December 17, 2010 (Bethesda, Maryland)— The US Food and Drug Administration has once again delayed a decision on the US approval of AstraZeneca's new antiplatelet drug ticagrelor (Brilinta) [1]. Yesterday was a tense waiting game, with the agency required to deliver a decision on the drug. But instead of the approval many expected, the FDA issued the company a "complete response" letter stating that it wanted "additional analyses" of the Platelet Inhibition and Patient Outcomes (PLATO) trial.
It remains unclear at this time exactly what other information the agency wants, but it has not requested additional studies as a prerequisite for the approval of ticagrelor, according to AstraZeneca. "We are still reviewing the letter, but we believe we can respond quickly and we remain confident," company spokesperson Sarah Lindgreen told heart wire.
While the holdup is widely believed to center on an anomaly in the PLATO findings from North American sites--where there was a statistically insignificant trend toward worse outcomes with ticagrelor vs clopidogrel (Plavix, Bristol-Myers Squibb/Sanofi-Aventis) among the 1800 subjects there--one expert who has been helping the FDA try to untangle the results told heart wire that the story is not that simple.
Aspirin Dose Unlikely to Be the Explanation
The 18 000-patient PLATO trial [2], conducted in 43 countries around the world, is the lynchpin of AstraZeneca's new drug application for ticagrelor and was sufficient to gain EU approval for the drug just last week. The overall results showed that ticagrelor plus aspirin reduced the primary end point--a composite of death from vascular causes, MI, or stroke--by 16% compared with clopidogrel plus aspirin over 12 months (9.8% of patients receiving ticagrelor suffered a primary end point event compared with 11.7% of those taking clopidogrel [p<0.001]).
But the outcome in 1814 patients in the US and Canada was worse in those taking ticagrelor than in those on clopidogrel, with a primary end point occurring in 11.9% of ticagrelor-treated patients compared with 9.6% of those on clopidogrel, although the difference was not significant.
If there is no convincing explanation forthcoming, then another trial is needed to resolve the disparity.
Dr Victor L Serebruany (HeartDrug Research, Johns Hopkins University, Towson, MD) says he was asked by the FDA to "try to help figure out what happened," after the FDA Cardiovascular and Renal Drugs Advisory Committeevoted seven to one to recommend approval of ticagrelor in July. He submitted his report to the agency at the end of October.
One of the explanations previously put forward is that North America is the one region in which higher doses of aspirin were used in the PLATO study, at 325 mg per day, so there have been suggestions that this could be an explanation for this divergent finding; other suggestions have included different body-mass indexes (BMIs) among North American patients, while others have said it could just have been a chance finding.
But Serebruany says, "Let's assume the aspirin theory is true and that it mattered for the US, then Europe--which is under uniform guidelines for low-dose aspirin--should see similar outcomes across all the countries there, but that's not true; the results there are all over the place."
The Advisory Panel Replies
In correspondence with heart wire, the chair of the advisory panel, Dr Sanjay Kaul (Cedars-Sinai Medical Center, Los Angeles, CA) argues that random variability could easily account for the different outcomes in different countries in Europe. He points out that country-specific data are consistent with the overall effect observed in PLATO and that the intercontinental disparity in the ischemic outcomes did not appear in the bleeding outcomes.
If there is no convincing explanation forthcoming, then another trial is needed to resolve the disparity.
If there was a real difference caused by differences in the aspirin doses, "then the efficacy and bleeding outcomes should have tracked together--ie, reduced efficacy in the US should have been associated with reduced bleeding and vice versa," Kaul says. "Absent any evidence of similar disparities in pharmacokinetic/pharmacodynamic variables, I infer that the disparities are as a consequence of confounding. Perhaps aspirin dose is a marker of differences in the process of care across regions."
Dr Mori Krantz (University of Colorado, Denver), who was the one "no" vote on the advisory panel, told heart wire,"The focus on aspirin is too simplistic. Beyond regional variation in aspirin dose, there were key process inconsistencies in North America, including lower adherence, more drug-eluting stents, and shorter length of stay, which could lead to divergent results."
However, Krantz said, "Ticagrelor is unique among ADP-antagonists in lowering mortality vs an active comparator, so I understand the sponsor's reluctance to just recapitulate PLATO in the US, even if it's a smaller, noninferiority design." Instead of repeating PLATO, Krantz suggested the sponsor demonstrate ticagrelor's effectiveness in noncoronary vascular disease. For example, peripheral vascular disease is the most aggressive form of atherosclerosis, and clopidogrel appears to be more effective than aspirin at managing it, so "if ticagrelor compared favorably with clopidogrel in a US [peripheral vascular disease] population, this could cement a strong biologic rationale for its use in acute coronary syndrome based on consistency of benefits," Krantz said.
