the percent stenosis between the estrogen group and the control group was 0.29 percentage point (95 percent confidence interval, –1.88 to 2.46), and the mean difference between the estrogen–progestin group and the control group was –0.65 (95 percent confidence interval, –2.87 to 1.57).
Conclusions In older postmenopausal women with established coronary-artery atherosclerosis, 17
-estradiol (the endogenous estrogen molecule) alone or administered sequentially with medroxyprogesterone acetate can slow the progression of atherosclerosis is unknown. -estradiol alone (estrogen group), or 17-estradiol plus sequentially administered medroxyprogesterone acetate (estrogen–progestin group). In all patients the low-density lipoprotein (LDL) cholesterol level was reduced to a target of less than 130 mg per deciliter. The primary outcome was the average per-participant change between base-line and follow-up coronary angiograms in the percent stenosis measured by quantitative coronary angiography. -estradiol either alone or with sequentially administered medroxyprogesterone acetate had no significant effect on the progression of atherosclerosis.
-estradiol (the endogenous estrogen molecule) would have effects on the progression of atherosclerosis different from those of other formulations of estrogen that have been studied.10
We present the primary results of the Women's Estrogen–Progestin Lipid-Lowering Hormone Atherosclerosis Regression Trial (WELL-HART), a randomized, double-blind, placebo-controlled trial that was designed to test the effects of oral micronized 17-estradiol with or without sequentially administered medroxyprogesterone acetate on the progression of atherosclerosis in postmenopausal women with angiographically documented coronary artery disease.
Methods
Study Participants
Women who either had or did not have a uterus were eligible for the study if they were postmenopausal (as indicated by a serum estradiol level below 20 pg per milliliter), were 75 years of age or younger, had a low-density lipoprotein (LDL) cholesterol level of 100 to 250 mg per deciliter (2.59 to 6.46 mmol per liter) and a total triglyceride level of less than 400 mg per deciliter (4.52 mmol per liter), and had at least one coronary-artery lesion occluding 30 percent or more of the luminal diameter. Women who had undergone percutaneous transluminal coronary angioplasty were eligible if they had at least 20 percent stenosis in a segment of a coronary artery that was not crossed by the guidewire used for angioplasty. Women who had undergone coronary-artery bypass grafting were eligible if they had at least 20 percent stenosis in a segment of a coronary artery that was not proximal to a patent graft. Women were excluded if they smoked more than 15 cigarettes per day, had received a diagnosis of breast cancer or gynecologic cancer within the five years before screening, had a life-threatening disease and a projected survival of less than five years, had a diastolic blood pressure of more than 110 mm Hg, had a fasting serum glucose concentration of more than 200 mg per deciliter, had thyroid disease, had a serum creatinine level of more than 2.5 mg per deciliter (220 µmol per liter), had congestive heart failure (Killip class III or IV and an ejection fraction below 30 percent), had more than five hot flashes per day that interfered with their daily activities, had plans to undergo a coronary-artery revascularization procedure within six months after the first screening visit, had a base-line coronary angiogram that had been obtained before or less than six months after a revascularization procedure, or had had a myocardial infarction less than six weeks before the first screening visit. All participants gave written informed consent, and the study was approved by the institutional review board at the University of Southern California.
Screening, Randomization, and Treatment
The study was conducted from June 1995 to October 2000. Participants were recruited from five sites. Eligible participants were randomly assigned to one of the three treatment groups, with stratification according to the presence or absence of diabetes mellitus. The data coordinating center performed the randomization with the use of a computerized random-number generator. Treatment-group assignment was carefully monitored, and adaptive randomization11 was used to adjust for imbalances among the treatment groups in the total cholesterol level. The participants, gynecologists, clinical staff, and image analysts were unaware of the treatment-group assignment.
Participants randomly assigned to the estrogen group received 1 mg of oral micronized 17-estradiol (Estrace, Mead Johnson) daily, plus a placebo tablet matching medroxyprogesterone acetate for 12 consecutive days of every month. Participants randomly assigned to the estrogen–progestin group received 1 mg of oral micronized 17-estradiol daily, plus 5 mg of medroxyprogesterone acetate (Provera, Upjohn) daily for 12 consecutive days of every month. Participants randomly assigned to the control group received two placebo tablets, one matching the 17-estradiol and taken daily and the other matching the medroxyprogesterone acetate and taken for 12 consecutive days of every month. The LDL cholesterol level was reduced to a target of less than 130 mg per deciliter (3.36 mmol per liter) by means of dietary intervention (25 percent of calories from fat and 7 percent from saturated fats; less than 200 mg of dietary cholesterol per day) and lipid-lowering therapy (primarily with a hydroxymethylglutaryl coenzyme A [HMG-CoA] reductase inhibitor).
