High-Dose Atorvastatin vs Usual-Dose Simvastatin for Secondary Prevention After Myocardial Infarction:  The IDEAL Study: A Randomized Controlled Trial FREE

Lowering of low-density lipoprotein cholesterol (LDL-C) with statins has in the last decade become part of the standard treatment regimen in patients with established coronary heart disease (CHD). Following the publication of the results of the Scandinavian Simvastatin Survival Study (4S) in 1994,¹ the most common treatment regimen for such patients in northern Europe has been simvastatin, 20 to 40 mg/d. In a recent trial among patients with acute coronary syndromes, incremental benefit was demonstrated with more intensive lowering of LDL-C to well below 100 mg/dL.² In the Treating to New Targets (TNT) study comparing high and low doses of atorvastatin in stable nonacute CHD, a significant improvement in prognosis with respect to cardiovascular disease was demonstrated.³ In that study, however, the benefit of reduced cardiovascular mortality appeared to have been offset by a higher number of deaths due to noncardiovascular causes. Although this difference did not reach statistical significance and could well be due to the play of chance, it led to a call for further safety information on the use of atorvastatin at a dose of 80 mg/d.⁴

The Third Joint Task Force of European and Other Societies on Cardiovascular Disease Prevention in Clinical Practice⁵ in 2003 recommended an LDL-C target level of less than 100 mg/dL (2.5 mmol/L) for CHD patients. The Third Report of the National Cholesterol Education Program (NCEP) Adult Treatment Panel recently introduced a new target of less than 70 mg/dL (1.8 mmol/L) for patients at very high risk.⁶ The main hypothesis of the current study, the Incremental Decrease in End Points Through Aggressive Lipid Lowering (IDEAL) study, was that intensive lowering of LDL-C with atorvastatin at the highest recommended dose would yield incremental benefit compared with the moderate, most widely used dose of simvastatin.

The IDEAL study was a multicenter, prospective, randomized, open-label, blinded end-point classification trial (the so-called PROBE design⁷) carried out at 190 ambulatory cardiology and private specialist centers in Denmark, Finland, Iceland, the Netherlands, Norway, and Sweden. A detailed description of the study design and baseline characteristics of the patients has been published elsewhere.⁸ In brief, recruitment and randomization took place from March 1999 to March 2001 and patients were followed up until March 2005. Records of patients previously treated at the centers were screened for the main eligibility criteria. Potentially eligible patients were invited for a screening visit. Written informed consent was obtained from all patients, and the study was approved by the national or regional review board in all countries and by governmental reimbursing institutions in countries where the main sponsor did not cover all costs.

Men and women aged 80 years or younger with a history of a definite myocardial infarction and who qualified for statin therapy according to national guidelines at the time of recruitment were eligible. The main exclusion criteria were any known contraindications to statin therapy, previous intolerance to statins in low or high doses, liver enzyme levels more than 2 times the upper limit of normal, pregnancy or breastfeeding, nephrotic syndrome, uncontrolled diabetes mellitus, uncontrolled hypothyroidism, plasma triglyceride levels higher than 600 mg/dL (6.8 mmol/L), congestive heart failure (New York Heart Association classification IIIb or IV), hemodynamically important valvular heart disease, gastrointestinal conditions affecting absorption of drugs, treatment with other drugs that seriously affect the pharmacokinetics of statins, and treatment with other lipid-lowering drugs. Patients previously treated with statins qualified if they had not already had titration to a dose higher than the equivalent of 20 mg/d of simvastatin.

After dietary counseling, patients fulfilling the eligibility criteria were randomized to receive simvastatin, 20 mg/d, or atorvastatin, 80 mg/d (Figure 1). Study medication was assigned via a central interactive voice response system (ClinPhone, Nottingham, England). Allocation numbers were given out in blocks of 24. Allocation was balanced by center; no other stratification was used. There was no run-in or washout period. Study medication was provided by prescription except in Finland, where it was dispensed at the expense of the sponsor.

