Author + information
- Received November 26, 1997
- Revision received March 17, 1998
- Accepted April 9, 1998
- Published online July 1, 1998.
- Sandra J. Lewis, MD, FACCa,
- Frank M. Sacks, MD∗,* (, )
- Jayne S. Mitchell, MS, ANPa,
- Cara East, MD†,
- Stephen Glasser, MD, FACC‡,
- Sheren Kell, MD§,
- Rebecca Letterer, RN, BSN∥,
- Marian Limacher, MD, FACC#,
- Lemuel A. Moye, MD, PhD∗∗,
- Jean L. Rouleau, MD, FACC††,
- Marc A. Pfeffer, MD, PhD, FACC∗,‡‡,
- Eugene Braunwald, MD, FACC∗,
- for the CARE Investigators
- ↵*Address for correspondence: Dr. Frank Sacks, Nutrition Department, Harvard School of Public Health, 665 Huntington Avenue, Boston, Massachusetts 02115
Objectives. We sought to determine the effect of pravastatin on recurrent cardiovascular events in women with average cholesterol levels after myocardial infarction (MI).
Background. Little information is available on the effectiveness of lipid lowering in secondary prevention of coronary heart disease (CHD) in women; in particular, those with CHD and average cholesterol levels.
Methods. In the Cholesterol and Recurrent Events (CARE) trial, 576 postmenopausal women, between 3 and 20 months after MI, with a total cholesterol level <240 mg/dl and a low density lipoprotein cholesterol level 115 to 174 mg/dl, were randomized to receive pravastatin 40 mg/day or matching placebo for a median follow-up period of 5 years. The main outcome measures were combined coronary events (coronary death, nonfatal MI, percutaneous transluminal coronary angioplasty [PTCA] or coronary artery bypass graft surgery [CABG]), the primary trial end point (coronary death or nonfatal MI) and stroke.
Results. Women treated with pravastatin had a risk reduction of 43% for the primary end point (p = 0.035), 46% for combined coronary events (p = 0.001), 48% for PTCA (p = 0.025), 40% for CABG (p = 0.14) and 56% for stroke (p = 0.07). The 3,583 men in the CARE trial also showed a reduction in risk, but the magnitude tended to be less. Pravastatin improved plasma lipids similarly in men and women. There were no differences in risk of coronary events in the placebo group between men and women. Minor differences between men and women were present in baseline characteristics and treatment for MI, in general, conferring a higher risk status and a lower incidence of CABG in the women.
Conclusions. Pravastatin led to significant early reduction of a wide range of cardiovascular events in post-MI women with average cholesterol levels.
Although cardiovascular disease remains the major cause of death in women (1–4), the exclusion or underrepresentation of women in clinical trials of lipid-lowering treatment has hampered the development of evidence-based standards of care for women with coronary heart disease (CHD) (5). A total of 402 women were included in secondary prevention trials that used therapies other than the hydroxymethylglutaryl coenzyme A (HMG CoA) reductase inhibitors (6–9). Overall, these trends showed a tendency toward fewer recurrent coronary events in the women (10,11), although no single trial found a significant benefit. Thus, guidelines for cholesterol treatment in women were not evidence based but were extrapolated from data in men (12,13), or it was recommended that guidelines be deferred (14).
The greater efficacy of HMG CoA reductase inhibitors than diet or previously available pharmacologic therapies and the policy of newer secondary prevention trials to enroll women have provided the much-needed opportunity to investigate whether contemporary lipid lowering therapy benefits women. The Scandinavian Simvastatin Survival (4S) trial demonstrated that hypercholesterolemic women with a myocardial infarction (MI) or angina treated with an HMG CoA reductase inhibitor experienced significant reduction in coronary events (15). However, the majority of women with CHD do not have hypercholesterolemia, and information is needed on the effects of lipid treatment in women with average cholesterol concentrations.
