Author + information
- Received March 17, 1999
- Revision received October 15, 1999
- Accepted November 19, 1999
- Published online March 15, 2000.
- ↵*Reprint requests and correspondence: Dr. Lars Kjøller-Hansen, Heart Center 2142, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Denmark
This study was performed to assess the effect of treatment with ramipril on the incidence of cardiac events after invasive revascularization in patients with asymptomatic moderate left ventricular dysfunction.
In patients with angina pectoris and left ventricular dysfunction, both invasive revascularization and treatment with angiotensin-converting enzyme inhibitors reduce cardiac mortality and morbidity. Whether there is a benefit from combining the two treatment strategies has never been evaluated prospectively.
After invasive revascularization, 159 patients with preoperative chronic stable angina pectoris, left ventricular ejection fraction between 0.30 and 0.50 and no clinical heart failure were randomly assigned to receive double-blind treatment with either ramipril or placebo and subsequently followed for a median of 33 months.
Ramipril reduced the incidence of the triple-composite end point of cardiac death, acute myocardial infarction or clinical heart failure (risk reduction 58%; 95% confidence interval 7% to 80%, p = 0.031). The incidence of the quadruple-composite end point of cardiac death, acute myocardial infarction, clinical heart failure or recurrent angina pectoris was not altered with ramipril. These findings were consistent across subgroups with respect to left ventricular ejection fraction below or above 0.40, and whether coronary artery bypass grafting or percutaneous transluminal coronary angioplasty was performed.
In patients with angina pectoris and asymptomatic moderate left ventricular dysfunction, long-term treatment with ramipril after invasive revascularization significantly reduced the incidence of the composite end point of cardiac death, acute myocardial infarction or clinical heart failure, indicating that the beneficial effects of angiotensin-converting enzyme inhibitor treatment may be extended to include treatment of this patient group.
Patients with ischemic heart disease and reduced left ventricular ejection fraction (LVEF) have an adverse prognosis with respect to cardiac mortality and morbidity (1–5). Although invasive revascularization improves prognosis in patients with severe angina pectoris and LVEF below 0.50 (1,3,5), these patients still have a worse prognosis than patients with normal LVEF.
Treatment with angiotensin-converting enzyme inhibitor (ACE-I) improves survival and reduces the incidence of cardiac events in patients with left ventricular dysfunction in many circumstances (6–9). However, extrapolation of the results of the major ACE-I trials (6–9)to apply to the sizable population of patients with asymptomatic and only moderate left ventricular dysfunction undergoing invasive revascularization for chronic stable angina pectoris is as yet not justified based on evidence. First, ACE-I should be used with caution in patients with severe angina pectoris (10–12). Second, Studies of Left Ventricular Dysfunction (SOLVD) prevention trial failed to show survival benefit with ACE-I treatment, and the morbidity benefit in the same study was confined to patients with lower LVEF fractiles (6,13). Finally, invasive revascularization improves prognosis per se (2,3,14).
Accordingly, the present study was designed to assess whether treatment with ramipril reduces cardiac mortality and cardiac morbidity in patients with asymptomatic moderate left ventricular dysfunction undergoing invasive revascularization for chronic stable angina pectoris
The Angiotensin-converting Enzyme Inhibition Post Revascularization Study (APRES) was a randomized, double-blind placebo-controlled monocenter study carried out at the Heart Center, Rigshospitalet, Copenhagen, Denmark. The study was approved by the local Ethics Committee (02-149/93) and the National Board of Health (5312-291-1993). Patients undergoing coronary angiography were found eligible if they were aged 18 to 75 years, had no prior cardiac surgery, had LVEF between 0.30 and 0.50 as determined by ventriculography or echocardiography and were referred for invasive revascularization with coronary artery bypass grafting (CABG) or percutaneous transluminal coronary angioplasty (PTCA) for angina pectoris after coronary angiography and clinical evaluation. Patients excluded from the study were those with a history of recent acute myocardial infarction (AMI) (<3 months) and/or clinical heart failure, i.e., history of dyspnoea relieved by diuretic therapy, ongoing ACE-I treatment due to evidenced indications, concomitant valvular disease or geographic restrictions to complete follow-up. Other exclusion criteria were participation in another investigational drug trial, known intolerance to ACE-I therapy, childbearing potential and medical conditions (including periprocedural complications) that could have major influence on outcome or known to contraindicate use of the test drug.