Patients were followed up at the centers after 12 and 24 weeks and every 6 months thereafter. If, at 24 weeks, plasma total cholesterol level was higher than 190 mg/dL (5.0 mmol/L), the dose of simvastatin could be increased to 40 mg/d. The dose of atorvastatin could be decreased to 40 mg/d for adverse events. If LDL-C decreased to less than 39 mg/dL (1.0 mmol/L), an investigator would be notified and could consider reducing the statin dose.

All lipid and lipoprotein levels were measured from fasting blood samples. Such measurements, along with liver enzymes and other laboratory measurements, were made at baseline, at 12 and 24 weeks, at 1 year, and yearly thereafter. All measurements were made at a central laboratory. The results of lipid and lipoprotein measurements were not revealed to study personnel during the study except in cases of titration of simvastatin at 24 weeks.

The primary clinical outcome was time to first occurrence of a major coronary event, defined as coronary death, hospitalization for nonfatal acute myocardial infarction, or cardiac arrest with resuscitation. Potential myocardial infarction cases were adjudicated according to current guidelines of the Joint European Society of Cardiology/American College of Cardiology. There were 3 prespecified composite secondary outcomes: (1) major cardiovascular events (any primary event plus stroke; the diagnosis of stroke required evidence of a neurological deficit, usually localized, lasting ≥24 hours or until death, usually confirmed by diagnostic imaging); (2) any CHD event (any primary event, any coronary revascularization procedure, or hospitalization for unstable angina); (3) any cardiovascular events (any of the former plus hospitalization with a primary diagnosis of congestive heart failure and peripheral arterial disease, defined as new clinical diagnosis or hospitalization for such disease). In addition, individual components of the composite end points were also prespecified as secondary outcomes, as was all-cause mortality.

In addition to per-protocol reporting by investigators, monitors reviewed patient records at regular intervals to search for potential end points. An end-point classification committee blindly reviewed reports on potential end points and adjudicated outcomes at regular meetings. All reports were first screened by an independent center (Inveresk, Raleigh, NC) for blinding of treatment allocation.

Based on previous experience, it was projected that simvastatin therapy would produce a mean 35% reduction in LDL-C from untreated levels, while the reduction with atorvastatin, 80 mg/d, would be at least 55%, creating a difference in plasma levels of about 40 mg/dL (1 mmol/L). The trial was designed to have 90% power to detect an anticipated 21% relative risk reduction (from 10% to 7.9%) in the primary outcome variable with atorvastatin over 5 years using a 2-tailed α level of .05. Because the risk of the patients first recruited was recalculated to be lower than first anticipated, the originally planned sample size of 7700 patients was increased to 8888. A data and safety monitoring board performed interim analyses when approximately 50% and 75% of the predetermined final number of 774 patients had experienced a primary end point.

The study was initiated by the investigators and scientifically led by a steering committee consisting of independent researchers and investigators and 3 members employed by the sponsor. Monitoring of data collection was provided by the sponsor. A contract research organization (Covance, Horsham, England) reviewed the data for errors and inconsistencies and sent queries to the investigators for clarification. This organization provided the interim reports for the data and safety monitoring board. The final statistical analysis was performed by the sponsor. An independent academic statistician of the steering committee (I.H.) had full access to all of the data at the completion of data collection and verified the analyses of the results.

Kaplan-Meier hazard rates were used to examine incidence over time and the log-rank test was used to assess group differences. Cox proportional hazards models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs).⁹ Statistical analysis was performed using SAS, version 8.2 (SAS Institute Inc, Cary, NC). All analyses in this report are based on the intention-to-treat principle including all randomized patients. Two-sided P values of <.05 were regarded as statistically significant.

After screening of patients’ records, 9689 potentially eligible patients were called in for a screening visit. Of these, 8888 met the eligibility criteria and were randomized (Figure 1). The main reasons for exclusion were patients’ use of higher statin doses than the equivalent of 20 mg/d of simvastatin, unwillingness to participate, and previous adverse experience with statins.