The Cholesterol and Recurrent Events (CARE) trial (16)investigated whether reducing average cholesterol concentrations with pravastatin in patients with an MI would prevent recurrent cardiac events. It showed that in post-MI patients with a cholesterol concentration <240 mg/dl, pravastatin reduced the risk of CHD death or recurrent MI by 24% (p = 0.003), coronary artery bypass graft surgery (CABG) or percutaneous transluminal coronary angioplasty (PTCA) by 27% (p < 0.001) and stroke by 31% (p = 0.03). Of the 4,159 patients, 576 were women. Pravastatin therapy reduced the risk of total coronary events, defined as fatal CHD, nonfatal MI, CABG and PTCA, by 46% in women (p = 0.001) and by 20% in men (p = 0.001).
The present report sought to determine the effect of pravastatin on the risk of individual cardiovascular events, including MI, CABG, PTCA and stroke, in the women and men in the CARE trial. In addition, their clinical characteristics, including risk factors and severity of coronary disease, and medical therapies were examined in relation to coronary event reduction. Finally, we compared the treatment given women and men with a recent MI in the CARE trial.
Study design and patients
The design and results of the CARE trial have been described in detail elsewhere (16,17). The study was a randomized, double-blind, placebo-controlled trial with median follow-up period of 5 years and was carried out in 80 participating centers throughout the United States and Canada (16). Institutional review board approval was obtained at all centers. Patients were eligible if they were 21 to 75 years old and had an acute MI between 3 and 20 months before randomization, a plasma total cholesterol level <240 mg/dl, a low density lipoprotein (LDL) cholesterol level 115 to 174 mg/dl and a fasting triglyceride level <350 mg/dl. Women were required to be post menopausal or surgically sterile. All participants received diet counseling using the National Cholesterol Education Program Step 1 guidelines (13)and were randomized to treatment with either pravastatin (40 mg/day) or matching placebo. The patient was the unit of randomization; blinded treatment assignment was given by telephone by a member of the Data Coordinating Center (University of Texas School of Public Health, Houston, Texas) to a staff member of a clinical center at the randomization visit. The prespecified primary outcomewas defined as fatal CHD or confirmed nonfatal MI. For analysis of the treatment effects in subgroups, an expanded end point was used and was defined as fatal CHD, nonfatal MI, CABG or PTCA (“combined coronary events”).
All analyses were performed on an intention to treat basis. A p value (two-tailed) <0.05 was considered statistically significant. Baseline characteristics in the two treatment groups were determined using standard zscores (18)for continuous variables and chi-square tests for categoric variables. All hypothesis testing and all risk reductions with their confidence intervals were assessed using the Cox proportional hazards model. Kaplan-Meier survival curves for the control and pravastatin groups were calculated (19).
Cox proportional hazards analysis was used to assess the relation between baseline characteristics and the occurrence of end points. In this analysis, risk factor status was assessed at baseline. Patients were then tracked at 3-month intervals.
The assessment of a potential difference between the therapeutic effect in women and men was addressed by adding an interaction term (the product of an indicator variable for therapy and an indicator variable for gender) in a model that included an effect for therapy and an effect for gender. If the effect of pravastatin on the end point was different in men than in women, this interaction term would be statistically significant. Because a possible explanation for the statistically significant interaction might be the influence of other baseline risk factors (e.g., previous MI, time since the index MI, baseline LDL cholesterol or diabetes), these, as well as other baseline risk factors, were added to this interaction model.
Reasons for exclusion before randomization
The clinical records from hospitals and ambulatory practices for a total of 11,207 patients (2,132 women [19%], 9,075 men [81%]) were screened, and 4,159 qualified for the CARE trial. Of the 2,132 women screened, 576 (27%) qualified for the study. Women were more likely to be excluded than men (73% vs. 61%, p < 0.001). Baseline characteristics for exclusion differed between women and men, respectively, as follows: 1) total cholesterol or triglyceride levels too high (15% vs. 12%); 2) LDL cholesterol levels too low (7% vs. 17%); and 3) criteria for MI not met (63% vs. 54%, p < 0.001 for all comparisons). No significant differences were found for concomitant medical conditions, disapproval by personal physician or refusal by patient.
