Diagram illustrating the search strategy for meta-analysis of long-term administration of angiotensin-converting enzyme inhibitors in patients with coronary artery disease (CAD) and no left ventricular (LV) systolic dysfunction. CABG indicates coronary artery bypass grafting; MI, myocardial infarction; and PCI, percutaneous coronary intervention.
All-cause mortality (A) and cardiovascular mortality (B) in patients with coronary artery disease and no left ventricular systolic dysfunction randomized to long-term angiotensin-converting enzyme inhibitor therapy or placebo. CAMELOT, Comparison of Amlodipine vs Enalapril to Limit Occurrences of Thrombosis11; CI, confidence interval; EUROPA, European Trial on Reduction of Cardiac Events With Perindopril in Stable Coronary Artery Disease4; het, heterogeneity; HOPE, Heart Outcomes Prevention Evaluation study3; MA, meta-analysis; OR, odds ratio; PART-2, Prevention of Atherosclerosis with Ramipril13; PEACE, Prevention of Events With Angiotensin-Converting Enzyme Inhibition5; QUIET, QUinapril Ischemic Event Trial10; SCAT, Simvastatin/Enalapril Coronary Atherosclerosis Trial.12
Nonfatal myocardial infarction (A) and stroke (B). Occurrence of stroke was not reported in the QUIET Trial. For an explanation of the abbreviations see the legend to Figure 2.
Hospitalization because of congestive heart failure (CHF) (A) or unstable angina (UA) (B); hospitalization for subsequent myocardial revascularization in patients with coronary artery disease and no left ventricular systolic dysfunction randomized to receive angiotensin-converting enzyme inhibitors (C); and development of diabetes mellitus in patients without diabetes mellitus at inclusion in the study (D). For an explanation of the abbreviations see the legend to Figure 2.
Danchin N, Cucherat M, Thuillez C, Durand E, Kadri Z, Steg PG. Angiotensin-Converting Enzyme Inhibitors in Patients With Coronary Artery Disease and Absence of Heart Failure or Left Ventricular Systolic DysfunctionAn Overview of Long-term Randomized Controlled Trials. Arch Intern Med. 2006;166(7):787-796. doi:10.1001/archinte.166.7.787
Results of randomized trials of angiotensin-converting enzyme inhibitors in patients with coronary artery disease (CAD) and preserved left ventricular function are conflicting. We undertook this study to determine whether long-term prescription of angiotensin-converting enzyme inhibitors decreases major cardiovascular events and mortality in patients who have CAD and no evidence of left ventricular systolic dysfunction.
We searched MEDLINE, EMBASE, and IPA databases, the Cochrane Controlled Trials Register (1990-2004), and reports from scientific meetings (2003-2004), and we reviewed secondary sources. Search terms included angiotensin-converting enzyme inhibitors, coronary artery disease, randomi(s)zed controlled trials, clinical trials, and myocardial infarction. Eligible studies included randomized controlled trials in patients who had CAD and no heart failure or left ventricular dysfunction, with follow-up οf 2 years or longer. Of 1146 publications screened, 7 met our selection criteria and included a total of 33 960 patients followed up for a mean of 4.4 years.
Five trials included only patients with documented CAD. One trial included patients with documented CAD (80%) or patients who had diabetes mellitus and 1 or more additional risk factors, and another trial included patients who had CAD, a history of transient ischemic attack, or intermittent claudication. Treatment with angiotensin-converting enzyme inhibitors decreased overall mortality (odds ratio, 0.86; 95% confidence interval, 0.79-0.93), cardiovascular mortality (odds ratio, 0.81; 95% confidence interval, 0.73-0.90), myocardial infarction (odds ratio, 0.82; 95% confidence interval, 0.75-0.89), and stroke (odds ratio, 0.77; 95% confidence interval, 0.66-0.88). Other end points, including resuscitation after cardiac arrest, myocardial revascularization, and hospitalization because of heart failure, were also reduced.
Angiotensin-converting enzyme inhibitors reduce total mortality and major cardiovascular end points in patients who have CAD and no left ventricular systolic dysfunction or heart failure.