The median follow-up time was 4.8 years (range of surviving participants, 4.0-5.9 years). Six patients were lost to follow-up and 48 patients withdrew consent prior to study close, but vital status was known for 35 of these at the close of the study. Data for these patients have been included in the analysis for the period prior to loss or withdrawal. Vital status at the end of the trial is thus unknown for 19 patients.

Baseline characteristics were well balanced between the 2 treatment groups (Table 1). The median time since last myocardial infarction was 22 months in the simvastatin group and 21 months in the atorvastatin group.

At 24 weeks of follow-up, 900 patients (21%) in the simvastatin group had their dosage increased to 40 mg/d. At the end of the study, 1034 (23%) were prescribed simvastatin, 40 mg/d. In patients allocated to receive atorvastatin, 80 mg/d, 250 (6%) had their dosage reduced to 40 mg/d by 24 weeks; in 587 patients(13%), the final dose was 40 mg. Overall adherence, defined as total study medication exposure as a percentage of total follow-up time, was 89% in the atorvastatin group and 95% in the simvastatin group. By the end of the study, 14% of the atorvastatin-allocated and 7% of the simvastatin-allocated patients had permanently discontinued study medication. Most patients who stopped taking the study drug switched to a different statin. In the simvastatin group, 360 patients took a different statin at some point; in 123 (2.8%), it was atorvastatin. In the atorvastatin group, 645 patients took a different statin; in 364 (8.2%), it was simvastatin.

Since the majority of patients allocated to receive simvastatin therapy were already taking simvastatin, 20 mg/d, or another statin at an equivalent dose at the time of randomization, the changes in lipid and lipoprotein levels for the group as a whole were small (Table 2). Patients in the simvastatin group who were not taking a statin at the time of randomization had, on average, a reduction in LDL-C of 33% after 12 weeks. In the group allocated to atorvastatin, 80 mg/d, statin-naive patients had a mean reduction in LDL-C of 49%. During treatment, mean (SE) LDL-C levels were 104 (0.3) mg/dL (2.7 [0.008] mmol/L) in the simvastatin group and 81 (0.3) mg/dL (2.1 [0.008] mmol/L) in the atorvastatin group. Total cholesterol and triglyceride levels were also significantly lower in the atorvastatin group compared with the simvastatin group, whereas high-density lipoprotein cholesterol (HDL-C) levels were slightly but significantly higher in the simvastatin group. Apolipoprotein levels changed correspondingly (Table 2).

In December 2004, reports of 702 patients with a confirmed primary end point had arrived at the coordinating centers. Based on this information, the steering committee decided that the study close-out procedures should be concluded by April 2005, when it was anticipated that the protocol-specified target of 774 patients with a primary end point would have occurred. When all study close-out visits had been performed, a total of 874 patients had actually experienced a primary end point.

The primary end point of coronary death, acute myocardial infarction, or cardiac arrest with resuscitation occurred in 463 patients (10.4%) in the simvastatin group and in 411 (9.3%) in the atorvastatin group (Table 3). This corresponds to a relative risk reduction of 11% with atorvastatin, 80 mg/d (HR, 0.89; 95% CI, 0.78-1.01; P = .07) (Figure 2). A post hoc Cox regression analysis of the primary end point with adjustment for sex, age, statin use at randomization, duration since last myocardial infarction, total cholesterol, and HDL-C resulted in an HR of 0.87 (95% CI, 0.76-0.99; P = .04). A preliminary analysis of prespecified subgroups defined by sex and age did not reveal any statistically significant treatment group interactions.

