March 10th, 2013

ACC.13: Is Cangrelor an Antiplatelet CHAMPION?

, and

CardioExchange’s Rick Lange and David Hillis ask Deepak Bhatt, lead author of the CHAMPION PHOENIX trial, about his study group’s findings and their implications for clinical practice.


Cangrelor, an intravenously administered ADP-receptor antagonist with rapid onset and offset of action, has been shown to reduce periprocedural PCI complications, compared with clopidogrel, in patients with ACS or stable angina. The finding was demonstrated in the randomized trial known as CHAMPION PHOENIX. Incidence of the primary composite endpoint — death, MI, ischemia-driven revascularization, or stent thrombosis at 48 hours — occurred in 4.7% of the cangrelor group versus 5.9% of the clopidogrel group (odds ratio, 0.78; P=0.005).


Lange and Hillis: In the previously published CHAMPION-PLATFORM and CHAMPION-PCI trials, cangrelor failed to reduce the rate of periprocedural ischemic events, compared with placebo or clopidogrel. Why are the results of CHAMPION PHOENIX different?

Bhatt: That is a good question. Although neither CHAMPION PCI nor CHAMPION PLATFORM met the primary endpoint, there was a statistically significant reduction in the secondary endpoint of stent thrombosis in CHAMPION PLATFORM, as well as in a prespecified pooled analysis of the two trials. The CHAMPION Executive Committee considered stent thrombosis to be a very sensitive measure of antiplatelet efficacy (or inefficacy). Granted, secondary endpoints must always be interpreted cautiously when the primary endpoint is not met. That is why we launched CHAMPION PHOENIX: We believed that the reduction in stent thrombosis was real and wanted to test it prospectively.

The older two CHAMPION trials used a definition of periprocedural MI that did not allow for discrimination of reinfarction in patients presenting for PCI shortly after being admitted with a biomarker-positive ACS. The time from hospital admission to randomization was very short in the previous two trials and also in CHAMPION PHOENIX (median time was a bit over 4 hours). This reflects the high-volume centers involved in the trials — the time from presentation to admission was so brief that it was not easy to determine whether ACS patients had a reinfarction or not. For example, if only one biomarker measurement was made before the procedure, a subsequent elevation might have been due to a complication of PCI or, instead, just the natural history of an evolving ACS. Of course, taking two preprocedure samples would clarify that matter, but because of the short times to catheterization, that was not always possible.

In CHAMPION PHOENIX, the definition of periprocedural MI required very careful assessment of baseline biomarker status (detailed in Table S1A and S1B of the NEJM online supplement). Duke Clinical Research Institute carefully adjudicated the clinical endpoints in a blinded fashion. In addition, the use of a blinded angiographic core lab (run by the Cardiovascular Research Foundation) allowed objective determination of intraprocedural complications. In Table S6 of the online supplement, we summarize the differences between the older and the new CHAMPION trials.

Lange and Hillis: In CHAMPION PHOENIX, all patients randomized to cangrelor received a bolus and infusion before PCI, but the clopidogrel recipients may have undergone PCI before the drug’s full antiplatelet effects were evident (i.e., 37% received clopidogrel after or during PCI). Was CHAMPION PHOENIX a comparison of optimal vs. suboptimal antiplatelet therapy? 

Bhatt: I am glad you raised that point. First of all, in the 63% of patients who received clopidogrel immediately before PCI, a statistically significant benefit of cangrelor was found (odds ratio, 0.80). This was similar to the effect in the 37% who received it immediately after the procedure (odds ratio, 0.79).

CHAMPION PHOENIX was a trial of a drug, but also of a strategy. Sites that participated (rather enthusiastically with brisk enrollment, I might add, including 37% from the United States) did not uniformly pretreat patients with clopidogrel in their clinical practices. That is because many centers are still concerned about the patients who are pretreated with clopidogrel and then undergo catheterization only to reveal surgical disease. At that point, many surgeons insist on delaying bypass surgery, which at a minimum greatly inconveniences the patient and the patient’s family — and has the potential to create problems in unstable patients who would benefit from earlier surgery.

Of course, some centers have moved to routinely pretreating all their patients with clopidogrel. This is especially common at sites where urgent surgery never happens because of logistical reasons. Also, some cardiac surgeons are comfortable operating with aspirin plus clopidogrel, although the recommendations still state to wait 5 days. These centers may believe that the results of CHAMPION PHOENIX do not apply to them, and in a literal sense that is true because we did not test cangrelor against clopidogrel given several hours before the procedure. However, I would also point out — and this may evoke strong opinions — that the prospective randomized assessments of clopidogrel pretreatment, namely CREDO and PRAGUE-8, have been negative. While observational analyses of randomized clinical trials and registries (some of which I have contributed to) do show an association between clopidogrel pretreatment and a lower rate of ischemic events, and although it does make biological sense, this does not represent the highest level of evidence — namely, prospective assessment in a randomized clinical trial.

