Numerous large randomized clinical trials have demonstrated the efficacy of secondary prevention of coronary artery disease (CAD). Effective lifestyle interventions include smoking cessation (1), dietary intervention (2- 3), and exercise (4- 5). Effective pharmacologic interventions include aspirin (6), clopidogrel (6), beta-blockers (7- 8), angiotensin converting enzyme inhibitors (9- 10), and statins (11). Studies of control of blood pressure (12), low-density lipoprotein (LDL) cholesterol (11,13- 16), and—less consistently—blood glucose (17- 19) serve as the basis for current secondary prevention guidelines. Although most secondary prevention trials have tested the impact of a single risk factor intervention, the American College of Cardiology (ACC), the American Heart Association (AHA), and the Canadian Cardiovascular Society recommend comprehensive lifestyle and pharmacologic interventions with specific risk factor targets (20- 22). Few clinical trials have included multiple risk factor intervention with behavioral and pharmacologic therapy as recommended by practice guidelines (23- 24).

Previous randomized trials that compared percutaneous coronary intervention (PCI) with medical therapy in patients with stable CAD failed to apply medical therapy that was multifaceted, aggressive, and provided equally to both treatment arms (25). The COURAGE (Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation) trial tested the impact of comprehensive intensive lifestyle and pharmacologic interventions (optimal medical therapy [OMT]) with or without PCI in 2,287 patients with stable CAD. There was no difference in the primary outcome of death or myocardial infarction (MI) during 4.6 years of follow-up (26), and the PCI group had a small but significant incremental benefit in angina control that disappeared by 36 months (27). This report analyzes the completeness and success of medical therapy in the COURAGE trial.

The methods, patient eligibility criteria, baseline characteristics, and main results of the COURAGE trial have been described previously (26,28). Data management and analyses were performed solely by the data coordinating center with oversight of the trial executive committee, which had full access to the data and analyses and vouches for their accuracy and completeness. Patients were enrolled from June 1999 to January 2004, and the study closed on June 30, 2006.

Risk factor goals were based on the 1995 ACC/AHA secondary prevention guidelines (Table 1) (29), although the blood pressure and LDL cholesterol goals were more aggressive in anticipation of what might become practice guidelines by the end of the trial. The LDL cholesterol goal—60 to 85 mg/dl—was established by the steering committee in 1997 when the National Cholesterol Education Program goal for CAD patients was <100 mg/dl. In July 2004, the National Cholesterol Education Program established an optional LDL cholesterol goal of <70 mg/dl for “very high risk patients” (30), and the COURAGE steering committee adopted that goal for the remainder of the trial.

Each patient met with a nurse case manager at baseline and 1, 2, 3, and 6 months, then every 6 months until the study ended. At each visit lifestyle and medication adherence were assessed, and body weight, blood pressure, fasting glucose, and lipid values were measured. Lipids were analyzed in a core laboratory; glucose and hemoglobin A1c (HbA1c) were analyzed locally. Diabetes was defined as having a history of diabetes or taking hypoglycemic medication at baseline. The HbA1c was collected at baseline in most patients but during follow-up only in patients defined as having diabetes at baseline.

Lifestyle intervention was delivered equally to both treatment groups at each visit. Case managers were trained before trial launch and at annual meetings to assess patient behaviors and provide behavioral counseling focused on smoking cessation, nutrition, physical activity, and weight management. Each patient's readiness to change health-related behaviors was assessed with a system developed by the Patient-centered Assessment and Counseling for Exercise and nutrition (PACE) project (31). Lifestyle counseling was standardized across sites with written materials developed by the PACE project, including scripts to instruct patients regarding lifestyle change (32). These materials were designed for brief, practical, focused interventions by a physician or nurse in an outpatient setting. A more detailed description of the lifestyle intervention and training for nurse case managers is provided in the Online Appendix.