There were 178 coronary deaths (4.0%) in the simvastatin group vs 175 (3.9%) in the atorvastatin group (HR, 0.99; 95% CI, 0.80-1.22; P = .90). Nonfatal myocardial infarction occurred in 321 patients (7.2%) in the simvastatin group and in 267 (6.0%) in the atorvastatin group (HR, 0.83; 95% CI, 0.71-0.98; P = .02). The composite secondary end point of a major cardiovascular event including stroke was reduced in the atorvastatin group (HR, 0.87; 95% CI, 0.78-0.98; P = .02). Similarly, there were reductions in the risk of nonfatal myocardial infarction, any CHD event, and any cardiovascular event. Hemorrhagic strokes occurred in 6 patients in each treatment group. Kaplan-Meier hazard rates for selected components of the secondary end points are shown in Figure 3.

The risk of death from any cause was similar in both study groups (HR, 0.98; 95% CI, 0.85-1.13; P = .80) (Table 3). There were no significant differences in noncardiovascular deaths between the treatment groups. In particular, mortality due to cancer was similar in the 2 groups, at 112 (2.5%) in the simvastatin group and 99 (2.2%) in the atorvastatin group. There were no significant differences in cancer mortality for any particular body system.

There were no significant differences in the frequency of serious clinical adverse experiences between the 2 groups. There were, however, more patients in the atorvastatin group who discontinued study medication because of investigator-reported adverse effects (Table 4). Elevation of liver enzyme levels occurred more frequently in the atorvastatin group, but this was not related to any increased incidence of clinical liver disease. Myalgias occurred more frequently in the atorvastatin group, but myopathy rates were exceedingly low in both the atorvastatin and simvastatin groups.

The majority of trials on statin therapy in the last decade have examined the effect of lowering LDL-C by 25% to 40%. In the 4S study, the mean LDL-C reduction was 35%.¹ The IDEAL study intended to explore whether a strategy of reducing LDL-C even further would yield incremental benefit. The prespecified primary end point did not achieve statistical significance (P = .07), but there were significant reductions in nonfatal acute myocardial infarction and in the secondary composite end points of any CHD event and major or any cardiovascular events. There was no difference in all-cause or cardiovascular mortality. Thus, although the reductions were more modest than expected and although the primary end point was not met, these results indicate that more intensive lowering of LDL-C than usual in patients with previous myocardial infarction might prevent 68 first cardiovascular events (95% CI, 39-97) per 1000 patients over 5 years.

There are several possible reasons why statistical significance was not reached for the primary end point. One explanation might be insufficient difference in levels of LDL-C between the groups since the observed difference was slightly smaller than projected. The initial LDL-C reduction of 49% in statin-naive patients taking atorvastatin, 80 mg/d, was less than expected, and although adherence with this therapy was excellent, it was not as good as in the simvastatin group.

A second possible explanation was that the follow-up duration was only a median of 4.8 years even though the study protocol anticipated the prespecified number of primary end points to be reached after a median of 5.5 years. A third possibility is that the effect of simvastatin on HDL-C would attenuate the difference produced by the improved effect of atorvastatin on LDL-C. However, the impact of statins on HDL-C has not yet been shown to influence patient outcomes.

The IDEAL study was carried out with the PROBE design and, thus, did not have the advantages of a double-blind trial. However, the end-point classification was conducted by a blinded clinical end-points committee with the idea of minimizing bias. The open-label design with prescription of study medication had the advantage of being more like a “real-world” setting, but the possibility of bias for some of the physician-initiated end points, such as coronary revascularization and hospitalization for unstable angina, cannot be excluded. The fact that most patients had to pay part of the cost of the study drug apparently did not affect prescription rates, because the cost for the patients of the 2 study drugs was identical.

The apparent adherence to atorvastatin was high and better than that in other comparable trials. The adherence in the simvastatin group was, however, exceptional (95%). The higher adherence to study medication in the simvastatin group than in the atorvastatin group may be explained by the fact that 51% of the patients had been taking simvastatin prior to randomization and were probably comfortable with it, while in an open-label design a high dose of atorvastatin might have led to hesitation by some patients and investigators, especially early in the trial. This also makes it difficult to make reliable comparisons of reported adverse experiences between the 2 treatment regimens.