Baseline characteristics of women and men
The average time from MI to randomization was 323 days for women and 306 days for men (p < 0.05) (Table 1). Women were older than men at randomization (61 vs. 58 years). The prevalence at baseline of the following risk factors was significantly greater in women than in men, respectively: hypertension (54% vs. 41%), diabetes (20% vs 13%), current smoking (30% vs. 20%) and family history of CHD (45% vs. 40%). Women were more likely than men to have multiple risk factors for CHD.
Women had a higher mean total cholesterol concentration than men (215 vs. 208 mg/dl), resulting from a higher high density lipoprotein (HDL) cholesterol concentration (45 vs. 38 mg/dl, p < 0.05) (Table 2). LDL cholesterol, very low density lipoprotein cholesterol and triglyceride levels were similar in women and men.
Fewer women than men in the CARE trial received thrombolytic therapy for their MI (38% vs. 43%), and fewer underwent CABG between the time of their qualifying MI and randomization (15% vs. 23%, p < 0.05). The use of other coronary procedures, including angiography and PTCA, was not significantly different. Women more frequently had evidence of clinical congestive heart failure at baseline, despite a higher mean left ventricular ejection fraction (both p < 0.05).
Therapy received at the time of randomization also differed significantly. Fewer women than men were given aspirin (80% vs. 84%), and beta-adrenergic blocking agents (36% vs. 41%), whereas more received calcium channel blocking agents (46% vs. 38%) and diuretic drugs (21% vs. 10%). Only 10% of women were receiving estrogen replacement therapy, and only 3% progesterone, at study entry.
Effects on plasma lipids
Pravastatin had a similar effect on plasma lipids in women and men: Averaged over 5 years, total cholesterol was decreased by 19% to 20%, LDL cholesterol by 28%, triglycerides by 13% to 14%; HDL rose by 4% to 5% (Table 2). The reduction in LDL was similar for the 5-year duration.
The risk of a recurrent event in the placebo-treated patients was similar in women and men (Fig. 1). Women in the pravastatin group experienced an early reduction in risk for major coronary events, with the treatment curve beginning to separate from the placebo curve at 1 year (Fig. 2). The overall risk reduction was 46% for women versus 20% for men (Fig. 2, Tables 3 and 4). ⇓⇓The test for significance for an interaction between therapy and gender showed a p value of 0.048 (Table 5).
Pravastatin therapy caused a substantial reduction of all major cardiovascular events in women (Table 3): The primary end point of CHD death or nonfatal MI was decreased by 43% (p = 0.035), fatal or non-fatal MI by 56% (p = 0.009), PTCA by 48% (p = 0.025), CABG by 40% (p = 0.14) and stroke by 56% (p = 0.07). Men in the pravastatin group also showed a significant reduction in risk for the primary and combined coronary end points and in revascularization procedures, but the magnitude of the reduction was lower than that in women (Table 4).
Cancer was reported in 12 women in the placebo group and 23 in the pravastatin group (5 vs. 2 for gastrointestinal, 1 vs. 12 for breast, 3 vs. 4 for genitourinary, 3 vs. 4 for respiratory, 0 vs. 1 for “other,” placebo vs. pravastatin, respectively). Of the 12 cases of breast cancer in the pravastatin group, 3 occurred in patients with previous breast cancer, 1 was ductal carcinoma in situ, and 1 occurred in a patient who took pravastatin for only 6 weeks. The one case in the placebo group occurred in a woman with previous breast cancer. Total cancer occurrence in men was 133 for the placebo group and 124 for the pravastatin group (colon 12 vs. 6, gastrointestinal 16 vs. 16, nervous system 2 vs. 4, respiratory 35 vs. 31, genitourinary 58 vs. 56, melanoma 3 vs. 3, “other” 7 vs. 8, placebo vs. pravastatin, respectively).
The CARE trial compared the effect of pravastatin and placebo in 576 post-MI women with average cholesterol levels who received treatment typical of North American patients with this condition. After an average duration of treatment of 5 years, the rates of the major coronary events were reduced by 40% to 55% in women. The risk of stroke was reduced 56% in the pravastatin group (p = 0.07), a new finding in women with lipid-lowering treatment.
