This article is more than 5 years old. Information may no longer be current.
Commentary: An in-depth look at the IMPROVE-IT trial
Click Here to Manage Email Alerts
by Seth S. Martin, MD; and Roger S. Blumenthal, MD
When the Improved Reduction of Outcomes: Vytorin Efficacy International Trial, more commonly known as IMPROVE-IT, began in 2005, the time was much different than today. Consumers did not have iPhones. Taylor Swift hadn’t yet released her first album. Seth Martin had just started medical school. Roger Blumenthal’s son, Ross, was only 6 years old and not yet a New England Patriots fan. It may give perspective for you too to think about what your life was like back then — and what you know and perhaps have forgotten.
Seth S. Martin
Roger S. Blumenthal
Nearly a decade later, we live in a different world. A world in which there is disagreement across oceans, and even within our own borders, about treating patients to LDL cholesterol goals of <100 mg/dL or <70 mg/dL. While the American College of Cardiology and American Heart Association have continued to recognize the central role of LDL in atherosclerosis, some in our community have questioned it.
Confusing evidence emerged during the past decade from serial carotid intima-media thickness surrogate endpoint trials using ezetimibe (Zetia, Merck), such as ARBITER and ENHANCE. Although this drug continued to generate billions of dollars in annual sales, we saw a series of disappointing lipid trials with other agents focused on raising HDL cholesterol. These trials casted doubt on whether lipid-modifying therapies beyond statins could show benefit. In contrast to those other trials, while the world was changing, the IMPROVE-IT trial stayed with a tried-and-true approach: LDL lowering.
To lower LDL, ezetimibe inhibits Niemann-Pick C1-Like 1 (NPC1L1) protein, which is predominantly located in the epithelial brush border of the gastrointestinal tract. Inhibition of NPC1L1 decreases cholesterol absorption and when added to statin therapy further lowers LDL by approximately 20%. For comparison, a doubling in a statin dose generally provides about 6% more LDL lowering.
Peter P. Toth
Recently, Peter P. Toth, MD, PhD, JoAnne M. Foody, MD, and colleagues reported better LDL control in high-risk patients with CHD who switched to ezetimibe/simvastatin in combination (Vytorin, Merck) compared with statin up-titration. In the retrospective, observational study of 27,919 patients, 81% of those who switched to combination therapy attained an LDL <100 mg/dL vs. 72% assigned statin up-titration. For LDL <70 mg/dL, the respective proportions were 35% and 23%.
It still needed to be proven that better LDL control through ezetimibe treatment improved clinical outcomes.
IMPROVE-IT trial design
IMPROVE-IT was a randomized, multicenter, active-control, double blind trial that set out to formally test the clinical benefit of ezetimibe and simvastatin in combination (10/40 mg/day single table) compared with simvastatin monotherapy (40 mg/day; Zocor, Merck). The primary endpoint was first occurrence of CV death, nonfatal MI, rehospitalization for unstable angina, coronary revascularization or stroke, with adjudication by a blinded clinical events committee. The trial was event-driven, designed to continue until a minimum of 5,250 participants sustained a primary endpoint event, and each participant was followed for a minimum of 2.5 years.
At 1,158 sites across 39 countries, recruitment spanned from Oct. 26, 2005, through July 8, 2010. The trial recruited 18,144 high-risk patients with ACS who were clinically stable within 10 days following hospital admission. Participants were aged 50 years and older with at least one of the following high-risk features: new ST-segment change, elevated troponin, diabetes, prior MI, peripheral arterial disease, cerebrovascular disease, prior CABG (>3 years) or multivessel CAD. Major exclusions included CABG for treatment of qualifying ACS, current statin treatment of intensity exceeding simvastatin 40 mg/day, creatinine clearance <30 mL/min or active liver disease.
Patients not taking a statin had to have a baseline LDL level of 50 mg/dL to 125 mg/dL and those already taking lipid-lowering therapy were required to have a baseline LDL of 50 mg/dL to 100 mg/dL. The rationale for not enrolling patients with higher baseline LDL levels was so that the simvastatin-monotherapy group could attain what were the guideline-recommended LDL treatment goals at the time. In the original protocol, if the LDL response was inadequate in either trial arm, simvastatin could be increased to 80 mg/day. However, a protocol amendment was enacted in June 2011 to recommend against new increases of simvastatin dose to 80 mg/day in conjunction with an FDA warning regarding the higher risk for myopathy with that regimen.
