FDA is positing its own explanation for why AstraZeneca's Brilinta showed an overall benefit in its pivotal international trial but did not do so for the North American subgroup: that ticagrelor has a delayed onset of effect that is exacerbated by differences in medical practice.

As the drug heads to an advisory committee review July 28, FDA briefing materials show that while the primary reviewer discredits AstraZeneca's explanation for the findings, the agency is considering alternative explanations. That may be an encouraging note for the first-in-class CPTP anti-platelet agent, which is under review for treatment of acute coronary syndromes.

Perhaps unsurprisingly, FDA's materials indicate that the pivotal point for discussion by the Cardiovascular and Renal Drugs Advisory Committee will be the "quandary of the unfavorable results in the U.S. that are inconsistent with the results in most other countries represented in PLATO," the sole pivotal trial supporting AstraZeneca's bid for Brilinta's approval.

In the 18,624-patient trial, relative to the overall findings, patients in the U.S. had worse results for all efficacy measures, including the composite MACE endpoint of myocardial infarction, stroke and cardiovascular death as well each of the components individually and all-cause mortality (Also see "Brilinta's North America Anomaly Stymies AstraZeneca, Worries Analysts" - Pink Sheet, 7 September, 2009.).

FDA's review team was not swayed by AstraZeneca's proposed explanation - that differences in aspirin dosage contributed to the discordant findings. But the medical team leader on the review, Thomas Marciniak, offered up the delayed onset of effect as an alternate possible reason in a June 29 memo to the advisory panelists.

Outlook: Tough Debate, But Approval Possible

There is certainly indication within the briefing documents that the agency sees the benefits offered by ticagrelor and considers it a potentially useful addition to the anti-platelet armamentarium.

Common limitations of approved drugs in the thienopyridine class -including Lilly/Daiichi's Effient (prasugrel) and Bristol-Myers Squibb/Sanofi-Aventis' Plavix (clopidogrel) -are that they are administered as prodrugs requiring metabolic activation for effect and have irreversible binding, Marciniak points out. But these do not apply to ticagrelor. It does not require metabolic activation and binds reversibly.

Marciniak expressly states that "clopidogrel may be less effective in some patients because of reduced activation due to genetic or drug-interaction factors, and prasugrel is associated with higher rates of bleeding and a question of cancer promotion." He adds that "a novel drug without these limitations would be a therapeutic advance."

The in vitro evidence also points to ticagrelor having greater platelet inhibition than clopidogrel (at the doses used in PLATO) and there is evidence of a long-term benefit that is distinct from the effect of the thienopyridines.

But the clearest sign that at least some within FDA are giving ticagrelor a positive eye is that the agency came up with an alternative explanation for the negative PLATO U.S. results after discrediting the sponsor's proposed explanation. That alternative will be presented to the advisory committee.

It's Not About The Aspirin

AstraZeneca's aspirin theory was based on the differences in clinical practice between the U.S. and Europe, as well as other countries, which is a factor FDA does find influential in many cardiovascular settings. In the U.S, aspirin use was split between 325 mg and 82 mg doses for trial participants, whereas outside the U.S. the "vast majority" of dosing was at 75 mg or 100 mg.

The sponsor argued that ticagrelor patients did worse with the 325 mg dosage (Also see "AstraZeneca: Aspirin May Be To Blame For Poor Results In Some Brilinta Trial Patients" - Pink Sheet, 29 October, 2009.). Various analyses were conducted, but in the end, FDA reviewers determined that because of multiple problems with the analyses (including the high correlation of aspirin dosage and region, the sensitivity of the analyses to reclassification of small numbers of cases regarding loading vs. maintenance aspirin dosing and events in high-dose aspirin outside the U.S., and biological plausibility) "aspirin dosing does not explain the disparate results."

Could Delayed Onset Be Behind It All?