Paucity of information on cholesterol-lowering treatment in women
Exclusion of women from studies of cholesterol lowering has hampered the development of evidence-based standards of care for women with CHD (20). The need for information in women regarding the efficacy of cholesterol treatment is compelling because cardiovascular disease is the leading cause of death in women. Moreover, at the time the CARE trial was designed, data from the late 1980s showed that mortality rates from MI were higher in women than in men, and it was suspected that this result might be related to less aggressive care, fewer procedures and, possibly, less optimal medication (21–26)received by women who were shown to be at higher risk than men for complications from invasive cardiovascular procedures (27–30). There was considerable uncertainty as to whether gender differences reflected treatment bias or appropriate clinical judgment (31). This uncertainty underscored the need for effective treatments for secondary prevention in women. The CARE study had, as a prestated goal, the inclusion of large numbers of women to facilitate the development of evidence-based recommendations.
Enrollment of women
The ability to randomize large numbers of women in a long-term clinical trial remained a challenge. A large series of patients with MI showed that 26% of those <75 years old were women (32). In the CARE trial with its upper age limit of 75 years, only 19% of the screened population were women, possibly reflecting a lower proportion of women in the CARE hospitals than reported elsewhere (33)or perhaps reluctance of research staff to approach elderly women (e.g., 70 to 75 years old) for participation in research because of practical concerns related to adherence and retention; barriers such as transportation and physical limitations; and dependence on other care givers for compliance. Whereas 19% of the screened population were women, only 14% of randomized patients in the CARE trial were women. Review of screening data and exclusion criteria suggests that women were more likely to be excluded because of failure to meet the strict MI criteria or because their total cholesterol or triglyceride levels were too high. They were not excluded more often than men for their or their physician’s disapproval of the study. Future trials should consider the need for specific eligibility criteria for women that reflect their inherent differences, with the goal of enrolling a cohort of women representative of the general female population (33).
Risk factors in women and men
There were small but significant differences between women and men in the severity of CHD and in CHD risk factors at baseline. These differences include age and a higher incidence of congestive heart failure, angina, smoking, hypertension and multiple risk factors. The only risk factor that would favor the women was a higher HDL cholesterol level (45 vs. 38 mg/dl), a level still low for women and likely to contribute to their risk. Nonetheless, the incidence of cardiovascular events in the placebo-treated women was not different from that in men. Some of the risk in the early post-MI period, which tends to be greater for women than for men, was avoided in the CARE trial by enrolling patients at least 3 months after MI. Perhaps the proportion of women in the CARE trial was lowered and the subsequent event rate modified because fewer very high risk women were randomized because of death or complications before screening.
Coronary events in women and men
Pravastatin treatment significantly reduced coronary events in women as well as in men. The percent reduction in the primary end point (CHD death or nonfatal MI) or in combined coronary events was approximately twice as great in women as that in men. However, we recommend caution in interpreting these differences because we did not hypothesize, a priori, that women would experience more benefit than men, and the p value is borderline (0.048) for a difference in event reduction between women and men. Also, participants were not stratified by gender in the randomization. Although there were minor but significant differences in baseline characteristics between women and men, they did not explain the difference in event reduction. In the 4S trial, the degree of event reduction was similar in women and men. Because a greater proportion of women than men in the 4S trial (15)qualified for the trial by a diagnosis of angina rather than MI, it is possible that the effects of treatment in the women could have been diluted by women who did not have significant structural coronary disease. The trials were also different with regard to lipid entry criteria. The North American patients in the CARE trial also had more medical and interventional treatments than their Scandinavian counterparts, which could have affected the action of cholesterol-lowering treatment. In conclusion, until information becomes available from other trials that have enrolled women, it remains uncertain whether there is a true gender difference in the efficacy of cholesterol treatment in general or in the effect of any particular agent.