Angiotensin-converting enzyme (ACE) inhibitors are an undisputed treatment in patients who have congestive heart failure or coronary artery disease (CAD) and concomitant left ventricular (LV) dysfunction.1,2 In patients who have CAD without heart failure or frank LV dysfunction, however, randomized trials have yielded discrepant results.3- 5 Inasmuch as other classes of medications are beneficial in patients with CAD6- 8 and because only limited resources can be allocated to secondary prevention in these patients, it is important to determine whether routine use of ACE inhibitors in this patient population is useful in preventing cardiovascular events and if current recommendations9 are appropriate. The purpose of this analysis was to assess the long-term effects of ACE inhibitor therapy in patients who have CAD and no signs of heart failure or severe LV dysfunction. The general goal of the analysis was to provide information complementary to that provided by previous meta-analyses in patients who have CAD and either signs of heart failure or impaired systolic function.1,2
The present meta-analysis focuses on all-cause mortality and major cardiovascular end points (ie, cardiovascular death, myocardial infarction [MI], and stroke). In addition, we studied the effects of ACE inhibitor therapy on “softer” end points, such as myocardial revascularization, hospitalization because of angina or congestive heart failure, and the development of diabetes mellitus.
To identify randomized controlled trials of ACE inhibitor therapy vs placebo in patients who had stable CAD, we conducted a systematic search of the MEDLINE database (National Library of Medicine, Bethesda, Md; 1990-2004), limited to studies in human beings, EMBASE (through 2004), and the Cochrane Controlled Trials Register (through 2004); and we reviewed secondary sources. The search was performed using Web-based tools (PubMed, Embase.com, International Pharmaceutical Abstracts, Ovid, Medscape, and Scholar Google). Relevant articles were selected on the basis of their titles or abstracts and were searched manually, particularly for the REFERENCES cited in reviews, commentaries, and other selected publications. National and international colleagues were contacted to limit the risk for selection bias. We searched presentations at scientific meetings (American Heart Association, American College of Cardiology, and European Society of Cardiology) in 2003 and 2004. Search terms included angiotensin-converting enzyme inhibitors, coronary artery disease, randomi(s)zed controlled trials, clinical trials, and myocardial infarction.
We included all placebo-controlled randomized trials with a follow-up of 2 years or longer performed in patients who had stable CAD and either no signs or symptoms of heart failure or no documented LV dysfunction (defined as left ventricular ejection fraction [LVEF] <0.35). The rationale for all trials was to determine whether ACE inhibitors would have an effect on the course of atherosclerotic disease. As for other classes of medications, such as statin drugs, a sufficient duration of exposure to the study drug is required to document a protective effect. Because the events curves in HOPE (Heart Outcomes Prevention Evaluation Study)3 and EUROPA (European Trial on Reduction of Cardiac Events With Perindopril in Stable Coronary Artery Disease)4 diverged only after 12 to 24 months, we set the threshold for treatment duration at 2 years. However, we found no trial that met our inclusion criteria and that had a follow-up of 1 to 2 years. All trials included either patients with documented CAD4,5,10- 12 or a high proportion with CAD,3,13 with the rest having other forms of atherosclerotic disease or being at high risk for cardiovascular disease. Risk in these latter populations was similar to that in patients with documented CAD in the corresponding trials; in addition, in HOPE, the effect of ACE inhibitor therapy was also highly significant when the analysis was restricted to the subset of patients with documented CAD. The ejection fraction threshold for defining clinically relevant LV dysfunction is usually 0.35 or, more frequently, 0.40. We accepted 0.35 as the lower limit and excluded trials in which an upper limit for LVEF was set (APRES [Angiotensin-converting Enzyme Inhibition Post Revascularization Study]).14 In both HOPE and EUROPA, measurement of LVEF was not a prerequisite for inclusion. However, subsequent analyses in patients in whom LVEF had been determined showed that in both trials LVEF was greater than 0.40 in more than 90% of patients.
The qualifying studies were checked for adequate blinding of randomization, completeness of follow-up, and methods of qualification of outcome events. All trials had to report information on the prespecified principal outcomes (all-cause mortality, cardiovascular death, or MI). We also analyzed secondary end points when reported similarly in at least 2 trials. These included stroke, cardiac arrest, myocardial revascularization, hospitalization because of unstable angina, hospitalization because of heart failure, and onset of diabetes mellitus in patients previously without diabetes mellitus. A QUOROM (Quality of Reporting of Meta-analyses)15 diagram of the study selection process is shown in Figure 1. Among articles excluded on the basis of their titles or abstracts, 49% were excluded because they were not randomized, 32% because they did not involve ACE inhibitors, 12% because the target population was different, and 7% because there were no placebo controls; many studies were excluded for multiple causes.
