The effects of ezetimibe vs. niacin on measures of mean carotid IMT in HALTS trial

American Heart Association Scientific Sessions 2009

In today’s online issue of the New England Journal of Medicine, Taylor et al described the results of an elegant study on the effects of ezetimibe vs. extended-release niacin on measures of mean common carotid intimal media thickness. The HALTS trial included adults treated with statin therapy who had documented atherosclerotic vascular disease, diabetes, a high 10-year Framingham risk estimate or high coronary calcification score with baseline LDL <100 mg/dL and HDL <50 mg/dL (men) or <55 mg/dL (women).

The patients were randomly assigned to two open-label adjunctive cholesterol-modifying medications (niacin vs. ezetimibe) and followed for 14 months. Because the primary endpoint of reduction in carotid intimal media thickness (IMT) had been reached in a prespecified interim analysis — though 363 patients were initially enrolled in the study — only 208 had complete data at the time the trial was stopped.

The researchers found that the addition of 2 g of extended-release niacin to statin therapy reduced common carotid IMT by an average of 0.014 mm during a period of 14 months. The addition of ezetimibe did not result in a net change in carotid IMT. These data were similar to the results from a recently published study from an Oxford research group using carotid MRI that showed that the addition of niacin resulted in a net improvement of 2 mm in plaque size.

Statin trials

Results of multiple randomized controlled trials have shown the importance of statin use in CV risk reduction with an attenuation of that risk of up to 45%. Efforts to address residual risk have concentrated on pharmacologic agents that lead to further reductions in LDL and triglyceride levels or increases in HDL. The addition of niacin to statin therapy in prior small studies indeed led to large relative risk reductions in CV events and to either stabilization or modest regression of coronary or carotid atherosclerosis.

Results from the HATS and FATS trials, in which researchers studied fewer than 100 patients assigned to niacin, showed a modest reduction in the number of cardiac events in those treated with statin plus niacin as compared with placebo. Underlying this benefit were promising results revealing both biochemical profile improvement as well as decreased coronary artery stenosis on angiography. The researchers of these two trials however did not compare statin alone to statin plus niacin.

Addressing treatment intensification with statin therapy, ARBITER-2 was a small, double blind, randomized placebo-controlled trial conducted to compare niacin with placebo in patients already assigned to stable statin therapy. In this study, clinical endpoints were few and not significantly different between groups. As a matter of fact, IMT increased by 0.014 mm in the 1,000-mg niacin group; this was the mean improvement in the HALTS trial with 2 g of niacin. Thus, ezetimibe seemingly had better results in HALTS than 1,000 mg of niacin in HALTS.

Prior lipid-lowering strategies demonstrated the requirement for both surrogate and hard endpoints to get a full picture of the pharmacologic effect of the study agent. Despite substantial increases in HDL levels conferred by treatment with the CETP inhibitor torcetrapib, this agent did not translate to any benefit, but rather, increased harm was shown in both surrogate and clinical outcomes in the ILLUMINATE and RADIANCE trials. Unpredictably, elevated BP related to increases in aldosterone may have negated the favorable changes in lipids.

In Arbiter 6-HALTS, the endpoint of carotid IMT was chosen as a marker for atherosclerosis. B-mode ultrasound imaging is a sensitive and reproducible method for assessment of carotid IMT, and in observational studies, the thickness of the tunica intima and media predicts future CVD events. Supporting the findings of the Arbiter 6-HALTS trial, the researchers from the recent Oxford study assessing carotid atherosclerosis by MRI found a very modest reduction in atherosclerotic plaque area in niacin-treated patients. In both of these trials, however, it was not evident whether the changes seen in carotid IMT, the surrogate marker for CVD, were clinically significant or relevant.

Carotid IMT is a distinct form of coronary atherosclerosis and may not fully mirror CVD events as a predictor of risk or as a surrogate endpoint. Coronary artery calcification is generally more strongly predictive of CVD events than carotid IMT. We await the results of ongoing trials conducted to conclusively determine if a change in carotid IMT predicts events in subsets of symptomatic and asymptomatic adults.

Affect on practice

The implications of Arbiter 6-HALTS have the potential to change medical practice. However, there are a number of areas of concern that need to be addressed. For one, the decision to prematurely terminate the trial was unfortunate. There are no formal stopping boundaries for carotid IMT trials (a surrogate marker). By cutting the trial short, more than 100 patients did not undergo 14-month carotid IMT measurements, the primary endpoint. A larger sample size may have either strengthened the provocative results for the major adverse CV events or, alternatively, reduced any evidence of meaningful clinical differences.

