ENHANCE and ONTARGET

These are 2 recently published studies, drug vs drug or combinations, in which there were no significant differences in the primary outcome but in which there was considerable interest in the results. One involves the principle of noninferiority, still somewhat controversial in its implications, and the other raises the question of the significance of relatively short-term surrogate end points. Summary. A randomized, double-blind 24-month study comparing the effects of high-dose simvastatin (80 mg/d) with placebo or with ezetimibe (10 mg/d) on carotid intima-media thickness (IMT) changes showed no significant differences in the latter despite significant decreases in low-density lipoprotein cholesterol (LDL-C) and C-reactive protein (CRP) in the combined therapy group. Rationale. Since the introduction of ezetimibe, which lowers LDL-C by a different mechanism than statins, considerable interest has been focused on the possibility of enhanced benefit of a statin/ezetimibe combination on atherogenic processes because of the incremental decrease in LDL-C by 10% to 20% compared with a statin alone. Patients with familial hypercholesterolemia were studied because of increased progression of atherosclerotic disease leading to early risk of cardiovascular (CV) events. The noninvasive approach of assessing change in carotid IMT allowed a relatively simple method of assessing progression of atherosclerosis. Not a small factor of interest in this study outcome has been the widespread use of a combination of simvastatin with ezetimibe with little clinical trial background on the efficacy of such a combination compared with a statin alone. Study Design. The study was conducted in 18 centers in North America, Europe, and South Africa. Men aged 30 to 75 with heterozygous familial hypercholesterolemia were eligible for enrollment if they had untreated LDL-C levels of ≥210 mg/dL. Excluded were candidates with high-grade stenosis of a carotid artery, significant congestive heart failure (CHF), cardiac arrhythmias, angina, or recent CV events. Enrollees were distributed equally into each of the 2 groups. Duplicate ultrasonography of the carotid and femoral arteries was accomplished at baseline and at 24 months, 1 week apart at each time. Primary outcome was change in far-wall mean carotid artery IMT of the right and left common carotid arteries, carotid bulbs, and internal carotid arteries. Secondary outcomes included proportions of patients showing regressive changes in carotid thickness, development of new carotid plaques, and combined changes in IMT of the carotid and femoral arteries. Results. Of the initial 720 patients randomized, 642 completed the trial: 320 in the simvastatin only group and 322 in the combined group. The average age in each group was 46 years, and there was no difference between groups in history of diabetes (<3%), hypertension (mean, 16%), smoking (28%), or previous myocardial infarction (MI) (mean, 6%). A high percentage (80%) had previously been using statins. After 24 months of treatment, the primary outcome (mean change in carotid artery IMT) was no different among groups (+0.0058 mm in statin only and +0.0111 in the combined group; P=.29) (Figure 1). Similarly, there were no significant differences in changes in femoral artery IMT or in comparisons of the common carotid arteries, carotid bulbs, and internal carotid arteries. These results were similar despite a significant decrease from baseline in the combined group vs the statin only group in LDL-C (−60% vs −40%) and CRP (−49.2% vs −23.5%). In terms of adverse events, no differences were found for discontinuation of medications due to elevated liver enzymes (2.2% with statin only, 2.8% with combination) or in elevated creatine kinase (2.2% with statin only, 1.1% with combination). Reported CV events developed in 7 patients (2.2%) in the statin only group and 10 patients (3.1%) in the combination group, which was also not significantly different. The Ezetimibe and Simvastatin in Hypercholesterolemia Enhances Atherosclerosis Regression (ENHANCE) study. Changes in carotid intima-media thickening between the 2 groups. The results indicated no significant benefit of the combination compared with simvastatin alone. Reprinted with permission from the Massachusetts Medical Society from Kastelein et al.1 Comments. Although this study did not evaluate clinical end points, there is ample evidence that carotid IMT correlates with CV outcome events.2 The finding that the observed change in intima-media