Although AstraZeneca has not revealed what FDA is asking for in its complete response letter, Kaul believes that the "fly in the ointment" is the regional disparity observed in PLATO. He points out that FDA officials have been critical of the general quality of data from trials conducted outside the US and are therefore scrutinizing these trials carefully. For example, the MERIT-HF trial of the beta blocker metoprolol in heart failure showed mortality was significantly reduced outside the US, but a trend in the opposite direction occurred within the US. FDA approved it despite this disparity, because the drug did reduce hospitalizations among the US patients. But, Kaul points out, PLATO does not have any end points that can "rescue" ticagrelor the way hospitalization rescued metoprolol.
"All outcomes, including the primary end point--all-cause mortality, cardiovascular death, MI, or stroke--go in the wrong direction in the US, so it is not surprising that the FDA is taking its time to understand these issues and allowing the sponsor ample opportunity to provide a good explanation," Kaul said. "If there is no convincing explanation forthcoming, then another trial is needed to resolve the disparity. If a postapproval trial is not feasible, then a preapproval trial needs to be done."
Many Issues to Resolve
Serebruany says there are a number of other major issues that concern him. First, whereas AstraZeneca monitored sites in most countries in the world included in the PLATO trial, it did not monitor the PLATO sites in the US, Russia, and Georgia--all locations in which those on ticagrelor did not do as well as those on clopidogrel. The company hired third parties to monitor sites, and it seems "the FDA doesn't even have a list of who was monitoring PLATO in individual countries, and that makes me feel very nervous," he commented.
The FDA doesn't even have a list of who was monitoring PLATO in individual countries, and that makes me feel very nervous.
And in fact, North America is not the only place where patients on ticagrelor fared worse than those on clopidogrel, he says. "The thesis that it is the US [and Canada] vs the rest of the world is completely wrong. In fact, in about a third of the countries ticagrelor showed no benefit over clopidogrel," he notes, citing Denmark, Australia, Taiwan, and Norway off the top of his head as examples, in addition to Russia and Georgia.
In addition, just two European countries, Poland and Hungary--which between them had 21% of the total patients enrolled in PLATO--"bring almost half of the benefit, 46% of that seen in the overall trial," he notes. Results in Brazil and India came out heavily in favor of ticagrelor as well, he says.
Kaul counters that the analysis by the FDA shows that if the data from Poland, Hungary, and Turkey are removed, the results still are consistent with the overall treatment effect.
But Serebruany insists that inconsistencies are evident in the mortality data that "drive the study findings." In PLATO, the overall all-cause death rate in the clopidogrel arm was 5.9%, which is even higher than mortality in the CURE trial (5.7%) conducted 10 years ago. This outcome is "ridiculously high" when compared to similar contemporary trials such as TRITON-TIMI 38, says Serebruany. Further, when looking at only vascular mortality in the clopidogrel arm of PLATO, he notes that it was "enormously high--6.1% in STEMI patients and 5.4% overall."
Yet in North America, the numbers for mortality concur very well with previous evidence, he says. For example, the death rate seen with clopidogrel in PLATO in the US was 3.2%, which is "identical to that seen with clopidogrel in TRITON-TIMI 38," and the mortality rate with ticagrelor in North America was 3.84% in PLATO, similar to that seen in all three arms of the ACUITYtrial at one year (range 3.6%-3.9%).
Another issue in PLATO is the timing of the deaths, he says. Usually in such studies, deaths occur early, but in PLATO "there was a bubble between three and six months, which is very strange. No other trials show delayed death patterns like this."
"I have no explanations. The mortality data are paradoxical and unnatural and need to be independently verified," says Serebruany.
PLATO investigator Dr Bob Harrington (Duke University, Durham, NC) told heart wire he is "not especially interested in Dr Serebruany's perspective unless it's in the peer-reviewed literature" and added that the academics involved with the trial will soon be submitting for publication a series of analyses addressing the differences in outcome observed in the US PLATO cohort.
Did the FDA Panel Have Sufficient Time to Consider the Evidence?
Despite these anomalies, the FDA Cardiovascular and Renal Drugs Advisory Committee was satisfied that the results from the totality of clinical data showed that ticagrelor is sufficiently safe and effective to merit US approval.
Serebruany, however, says the panel did not have nearly enough time to properly study the issues. They received the FDA briefing documents on ticagrelor, some 364 pages, just a day or so before the meeting, he says.