Follow-up
Participants had follow-up visits every month for the first 6 months and every other month for the remainder of the trial (36 months). Dietary intake was monitored with the use of a three-day dietary diary (Nutrition Scientific); compliance with study medication was assessed with the use of pill counts and measurement of serum estradiol; and the use of nonstudy medications and dietary supplements was ascertained. Vital signs and clinical events were recorded at each visit. Blood samples were drawn after an eight-hour fast every six months. Electrocardiography, mammography, a Papanicolaou smear, and a pelvic examination with transvaginal ultrasonography were performed yearly. Uterine biopsy was performed if the endometrial thickness was greater than 5 mm.
Acquisition and Evaluation of Coronary Angiograms
Coronary angiography was performed with the percutaneous femoral technique, and right and left anterior oblique views were obtained in order to show all lesions.12,13,14,15 Follow-up angiography was scheduled three years after the base-line angiogram was obtained and was performed according to the same protocol followed at base line. A clinically indicated coronary angiogram obtained within six months before the scheduled final angiography or before a revascularization procedure followed the same protocol used at base line and was used as the final angiogram if one was not available.
With the treatment-group assignment masked, all readable pairs of angiograms showing identical views of the coronary arteries were evaluated by an expert panel of two angiographers and a moderator.12,13,14,15,16 The panel reached a consensus on a global change score (indicating regression, no change, or progression) that integrated the visual changes that they observed. All lesions in native arteries (excluding those proximal to grafts) and all lesions in grafts were assessed.
Quantitative coronary angiographic analyses were performed according to validated methods by a single technician who was unaware of the treatment-group assignment.12,13,15,17,18 Arterial segments were defined as extending from branch to branch. All lesions in native arteries (excluding those proximal to grafts) and all lesions in grafts were analyzed. The percent stenosis and the minimal luminal diameter were measured at the site of lesions identified by the panel, by the imaging analyst, or both.
Biochemical Variables
Lipoproteins and estradiol were measured with the use of standardized enzymatic19 and radioimmunoassay3 methods, respectively.
Statistical Analysis
The primary end point was the average (per-participant) change from base line in the percent stenosis in all lesions evaluated by quantitative coronary angiography. According to the power calculations, a sample of 61 participants in each treatment group was required in order to detect a treatment-effect size (the mean difference between the estrogen group and the control group or between the estrogen–progestin group and the control group in the percent stenosis divided by the standard deviation of the difference) of 0.51 or greater with 80 percent power at the 0.05 level of significance (two-sided). Two secondary end points were the average (per-participant) change in minimal luminal diameter (on quantitative coronary angiography) and the global change score. Analyses of the changes in the percent stenosis and the minimal luminal diameter were specified a priori for subgroups defined according to the presence or absence of diabetes and subgroups defined according to the percent stenosis at base line (<50 percent or 
50 percent).
Analyses included participants with a scheduled follow-up angiogram or an early clinically indicated follow-up angiogram; participants without a follow-up angiogram were excluded. Analysis of variance and the chi-square test were used to compare the treatment groups at base line and during the study in terms of changes in lipoprotein and estradiol levels and the incidence of adverse events. Analysis of covariance (with covariates for the presence or absence of diabetes and for variables that were unevenly distributed among the groups at base line) was used to compare the differences among the treatment groups in the percent stenosis and the minimal luminal diameter. The treatment groups were compared in terms of the global change scores with the use of chi-square methods, with and without adjustment for diabetes status. Statistical analyses were conducted with the use of SAS software (SAS Institute), and a P value of 0.05 was considered to indicate statistical significance. No interim analyses were performed.
Results
Base-Line Characteristics
Of the 1369 women who were prescreened, 773 had qualifying base-line angiograms and were invited for screening (Figure 1). Of these women, 68 (9 percent) were excluded for medical reasons, and 479 (62 percent) were excluded on the basis of other trial criteria; the remaining 226 women (29 percent, 115 [51 percent] of whom had diabetes and 111 [49 percent] of whom did not have diabetes) underwent randomization.