The behavioral counseling was based on Social Cognitive Theory (33) and the Transtheoretical Model (34). For each behavioral risk factor, patients were categorized into pre-contemplators (those who do not wish to change), contemplators (those who are willing to change), and actives (those who are already meeting the recommended goal). Nurse case managers were trained to provide messages tailored to the patient's stage of change for each behavioral risk factor. Each stage-based message focused on known mediators of behavior change (e.g., goal setting, increasing social support, and self-efficacy). Print-based patient materials were used to assess and counsel patients (see examples in the Online Appendix).

Smokers who were pre-contemplators were given advice and rationale to quit. Smokers who were contemplators were provided with stage-based counseling, pharmacologic therapy, and/or referral to a formal smoking cessation program.

Diet was assessed with MEDFICTS, a questionnaire recommended by the National Cholesterol Education Program to assess dietary fat (35- 36). The questionnaire can be self-administered in 3 to 5 min and scored by a health care provider in 2 min. Nutritional counseling was designed to achieve the dietary guidelines of the National Cholesterol Education Program (35,37). With the same stage-based approach, patients were advised to reduce calories, total fat, and saturated fat and increase consumption of fruits and vegetables.

Subjects were counseled to gradually increase moderate-intensity physical activity to 30 to 45 min 5 times/week. Typically this was moderate-intensity walking. Counseling was tailored for weight loss if needed.

Pharmacologic therapy conformed with ACC/AHA guidelines for secondary prevention and management of angina (20- 21). All patients received the same medical therapy for secondary prevention, regardless of treatment assignment (Table 2). The intensity of anti-anginal therapy varied according to angina severity. Patients received medications at no cost except for aspirin, thiazide diuretics, fenofibrate, and omega-3 fatty acids.

Among patients randomized to PCI with no history of MI, an attempt was made to discontinue long-acting metoprolol, amlodipine, and/or isosorbide mononitrate if they had no angina 3 to 6 months after PCI. Patients undergoing PCI received aspirin and clopidogrel according to prevailing treatment guidelines (38). If a patient was not at blood pressure or lipid goals, pharmacologic therapy was adjusted according to study protocols (see treatment algorithms in the Online Appendix).

For LDL cholesterol control, all patients were prescribed simvastatin according to a dosing algorithm. If simvastatin was not tolerated or was not effective, patients were free to use other statins. Extended-release niacin or ezetimibe was added if needed to reach goal. After the LDL cholesterol goal was achieved, an attempt was made to raise high-density lipoprotein cholesterol above 40 mg/dl and lower triglycerides below 150 mg/dl with lifestyle change, extended-release niacin, fibrates, or omega-3 fatty acids.

Adherence to medical therapy was assessed by the 4-item self-report scale developed by Morisky et al. (39). Scores range from 0 to 4; higher scores indicate worse adherence. A score of 0 to 1 was defined as adherent.

We assessed patient satisfaction at baseline and 1, 3, 6, and 12 months and annually thereafter with the Seattle Angina Questionnaire Treatment Satisfaction scale, which ranges from 0 to 100 (40). Higher scores indicate greater patient satisfaction with therapy.

The data are presented as percentage of individuals in each treatment group. Continuous variables are described as median ± standard error of the median (41). Missing data were not included in the calculation of medians or percentages. For comparisons between values at baseline and 5 years, McNemar's test was used for discrete variables, and Wilcoxon signed rank test was used for continuous variables (lifestyle changes, therapeutic targets). Proportions of patients taking various medications assigned to the PCI group compared with the OMT group are compared at each visit with the chi-square test (pharmacologic therapy). Comparisons of continuous variables between treatment groups at each visit were made with the median test (therapeutic targets). For over-time comparisons between treatments, sequential differences were modeled with baseline, interval, time, and (time)2 in addition to treatment (therapeutic targets, treatment satisfaction). The intention-to-treat principle was used in all analyses.

Baseline characteristics were similar between groups (Table 3).

