Ticagrelor: The Story Behind the Approval
The US Food and Drug Administration (FDA) has just approved ticagrelor, a new oral antiplatelet agent, for prevention of thrombotic events in patients with acute coronary syndromes. The drug had previously been approved, in December 2010, by the European Commission for use in the European Union. Laurie Scudder, DNP, NP, spoke with Norman Stockbridge, MD, PhD, Director of the Division of Cardiovascular and Renal Products within the FDA Center for Drug Evaluation and Research, about this approval and the implications for practice.
Medscape: In December 2010, the FDA declined to approve this agent which was approved by the European Commission just weeks later. The manufacturer submitted new data for consideration, which contributed to this new decision. Can you describe some of that new data? What information was most conclusive in this change?
Dr. Stockbridge: In some sense there were no new data. There is only 1 trial -- the PLATO (PLATelet Inhibition and Patient Outcomes) study. The European Medicines Agency (EMA) had no problem approving based on the available data because the problem that limited the FDA's initial approval was the observation that the product appeared to be adverse compared with clopidogrel in the United States. This finding would not have disturbed the EMA any more than it would have disturbed us if the Australian result had gone in the wrong direction.
Following that December 2010 decision, the FDA asked the sponsor to provide some further exploration of a hypothesis proposed in response to the observation that the results were different in the United States, actually in all of North America, vs the rest of the world. They had suggested that the difference was based on the usual dose of aspirin used in North America in comparison to Europe. PLATO was not specifically set up to look at the issue of aspirin, and so the manner in which aspirin dosing information was collected did not generate a detailed picture of an individual's aspirin usage. There were, however, a variety of ways to impute aspirin dose. In addition, there was no prospective plan to analyze the aspirin dose and its effect on the outcome measures of cardiovascular (CV) death, myocardial infarction (MI), and stroke.
Rather than just accepting the analyses initially proposed to us or the additional analyses that our review team had performed, we asked the sponsor to conduct more detailed sensitivity analyses to examine more thoroughly the various ways that aspirin dose tracked with the apparent difference in the effects of ticagrelor and clopidogrel. This detailed analysis, then, is what they provided and what we have spent the last 6 months examining.
The original analysis of data from PLATO found a very significant 16% reduction in the combined endpoints of CV deaths, stroke, and MI in patients treated with ticagrelor as compared with those treated with clopidogrel. That was the effect in the trial as a whole and this outcome was the primary analysis. The original trial design was not intended to look at the effects stratified by aspirin dose or geographic region. It was only when a variety of baseline factors were examined that some concern about discrepant effects in North America, including Canada, were recognized. Indeed, the results in this region looked adverse compared with clopidogrel. The 1400 US patients actually had been observed to be at greater risk for a clinical event.
In the original submission, the sponsor had identified aspirin as the likely basis for the US vs outside-US results. Our advisory committee did not endorse the aspirin hypothesis and expressed varying levels of concern about the discrepancy in results between the United States and the rest of the world. They opined that the difference between the results in North America and the rest of the world was likely to be a chance finding. However, the aspirin hypothesis was at least on the table at the time of our original complete response letter and was the reason for the request for some further sensitivity analyses to determine whether this was a reasonably plausible explanation for the varying results. There could have been any number of factors that were responsible for the discrepant results. It is still possible that the discrepant results were the result of chance, though we believe that is unlikely, in part because this statistically significant difference was seen in the composite endpoint as well as the major components of that composite, but not major bleeding.
Medscape: The new approval comes with some significant caveats for use of this agent, including a box warning stating that use of this agent is not recommended in patients taking more than 100 mg/day of aspirin. However, Health Canada recently approved ticagrelor and specified that the drug be coadministered with a low maintenance dose of aspirin of 75-150 mg. Can you discuss some of these specific issues and their clinical implications?
Dr. Stockbridge: The sponsor and the FDA advisory committee explored a number of potential baseline covariate differences that might underlie the discrepant regional results. There are differences in the patient populations in North America and the rest of the world -- for example, more people going to percutaneous intervention in the United States. There were, conceivably, other practice-related differences between the United States and the rest of the world. The only one in a long list of univariate and multivariate analyses that were done by the sponsor and by our team that appeared to have explanatory power was the aspirin hypothesis. The aspirin dose was usually less than 100 mg in Europe and it was usually more than 100 mg in the United States. However, there was enough use of high- and low-dose aspirin in both places that the results by aspirin dose are very similar. In both the United States and the rest of the world, patients on low-dose aspirin have better outcomes using ticagrelor. In both regions, patients on high-dose aspirin do better with clopidogrel. That result was seen for the overall composite endpoint and the major components of CV death and MI.