At baseline, the patients’ mean age was 64 years and three-quarters were men. Twenty-seven percent of patients had diabetes and 21% had a prior history of MI. With regard to the index hospitalization, the distribution of ACS-type was 29% for STEMI, 47% non-STEMI and 24% unstable angina. ACS management included cardiac catheterization in 88% and PCI in 70%. About one-third of patients were taking prior lipid-lowering therapy. The median baseline LDL level was 95 mg/dL in both groups.
During the trial, patients had follow-up visits on day 30 and every 4 months thereafter. Up-titration to simvastatin 80 mg/day occurred in 6% of the ezetimibe/simvastatin group and 27% of simvastatin-monotherapy group. The proportion of patients in whom there was premature study drug discontinuation was remarkably high, but similar in the two groups: 42%. Median follow-up was 6 years; the completeness of follow-up was excellent, with only 0.1% of patients lost to follow-up per year and more than 90% of patients followed up for the primary endpoint and survival. After the prespecified follow-up and event requirements were fulfilled, the database was locked just a few weeks before presentation at the American Heart Association Scientific Sessions.
As expected, predictions ahead of time were mixed. Swayed perhaps by the CIMT data, other disappointing recent trials or low LDL in the both groups, many of our colleagues were anticipating a neutral trial result. Probably fewer folks were concerned that possible harms from ezetimibe would be substantial enough to produce net harm since the study had not been stopped early. Others, however, were more optimistic that the drug would be safe and efficacious, and would show what it set out to show.
New data
Christopher P. Cannon
At noon on Nov. 17 during the AHA Scientific Sessions in Chicago, a packed room in the main tent listened to a presentation of the IMPROVE-IT results by Christopher P. Cannon, MD. The intention-to-treat analysis of IMPROVE-IT showed that the primary endpoint occurred in 32.7% of patients assigned ezetimibe/simvastatin vs. 34.7% assigned simvastatin monotherapy (P=.016). Thus, there was an absolute risk difference of 2%, yielding a number needed to treat (NNT) of 50 during the trial duration of 6 years. That means that 50 patients would need to be treated with ezetimibe to prevent one CVD event over 6 years. The relative risk reduction was 6.4% (95% CI, 1.2-11.3).
All of the secondary composite outcomes were statistically significant, and analysis of individual endpoints showed that the findings were driven by reduction of MI and ischemic stroke — not by repeat revascularization. Thus, ezetimibe protected against the “hard” atherosclerotic CVD outcomes that are the focus of the latest ACC/AHA guidelines, although we do not yet know whether the decrease was predominantly in spontaneous MIs or perioprocedural MIs.
Impressive findings in patients with diabetes
Considering the major prespecified subgroups, statistically significant effect modification was only seen in patients with diabetes compared with those without diabetes, with a greater protective effect in the diabetes group. In patients with diabetes, the primary outcome occurred in 45.5% of those assigned simvastatin monotherapy vs. 40% of those assigned ezetimibe/simvastatin, yielding a 5.5% absolute risk reduction and NNT of 18.
The safety data were not only reassuring to the many people who have been taking and prescribing ezetimibe, but they looked pristine. There were no differences between the trial groups in incidence of cancer, muscle- or gallbladder-related events. An upside of IMPROVE-IT taking longer than expected was that we obtained longer-term safety data, in fact longer than for many of the statin trials.
The IMPROVE-IT investigators and participants should be congratulated on a truly monumental effort that is undoubtedly a landmark trial. This is the first randomized controlled trial to show a conclusive benefit of a nonstatin agent on a background of statin therapy. This is especially notable since the first statin was approved by the FDA in 1987 (lovastatin). The exceptional IMPROVE-IT trial team persevered during the past decade to bring the world rigorous scientific evidence we needed to feel confident that ezetimibe benefits our patients.
Impact of results on recent guidelines
Neil J. Stone
As noted by the trial discussant, Neil J. Stone, MD, these findings fit within the framework of the ACC/AHA 2013 Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults. The guideline states: “Clinicians treating high-risk patients who have a less-than-anticipated response to statins, who are unable to tolerate a less-than-recommended intensity of a statin or who are completely statin intolerant may consider the addition of a nonstatin cholesterol-lowering therapy.” The guideline further advises: “In this situation, this guideline recommends clinicians preferentially prescribe drugs that have been shown in [randomized controlled trials] to provide [atherosclerotic CVD] risk-reduction benefits that outweigh the potential for adverse effects drug-drug interaction, and patient preferences.”
After meeting its atherosclerotic CVD risk-reduction endpoint in a high-quality randomized controlled trial, ezetimibe can now be considered a drug that deserves preferential use, especially in those individuals who cannot tolerate daily doses of a high-potency statin.