Marciniak reports that a new possible explanation arose late in the review. Because the issue was identified after the primary reviewers had completed their work, relevant parameters had not yet been incorporated into the agency's analyses - but they will be presented at the meeting.

Marciniak noted that ticagrelor could have a delayed onset of effect. In his analysis of the time to first MACE, Marciniak notes ticagrelor has a different time course than is seen with the thienopyridines in ACS, for which there is an "almost immediate benefit that rapidly accrues during the early days" but "benefit beyond 30 days is harder to establish." With ticagrelor, there was little variation early on, and beneficial effects were shown longer term (for all three types of events: STEMI, NSTEMI and unstable angina).

In looking at patients who needed subsequent percutaneous intervention, there seems to be a detrimental effect in patients with PCIs on day one, which led Marciniak to conclude that "ticagrelor may have a delayed onset of platelet inhibition compared to clopidogrel."

The pharmacokinetic and pharmacodynamic evidence do not give any clear indication on the delayed onset issue; it takes clearer form in the clinical data. The clinical pharmacology review determined the median Tmax after oral administration is 2.65 hours, which "suggests some patients could be at risk for delayed effect." That review also notes that the rate of offset of inhibition of platelet aggregation is faster for ticagrelor than for clopidogrel in patients on aspirin, but "given the higher antiplatelet activity and longer half-life of ticagrelor and its active metabolite," there isn't really a difference in the time to conduct surgery after stopping use versus clopidogrel (an issue for medical practice), about five days. Ticagerlor's half-life is about eight hours, which doesn't explain the delayed offset.

"Because we do not understand the delayed offset, I question whether we really understand the timing of onset as well," Marciniak states in his memo to the advisory committee members.

Geographical differences in medical practice are still the issue in the alternative explanation. "Many U.S. practitioners prefer to delay giving antiplatelet drugs until after angiography to determine whether the coronary anatomy is suitable for PCI or for coronary artery bypass surgery. European practitioners prefer to administer the antiplatelet drugs early so that platelet inhibition is maximal at the time of PCI," Marciniak explained.

"If U.S. investigators delayed study drug administration and if ticagrelor does have a delayed onset relative to clopidogrel, I would expect to see the results that PLATO has produced, including the U.S-O.U.S. disparity."

He planned to validate this theory by re-analyzing the U.S. versus ex-U.S. results to consider all clopidogrel use prior to percutaneous intervention, timing of study drug administration relative to PCI, and early PCI vs. no PCI.

Panel Will Also Weigh In On Single-Trial Strategy

AstraZeneca's regulatory strategy will also be examined - and the resulting opinions could have wide-reaching implications for other development programs. AstraZeneca conducted PLATO as a broad, encompassing trial, whereas "other anti-platelet development programs have sometimes had separate studies for patients whose initial therapeutic strategy involved PCI or medical management and also separate strata for UA/NSTEMI and STEMI."

FDA is asking the committee to consider whether the sponsor's application covers multiple indications that should be considered together, or if there are separate indications (and if so, if PLATO supports them all).

That FDA is raising that issue for advisory committee discussion might reveal internal frustration with trends in cardiovascular trials.

In an amended memo sent to advisory committee members July 16, Marciniak observes that "I could argue that PLATO tried to do too much: a priori it attempted to evaluate three different conditions (i.e., STEMI, NSTEMI and UA), two different management approaches (invasive and non-invasive), and two different pre-treatments (prior antiplatelet and no prior antiplatelet use)." That translates into 24 different possible subgroups, which are further aggravated by looking at the discrepancy between the U.S. and non-U.S. populations and other issues, like concomitant medications.

That might be too much variation for a discrete understanding of the drug's effect. "As to the date of this review, we have not successfully identified clear patterns among this complexity," Marciniak notes.

-Mary Jo Laffler (m.laffler@elsevier.com)

FDA Will Present Brilinta Panel With Alternative Explanation For Negative U.S. Data

More from Archive

More from Pink Sheet