On that basis, we concluded that it was pretty likely that the results seen by region were real and that the most likely explanation had to do with the use of generally a higher dose of aspirin in the United States. I do, however, want to point out that all of this was based on post-hoc analyses. Many such analyses are possible, of course, so P values from them are hard to interpret, especially if they are based on small sample sizes or are sensitive to the details of how you do the analysis. This is in sharp contrast to the findings of a study's prospective analysis plan, which yields a P value more easily translated into the likelihood of seeing a similar result in a subsequent study. In this case, the highly reliable findings were the overall effects on the composite endpoint and on the components of CV death and MI.
There will continue to be questions about the specific outcome with respect to the aspirin dose. We did not have complete uniformity of opinion, as will be clear when the various reviews are made available on the FDA Website.
Medscape: Do you believe that the difference between a high dose of 150 mg and high dose of 100 mg is significant?
Dr. Stockbridge: It is probably not, and I think there is no compelling basis for where you make the cutpoint. The observation is that patients receiving the lowest dose seemed to do the best on ticagrelor. Results for patients receiving the middle doses, around 150 mg, seemed to be about the same on ticagrelor and clopidogrel. Patients on the high dose had the poorest outcome on ticagrelor. Where you cut is not clear in a setting in which results vary slightly depending on exactly how the analyses are conducted. It is not surprising that various regulatory agencies interpret the results differently and differ on where they would cut the aspirin dose.
Medscape: Does the approval specify coadministration of aspirin? What about patients not on aspirin?
Dr. Stockbridge: That is a really good question. The fact is that essentially everybody in the PLATO trial was on aspirin. I don't think it is a ridiculous question to ask whether these data might suggest to somebody that if you are on ticagrelor you might be well off enough without aspirin.
I think many people will feel very uncomfortable doing that, particularly for patients with a stent. We don't really know what role having any aspirin on board had. It is not likely that this question is going to receive further attention through an additional trial. The FDA did not ask the sponsor to conduct such an investigation.
Medscape: Some experts have voiced concern that the offset effect of ticagrelor is so fast that missing a single dose could potentially be dangerous. Patients who were deemed unlikely to comply with the twice-daily regimen of ticagrelor were excluded from the PLATO study. Can you speak to this issue?
Dr. Stockbridge: First of all, it is never a good idea to incorporate people in a trial who aren't going to take the drug. I think the relevant information about rapidity of onset and offset is found in this Figure (see below), which is also included in the prescribing information and illustrates platelet aggregation activity for placebo, for ticagrelor at the dose that was studied here, and for clopidogrel.
What you will see there is that there is greater platelet inhibition with ticagrelor when compared with clopidogrel. Although the rate of restoration of platelet activity is somewhat faster on ticagrelor, because you get to a higher peak it still takes several days for recovery to occur. The exact time is not clear because that is going to depend largely on the particulars of the platelet assay. It takes a couple of days before the more rapidly declining platelet inhibition with ticagrelor drops below the level of platelet inhibition seen with clopidogrel.
Figure. Mean inhibition of platelet aggregation following 6 weeks on placebo, ticagrelor 90 mg twice daily, or clopidogrel 75 mg daily.
Brilinta™ (ticagrelor). Package insert. Wilmington, Del: AstraZeneca; 2011. Available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2011/022433s000lbl.pdf
That means that the concern over missing a dose is not a very big one. Missing a dose will not result in a platelet activation level that is any lower than if you took clopidogrel without missing a dose. In some ways, this is a disappointing finding. The potential that this agent would be rapidly reversible, and therefore useful for patients requiring an invasive procedure like a coronary artery bypass graft, was not realized. As recommended in the label, clinicians should ideally discontinue ticagrelor 5 days before surgery.
Medscape: Are there any specific populations of patients for whom this agent would be a particularly good choice?
Dr. Stockbridge: All of these drugs are associated with bleeding. It is the price you pay to prevent an ischemic event associated with activated platelets. It is contraindicated in those with a history of intracranial hemorrhage, active pathological bleeding or severe hepatic impairment. This contraindication does not extend to any other population.
Medscape: The drug was approved with the requirement that the manufacturer develop a Risk Evaluation and Mitigation Strategy (REMS) to help ensure that the drug's benefits outweigh its risks. This REMS must include educational outreach to prescribers to alert them to the risks associated with using higher doses of aspirin. Can you describe the anticipated components of that REMS?
Dr. Stockbridge: The REMS was largely intended to address the use of low-dose aspirin with this product. In addition to inclusion of a boxed warning, we have required a communications plan that includes outreach to healthcare providers through a series of letters warning them that they need to pay attention to the aspirin dose. The only additional data that we have asked from the sponsor relates to the effectiveness of that communication in conveying this information to healthcare providers.