The design of the study has previously been presented (15). After uncomplicated invasive revascularization, eligible patients were enrolled in the study after giving written consent. A test dose of ramipril 2.5 mg was given and, if tolerated, patients were assigned randomly to ramipril 5 mg o.d. or matching placebo according to a computer-generated assignment scheme in blocks of four (two ramipril, two placebo), and with stratification according to type of invasive revascularization (CABG or PTCA). Study drug treatment was commenced immediately after randomization. The intended randomization time was five to seven days after CABG, and one to two days after PTCA. The ramipril dose was increased to 10 mg o.d or matching placebo after one month. If this dose was not tolerated, patients were given the highest tolerable dose.
Follow-up visits were scheduled at one, three and six months after randomization and every six months thereafter. Patients were instructed to contact the study physicians between scheduled visits in case of health-related problems. A median follow-up time of 2.5 years was intended with a minimum follow-up of one year. At each visit, occurrence of any study outcome, use of medication, adverse events and patient compliance with the test drug were recorded.
The main outcome measures were time to cardiac events. The primary end points were the composite quadruple end point of cardiac death, AMI, development of clinical heart failure or recurrent angina pectoris; and the composite triple end point of cardiac death, AMI or development of clinical heart failure. The limited number of potentially eligible patients at our disposal necessitated the use of composite end points. Thus, the triple-composite end point was constituted of single events on which ACE-I previously had shown benefit, i.e., cardiac death, AMI and heart failure. The quadruple-composite end point additionally included recurrent angina pectoris as a clinical expression of an anti-ischemic and antiatherosclerotic efficacy.
Several secondary end points were predefined. These were double composite end points of cardiac death and AMI, or cardiac death and development of clinical heart failure, and the following single events: mortality due to all causes, cardiac death, development of clinical heart failure, AMI and recurrent angina pectoris.
For patients who died, dates of death were obtained from the Danish Central Person Registry. Causes of death were acquired from hospital records and/or personal interview with relatives. Determination and classification of causes of death were made by the investigators before the treatment allocation code was broken. In patients who suffered AMI during the study period, the diagnosis was confirmed from hospital records. Development of clinical heart failure was defined as the need for open ACE-I treatment, use of diuretics for clinical heart failure as judged by a study physician or hospitalization for heart failure. Recurrent angina pectoris was defined as the need for antianginal drugs for angina pectoris and/or angina pectoris of at least Canadian Cardiovascular Society functional class II in a patient initially free of angina after revascularization. A discussion of the relevance of the end points has been presented previously (15). Furthermore, substudies were carried out, including an echocardiographic study, exercise stress testing and quality of life (15).
Based on an expected cumulative event rate of 30% and a reduction in cardiac events of 40% during ramipril treatment, a significance level of 5% and a power of 80%, power calculation revealed that a sample size of 400 patients was required.
Statistical analyses were carried out using Statistica for Windows version 5.1 software (StatSoft Inc, Tulsa, OK). Analyses were carried out according to the principle of intention to treat, and all p values were two sided. Kaplan-Meier estimates of time from randomization to event of interest were computed, and log-rank test was used for comparison of the two treatment groups, ramipril and placebo. For composite end points, only the first occurrence of any end point was counted in the study. The relative risk was calculated as a hazard ratio and the percent risk reduction was calculated as (1 − relative risk). Continuous and categorical variables were compared with unpaired ttests and chi-square tests, respectively.
To ascertain the uniformity of the effects with ramipril treatment across the patient population, univariate and multivariate Cox proportional hazard regression analyses were carried out with end points as dependent variables. As explanatory variables were selected: treatment group, CABG versus PTCA, LVEF (baseline ejection fraction >0.40 or ≤0.40), triple-vessel disease, complete revascularization, anterior Q waves on electrocardiogram at randomization, use of beta-blockers, use of calcium-antagonists, history of hypertension, diabetes mellitus, smoking, age (≥65 or <65 years), and gender. Two multivariate models were carried out using explanatory variables that turned out as significant predictors in the univariate model; and additionally forcing age, hypertension, diabetes mellitus, gender, type of revascularization and LVEF into the model.