In addition to the large number enrolled who did not get to complete the trial, about 10% withdrew because of adverse effects. There was a differential dropout in the niacin and ezetimibe treatment groups. If less-compliant individuals could not handle the mild- to moderate-nuisance adverse effects of niacin, they might also be less compliant with their other secondary prevention measures. By removing these potentially less-compliant patients and only using the “completers,” the researchers may have selected a more compliant, healthier cohort of patients left in the niacin group. This could favor the niacin treatment strategy and be a potential source of bias.

It is surprising that adherence to study medication by pill count was so high in the niacin cohort. The final daily achieved dose of 2 g was reached in 75% of patients, with 3% assigned to 1,500 mg per day. The remaining 22% were assigned to the more commonly tolerated dosages of 500 mg or 1,000 mg. In real-world clinical practice, the strong flushing phenomenon associated with cutaneous vasodilation limits most patients to the 1,000 mg dose.

Statin therapy reduces CV morbidity but it does not slow the progression of another surrogate marker, coronary artery calcification. The paradoxical (nonprespecified) finding of a greater reduction of LDL with ezetimibe associated with carotid IMT progression was highlighted prominently in the abstract; however, post-hoc analyses should not be more than exploratory.

Of note, the baseline carotid IMT values in the prior ENHANCE trial were much lower than in HALTS. HALTS results did show that with higher baseline IMT values regression is indeed possible. However, the number of patients in the ezetimibe arm may be too small to define whether the change in LDL during 14 months is sufficient to reduce IMT. Perhaps a longer follow-up would have led to a modest reduction in carotid IMT.

The researchers’ discussion on the potential negative consequences of ezetimibe may be misleading. It is based on in vitro assessments that likely bear little relevance to the human body. It should be remembered that rosuvastatin (Crestor, AstraZeneca) did not reduce IMT in a large IMT trial, but it certainly reduced events in a subsequent large clinical trial.

We still do not know if more aggressive LDL reduction with a more potent statin as monotherapy would be more effective than adding a second lipid-modifying agent to a less-potent statin. For most high-risk patients, reduction of >50% of LDL is needed to reach an LDL of <70 mg/dL and a non-HDL <100 mg/dL. All of the patients in HALTS would have an optional LDL goal of <70 mg/dL, but the mean LDL in this trial was 82 mg/dL. There was no statin monotherapy that aimed for an on-treatment LDL of <70 mg/dL. Thus, we do not know if an LDL <70 mg/dL is better than simply an LDL <100 mg/dL reached with a weaker statin plus another agent.

Clearly, the evidence supports the initial use of a statin titration as needed to reach LDL. An evidence-based agent should be chosen when selecting a second agent after a potent statin is used at maximal tolerated dose. If a patient cannot reach their lipid goals with a potent statin alone, the choices remain a fibrate; niacin; a bile acid sequestrant; omega-3 fatty acid supplements; or ezetimibe as add-on therapy.

Unfortunately, the premature termination of this HALTS trial precludes us from conclusively making niacin the most evidence-based agent. Nevertheless, it appears preferable at this stage to add niacin as a second agent and reserve ezetimibe for third- or fourth-line status in most cases. Two trials likely to clarify the value of augmenting statin therapy with niacin are the AIM-High and HPS2-THRIVE studies, which should be reported in the next two to three years.

The use of surrogate markers (eg, changes in IMT) as reliable indicators of future CVD events still needs to be viewed with some degree of healthy skepticism until the above three definitive trials are reported. However, the secondary choices for optimizing cholesterol-lowering therapy should not overshadow the importance of optimizing antiplatelet therapy, BP, dietary and exercise habits.

Gabriel Brooks, MD, and Rhondalyn McLean, MD, are both from Johns Hopkins Ciccarone Center for the Prevention of Heart Disease.

For more information:

• Lee JMS. J Am Coll Cardiol. 2009;54:1787-1794.

HALTS is a well-done study whose results support the finding from a recent Oxford carotid MRI study that niacin should be considered the preferred second lipid-lowering agent after a potent statin is used. However, mean benefits of 0.014 mm in the former and 2² mm in the latter are not definitive or sufficient to change national lipid-lowering guidelines. Fortunately, we have two large pivotal clinical trials using niacin on top of statin therapy vs. statin therapy alone (HPS2-THRIVE and AIM-HIGH) that will be finished in two to three years. We also have a crucial study with ezetimibe on top of statin therapy vs. statin alone in patients with acute coronary syndrome (IMPROVE-IT). Thus, we will know by 2012 if the provocative findings from HALTS will coincide with better clinical outcomes. In the words of the legendary actress and author, Patty Duke, “The truth should out. It almost always does.”

– Roger S. Blumenthal, MD

Director, Johns Hopkins Ciccarone Center for the Prevention of Heart Disease