High-Intensity Statin Treatment for Coronary Heart Disease

In this issue of THE JOURNAL, Nissen and colleagues¹ report the findings of the Reversal of Atherosclerosis with Aggressive Lipid Lowering (REVERSAL) study, a clinical trial comparing the effects of 2 statins, pravastatin (40-mg dose) and atorvastatin (80-mg dose). The REVERSAL study is a first in 2 respects. First, it showed that a statin regimen using the most efficacious drug (atorvastatin) and dose available for lowering low-density lipoprotein cholesterol (LDL-C) when the trial began improved coronary atherosclerosis more than pravastatin, which represented a statin regimen that had been fully established to safely reduce cardiovascular disease and mortality.^{2 - 6} Second, the advantage of atorvastatin was demonstrated using an intravascular ultrasound examination of the coronary arteries, a technique that quantifies plaque by visualizing the coronary arterial intima. While this study is an important achievement, it is equally important to determine what it adds to current knowledge of lipid therapy in atherosclerotic cardiovascular disease and how it should affect clinical practice.

All told, nearly 70 000 patients have participated in large placebo-controlled clinical trials of statins of at least 3 years in duration. The study populations have been diverse, including those without clinical evidence of coronary artery disease and those with various presentations of coronary artery disease prior to statin treatment. Risk reduction has been similar (range, 20%-30%) across most of these trials in mortality and in virtually every atherosclerotic cardiovascular event that has been evaluated. All of the statins that have been studied in these trials, namely pravastatin (40-mg dose),^{2 - 6} simvastatin (20- to 40-mg dose),^{7 - 8} lovastatin (40-mg dose),⁹ fluvastatin (80-mg dose),¹⁰ and atorvastatin (10-mg dose),¹¹ have produced risk reduction in serious coronary vascular end points. Although there is some variation between the trials in the results of the population subgroups, considering the results across all trials, it now seems true that statins produce similar relative risk reduction across a wide variety of patient types and clinical presentations. Therefore, health benefit (ie, absolute risk reduction), is principally dependent on the underlying level of risk. This is explicitly recognized by national guidelines for lipid treatment.^{12 - 13}

The success of statin therapy used at initial or moderate doses leads to an important next step to evaluate whether a higher dose is better. The results of the REVERSAL trial¹ extend similar findings in carotid atherosclerosis to coronary atherosclerosis. In one trial, carotid artery intima-media thickness increased more with 40 mg of simvastatin than with 80 mg of atorvastatin; the difference became evident at 1 year and was highly statistically significant by 2 years.¹⁴ A subsequent trial reported similar results—a borderline significant difference between atorvastatin and pravastatin at 1 year.¹⁵ Thus, there is a consistency developing that maximal statin therapy produces a superior result on measurements of atherosclerosis in clinically important sites.

A potential limitation of REVERSAL is the loss of participants to follow-up. Twenty-three percent of the total participants randomized were lost to follow-up, which is similar to some trials that used coronary arteriography for the outcome,^{16 - 17} although others had lower rates of between 9% and 10%.^{18 - 19} Although the primary analysis of REVERSAL included just the participants who completed the treatment and had intravascular ultrasound at the end, the authors also presented the results with all randomized patients, using baseline data when outcome data could not be obtained, as recommended.²⁰ The results of the 2 approaches are consistent with each other, strengthening the conclusion that when compared with a 40-mg dose of pravastatin, an 80-mg dose of atorvastatin favorably affected coronary atherosclerosis.

Why should a complete analysis of randomized patients, as presented by Nissen et al, be so important? The randomized controlled trial is a simple and elegant solution to problems inherent in uncontrolled trials and in observational studies of populations. Randomization creates a control group that at the start of the study is identical to the experimental group. This, and complete ascertainment of outcome in all patients ensure that selection bias and other confounding factors do not affect the results. When there are substantial losses to follow-up, the accuracy of the result can be biased because the groups that were identical at the start become different to an unknown extent.²⁰ In this way, a randomized trial takes on the limitations of observational studies. Clinical investigators, understandably experiencing frustration that the treatment effect may be diluted by including all randomized patients, are at times tempted to present the data of only the patients who completed the study. Trials should have sufficient sample size to maintain adequate statistical power with all randomized patients—essential to preserve the validity of the randomized trial, which is the criterion standard for evidence-based medicine.

The authors defended their selection of 40 mg of pravastatin and 80 mg of atorvastatin as a fair test of their hypothesis on intensity of statin therapy, and their reasoning is appropriate. However, the reductions in the pravastatin group in LDL-C of 25% and 5% in C-reactive protein are less than those established by larger trials with high adherence in which LDL-C was reduced by 32% over 5 years² and C-reactive protein by between 13% and 14% at 6 months²¹ and 17% and 18% at 5 years.²² In REVERSAL, adherence was not reported, and it is possible that diminished adherence could have affected the response in the pravastatin group.