Effect of participants going off treatment regimen
IMPROVE-IT achieved its positive result despite 42% of patients in both groups going off therapy. Why did so many patients stop the study drug? The longer trial duration created a challenge, but, beyond that, concerns arose regarding efficacy based on subclinical carotid atherosclerosis endpoint trials such as ARBITER and ENHANCE, and there was concern about a cancer signal in the SEAS trial. Then, in 2011, the FDA released a simvastatin label change regarding the higher risk for myopathy with the 80-mg daily dose, requiring down-titration of some patients to 40 mg/day.
Many patients going off treatment could dilute treatment effects and erode statistical power. So a prespecified, on-treatment analysis was performed to address this concern. The investigators excluded participants who never took study drug, and in others they conservatively censored at a minimum of 30 days after the last dose or last complete endpoint ascertainment of clinical events. There were 60,298 patient-years of follow-up compared with 80,286 patient-years in the intention-to-treat analysis.
The mean age of patients who completed the trial on study drug was 62.8 years (n=10,573) and 63.8 years in those who did not complete it on study drug (n=7,133). There were 22% and 28% of women in the two groups, respectively, and 38% and 32%, respectively, had been on prior lipid-lowering therapy. In both groups, 27% of patients had diabetes, 88% had undergone cardiac catheterization and baseline LDL levels were remarkably similar at about 95 mg/dL. The time-averaged LDL levels and difference between the groups was similar to the intention-to-treat analysis, which makes sense since LDL levels were only monitored in those on treatment.
Analysis of the primary endpoint demonstrated a 7.6% reduction (95% CI, 1.7-13.2) in the primary endpoint. The endpoint occurred in 29.8% of patients assigned ezetimibe/simvastatin vs. 32.4% assigned simvastatin-monotherapy arm, yielding an absolute risk difference of 2.6% and NNT of 38. All of the secondary composite outcomes showed point estimates that were even more favorable than the intention-to-treat analyses.
Importantly, data matched up with the known relative risk reduction in occlusive CV events of 22% per 1 mmol/L (39 mg/dL) lowering in LDL from the Cholesterol Treatment Trialists (CTT). The confidence intervals around the point estimates for benefit with major vascular events, major coronary events, nonfatal MI, coronary death, coronary revascularization, stroke or ischemic stroke all overlapped the known 22% relative benefit.
We need to synthesize IMPROVE-IT with prior literature to best understand the findings. As the late CV epidemiologist Trudy Bush, PhD, MHS, said, “Every time we do a study or write a new guideline, it needs to be evaluated in the context of all the studies that came before. New findings and guidelines bring us closer to the truth but we may never reach it entirely.” Data from the SHARP investigators in 2009 further reinforce the confidence in this result. Although the control was placebo, not statin monotherapy, benefit was shown in SHARP in a chronic kidney disease population with simvastatin and ezetimibe in combination. There, the proportional reduction in CVD events was proportional to the LDL reduction expected based on prior CTT data of statin-only studies.
Pleiotropic effects
So what does this mean for the proposed pleiotropic effects of statins?
On the one hand, it suggests that those pleiotropic effects may be largely lipid dependent — that is, they are mediated through lipid lowering. It is known that there is an important pathophysiologic intersection between lipids and inflammatory pathways. Although some prior studies may seem somewhat contradictory in suggesting faster reduction in risk than we would expect from lipid lowering alone, the lipid-mediated effects translated through inflammatory pathways to stabilize plaque may occur quickly, while atherosclerosis slowing or even regression is the longer-term effect.
Editor’s note: This is part one of a two-part commentary. Stay tuned for the conclusion, which will explore the relevance of the IMPROVE-IT trial to clinical practice, as well as the varying reactions to the results from experts in the field. Share your thoughts on IMPROVE-IT in the comments section below.
For more information:
Cannon CP. LBCT.02: Anti-Lipid Therapy and Prevention of CAD. Presented at: American Heart Association Scientific Sessions; Nov. 15-19, 2014; Chicago.
Toth PP. J Clin Lipidol. 2014;8:107-116.
Seth S. Martin, MD, is a clinical fellow at The Johns Hopkins Ciccarone Center for the Prevention of Heart Disease and a member of the Cardiology Today Fellows Advisory Board. Roger S. Blumenthal, MD, is director of the Ciccarone Center for the Prevention of Heart Disease and is the CHD and Prevention Section Editor of Cardiology Today.
Disclosure: The authors report no relevant financial disclosures.