NCHR Testimony at FDA on IMProved Reduction of Outcomes: Vytorin Efficacy International Trial (IMPROVE-IT)

Thank you for the opportunity to speak today. My name is Dr. Tracy Rupp. I was previously a clinical pharmacist at Duke University Medical Center and am now the Director of Public Health Policy Initiatives at the National Center for Health Research. Our research center analyzes scientific and medical data and provides objective health information to patients, providers and policy makers. We do not accept funding from the drug or medical device industry and I have no conflicts of interest.

Clinical Significance for Patients is Not Established

IMPROVE-IT is a good illustration of the difference between statistical significance and clinical significance. In this trial, the addition of ezetimibe to simvastatin in patients with a recent acute coronary event was associated with a statistically significant reduction in the composite primary end point of death from cardiovascular (CV) disease, a major coronary event, or nonfatal stroke after 7 years. However, the result was not robust with an absolute risk reduction of only 2% for the composite primary outcome. The statistically significant outcome was primarily driven by reductions in heart attack and stroke, and was not related to a reduced risk of death. Is a 2% reduction in risk for heart attack and stroke clinically meaningful?

Since the number needed to treat to prevent one CV event is 50 patients who are at high risk for 7 years, there is about a 1 in 350 chance that an individual very high risk patient will be prevented from having a attack or stroke in any given year as a result of taking Vytorin. In other words, a patient’s chances of landing in the hospital are reduced for 1 out of 350 patients taking Vytorin for a full year, and the chance of dying is not reduced at all for any of those 350 patients. As we heard today, the chances of benefiting from Vytorin are even less for patients without diabetes and those less than age 75.

Of significant concern is the fact that the proposed indication states that adding ezetimibe to statin therapy will reduce the risk of CV events. This will probably cause many providers to think that risk of death is reduced when this is not the case.

Of course, it is better to not have a heart attack or stroke than to have one, but the small absolute risk reduction found in IMPROVE-IT was one of the smallest effects ever observed in statin trials. In fact, this study achieved statistical significance simply because of the enormous sample size. In order to show such a small difference, the investigators had to study more than 18,000 patients with a very high baseline risk (35% for the primary outcome and 15% for death). In fact, the protocol had to be amended to increase the sample size and extend the length of the trial.  If they hadn’t made that post hoc change, the tiny absolute benefit of Vytorin would not have been statistically significant. And it is possible that the results would not have been statistically significant if we had the data from those participants with missing follow-up time. As we heard today, more patients in the Vytorin group were missing follow-up time in the first year after randomization, which is when a CV event would have been most likely to occur. So we do not know if the results would still be statistically significant if we had this data.

It’s important to remember that, before the trial began, the investigators decided an absolute risk reduction of 3% was needed to show that Vytorin was beneficial compared to simvastatin alone.  But, after it became obvious that Vytorin wasn’t that effective, the investigators changed the criteria to a 2% improvement for patients taking the drug for 7 years.  The investigators had not initially believed that such a small difference was clinically meaningful.  One has to assume that they changed the study design to gain FDA approval by showing a statistical difference, not by making a difference in patients’ lives.

Benefit to the Healthcare System is Questionable

The benefit is minimal for Vytorin, especially compared to high potency statins such as atorvastatin and rosuvastatin.  And since Vytorin is already on the market, we know that the cost of preventing one event is expensive — just under $900,000. The cost of preventing one event is much higher when Vytorin is taken by patients at low or moderate risk instead of those at high risk.  We know that Vytorin will be prescribed to low-risk outpatient populations for primary prevention, as has occurred with other statins and with Vytorin itself. If Vytorin is used in lower risk patients, the number needed to treat to prevent one heart attack or stroke, as well as the cost, will increase exponentially – money that would be better spent on more effective treatments.

FDA is not supposed to take cost into account, but they should take treatment options into account.  And this panel should consider whether Vytorin represents the kind of innovation we want for our health care system, or the kind of costly mediocrity that we need to avoid?

Beneficial Effects of Adding Ezetimibe Not Proven to be Due to LDL Lowering

The results of IMPROVE-IT have led some to proclaim that use of LDL cholesterol as a surrogate endpoint for cardiovascular events has been proven. Although the patients in the simvastatin plus ezetimibe group did experience lower LDL cholesterol levels and a slightly lower rate of the primary endpoint, the trial does not prove that the effect was mediated by the lowering of LDL levels.

Some studies have suggested that ezetimibe may have beneficial effects unrelated to LDL lowering that are responsible for its modest decrease in coronary events. Some of these non-LDL effects may have been responsible for the small beneficial effect of ezetimibe seen in IMPROVE-IT, since levels of high-sensitivity C-reactive protein were significantly lower in the simvastatin plus ezetimibe group than in the simvastatin-only group. Similarly, since patients with diabetes are particularly likely to benefit from reductions in triglycerides, the greater decrease in triglyceride levels in the simvastatin plus ezetimibe group may explain why patients with diabetes experienced more improvement than patients without diabetes with the combination drug.

We also cannot extrapolate any benefit from simvastatin plus ezetimibe to high-potency statins. Some physicians might assume that adding Zetia to a high-potency statin would be a good idea, but in the absence of any evidence, we don’t know what the risks or benefits of such a combination would be.

Side Effects of Statins

Lastly, we all know that statins have painful side effects that are not tolerable to some patients. Since Vytorin contains a statin, it does not eliminate the risk of developing those painful side effects.


In summary, the very small chance of any potential benefit and the high cost of preventing one CV event lead us to conclude that ezetimibe is not the cure most patients desire. We recommend revising the drug label to limit the indication for Vytorin and Zetia to patients with diabetes and patients older than age 75. The benefits in this group should be confirmed with a study specifically designed for this purpose. The label should also make it clear that an effect on mortality has not been found.

IMPROVE-IT also cannot tell us anything about the validity of LDL cholesterol levels as a surrogate endpoint for CV events, or whether patients who are already on high-potency statins will benefit from the addition of ezetimibe.

Thank you for the opportunity to comment today and for consideration of our views.

FDA Advisory Committee vote results: 10 no, 5 yes

Update February 16, 2016:
FDA denied Merck’s application to expand the indication of Vytorin and Zetia. The proposed new indication was for the reduction of cardiovascular risk in people with heart disease. The drugs are approved for the reduction of LDL cholesterol in people with hyperlipidemia.