Researchers studying the effects of lipid therapy have bemoaned the limited information provided by contrast arteriography that quantifies only the lumen. Clinical coronary disease has gradual and abrupt presentations. Gradual is often associated with luminal narrowing of more than 75% and chronic angina and abrupt is associated with less luminal narrowing but lipid-rich "vulnerable" plaque with active inflammation that causes acute coronary syndromes.²³ Intravascular ultrasound can identify morphological characteristics of vulnerable plaque, but not of inflammatory plaque activity, which requires other investigational "activity imaging" techniques.²³ Nonetheless, intravascular ultrasound is an exciting step forward and will probably continue to be refined for use in clinical trials to identify potentially vulnerable plaques and quantify changes due to therapy. New lipid therapies and other antiatherosclerosis drugs could be evaluated by intravascular ultrasound for improving specifically these most worrisome lesions, and the results would have immediate clinical meaning especially if coupled with assessment of plaque inflammation. However, intravascular ultrasound may be limited in use by restrictive vessel characteristics or by a high percentage of patients who decline intravascular ultrasound or who do not complete follow-up studies.

The critical remaining question is to what extent the additional improvement in atherosclerosis measurements caused by maximal compared with standard statin therapy translates into more clinical benefit, such as prevention of unstable angina, myocardial infarction, or coronary death. It is reasonable to expect benefit, although it is difficult to predict how much from the REVERSAL study in view of the newness of intravascular ultrasound. The authors' assertion made in the article that "intensive treatment halted progression of atherosclerosis"¹ is somewhat troubling. This provokes an expectation for a dramatic reduction in coronary events when statin therapy is intensified. The likelihood of this is belied by the small difference in atheroma volume between the groups—1%—and by the progression in atheroma in many patients in the intensive group, as noted by the authors. A more conservative and perhaps realistic projection can be made from the effects of the 2 statins on LDL-C. In the Heart Protection Study,⁸ an approximately 40-mg/dL (1-mmol/L) reduction in LDL-C corresponded with a 25% reduction in coronary event incidence, and this relationship also fits the results of most other statin trials.^{2 - 11} In the REVERSAL trial, the difference in LDL-C between the treatment groups was 30 mg/dL (0.78 mmol/L) (80 mg/dL [2.07 mmol/L] for atorvastatin vs 110 mg/dL [2.85 mmol/L] for pravastatin), projecting a 19% additional relative risk reduction from intensive compared with conventional statin therapy.

Could this "cholesterol-centric" projection of clinical benefit of an 80-mg dose of atorvastatin compared with conventional statin therapy be amplified by nonlipid-mediated effects? Perhaps, but in this projection, a 40-mg/dL (1.04-mmol/L) reduction in LDL-C is accompanied by reductions in other risk factors including other atherogenic lipoproteins²⁴ and inflammatory mediators.^{21 - 22 ,25 - 26} Together they correspond to the 25% relative reduction in coronary incidence. How much reduction in coronary heart disease during statin therapy is due to LDL-C lowering or to other effects is difficult to determine, although the LDL-C theory still dominates, justifiably so because of many levels of strong evidence.¹² A good case has been made for statins reducing stroke by enhancement of nitric oxide production and endothelial function,²⁷ although this mechanism lacks direct proof by clinical trial. Recently, new important data on this issue was reported in a statin trial on symptomatic nonischemic dilated cardiomyopathy.²⁸ Treatment for only 14 weeks with a low dose of simvastatin (5-10 mg) compared with placebo improved left ventricular ejection fraction and functional class, while lowering substantially several inflammatory mediators, tumor necrosis factor α, IL-6, and brain natriuretic peptide, and improving endothelial function. It seems highly unlikely that such rapid clinical improvement in nonischemic cardiomyopathy could have been caused by the small (16%) reduction in LDL-C. Would these non-LDL-C–mediated effects track proportionately with LDL-C reduction with higher statin doses? Relatively equipotent doses of statins (eg, 10 mg of atorvastatin, 20 mg of simvastatin, or 40 mg of pravastatin) have similar effects on C-reactive protein,²⁶ and there is a dose-related effect.^{1 ,25} Thus, nonlipid effects of statins probably contribute to the cardiovascular benefits as a function of intensity of therapy.

Is this reasoning compelling enough to change clinical practice without the need for a clinical outcome trial, as has been urged recently?²⁹ The results of the 3 atherosclerosis studies with ultrasound^{1 ,14 - 15} add support. However, a policy to use the highest statin doses as standard therapy has implications for cost and potential adverse effects.³⁰ The vast majority of clinical and research experience is with conventional statin doses, which have provided a deservedly positive view of statin therapy. To put maximal statin therapy, such as with 80 mg of atorvastatin or 40 mg of rosuvastatin, on the same footing of clinical confidence requires results from large long-term clinical trials. Fortunately, several large trials comparing intensive with conventional statin therapy are near completion and results are expected soon.

Until the results of these studies are available, it is prudent to treat any high-risk patient, as defined by national guidelines,^{12 - 13} with a statin at an intensity appropriate to achieve the recommended goals for LDL-C. In addition, with this focus on LDL-C reduction, clinicians must not lose sight of the need to manage other established cardiovascular disease risk factors including hypertension and atherogenic dyslipidemia, manifested by low levels of high-density lipoprotein cholesterol and high levels of triglycerides.¹² At least as important, this concentration on drug treatment should not deflect attention from diet and lifestyle interventions that have the potential even with moderate improvements to reduce cardiovascular disease incidence by between 75% and 80%.³¹