To assess study quality, we evaluated trials for the adequacy of allocation concealment, blindness of patients and physicians to the treatment, and blind assessment of the outcome of interest. We used the criteria recommended by Altman and Schulz16 and Juni et al17 to decide whether treatment allocation was adequately concealed. Two physicians (N.D. and M.C.) graded each of the trials included in the meta-analysis. We did not use a summary score to identify trials of low or high quality, or perform weighting by quality scores because this practice has been discouraged by some investigators.17- 19
Because of lack of access to individual patient data from the trials, we used the figures reported in the articles describing the trial results, on intention-to-treat analyses. We used EasyMA 2001 software20 (developed by M.C.) for the analyses. The meta-analysis was performed using odds ratios (ORs) as the parameter of efficacy. Odds ratios were combined using inverse variance-weighted averages of their logarithmics in fixed-effects models. Other methods (relative risk random model, relative risk Greenland-Robins, Mantel-Haenszel test, or Peto's method for OR) provided similar results. In the case of absence of an event in one group of patients, a pseudo-count method was used to calculate the OR, adding a value of 0.25 event in each group.
Between-study heterogeneity was analyzed using standard χ2 tests (Cochran), with P<.05 deemed statistically significant. Where no significant statistical heterogeneity was identified, the fixed-effect estimate was used preferentially as the summary measure. In addition, sensitivity analyses were performed to assess the contribution of each study by excluding individual trials one at a time and recalculating the pooled ORs for the remaining trials. Publication bias was assessed graphically using a funnel plot of the logarithm of effect size vs the standard error for each trial, and mathematically using an adjusted rank correlation test, according to the method of Begg and Mazumdar.21
We identified 7 trials that met our inclusion criteria.3- 5,10- 13 Five trials4,5,10- 12 enrolled only patients with documented CAD. One trial3 enrolled patients with coronary disease or patients with diabetes mellitus who were 55 years or older and who had at least 1 additional risk factor. In this trial, however, 80% of the population enrolled had documented CAD. The last trial13 enrolled patients who had CAD (68% of the trial population), intermittent claudication, or transient ischemic attack. The trial names, acronyms, designs, and main baseline characteristics of the patients included are summarized in Table 1 and Table 2. Five ACE inhibitors were tested, and 2 of the trials3,4 used doses higher than those usually necessary for antihypertensive therapy. The study populations ranged from 460 to 12 218 patients (total, 33 960 patients), and the duration of follow-up ranged from 2 to 5 years (mean, 4.4 years). One trial11 had a 3-arm treatment design and compared the effectiveness of enalapril with amlodipine and placebo; for the purpose of our analysis, only data from the enalapril and placebo treatment arms were considered. Another trial12 had a 2 × 2-factorial design that compared enalapril treatment with placebo and simvastatin treatment with placebo.
The primary end point of each trial differed, but all trials reported all-cause mortality, cardiovascular mortality, and MI. Cardiovascular mortality in the PEACE (Prevention of Events With Angiotensin Converting Enzyme Inhibition) trial5 did exclude death from unknown causes. Stroke was not a reported end point in QUIET (Quinapril Ischemic Event Trial),10 and a combination of stroke and transient ischemic attack was reported in the CAMELOT (Comparison of Amlodipine vs Enalapril to Limit Occurrences of Thrombosis) trial.11
The overall quality of the trials was assessed by analysis of data given in the protocol and design publications, and the main publications presenting the results of the trials.3- 5,10- 13,22- 27 Both graders (N.D. and M.C.) were concordant in attributing a score of 5 to all trials. We found no consistent visual or statistical evidence of publication bias (P value of Begg and Mazumdar21 and test range from 0.18-0.65).
All-cause mortality (Figure 2A) was lower in the treatment arms compared with the placebo arms of all trials except one,11 and the reduction was statistically significant in HOPE.3 There was no heterogeneity among the trials, and the meta-analysis showed a 14% reduction in mortality (P<.001). Likewise, cardiovascular mortality (Figure 2B) was consistently reduced by ACE inhibition across all trials except CAMELOT,11 with a significant reduction in HOPE and a highly significant 19% reduction in the meta-analysis (P<.001). Sensitivity analyses were performed after individual exclusion of HOPE, EUROPA, and PEACE. After exclusion of HOPE, the OR for all-cause mortality was slightly less at 0.88 (95% confidence interval [CI], 0.79-0.98) but remained statistically significant, and the OR for cardiovascular mortality was 0.88 (95% CI, 0.76-1.01; P = .07). After exclusion of EUROPA, the respective ORs were 0.85 (95% CI, 0.77-0.94) and 0.79 (95% CI, 0.70-0.90). Conversely, exclusion of PEACE yielded slightly higher risk reductions in total and cardiovascular mortalities.