Kaul rejects the suggestion that the panel was unprepared to review the PLATO results. "The panel members received the FDA briefing documents sufficiently in advance to properly review all of these issues. And the panel devoted plenty of time to a very thorough discussion on the issue of regional disparity in outcomes," he said. He points out that most of the committee recommended that the FDA require AstraZeneca to conduct a postapproval trial to solve the mystery of why subjects of different nationalities had such different results with ticagrelor, because "the committee was not able to identify confidently a plausible hypothesis for the regional disparity. And not being able to do so did not alleviate the committee’s concern that the difference was real."
The agency's decision on ticagrelor had been due in September, but it extended its review, shifting the target date for a decision from September 16, 2010 to December 16, 2010 in accordance with the prescription drug user fee rules.
Other Reaction to the FDA Letter
European reaction to the news of the FDA decision was mixed this morning. The stock of British-based AstraZeneca fell by almost 5% when the markets opened in London. Some analysts expressed disappointment, while others said they viewed this as a temporary setback and were not likely to change their forecasts on AstraZeneca.
"We believe the company has been negotiating strongly on the label and probably pushed it too far, so the agency responded with a request of further analysis. . . . We view today's news as a delay of up to nine months . . . we are not making any changes to forecasts at this stage," Panumure Gordon analyst Savvas Neophytou said in a note.
Ticagrelor is currently being reviewed by regulatory agencies in a number of territories around the world, including Canada.
The drug will be marketed as Brilique in the EU; it is not expected to be available in most European countries until the second half of 2011 due to pricing and reimbursement negotiations. The UK's National Institute of Health and Clinical Excellence is currently reviewing ticagrelor and is expected to produce guidance on the drug next year.
Serebruany has received funding for research studies on clopidogrel and consulting fees from manufacturers of both clopidogrel and ticagrelor. Kaul chaired the advisory panel on ticagrelor but states that his opinions are his own and don't reflect the FDA’s opinions. He also reports equity interest in Johnson & Johnson and consulting for Hoffman La Roche and Novo Nordisk. Harrington discloses that AstraZeneca has research grant contracts with the Duke Clinical Research Institute and a consulting relationship with him. All other disclosures are available here .
December 17, 2010 (Bethesda, Maryland) — The US Food and Drug Administration has once again delayed a decision on the US approval of AstraZeneca's new antiplatelet drug, ticagrelor [1]. Yesterday was a tense waiting game, with the agency required to deliver a decision on the drug. But instead of the approval many expected, the FDA issued the company with a "complete response" letter stating that it wanted "additional analyses" of the Platelet Inhibition and Patient Outcomes (PLATO) trial.
It remains unclear at this time exactly what other information the agency wants, but it has not requested additional studies as a prerequisite for the approval of ticagrelor, according to AstraZeneca. "We are still reviewing the letter, but we believe we can respond quickly and we remain confident," company spokesperson Sarah Lindgreen told heart wire.
While the holdup is widely believed to center on an anomaly in the PLATO findings from North American sites--where there was a statistically insignificant trend toward worse outcomes with ticagrelor vs clopidogrel among the 1800 subjects there--one expert who has been helping the FDA try to untangle the results told heart wire that the story is not that simple.
Aspirin Dose Unlikely to Be the Explanation
The 18 000-patient PLATO trial, conducted in 43 countries around the world, is the lynchpin of AstraZeneca's New Drug Application for ticagrelor and was sufficient to gain EU approval for the drug just last week [2]. The overall results showed that ticagrelor plus aspirin reduced the primary end point--a composite of death from vascular causes, MI, or stroke--by 16% compared with clopidogrel (Plavix, Bristol-Myers Squibb/Sanofi-Aventis) plus aspirin over 12 months (9.8% of patients receiving ticagrelor suffered a primary end point event compared with 11.7% of those taking clopidogrel (p<0.001).
But the outcome in 1814 patients in the US and Canada was worse in those taking ticagrelor compared with those on clopidogrel, with the primary end point occurring in 11.9% of ticagrelor-treated patients compared with 9.6% of those on clopidogrel, although the difference was not significant.
The mortality data are paradoxical and unnatural, and need to be independently verified.
Dr Victor L Serebruany (HeartDrug Research, Johns Hopkins University, Towson, MD) says he was asked by the FDA to "try to help figure out what happened," after the FDA Cardiovascular and Renal Drugs Advisory Committee voted seven to one to recommend approval of ticagrelor in July. He submitted his report to the agency at the end of October.
One of the explanations previously put forward is that North America is the one region in which higher doses of aspirin were used in the PLATO study, at 325 mg per day, so there have been suggestions that this could be an explanation for this divergent finding; other suggestions have included different body-mass indexes (BMIs) among North American patients, while others have said it could just have been a chance finding.