|
|
Compliance
The mean level of compliance with study treatment was 93.6 percent with the estrogen-matching placebo and 98.4 percent with the progestin-matching placebo in the control group, 92.6 percent with estrogen and 99.9 percent with the progestin-matching placebo in the estrogen group, and 94.1 percent with estrogen and 96.1 percent with progestin in the estrogen–progestin group (P=0.61 for the comparison among the three groups with respect to the use of estrogen or estrogen-matching placebo; P=0.28 for the comparison among the three groups with respect to the use of progestin or progestin-matching placebo). At base line, there were no significant differences among the treatment groups in the serum estradiol level (P=0.74 for the comparison among the three groups). During follow-up, the serum estradiol levels and the changes in these levels were significantly different among the treatment groups (P<0.001 for the comparison among the three groups), with a significant increase from base line to follow-up in the mean serum estradiol level in the estrogen group (from 13.7 pg per milliliter to 39.5 pg per milliliter [50.3 pmol per liter to 145.0 pmol per liter], P<0.001) and the estrogen–progestin group (from 12.8 pg per milliliter to 43.4 pg per milliliter [47.0 pmol per liter to 159.3 pmol per liter], P<0.001). There was no significant change in the control group (from 13.2 pg per milliliter to 13.0 pg per milliliter [48.5 pmol per liter to 47.7 pmol per liter], P=0.84).
Coronary Angiographic Outcomes
The median duration of angiographic follow-up was 3.3 years (interquartile range, 3.1 to 3.5) and did not differ significantly among treatment groups (P=0.91). Table 2 presents the results of quantitative coronary angiography adjusted for diabetes status, race or ethnic group, and age. The percent stenosis and the minimal luminal diameter at base line did not differ significantly among the treatment groups. The mean change in these end points among all lesions (or among lesions stratified according to severity) did not differ significantly among the groups. The mean progression of stenosis in participants with diabetes was approximately twice as rapid as that in participants without diabetes. The angiographic end points did not differ significantly among the treatment groups, either among participants with diabetes or among those without diabetes.
|
|
There were no significant differences among the treatment groups in the lipoprotein levels at base line (Table 4). As compared with participants in the control group, participants in the estrogen and estrogen–progestin groups had significantly greater percentage increases in the high-density lipoprotein (HDL) cholesterol level and significantly greater percentage decreases in the LDL cholesterol level.
|
In total, there were nine deaths (four in the control group, two in the estrogen group, and three in the estrogen–progestin group). The causes of death included cardiovascular causes (in five participants), lung cancer (in one), sepsis (in one), cryptogenic cirrhosis (in one), and respiratory failure (in one). At least one "hard" cardiovascular event (death from cardiovascular causes, nonfatal myocardial infarction, unstable angina, cerebrovascular accident, transient ischemic attack, reversible ischemic neurologic deficit, deep venous thrombosis, or pulmonary embolism) occurred in 50 of the 226 participants (16 in the control group, 16 in the estrogen group, and 18 in the estrogen–progestin group; P=0.86 for the comparison among the three groups); when revascularization was included as a cardiovascular event (a "soft" cardiovascular event), the total was 67 participants with at least one event (23 in the control group, 19 in the estrogen group, and 25 in the estrogen–progestin group; P=0.50). There was no significant difference among the groups in the number of cardiovascular events occurring during the first year of study treatment.
Among the 126 participants who had a uterus, the endometrial thickness was measured as 5 mm or more in a total of 28 instances in 20 participants in the control group, in 43 instances in 21 participants in the estrogen group, and in 57 instances in 27 participants in the estrogen–progestin group. Simple hyperplasia without atypia occurred in two participants in the estrogen group. Two participants in the estrogen group underwent hysterectomy during the study, one because of complex hyperplasia with atypia and one because of preexisting uterine prolapse. Breast cancer was diagnosed in one participant in the control group; uterine cancer was not diagnosed in any of the participants.
Discussion
As an addition to lipid-lowering therapy, oral 17-estradiol alone and 17-estradiol with sequentially administered medroxyprogesterone acetate had no significant effect on the progression of coronary-artery atherosclerosis in older women with preexisting coronary artery disease who were studied an average of 18 years after menopause. The results of our study are consistent with those of other trials of hormonal interventions for atherosclerosis in women with preexisting cardiovascular disease,4,5,6,7 as well as with those of studies examining cardiovascular events,8,9,20,21,22 of which the imaging end point that we used is highly predictive.23 However, the results of WELL-HART are strikingly different from those of the Estrogen in the Prevention of Atherosclerosis Trial (EPAT),3 a sister study to WELL-HART that we also conducted. EPAT was a randomized, controlled trial that used protocols that were similar to those used in the current trial and that was conducted by the same personnel but that found that, relative to placebo, oral 17-estradiol alone slowed the progression of carotid intima–media thickness.3 Together, the two studies were designed to determine the effects of hormone therapy on the progression of atherosclerosis in postmenopausal women with preexisting cardiovascular disease (WELL-HART) and in postmenopausal women without such disease (EPAT). The divergent outcomes of the two studies may be related to the timing of the intervention relative to the stage of atherosclerosis, as reflected by the different imaging methods used. Carotid-wall thickness is a measure of early, subclinical, asymptomatic atherosclerosis, whereas coronary angiography is used to evaluate late-stage, symptomatic atherosclerosis.