Significant reductions in acute MI with ACE inhibitor therapy were found in both HOPE and EUROPA (Figure 3A).3,4 Angiotensin-converting enzyme inhibition seemed to have had no effect on the MI rate in PEACE,5 and nonsignificant reductions were observed in the other trials.10- 13 Overall, there was an 18% reduction in MI (P<.001) with ACE inhibitor therapy. Stroke was less frequent in patients receiving ACE inhibitors in all trials except PART-2 (Prevention of Atherosclerosis with Ramipril Trial-2)13 (Figure 3B), and the reduction achieved statistical significance in HOPE.3 Overall, ACE inhibitors resulted in a 23% relative reduction in the occurrence of stroke (or transient ischemic attack in CAMELOT) (P<.001). The sensitivity analyses after exclusion of either HOPE or EUROPA resulted in an attenuation of the benefit in terms of MI (OR, 0.85; 95% CI, 0.76-0.95, and OR, 0.84; 95% CI, 0.76-0.93, respectively), but the reduction remained statistically significant. For stroke, the OR was 0.85 (95% CI, 0.70-1.04) after exclusion of HOPE and 0.71 (95% CI, 0.60-0.83) after exclusion of EUROPA. Conversely, exclusion of PEACE resulted in slightly higher risk reductions.
Five trials3- 5,11,13 analyzed the occurrence of hospitalization because of congestive heart failure (Figure 4A) and showed a reduction in this end point in patients receiving ACE inhibitors. The reduction was significant in EUROPA4 and PEACE.5 Overall, the risk was reduced by 23% (P<.001). Hospitalization because of unstable angina was reported in 4 trials3,4,10,13 (Figure 4B), and the effect of ACE inhibitor therapy on this end point was neutral (3% risk reduction; P = .06). In the 6 trials3- 5,10- 12 in which rates of myocardial revascularization were reported (Figure 4C), the risk was reduced by 8% (P<.01). Cardiac arrest was a rare event in all trials, with a significant (P<.001) 42% risk reduction in the meta-analysis. Two trials3,4 reported the rate of new onset of diabetes mellitus in patients initially without diabetes mellitus (Figure 4D) and found an overall 23% reduction with ACE inhibitor therapy compared with placebo (P<.001).
In our meta-analysis of patients with either no heart failure or no LV systolic dysfunction, we found a significant reduction in all-cause mortality, cardiovascular mortality, MI, and stroke, as well as in several softer end points, such as hospitalization because of heart failure, myocardial revascularization, or new onset of diabetes mellitus (Table 3). These results confirm and extend those of a recently published meta-analysis of HOPE, EUROPA, and PEACE.28
The ability of ACE inhibitors to prevent cardiovascular events in patients with CAD was assessed first in patients with poor LV function or heart failure. The meta-analysis of the trials including such patients1 showed a reduction in death (OR, 0.80; 95% CI, 0.74-0.87), repeat MI (OR, 0.79; 95% CI, 0.70-0.89), and hospital readmission because of heart failure (OR, 0.67; 95% CI, 0.61-0.74) but no significant difference for stroke (OR, 0.96; 95% CI, 0.80-1.15). The magnitude of risk reduction in our meta-analysis was nearly identical to that observed in the subgroup of 1428 patients with LVEF greater than 35% in the previous meta-analysis of ACE inhibitors in patients who had coronary disease with poor LV function.1
Trials excluded from the meta-analysis are summarized in Table 4. In addition to the trials included in our systematic review, one trial14 included 159 patients with moderate LV dysfunction (LVEF, 0.30-0.50) randomized to receive ramipril or placebo for 33 months; ramipril therapy was associated with significantly fewer major cardiac events. Seven other trials29- 35 compared short-term (6-12 months) administration of ACE inhibitors vs placebo in patients who had undergone coronary artery bypass surgery or coronary angioplasty. Clinical event rates in any of these short-term studies did not differ between patients who did or did not receive ACE inhibitors.
In both QUIET and CAMELOT,10,11 the duration of follow-up may have been too short to demonstrate any benefit of ACE inhibitor therapy in terms of secondary prevention. In SCAT (Simvastatin/Enalapril Coronary Atherosclerosis Trial)12 and PART-2,13 follow-up was longer and there was a trend favoring ACE inhibitors; however, neither trial had enough statistical power to document a significant effect for clinical events. In contrast, HOPE,3 EUROPA,4 and PEACE5 had both a long follow-up and a large population. However, in PEACE, the initial primary end point (cardiovascular death or MI) was modified during the course of the trial to include softer end points.