But Serebruany says, "Let's assume the aspirin theory is true, and that it matters for the US, then Europe--which is under uniform guidelines for low-dose aspirin--should see similar outcomes across all the countries there, but that's not true, the results there are all over the place."
Many Issues to Resolve
Serebruany says there are a number of other major issues that concern him. First, whereas AstraZeneca monitored sites in most countries in the world included in the PLATO trial, it did not monitor the PLATO sites in the US, Russia, and Georgia--all locations in which those on ticagrelor did not do as well as those on clopidogrel. The company hired third parties to monitor sites and it seems "the FDA doesn't even have a list of who was monitoring PLATO in individual countries, and that makes me feel very nervous," he commented.
The FDA doesn't even have a list of who was monitoring PLATO in individual countries, and that makes me feel very nervous.
And in fact, North America is not the only place where patients on ticagrelor fared worse than those on clopidogrel, he says. "The thesis that it is the US [and Canada] vs the rest of the world is completely wrong. In fact, in about a third of the countries ticagrelor showed no benefit over clopidogrel," he notes, citing Denmark, Australia, Taiwan and Norway off the top of his head as examples, in addition to Russia and Georgia.
In addition, just two European countries, Poland and Hungary--which between them had 21% of the total patients enrolled in PLATO--"bring almost half of the benefit, 46% of that seen in the overall trial," he notes. Results in Brazil and India came out heavily in favor of ticagrelor as well, he says.
Also, there are inconsistencies in the mortality data, and the latter "drive the study findings," he says. In PLATO, the overall all-cause death rate in the clopidogrel arm was 5.9%, which is even higher than mortality in the CURE trial (5.7%) conducted 10 years ago. This outcome is "ridiculously high" when compared to similar contemporary trials such as TRITON-TIMI 38, says Serebruany. Further, when looking at only vascular mortality in the clopidogrel arm of PLATO, it was "enormously high--6.1% in STEMI patients and 5.4% overall."
Yet in North America the numbers for mortality concur very well with previous evidence, he says. For example, the death rate seen with clopidogrel in PLATO in the US was 3.2%, which is "identical to that seen with clopidogrel in TRITON-TIMI 38," and the mortality rate with ticagrelor in North America was 3.84% in PLATO, similar to that seen in all three arms of the ACUITY trial at one year (range 3.6-3.9%).
Another issue in PLATO is the timing of the deaths, he says. Usually in such studies, deaths occur early, but in PLATO "there was a bubble between three and six months, which is very strange. No other trials show delayed death patterns like this."
"I have no explanations. The mortality data are paradoxical and unnatural, and need to be independently verified," says Serebruany.
FDA Panel Did Not Have Sufficient Time to Consider All the Evidence
Despite these anomalies, the FDA Cardiovascular and Renal Drugs Advisory Committee voted seven to one to recommend approval of ticagrelor in July. The panel was satisfied that the results from the totality of clinical data showed that ticagrelor is sufficiently safe and effective to merit US approval.
But Serebruany says the panel did not have nearly enough time to properly study the issues. They received the FDA briefing documents on ticagrelor, some 364 pages, just a day or so before the meeting, he says.
The agency's decision on ticagrelor had been due in September, but it extended its review, shifting the target date for a decision from September 16, 2010 to December 16, 2010 in accordance with the prescription drug user fee rules.
Europe reaction to the news of the FDA decision was mixed this morning. The stock of British-based AstraZeneca fell by almost 5% when the markets opened in London. Some analysts expressed disappointment, while others said they viewed this as a temporary setback and were not likely to change their forecasts on AstraZeneca.
"We believe the company has been negotiating strongly on the label and probably pushed it too far, so the Agency responded with a request of further analysis . . . we view today’s news as a delay of up to nine months . . . we are not making any changes to forecasts at this stage," Panumure Gordon & Co analyst Savvas Neophytou said in a note.
Ticagrelor is currently being reviewed by regulatory agencies in a number of territories around the world, including Canada.
The drug will be marketed as Brilique in the EU; it is not expected to be available in most European countries until the second half of 2011 due to pricing and reimbursement negotiations. The UK's National Institute of Health and Clinical Excellence is currently reviewing ticagrelor and is expected to produce guidance on the drug next year.
Serebruany has received funding for research studies on clopidogrel and consulting fees from manufacturers of both clopidogrel and ticagrelor.
Heartwire from Medscape © 2010 Medscape, LLC
Cite this: No US Approval for Ticagrelor Yet; FDA Requests Further Analysis of PLATO - Medscape - Dec 17, 2010.
Comments