Autopsy studies (34)have shown that coronary lesions are different in women than in men, with MI in women more often caused by plaque erosion rather than by rupture. Erosion of endothelial cells occurs over plaque that has a predominance of smooth muscle cells and proteoglycan rather than macrophage foam cells and cholesterol esters, as in plaque that is prone to rupture. Plaque erosion also occurs in patients who have an average rather than an elevated serum cholesterol level (220 vs. 262 mg/dl for erosion vs. rupture, respectively) (15). Davies (35,36)hypothesized that the results of the CARE trial, in women with average cholesterol, indicate that cholesterol treatment prevents thrombosis over eroded plaque. Although an attractive concept, we caution that it is not known which mechanism was responsible for the clinical coronary events in the patients in the CARE trial—plaque erosion or rupture—and which process was prevented by pravastatin.
Total cancer incidence was not significantly different between the pravastatin and placebo groups. In the men, there were no differences between the groups in cancer incidence by site. Possible reasons for the higher number of breast cancer cases in the women in the pravastatin than those in the placebo group have been discussed (16)and include a very low incidence in the placebo group (no new cases) or an imbalance in risk factors at baseline. There is no known biological basis for suspecting a causal link. The clearest evidence that this finding is anomalous and due to chance comes from the results of the Long-Term Intervention with Pravastatin in Ischemic Disease (LIPID) trial (37), which studied 1,508 women for 6 years after pravastatin or placebo treatment. There was no excess of breast cancer in the pravastatin group (Tonkin A, oral presentation, 70th Annual Scientific Sessions of the American Heart Association, November 1997). Finally, the 4S trial (15)found no increase in breast cancer in women treated with simvastatin for 5 years.
Women with an MI show strong, early reduction in recurrent coronary events during therapy with pravastatin, despite having average cholesterol levels before treatment. Women experienced a major reduction in risk for coronary events and stroke of 40% to 55%, with benefit beginning by 1 year. Treatment will prevent an extensive range of subsequent cardiovascular events: It may be anticipated that for every 1,000 women entered into the CARE trial and treated for 5 years, 228 cardiovascular events would be prevented (16). Physicians now have the information for evidence-based medical practice of lipid-lowering therapy in women with CHD, in particular, for the majority of women with average levels of total and LDL cholesterol.
We acknowledge the efforts of members of the Gender and Age Writing Group who contributed to the conception and writing of the report: Jane Grant, RN, David Johnstone, MD, Gerald Timmis, MD and Barry R. Davis, MD.
↵‡‡ A complete list of the CARE Investigators appears in references.
☆ The Cholesterol and Recurrent Events trial was funded by an investigator-initiated grant from the Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, New Jersey.
- coronary artery bypass graft surgery
- Cholesterol and Recurrent Events
- coronary heart disease
- high density lipoprotein
- HMG CoA
- hydroxymethylglutaryl coenzyme A
- low density lipoprotein
- myocardial infarction
- percutaneous transluminal coronary angioplasty
- Scandinavian Simvastatin Survival Study
- Received November 26, 1997.
- Revision received March 17, 1998.
- Accepted April 9, 1998.
- American College of Cardiology
- American Heart Association
- Eaker E.D.,
- Cheseboro J.H.,
- Sacks F.M.,
- Wenger N.K.,
- Whisnant M.D.,
- Winston M.
- Buchwald H.,
- Varco R.L.,
- Matts J.P.,
- et al.
- 8.Five-year study by a group of physicians of the Newcastle Upon Tyne region: trial of clofibrate in the treatment of ischaemic heart disease. BMJ 1971;4:767–75.
- Research Committee of the Scottish Society of Physicians
- 12.↵Report of the National Cholesterol Education Program Expert panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults. Arch Intern Med 1988;148:36–69.
- 13.↵Second Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II). JAMA 1993;269:3015–22.
- Miettinen T.A.,
- Pyorala K.,
- Olsson A.G.,
- et al.
- Snedecor G.W.,
- Cochran W.G.
- Kaplan E.I.L.,
- Meier P.
- The Israeli SPRINT Investigators,
- Greenland P.,
- Reicher-Reiss H.,
- Goldbourt U.,
- Behar S.
- Kelsey S.F.,
- James M.,
- Holubkov A.L.,
- et al.
- Barrett-Connor E.
- Farb A.,
- Burke A.P.,
- Tang A.L.,
- et al.
- The Lipid Study Group