It has been suggested that the difference in outcomes in PEACE5 vs HOPE3 and EUROPA4 might be related to differences in baseline risk. Although mortality was higher in HOPE, both mortality and cardiovascular mortality were similar in EUROPA and PEACE (Table 1). Furthermore, in HOPE and EUROPA, outcomes were consistent throughout subgroups defined according to the patients' initial risk profile, including age, previous MI, or presence of peripheral vascular disease, or according to concomitant secondary prevention medications. The percentages of patients receiving antiplatelet agents, β-blockers, and lipid-lowering agents were similar in EUROPA and PEACE.
The effect of the active treatment on blood pressure was reported in 6 of 7 trials (Table 1). There was no obvious relationship between the magnitude of blood pressure reduction and clinical events in the trials. Blood pressure reduction per se is not necessarily associated with cardiovascular protection in patients who have stable CAD and preserved LV function. ACTION (A Coronary Disease Trial Investigating Outcome With Nifedipine gastrointestinal therapeutic system)56 failed to show a decrease in mortality and MI in patients receiving nifedipine, despite a reduction of 6 mm Hg in systolic blood pressure with this medication.
Inasmuch as 5 different ACE inhibitors were tested in the trials, the individual properties of the ACE inhibitors57- 59 may, at least in part, account for the differences in clinical outcome. However, there is no obvious relationship between the pharmacologic properties (tissue affinity and lipophilicity) of the individual medications and their clinical benefit in the trials.
The dosage used in each of the trials differed, with the highest doses in HOPE3 and EUROPA.4 In PEACE,5 the initial dose of trandolapril was low (2 mg) and could be maintained at 2 mg or increased to 4 mg after 6 months, at the discretion of the physician, so that fewer than 69% of the patients received the 4-mg dose at any time during the trial. Because the 50% inhibitory concentration for plasma ACE for ramipril, perindopril, and trandolapril is similar,60- 62 the 2- to 4-mg dose of trandolapril is comparatively less than the 10-mg dose of ramipril or the 8-mg dose of perindopril. Therefore, the 2 trials with definitely positive results used the highest doses. The clinical situation in which the therapeutic target is controlling the atherothrombotic process is different from that of heart failure, in which the level of activation of the renin-angiotensin system is such that even rather low doses of ACE inhibitors are effective. There is evidence that only the larger doses of ACE inhibitors can slow the progression of atherosclerotic disease. In SECURE (Study to Evaluate Carotid Ultrasound Changes in Patients With Ramipril and Vitamin E),63 2 doses of ramipril (2.5 mg and 10 mg) that yielded similar reductions in blood pressure were tested against placebo; only the 10-mg dose was associated with significant slowing of carotid atherosclerosis progression.
In this overview of randomized trials of ACE inhibitors for the long-term secondary prevention of CAD in patients without LV dysfunction or heart failure, active treatment was associated with a highly significant reduction in all-cause mortality and all major cardiovascular events. These results, along with those previously reported in patients who have CAD with LV dysfunction or heart failure, suggest that ACE inhibitor therapy should be systematically used in all patients with documented CAD and that continued efforts should be made to implement current secondary prevention guidelines.
Correspondence: Nicolas Danchin, MD, FESC, Department of Cardiology, Hôpital Européen Georges Pompidou, 20 rue Leblanc, 75015 Paris, France (email@example.com).
Accepted for Publication: October 5, 2005.
Financial Disclosure: Dr Danchin has been a speaker and has served on steering committees of studies sponsored by Sanofi-Aventis, Paris, France, which markets ramipril, and by Servier Medical, Neuilly-sur-Seine, France, which markets perindopril. Dr Cucherat has been a consultant for Servier Medical. Dr Thuillez has been a speaker for Sanofi-Aventis and Servier Medical and has received research grants from Servier Medical. Dr Steg has been a speaker and consultant for Sanofi-Aventis and for Servier Medical and has received research grants from Aventis Pharma, Paris. He was a member of the Critical Events Committee for the EUROPA trial.
Role of the Sponsor: This analysis was performed independently of all pharmaceutical firms that manufacture or market ACE inhibitors and was not supported, directly or indirectly, by grant money from sponsors.