In aggregate (both treatment groups combined), smoking decreased from 23% to 19% (p = 0.025). The proportion of subjects who reported the dietary goal of <7% of calories from saturated fat increased from 46% to 80% (p < 0.001). The proportion of subjects who achieved the physical activity goal in the form of walking ≥150 min/week increased from 58% to 66% (p < 0.001). When analyzed by treatment group, there were no significant differences between groups at any time for any lifestyle variables (Figure 1). Body mass index increased from 28.8 ± 0.13 kg/m² at baseline to 29.3 ± 0.23 kg/m² at 5 years (p ≤ 0.001).

(Figure 2) shows medication use during the trial. In aggregate, medication use increased from baseline to 5 years as follows: antiplatelets 87% to 96%; beta-blockers 69% to 85%; and renin-angiotensin-aldosterone system inhibitors 46% to 72%. Before randomization, 31% of patients took calcium channel blockers. By 6 months, 40% of the PCI group and 50% of the OMT group took these agents (p < 0.001), and this difference persisted though 5 years of follow-up (42% vs. 52%, respectively, p = 0.004). Before randomization, 58% took long-acting nitrates. By 6 months, 55% of the PCI group and 71% of the OMT group took these agents (p < 0.001); at 5 years, the proportions were 40% versus 57%, respectively (p < 0.001). Statin use rose from 64% to 93%, niacin use rose from <1% to 18%, and ezetimibe use rose from 0% to 33% (at the trial onset, ezetimibe was not approved for use). The use of any lipid medication rose from 67% to 97%.

At baseline, 51% of PCI group and 51% of OMT group patients took the combination of aspirin, beta-blocker, and lipid drug therapy. By 5 years, the proportions rose to 79% and 80%, respectively. At baseline, 27% of PCI group and 29% of OMT group patients took aspirin, beta-blocker, lipid lowering, and renin-angiotensin-aldosterone system inhibitor therapy. By 5 years, the proportions were 55% and 51%, respectively. No between-group differences were statistically significant.

At 6 months, average self-reported adherence to prescribed medications was 97% in both groups; adherence remained at 95% throughout the trial, with no significant difference between groups.

At baseline, 845 patients (37%) without the diagnosis of diabetes had fasting glucose <100 mg/dl, 635 (28%) had impaired fasting glucose (fasting glucose 100 to 125 mg/dl), and 766 (34%) had diabetes. During the study, 191 patients developed diabetes, 97 in the PCI group and 94 in the OMT group. At baseline, 59% of patients had metabolic syndrome.

(Figure 3) presents the changes in blood pressure, lipids, and HbA1c. Median systolic blood pressure was 131 ± 0.49 mm Hg at study entry and 123 ± 0.88 mm Hg at 5 years (p < 0.001). Before randomization, median LDL cholesterol was 101 ± 0.83 mg/dl. At 6 months the median LDL cholesterol was 82 ± 0.62 mg/dl and declined to 72 ± 0.88 mg/dl at 5 years (p < 0.001). There were no significant intergroup differences in LDL cholesterol levels at any visit during the study. During the trial high-density lipoprotein cholesterol rose from 39 ± 0.3 mg/dl to 41 ± 0.5 mg/dl (p < 0.001) and triglycerides fell from 146 ± 2 mg/dl to 126 ± 3 mg/dl (p < 0.001). Among patients with diabetes, HbA1c was 7.0 ± 0.1% at baseline and 7.1 ± 0.01% at 5 years (p = 1.0).

(Figure 4) shows the percentage of subjects who achieved blood pressure, lipid, and HbA1c goals. At randomization, 43% of patients were at the systolic blood pressure goal of <130 mm Hg; this increased to approximately 60% by 3 years and remained unchanged through the rest of the trial in both groups. At randomization, 28% of patients had LDL cholesterol levels at study goal (≤85 mg/dl); by 5 years, this proportion had increased to 70%. With the more aggressive LDL cholesterol target of <70 mg/dl proposed in 2004 (30), 14% of patients were below this target at randomization, and 46% achieved this target by 5 years.