thickening was not benefited by a reduction in LDL-C in the combined group could be explained by the possibility that ezetimibe does not have the pleiotropic effects of the statins to further influence limitation of atherogenesis. However, other studies have demonstrated that LDL-C reductions per se, independent of statins, were associated with CV risk reductions,3 and a meta-analysis has demonstrated that risk reduction with statins is not increased by effects beyond LDL-C reduction.4 The authors point out that the initial carotid IMT was lower than that in other studies of familial hypercholesterolemia and that the progression of IMT seen in the current study was less than that seen in another study of familial hypercholesterolemia (Atorvastatin vs Simvastatin on Atherosclerosis Progression [ASAP]),5 in which the simvastatin dose was half of that used in the current study. They postulate that the decrease in progression may be limited physiologically and that the initially relatively low IMT and more significant effects on progression had reached the limits beyond which ezetimibe would not be effective. Two accompanying editorials6, 7 provide perspectives on the study. To summarize, (1) the large percentage of those who were pretreated with statins before randomization may already have had plaque depletion, which would have affected the results; however, the 19% who did not have statin pretreatment did not have a better response with combination therapy; (2) it is possible that the absence of additional effects of the combination on high-density lipoprotein cholesterol may have been important; and (3) ezetimibe may trigger proatherogenic gene-regulatory mechanisms or promotion of other adverse gene expression. As many are aware, the results of this trial received great publicity before publication. Some results of the trial were released by the manufacturers of the combination drug ezetimibe/simvastatin in January 2008 at a news conference. The results require further assessment, including information on previous use of statins before randomization. Several studies now under way may shed further light on the efficacy of ezetimibe.8, 9 The evaluation of clinical end points, CV events, is the final arbiter of the efficacy of ezetimibe, and the results are clearly not yet available. Summary. This was an international randomized median 56-month trial of the use of an angiotensin receptor blocker (ARB) and angiotensin-converting enzyme inhibitor (ACEI) alone or in combination in 25,620 persons with vascular disease or diabetes with end-organ damage. The goal was to evaluate the primary outcome event of death from CV causes, MI, stroke, or hospitalization for CHF. The results indicated that telmisartan (80 mg/d) was equivalent to ramipril (10 mg/d) in primary outcome, although there were significant differences in side effects such as hypotensive symptoms, angioedema, syncope, and renal dysfunction. Rationale. The purpose of the study was to evaluate the noninferiority or superiority of an ARB to an ACEI in affecting CV outcomes in persons without CHF at high risk for events. The dosage of ramipril used had been shown previously to benefit CV outcomes. ARBs had been shown previously to reduce outcome events in patients with CHF and in high-risk patients with hypertensive left ventricular hypertrophy.11 Study Design. Patients with coronary artery disease, peripheral vascular disease, cerebrovascular disease, or diabetes and end-organ damage were enrolled. CHF was a reason for exclusion. The final dosages of medication were telmisartan 80 mg/d, ramipril 10 mg/d, or a combination. In addition to the primary outcome events evaluated above, secondary outcome events included the development of CHF and other individual CV events, diabetes, atrial fibrillation, cognitive decline, nephropathy, and revascularization. Recruitment was obtained in 733 centers in 40 countries throughout the world. The primary analysis used a time-to-event approach. Power calculations for the original planned sample size provided a power of 93% for superiority for a 13% difference among groups and 89% for noninferiority among groups for a hazard ratio (HR) of 1.00 among groups. Results. Baseline characteristics of the patients (8576 ramipril, 8542 telmisartan, and 8502 combination) were similar: mean age was 66 years; blood pressure was 142/82 mm Hg; approximately 75% had a history of coronary artery disease (49% MI, 49% angina); 21% had a history of stroke; 13% had a history of peripheral artery disease; 69% had