Review of Survival and Ventricular Enlargement (SAVE) Trial
Effect of captopril on mortality and morbidity in patients with left ventricular dysfunction after myocardial infarction
Background After a large heart attack, cardiac hemodynamics are altered, which leads to a series of changes in the heart muscle itself. The immediate consequence of a large heart attack is decreased myocardial contractility. This leads to a reduction in stroke volume and cardiac output. Compensatory mechanisms governed by the renin-angiotensin-aldosterone system and sympathetic nervous system become activated and may lead to ventricular dilation or “remodeling”, which has negative short and long-term consequence for the heart muscle.
By the mid-1980’s, angiotensin converting enzyme inhibitors (ACEi) were commonly used for patients with chronic systolic heart failure and laboratory work had shown they could improve ventricular remodeling, reduce heart failure and prolong survival in animal models of AMI.
Several studies had also shown promise in humans but they were too small to test hypotheses involving hard clinical endpoints. The Survival and Ventricular Enlargement (SAVE) trial sought to test the hypothesis that administration of captopril to patients with AMI complicated by left ventricular dysfunction but who did not have overt heart failure requiring vasodilatory therapy would reduce morbidity and mortality over long-term follow-up.
Patients Eligible patients were between 21 to 80 years of age with a definite myocardial infarction occurring 3 to 16 days prior to randomization with a left ventricular ejection fraction (EF) ≤40%, measured by radionuclide ventriculography. Patients were excluded with relative contraindications to ACEi or if an ACEi was indicated for treatment of symptomatic congestive heart failure or systemic hypertension. Other exclusion criteria included serum creatinine >2.5 mg/dl, other conditions limiting survival (unspecified) or who had an unstable course following AMI (also unspecified).
Baseline characteristics The average age of patients was 59 years and 82% were men. Approximately one third of patients had a prior MI, more than 20% had diabetes, 40% had hypertension and over 50% were current smokers. The average ejection fraction was 31% and over half of patients had an anterolateral Q wave MI. The mean time to randomization was 11 days. Prior to randomization approximately a third received thrombolysis, over 50% underwent coronary angiography, and approximately 25% underwent either percutaneous coronary angioplasty or coronary artery bypass surgery. At the time of randomization, the average blood pressure was 113/70 mmHg and heart rate was 78 beats per minute. Within 24 hours of randomization approximately one third of patients received a beta blocker and a quarter received digoxin.
Procedures There was a mini run-in period where all 2,250 eligible patients were given a test dose of 6.25 mg of captopril. This led to exclusion of 19 patients (3 for ischemic discomfort and 16 for symptomatic hypotension).
Patients received either captopril or placebo. The initial dose of the blinded study drug was 12.5 mg but could be administered at 6.25 mg to patients who had marked, yet asymptomatic, reduction in BP with the run-in dose. The target dose was 25 mg three times a day by the end of the in-hospital phase and was gradually increased to 50 mg three times a day unless side effects occurred. There was no prespecified level of blood pressure in the titration regimen.
Outpatient visits were scheduled 2 weeks following randomization and then every 3 months during year 1 and every 4 months thereafter. Compliance with the study drug was assessed by pill count.
Endpoints There were no prespecified hypothesis tests used to determine sample size. Prospectively defined measures of outcomes included all cause death, cardiovascular death, incidence of clinical congestive heart failure and first hospitalization for heart failure. Once the clinical endpoint committee was notified of a diagnosis of clinical heart failure the study medication was discontinued so open-label therapy with an ACEi could be started.
Results 2,231 patients were included in the final analysis; 1116 in the placebo group and 1115 in the captopril group. The mean follow-up was 3.5 years. Blood pressure increased from baseline in both groups but to differing extents. At 3 months, BP was 125/77 in the placebo group compared to 119/74 in the captopril group. The average heart rate was 74 in both groups.
Compared to placebo, captopril significantly reduced all cause death by 19% (20% vs 25%; p = 0.019) and cardiovascular death by 21% (17% vs 21%; p = 0.014). It also reduced clinical heart failure by 37% (11% vs 16%; p <0.001), heart failure hospitalization (14% vs 17%; p = 0.019), and nonfatal MI (12% vs 15%; p = 0.015).
Examination of subgroups for the endpoints of all-cause mortality and cardiovascular mortality reveals the effect of captopril was generally preserved across groups; however, given the size of the trial we feel that all subgroup analyses are limited.
At 1 year, 82% of patients in the placebo group and 79% of patients in the captopril group were still taking their assigned therapy and at the end of the study it was 73% and 70%. In the captopril group, 79% who were still taking the assigned drug were on the target dose by the end of the study. Side effects that occurred more often in the captopril group included dizziness (5%), altered taste (2%), cough (6%) and diarrhea (2%).
Conclusions In patients with AMI complicated by significant LV dysfunction but not overt clinical heart failure, captopril significantly reduced death over 3.5 years of follow-up with a number needed to treat of approximately 20 patients. Statistically significant reductions were noted across all prespecified outcomes including cardiovascular mortality, incident congestive heart failure, hospitalization for heart failure and nonfatal MI. No obvious treatment effect heterogeneity was noted across subgroups of patients but these analyses were limited due to the overall sample size.
We urge caution in accepting these results, which may seem robust on the surface. A 20% relative reduction and 5% absolute reduction in death may be unrealistic without much precedent from previous trials to base this on. Had the authors sought to test a more conservative estimate of mortality reduction, like 10%, the sample size would have needed to be over 4x higher. We estimate that 9,094 patients would be required to show a relative reduction of 10% or 2.5% absolute reduction at an alpha of 0.05 and power of 80%.
From a clinical translation standpoint, it should be noted that patients were not eligible for enrollment until at least 3 days following AMI and the average start date was day 11.
Patients over 80 years of age were excluded and a mini run-in period with a test dose of captopril was given which led to the exclusion of 16/2,250 eligible participants. Despite these limits, in many other ways, external validity was high, especially compared to earlier beta blocker trials. No limits were placed on blood pressure and heart rate at study entry and dose titration protocols were not strictly governed.