Accumulating data indicate that estrogen has little effect in reversing atherosclerosis once it is established, whereas it significantly reduces the extent of atherosclerosis if therapy is initiated at an early stage.24 In nonhuman primates, when the initiation of estrogen therapy is delayed for two years (equivalent to six years in humans) after oophorectomy, there is no effect on the extent of atherosclerosis.25 However, when estrogen is administered immediately after oophorectomy in primates that are fed an atherogenic diet, the development of atherosclerosis is significantly reduced.26,27,28 Our results are consistent with those of the first set of studies in primates, in which hormone therapy was initiated many years after atherosclerosis had developed. The results of EPAT are consistent with the latter set of studies in primates, in which estradiol therapy was initiated earlier in the atherosclerotic process. The time from menopause to randomization was approximately five years shorter in EPAT than in WELL-HART.
Since the progression of atherosclerosis is silent, women may have advanced but asymptomatic vascular disease many years after menopause. Given this fact, the timing of treatment relative to menopause or perimenopause may be important in slowing the progression of atherosclerosis. In humans, nondiseased coronary-artery vessels dilate in response to the administration of estrogen, whereas diseased vessels do not respond.29 The lack of estrogen-receptor expression in the presence of atherosclerosis could result in a decreased ability of vascular tissue to respond to estrogen.30 This lack of estrogen-receptor expression may result from methylation of the promoter region of the gene for estrogen receptor 
, which occurs in aging and diseased vessels.31
Comparison of the current results with those of EPAT indicates that other factors such as concomitant lipid-lowering therapy may overshadow the potential beneficial effects of estrogen on the progression of atherosclerosis. Although all participants in the WELL-HART study received lipid-lowering therapy, 40 percent of the participants in EPAT did not. Subgroup analyses in EPAT indicated that 17-estradiol therapy alone significantly slowed the progression of atherosclerosis relative to that in the women in the placebo group who did not receive lipid-lowering therapy but had no demonstrable effect relative to that in the women in the placebo group who did receive such therapy.3 Unopposed 17-estradiol alone and lipid-lowering therapy alone had similar effects on the progression of atherosclerosis.3 Similar results were observed in the Asymptomatic Carotid Artery Progression Study.32 These data indicate that estrogen slows the progression of atherosclerosis but that this effect is overshadowed by the effects of lipid-lowering therapy.
The trials that have been completed to date have not adequately tested the hypothesis that estrogen is cardioprotective. Observational studies indicate that there is cardiovascular benefit in women in whom hormone therapy is initiated early after the onset of menopause for the amelioration of hot flashes.1 EPAT has been the closest test of this pattern of hormone use, but it was not a complete test. Hot flashes are a result of vasomotor instability and may occur during a time when the vascular wall is still responsive to hormone therapy.
In contrast to previous trials using continuous, daily administration of a combination of conjugated equine estrogen and medroxyprogesterone acetate,7,8,9 our study found no increase in the rate of coronary events during the first year of hormone therapy. These results suggest that continuous, daily therapy with medroxyprogesterone acetate may initially have negative cardiovascular consequences. Although the statistical power to detect an early increase in the risk of cardiovascular events in our study was limited, other trials testing estradiol therapy have also not shown a significantly increased risk of coronary events during the first year of therapy.3,5,20,21,22 Participants in our study received an HMG-CoA–reductase inhibitor along with hormone therapy, and this may also account for the absence of early cardiovascular events. In subgroup analyses in the Heart and Estrogen/Progestin Replacement Study (HERS), the rate of coronary events during the first year of intervention was equivalent in the placebo and hormone-therapy groups among participants taking HMG-CoA–reductase inhibitors.33 Estrogen therapy increases the production and activity of matrix metalloproteinases, degradative enzymes that are important in the destabilization and rupture of plaque.34,35 The administration of an HMG-CoA–reductase inhibitor along with estrogen may reduce the instability of plaque and the risk of plaque rupture.36,37
In summary, our results are in agreement with those of previous randomized, controlled trials in elderly women with coronary artery disease studied an average of two decades after menopause. Our study also provides additional information regarding a particular estrogen compound and a particular hormone regimen. Our results extend the previous null findings to a population predominantly composed of members of minority groups and to a population of patients with diabetes mellitus. However, comparison with EPAT indicates that estrogen therapy may be effective in slowing the progression of atherosclerosis when it is initiated early in menopause, while the vascular wall remains responsive to estrogen. The difference in outcomes between WELL-HART and EPAT warrants further investigation.