Numbers needed to treat (NNT) were calculated as 1/(frequency of event in placebo group × risk reduction), which corresponds to 1/(frequency of event in placebo group − frequency of event in treatment group).
From February 1, 1994 to October 15, 1996, nearly 3,000 consecutive patients undergoing coronary angiography were screened. Despite a lower inclusion rate than estimated, it was decided to end study recruitment period after 32 months. A total of 213 patients were found eligible and 159 patients were randomized, 130 after CABG and 29 after PTCA. Study drug treatment was commenced at a median interval from invasive revascularization of seven and six days for the ramipril and placebo groups, respectively. At the end of the study, the vital status was available for all randomized patients. Among the enrolled patients, the survivors were followed for a median of 33 months (range 12 to 46 months). The baseline characteristics of patients in the two treatment groups are presented in Table 1.
Primary end points
Results of the main outcome measures are presented in Table 2, and time to cardiac event curves for each treatment group are presented in Figures 1–3. ⇓⇓⇓The quadruple composite end point of cardiac death, nonfatal AMI, development of heart failure or recurrent angina pectoris was reached in 36 patients in the ramipril group and in 41 in the placebo group (p = 0.63). With respect to the triple composite end point, eight patients in the ramipril group suffered cardiac death, nonfatal AMI or developed heart failure compared with 18 in the placebo group (p = 0.031). The time to cardiac event curves diverged early and continued to diverge for the triple composite end point (Fig. 2).
Secondary end points
Results regarding the secondary end points are presented in Table 2and Figure 3. The causes of deaths in the two treatment groups are presented in Table 3. There were two deaths (2.5%) in the ramipril group and eight in the placebo group (10%) (p = 0.053). There was a significant reduction in the prespecified secondary end point of cardiac death with ramipril treatment (p = 0.032). All cardiac deaths were classified as sudden death/fatal AMI (being sudden ischemic or sudden arrhythmic), including one patient who died within 6 h of onset of symptoms of AMI. Six cardiac deaths were out of hospital. In only one patient, cardiac death was related to progressive heart failure.
There were no significant differences between groups with respect to AMI or development of angina pectoris (Table 2). Twelve patients in the ramipril group were hospitalized with chest pain on suspicion of unstable angina pectoris compared with nine in the placebo group (p = 0.50). All cases of nonfatal AMI were non-Q-wave. Among the five patients who suffered AMI, two subsequently died and 2 developed heart failure. With respect to development of heart failure, 6 patients in the ramipril group and 12 in the placebo group developed heart failure (p = 0.10), while two patients in the ramipril group and 5 patients in the placebo group were hospitalized for heart failure (p = 0.23). The 18 patients who developed heart failure were followed for a median period of 19 months after this incident. Yet, only one patient developed persistent severe heart failure, i.e., New York Heart Association functional class III, despite medical treatment, and only this single patient died after development of heart failure.
For the quadruple composite end point, LVEF ≤ 0.40 was the single independent predictor of an adverse outcome. The relative risk (95% confidence interval [CI]) with LVEF ≤ 0.40 was 1.63 (1.04 to 2.56). For the triple composite end point, treatment group and LVEF ≤ 0.40 were independent predictors of an adverse outcome. The relative risk with ramipril treatment was 0.39 (95% CI 0.16 to 0.90). The relative risk (95% CI) with LVEF ≤ 0.40 was 3.29 (1.48 to 7.28).
Forcing age, gender, diabetes mellitus, hypertension, type of revascularization and LVEF into the multivariate model gave similar results. In this model, the relative risks with ramipril treatment were 0.91 for the quadruple composite end point (95% CI 0.58 to 1.42, p = 0.67), and 0.38 for the triple composite end point (95% CI 0.16 to 0.88, p = 0.024).