The Pravastatin or Atorvastatin Evaluation and Infection Therapy (PROVE-IT) study compared pravastatin, 40 mg/d, with atorvastatin, 80 mg/d, in patients with recent acute coronary syndromes.² In that study, the difference in LDL-C levels between the treatment groups averaged 33 mg/dL (0.86 mmol/L), which produced a 16% relative reduction in the HR of the primary composite end point, which included death, myocardial infarction, unstable angina, coronary revascularization, and stroke. Thus, in that study the patient population was different, since only 11% of the IDEAL population had a recent myocardial infarction (<2 months); in addition, the difference in LDL-C between the treatment groups was larger and the definition of the primary end point was different.

In the TNT study, the primary end point included stroke.³ When comparing the primary end point of the TNT trial with the same end point in the IDEAL study, the difference between treatment groups was smaller in IDEAL (HR, 0.78 vs 0.87), but the 95% CIs for the HRs overlap. Comparison of the broader end point of any cardiovascular events, however, reveals more similar HRs of 0.81 and 0.84 in the 2 trials, respectively.

A recent prospective meta-analysis of 14 cholesterol-lowering statin trials with more than 90 000 patients found a 23% proportional reduction in the incidence of major coronary events and a 21% proportional reduction in the incidence of major cardiovascular events per 1 mmol/L of LDL-C reduction (corresponding to a 5.5% and 5% reduction in incidences per 10 mg/dL reduction in LDL-C, respectively).¹⁰ Our results are consistent with these findings and are also in accordance with other meta-analyses and epidemiological data on the relationship of cholesterol levels and CHD risk^10- 14 and by findings from internal subgroup analyses of results of previous trials.^15- 16 Recent findings in trials of other comparative drugs^3,17 and in different clinical settings^2,18 have provided evidence of the same relationship.

The IDEAL trial was not powered to detect a significant difference in all-cause mortality. In the 4S study, the comparator was placebo, and in the placebo group 74% of the deaths were coronary.¹ In IDEAL, only 48% of the deaths in the simvastatin group had a coronary cause, which is considerably lower than the 61% of deaths having a coronary cause in the simvastatin group in 4S. This decline in coronary mortality may well reflect improvements in coronary prevention and care during the last decade. While this improvement must be welcomed, it has made it more difficult for trialists to demonstrate further benefit in survival.

In the IDEAL study, there was a small and nonsignificant excess of 13 more noncardiovascular deaths in the simvastatin group than in the atorvastatin group. In contrast, atorvastatin, 80 mg/d, was associated with a small and nonsignificant increase in noncardiovascular deaths compared with atorvastatin, 10 mg/d, in the TNT study. Such small differences are likely to have occurred by chance. There was no difference between the groups in the frequency of adverse events that were rated as serious. There were, however, more nonserious adverse events resulting in drug discontinuation in the atorvastatin group. This difference may reflect real nontolerance to a high dose of atorvastatin, but the possibility of reporting bias is present given the open-label design of the trial.

The proportion of patients developing liver enzyme elevation with atorvastatin, 80 mg/d, was low and is comparable with results of other similar trials. The proportion of patients in the simvastatin group who developed liver enzyme elevations was exceptionally small and was significantly lower than in the atorvastatin group. This reporting was not subject to bias, since it was performed by the central laboratory that transferred the results directly to the study database. The low frequency is readily explained by the fact that half of the patients had received simvastatin prior to randomization and were selected as “simvastatin-tolerant.”

In summary, when comparing standard and intensive LDL-C–lowering therapies in patients with previous myocardial infarction, there was no statistically significant reduction in the primary end point of major coronary events, but there was reduced risk of other composite secondary end points and nonfatal acute myocardial infarction. There were no differences in cardiovascular and all-cause mortality. The results indicate that patients with myocardial infarction may benefit from intensive lowering of LDL-C without increase in noncardiovascular mortality or other serious adverse reactions.