Supported by a grant (U01-HL-49298) from the National Heart, Lung, and Blood Institute and by the Office of Research on Minority Health.
Dr. Hodis reports having received lecture fees from Wyeth and grant support from Eli Lilly. Dr. Lobo reports having received consulting fees or grant support from Berlex, Novartis, Ortho, Solvay, and Wyeth. Dr. Shoupe reports having received consulting or lecture fees or grant support from Berlex, Eli Lilly, Novo Nordisk, Solvay, and Wyeth. Dr. Mehra reports having received lecture fees from Wyeth.
We are indebted to Mead Johnson Laboratories for supplying the micronized 17
-estradiol and matching placebo pills and to Pharmacia & Upjohn for supplying the medroxyprogesterone acetate and matching placebo pills.
Source Information
From the Atherosclerosis Research Unit, Keck School of Medicine (H.N.H., W.J.M., S.P.A., A.S., A.B.S., R.H.S.), the Department of Preventive Medicine (H.N.H., W.J.M., S.P.A., M.T.), the Department of Molecular Pharmacology and Toxicology, School of Pharmacy (H.N.H., A.S.), the Department of Obstetrics and Gynecology (D.S.), and the Division of Cardiovascular Medicine (H.N.H., A.O.M., R.R.), University of Southern California, Los Angeles; the Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, New York (R.A.L.); the Kaiser Permanente Medical Center, Los Angeles (P.R.M.); the Division of Cardiology, University of Chicago, Chicago (D.P.F.); Boston Heart Foundation, Boston (L.C.-H.); the Lakewood Regional Medical Center, Lakewood, Calif. (M.E.S.); the Division of Cardiology, Harbor–UCLA Medical Center, Torrance, Calif. (W.J.F.); Good Samaritan Hospital, Los Angeles (T.L.S.); Presbyterian Intercommunity Hospital, Whittier, Calif. (T.D.G.); Southern California Heart Specialists, Pasadena (K.H.V.); and the Jet Propulsion Laboratory, California Institute of Technology, Pasadena (R.H.S.).
Address reprint requests to Dr. Hodis at the Atherosclerosis Research Unit, 2250 Alcazar St., CSC132, Los Angeles, CA 90033, or at watcher{at}usc.edu.
References
The following persons were members of the WELL-HART Research Group: Steering Committee: H.N. Hodis (Chairman), S.P. Azen, D.P. Faxon, R.A. Lobo, W.J. Mack. Clinical Research Center Staff (University of Southern California): M. Charlson, C. Gesselman, T. Morales, M. Torres, F. Watcher, L. Zurbrugg. Recruitment Centers: Los Angeles County–University of Southern California Medical Center and University Hospital, Los Angeles — D.P. Faxon, A.O. Mehra, R. Rabbani, B. Agra, I. Coria, J. Ee, S. Johnson, L. Reyes, C. Stolicky; Kaiser Permanente Medical Center, Los Angeles — P.R. Mahrer, V.J. Aharonian, R.A. Browning, P. Jackimowicz, P.J. Scutella, A. Velasco, C. Williams; Harbor–UCLA Medical Center, Torrance, Calif. — W.J. French; Good Samaritan Hospital, Los Angeles — T.L. Shook, R.V. Matthews; Huntington Memorial Hospital, Pasadena, Calif. — K.H. Vogelbach, D.L. Ujiiye. Quantitative Coronary Angiographic Reading Laboratory: R.H. Selzer (Director), A.B. Shil, A. Shircore. Data Coordinating Center: W.J. Mack (Director), S.P. Azen, M. Ying-Lai, O. Morales, M. Xiang. Lipid/Lipoprotein Laboratory: A. Sevanian (Director), J. Hwang, O. Bolusan, A. Ramirez. Angiographic Panel Reading: L. Cashin-Hemphill (Moderator), M.E. Sanmarco (Moderator), D.P. Faxon, W.J. French, T.D. Gaarder, P.R. Mahrer, A.O. Mehra, R. Rabbani, T.L. Shook. Study Gynecologists: G. Mezrow, D. Shoupe. Reproductive Endocrinology Laboratory: F.Z. Stanczyk (Director). Gynecological Pathology Laboratory: J.C. Felix (Director). Mammography Reading Center: Y. Parisky (Director). Data Safety Monitoring Committee: B.G. Brown (Chairman), G. Chaudhuri, M. Malloy, T. Odom-Maryon, J. Wilson-Ross, D.J. Gordon (ex officio, National Heart, Lung, and Blood Institute).
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