During the first year, 4% of patients in each treatment group missed follow-up visits; during subsequent years, the proportion of patients who missed clinic visits gradually increased. By year 5, 12% of the PCI group and 10% of the OMT group had missed at least 1 visit during the trial (see the Online Appendix). Patients who attended every visit had better systolic blood pressure and LDL cholesterol-lowering than patients who missed visits (see Figs. A and B in the Online Appendix). There were no significant differences between treatment groups. Medication-adherent patients had lower systolic blood pressure and LDL cholesterol than patients who were not adherent (see Figs. C and D in the Online Appendix).

The Seattle Angina Questionnaire Treatment Satisfaction score was 88 for the PCI group and 86 for the OMT group at randomization. The scores gradually increased during the study, and at 5 years, they were 92 (p = 0.08) and 94 (p = 0.001), respectively, indicating that patients were satisfied with their treatment and their satisfaction increased during the trial. There were no differences between treatment groups over time (p = 0.91).

Unlike previous trials of revascularization versus medical therapy, the COURAGE trial applied a multifaceted strategy of intensive secondary prevention to both treatment arms. We succeeded in achieving multiple therapeutic targets and in applying OMT equally to both treatment groups. Moreover, this aggressive treatment strategy was associated with improved treatment satisfaction. Smoking, dietary fat intake, and physical activity improved. Although some behavioral improvements were modest, they were observed over several years and achieved with simple counseling techniques that can be replicated. We were not successful in assisting people with weight loss.

Self-reported long-term adherence to drug therapy and risk factor control in the COURAGE trial far exceeded what has been reported in surveys of CAD outpatients. The EUROASPIRE II (European Action on Secondary and Primary Prevention by Intervention to Reduce Events) trial, a European survey of CAD patients, found use of aspirin to be 86%, use of beta-blockers to be 63%, and use of lipid-lowering therapy to be 61% (42). Newby et al. (43) reported consistent use of aspirin to be 71%, use of beta-blockers to be 46%, use of lipid-lowering therapy to be 44%, and use of the combination of aspirin, beta-blockers, and lipid-lowering therapy to be 21%. The corresponding numbers at the end of the COURAGE trial were substantially higher: aspirin, 96%; beta-blockers, 85%; lipid-lowering therapy, 97%; and all 3 drug classes, 80%. Our success might be attributable to several factors. First, clinical trial volunteers are more likely to adhere to drug regimens. Second, drugs were provided free of charge. Third, treatment protocols were used to provide standardized care. Fourth, nurses who delivered the intervention were trained to improve lifestyle and medication adherence, shown previously to improve success in risk-factor control (23,44- 45).

In the EUROASPIRE III trial, 18% of CAD patients smoked, 39% achieved blood pressure <140/90 mm Hg in nondiabetic patients or <130/80 mm Hg in diabetic patients, and 54% attained total cholesterol <175 mg/dl (46). By the end of the COURAGE trial, 19% of patients smoked, 61% achieved our more stringent blood pressure goals, and 70% achieved our more aggressive lipid goal (LDL cholesterol ≤85 mg/dl). Hence, with the exceptions of smoking (where our positive effect was modest) and weight loss, we were quite successful in achieving and maintaining ambitious risk-factor goals in high proportions of patients. This endorses a secondary prevention model with nurse case managers with simple behavior assessment and counseling tools and treatment algorithms to effect positive lifestyle change, appropriate use of medications, and titration of medications to achieve treatment targets. This method is consistent with guidelines from the ACC, AHA, and the Canadian Cardiovascular Society (20- 21) and is responsive to what EUROASPIRE investigators describe as “a compelling need for more effective lifestyle management of patients with coronary heart disease” (46).