Compliance with treatment, withdrawals, adverse events, effects on blood pressure and concurrent medication
All patients tolerated the test dose. The target dose of 10 mg ramipril once daily or matching placebo was reached in 90% of patients in the ramipril group and in 96% in the placebo group. At the end of the study, 85% in the ramipril group were on study drug compared with 84% in the placebo group. Reasons for withdrawal from treatment with the test drug were open-label ACE-I treatment (five ramipril, seven placebo), loss of consent or lost to follow-up (five ramipril, four placebo), side effects (two ramipril, two placebo) and endocarditis requiring surgery (one ramipril).
No patients developed renal impairment or electrolytical derangement. Thirty patients in the ramipril group, compared with 27 in the placebo group, experienced minor adverse events, the most frequent being tiredness, gastrointestinal disturbances, dizziness and cough. A tendency towards a higher frequency of coughing, tiredness and dizziness in the ramipril group was found. From three months and beyond, the mean systolic and mean diastolic blood pressures were lower with ramipril than with placebo (128/75 vs. 136/81 at one year, p < 0.05), whereas the mean heart rates were equal (70 vs. 70 beats/min). The use of concurrent cardiovascular drugs was similar in the two treatment groups (Table 4). Four of the seven cases of sudden death/fatal AMI occurred during beta-blocker therapy. The use of diuretics reflects treatment for hypertension and the frequent administration of diuretics in the early phase after CABG due to volume overload postoperatively. After publication of the Scandinavian Simvastatin Survival Study (4S), patients were given lipid-lowering drugs according to the 4S criteria (16). The mean p-cholesterol values at one year were 5.7 mmol/liter for the ramipril group and 5.9 mmol/liter for the placebo group.
In the present study, we found that in patients with asymptomatic and moderate left ventricular dysfunction undergoing invasive revascularization for chronic stable angina pectoris, long-term treatment with ramipril, initiated shortly after invasive revascularization, reduced the incidence of the triple composite end point of cardiac death, AMI and development of clinical heart failure (Table 2and Fig. 2). In contrast, with the quadruple composite end point including recurrent angina pectoris, the study showed no significant benefit with ramipril therapy (Fig. 1). This finding may be attributed to a higher although not significant incidence of recurrent angina pectoris, observed in the ramipril group mainly during the first six months (Fig. 3d).
Among the single events that constituted the triple composite end point, ramipril significantly reduced the incidence of cardiac death, whereas the reduction in clinical heart failure and AMI was not significant (Table 2). However, in interpreting nonsignificance of the single events, one should bear in mind that the design and power of the study prescribed the use of composite end points.
The multivariate analysis revealed that the benefit with ramipril therapy on cardiac death and the triple-composite end point was not confined to patients with LVEF ≤ 0.40, as treatment group was an independent predictor. Likewise, the benefit was consistent according to whether patients were included after CABG or PTCA. The risk reduction with ramipril persisted over time as the cumulated event rate continued to diverge during follow-up time (Fig. 2). The risk reduction with ramipril treatment was substantial, but should be seen in the light of a wide CI due to the limited number of patients in the study. Nevertheless, the lower 95% CI limit of the risk reduction was comparable with that of larger studies (6,7). Also, the upper 95% CI of NNT with regard to the triple-composite end point was found within reasonable clinically relevant limits: 62 patients treated for two and a half years (Table 2).
Comparisons and dissimilarities to other studies
Despite obvious differences in inclusion criterions compared with other ACE-I trials (6–10,17–22), the efficacy of ACE-I treatment found in the present study falls in line with previous findings. Thus, in previous studies, ACE-I therapy was followed by reductions in just cardiac death, clinical heart failure and AMI (6–9,17), whereas treatment with ACE-I failed to prevent restenosis, atherosclerosis and recurrent angina pectoris after PTCA in the short term (18)as well as in the long term (19). Likewise, the effects of ACE-I treatment on stable angina pectoris were found to be variable (20–22)and even harmful (10)in previous clinical trials. Thus, with respect to symptomatic angina pectoris, clinical application of the theoretical antiatherosclerotic and anti-ischemic properties of ACE-I has as yet been disappointing.