hypertension; and 37% had diabetes. A high percentage had been receiving a statin (62%), a β-blocker (57%), or an antiplatelet agent (>80%). After a median of 56 months, the primary outcome developed similarly in the ramipril and telmisartan groups (16.5%and 16.7%, respectively; HR for telmisartan, 1.01; 95% confidence interval [CI], 0.94–1.09)(Figure 2). Similarly, in the combination group, the rate was 16.3%. In terms of noninferiority of telmisartan, the upper boundary of 1.09 was within the boundary of noninferiority to ramipril, but the lower boundary of 0.94 did not indicate superiority of telmisartan. As for secondary outcomes, there was no evidence of inferiority or superiority of telmisartan vs ramipril. Side effect analysis of medication discontinuation indicated that a significantly higher percentage of ARB users had hypotensive symptoms (2.7% vs 1.7%) and a higher percentage of ACEI users had cough (4.2%vs 1.1%). In comparison of combination therapy vs ramipril alone, a significantly higher percentage of the combination group had medication discontinued because of hypotensive symptoms (HR, 2.75), syncope (HR, 1.95), diarrhea (HR, 3.28) or renal impairment (HR, 1.58). The Ongoing Telmisartan Alone and in Combination With Ramipril Global End Point Trial (ONTARGET). Comparison of primary outcomes over a 5-year follow-up including death from cardiovascular causes, myocardial infarction, stroke, or hospitalization for heart failure. Reprinted with permission from the Massachusetts Medical Society from the ONTARGET Investigators.10 Comments. In conducting a trial comparing efficacy of a newer agent with an established agent in which a placebo is not used, evaluation of the newer agent is based not only on possible superiority of the newer agent but also on noninferiority. The concept of noninferiority of the newer agent in such a clinical trial requires a reference trial in which the older agent was proven superior to a placebo (in this case, the ACEI ramipril in the Heart Outcomes Prevention Evaluation [HOPE] trial),12 that the older drug is used in a dosage similar to that in the reference trial, that the patient population is similar, that the outcomes measured are similar to those of the reference trial, and that high adherence rates are achieved. In regard to this study, the HOPE trial had a similar population of high-risk patients, the dosage of ramipril was the same, and the adherence rates in ONTARGET were even higher than those in HOPE (85% vs 79% at the end of the study). The upper boundary of the 95% CI for telmisartan for the primary outcome (1.09) was lower than the 1.13 predefined outcome above which telmisartan would have been considered inferior to ramipril. It may be worthwhile to demonstrate how the number 1.13 was determined. In the initial statistical study planning, the investigators used the HR of 0.775 of ramipril vs placebo for the primary outcome in the HOPE trial as the basis for computation of the noninferiority boundary. They chose a more robust HR (0.794), which was computed as an excess risk of 1.26 of ramipril to placebo in HOPE. The number 1.13 would indicate that in the current trial, telmisartan would have at least one-half the effect of ramipril. From the study results applied by results from HOPE, telmisartan was found to preserve about 95% of the benefits of ramipril over placebo for the primary outcome and 105% of the benefits from the combined outcomes of deaths from CV causes, MI, or stroke. The use of inferiority standards instead of placebo in assessment of drug efficacy has been of concern,13 primarily in that by the statistically accepted criteria, noninferiority is established if as little as half the benefit of the comparison drug is preserved.14 In terms of side effects in this trial, hypotension and renal impairment were found in the combination group compared with ramipril alone, and despite lower blood pressures, CV events were not decreased with the drug combination. The lack of a benefit of further blood pressure lowering requires further investigation. The adverse effects of combined ACEI and ARB therapy were also found in the Valsartan in Acute Myocardial Infarction (VALIANT) trial,15 which evaluated a similar high-risk population, although with heart failure or left ventricular dysfunction. The results of this study suggest that an ARB can be substituted for an ACEI in this population with similar benefits; therefore, the use of an ARB would depend on cost considerations and evident side effects of the ACEI, primarily cough.