The design of the COURAGE trial precludes the opportunity to measure the impact of OMT on death, MI, or other major cardiovascular events, because OMT was provided equally to both treatment arms. With evidence from prior clinical trials, before the initiation of the COURAGE trial we projected a 3-year primary event rate of 16.4% in the PCI group and 21.0% in the OMT group (relative difference of 22%, absolute difference of 4.6%). We observed a 4.6-year primary event rate of 19.0% in the PCI group and 18.5% in the OMT group (26). Hence, we overestimated the event rate by a substantial margin. In part, this might have been due to the impact of OMT. The medical therapy delivered in the COURAGE trial has been criticized as not achievable in the real world (47). We cannot assess to what extent free medication influenced behavior, but we acknowledge that medication adherence and persistence would probably have declined if patients had been required to share the cost of medications (48- 50). However, the key medications used in the COURAGE trial are all available in more affordable generic forms. Furthermore, OMT was delivered by nurses who were provided with simple protocols and a modest amount of training to deliver behavioral counseling. The frequency of patient contact mandated by the protocol was within the norms of conventional practice in the U.S. and Canada. Hence, this quality of care is feasible in typical outpatient practice.

Medical therapy in the COURAGE trial was protocol-driven, delivered by nurse case managers, comprehensive in scope, intensive in its application, and provided equally to patients whether or not they received PCI. Significant behavior change was achieved, self-reported medication adherence and persistence was high, and therapeutic targets were reached in large proportions of patients. The delivery of OMT in the COURAGE trial is a model for secondary prevention in practice, with potential policy implications regarding the use of nurse case managers and free medications to optimally manage patients with chronic CAD.

Dr. Boden has received consulting fees and lecture fees from Kos Pharmaceuticals, PDL BioPharma, Pfizer, CV Therapeutics, and Sanofi-Aventis and grant support from Merck and Abbott Laboratories. Dr. O'Rourke has received consulting fees from King Pharmaceuticals, Lilly, and CV Therapeutics. Dr. Calfas has been a stockholder of SanTech, Inc. a company founded in cooperation with San Diego State University and exclusively licensed to disseminate the PACE materials. Dr. Mancini has received honoraria from Merck, GlaxoSmithKline, Sanofi-Aventis, AstraZeneca and Abbott. Dr. Spertus has received consulting fees from Amgen and United Healthcare and grant support from Amgen, Roche Diagnostics, and Lilly (and in the past, consulting fees and grant support from CV Therapeutics and has been owner of the copyright for the Seattle Angina Questionnaire, the Peripheral Artery Questionnaire, and the Kansas City Cardiomyopathy Questionnaire). Dr. Bates has received consulting fees from Sanofi-Aventis and AstraZeneca and lecture fees from Sanofi-Aventis. Dr. Shaw has received grant support from Bristol-Myers Squibb and Astellas Healthcare. Dr. Berman has received consulting fees, lecture fees, and grant support from Bristol-Myers Squibb and software royalties from Cedars Sinai Medical Center. Dr. Chaitman has received consulting fees from CV Therapeutics, Roche, Merck, Lilly, Sanofi-Aventis, Forest; lecture fees from CV Therapeutics; and research grants from Roche and CV Therapeutics. Dr. Weintraub has received consulting fees from Sanofi-Aventis and Bristol-Myers Squibb and grant support from Sanofi-Aventis. Dr. Teo has received grant support from Boehringer Ingelheim.

For a detailed description of training for nurse case managers to perform lifestyle assessment and behavioral counseling; samples of PACE materials; COURAGE treatment algorithms; and an analysis of clinic visit attendance, medication adherence, and risk factor control, please see the online version of this article.

Intensive Multifactorial Intervention for Stable Coronary Artery Disease: Optimal Medical Therapy in the COURAGE Trial

Intensive Multifactorial Intervention for Stable Coronary Artery Disease: Optimal Medical Therapy in the COURAGE Trial

Intensive Multifactorial Intervention for Stable Coronary Artery Disease: Optimal Medical Therapy in the COURAGE Trial

Intensive Multifactorial Intervention for Stable Coronary Artery Disease: Optimal Medical Therapy in the COURAGE Trial