The mechanism by which ACE-I reduces cardiac death is a matter of debate (23). The reduction in cardiac death without preceding clinically recognized heart failure, which we observed in the present study, supports the hypothesis that ACE-I treatment reduces cardiac death not only by preventing development and progression of clinical recognized heart failure, but probably additionally by reducing sudden death. The fact that the study population in APRES, compared with study populations in previous ACE-I trials (6–9), were less prone to develop severe heart failure may have revealed and accentuated this efficacy.
The finding of a reduction in sudden death and AMI but not in recurrent angina pectoris may seem counter-intuitive. However, this finding is in concordance with some previous findings (8), and differences in the mechanisms behind development of these conditions can be emphasized. Sudden death in a patient with documented ischemic heart disease is likely to be due to sudden ischemic or arrhythmic events (23). Neurohormonal modulation by ACE-I may reduce sudden vasoconstriction and arrhythmogenesis. In contrast, recurrent angina pectoris may be due to graft atherosclerosis, myointimal proliferation and reduction in perfusion pressure in patients with multiple coronary artery stenoses.
Acute myocardial infarction, severe angina pectoris, left ventricular dysfunction and clinical heart failure are definite risk factors in patients with ischemic heart disease. Previous ACE-I trials focused on patients with these conditions. In contrast, the patients included in APRES had no recent AMI and underwent uncomplicated invasive revascularization for chronic stable angina pectoris, the mean LVEF was 0.41 compared with 0.25 to 0.31 in other studies (6,7,9,13)and the patients included had no clinically recognized heart failure and were not on diuretic therapy for heart failure. Taking only these risk factors in account, the study population in APRES could be regarded as being at rather low risk compared with the study populations in previous ACE-I trials (6–9). However, the incidence of severe cardiac events in the placebo group in the APRES was not negligible. Compared with the study populations in other ACE-I trials, the study population in APRES may have had risk factors of more extended coronary artery disease and a longer history of ischemic heart disease. The fact that the event rate in the placebo group in APRES is representative for the studied population is supported by the fact that the risk of cardiac death and AMI found in APRES is similar to the risk found for surgical patients with ejection fraction < 0.50 in the Coronary Artery Surgery Study (CASS) (2).
The findings in APRES support the fact that patients with asymptomatic and only moderate left ventricular dysfunction who underwent uncomplicated invasive revascularization for chronic stable angina pectoris still carry a significant risk of cardiac death and other cardiac events (2,4). However, the results in APRES indicate that long-term treatment with ramipril may reduce this risk. Accordingly, the benefits with ACE-I treatment may be extended to include and recommend treatment of this patient group.
☆ This study was made possible with a grant from AstraZeneca Danmark A/S, Albertslund, Denmark and a grant from Rigshospitalet, Copenhagen, Denmark.
- Scandinavian Simvastatin Survival Study
- angiotensin-converting enzyme inhibitor
- acute myocardial infarction
- Angiotensin-converting Enzyme Inhibition Postrevascularization Study
- coronary artery bypass grafting
- coronary artery surgery study
- left ventricular ejection fraction
- numbers needed to treat
- percutaneous transluminal coronary angioplasty
- Received March 17, 1999.
- Revision received October 15, 1999.
- Accepted November 19, 1999.
- American College of Cardiology
- Alderman E.A.,
- Bourassa M.,
- Cohen L.,
- et al.
- Holmes D.R.,
- Detre K.M.,
- Williams D.O.,
- et al.
- Huikuri H.V.,
- Yli-Mäyry S.,
- Airaksinen K.E.,
- et al.
- Muhlbaier L.H.,
- Pryor D.B.,
- Scott Rankin J.,
- et al.
- Cleland J.G.F.,
- Hendersson E.,
- McLenachan J.,
- et al.
- Davies M.K.
- Hamer A.,
- Takayama M.,
- Abraham K.A.,
- et al.
- Kjøller-Hansen L.,
- Steffensen R.,
- Grande P.
- Yusuf S.,
- Pepine C.J.,
- Garces C.,
- et al.
- Marcator Investigators
- Gibbs J.S.R.,
- Crean P.A.,
- Mockus P.,
- et al.
- Vogt M.,
- Motz W.,
- Strauer B.E.