For those who find the above statements about clopidogrel pretreatment blasphemous, I point them to the ACCOAST trial. In that large randomized trial of prasugrel pretreatment (using a more potent agent than clopidogrel), the Data Safety Monitoring Board terminated the trial due to lack of efficacy as well as bleeding. Thus, if one sets aside preconceptions and examines the totality of data for oral pretreatment, the option of cangrelor may be appealing for many practices. 

Lange and Hillis: Dual antiplatelet therapy with aspirin and a platelet ADP-receptor antagonist is the standard of care in patients being considered for PCI. Clopidogrel usually is given to those with stable angina, whereas more-potent ADP-receptor antagonists (prasugrel or ticagrelor) are recommended for ACS patients. In your opinion, which patients should receive cangrelor rather than one of the other ADP-receptor antagonists?

Bhatt: Clopidogrel remains the most commonly used oral ADP-receptor antagonist worldwide. It is true that in patients at low bleeding risk, prasugrel is superior to clopidogrel in ACS patients who will definitely undergo PCI and that ticagrelor is superior to clopidogrel across the ACS spectrum. Use of those agents is likely to increase as familiarity with them grows. However, knowledge of the coronary anatomy is generally needed before administering prasugrel, so if a center is concerned about potential delays in patients who ultimately need bypass surgery, those concerns persist — and are amplified — with the more potent prasugrel. Ticagrelor is reversible, but in a practical sense, the recommendations still advise waiting until 5 days before bypass surgery (granted, the biology likely supports 3 days, but that still results in days of delay). For prasugrel, the TRITON event curves separated early, but the concerns about potential bypass surgery and bleeding cited above exist. For ticagrelor, the event curves in PLATO did not diverge early (for a variety of potential reasons), and the benefit seems to be largely from the chronic phase of therapy more so than peri-PCI.

So, if cangrelor were available, in my practice I would be likely to use it for the PCI and then decide which of the three oral agents to use based on the patient’s ischemic and bleeding risks. That would give me the flexibility of an intravenous agent during the procedure and the option to use whatever oral agent is best suited to the patient for long-term therapy. We have great data regarding transitioning cangrelor to clopidogrel (just give the loading dose at the end of the infusion). Those transitioning data with cangrelor are being accrued for prasugrel and ticagrelor right now and will hopefully be available soon. Then doctors will need to weigh all the data and determine the best fit for their specific practices and patients.

6 Responses to “ACC.13: Is Cangrelor an Antiplatelet CHAMPION?”

  1. Deepak….excellent responses. Thanks. Perhaps you can give our readers an idea of the cost (or projected cost)of cangrelor therapy. This will obviously weigh into whether or not it fits into or practice.

    • Several physicians have asked that question and it is an important one, not addressed by our paper. I don’t know the answer, as the drug is still investigational. Typically, companies set the price once a drug is approved and it is clear what the label states regarding indications and contraindications. In addition to the reduction in ischemic complications identified in CHAMPION PHOENIX, there may be other benefits of a strategy that uses cangrelor beyond what we directly studied in our trial. For example, use of cangrelor instead of pretreatment with an oral agent could prevent the situation where a patient with surgical anatomy needs to stay in the hospital an extra 5 to 7 days prior to bypass surgery. Depending on the health care system, there could be significantly decreased costs if that scenario plays out. Clearly, cost will be a factor in how often cangrelor would be used during PCI.

  2. The CHAMPION PHOENIX trial investigators are to be congratulated for the trial reaching its primary endpoint.
    The results are CLEAR and CONSISTENT. However, whether they are COMPELLING enough to change clinical practice or not is contingent on 2 BELIEFS:

    1. If one is a believer in “clopidogrel pretreatment” (about half of the interventionalists in the US and guideline writers are), then nothing will change as this trial did not compare pretreatment with clopidogrel vs cangrelor. As an aside, pretreatment hypothesis remains to be validated by high-quality, practice-changing data.

    2. If one does not believe in periprocedural biomarker elevations being clinically relevant and therefore does not routinely measure biomarkers post-procedure, then nothing will change as the benefit was driven by biomarker MIs (94% of total MIs, 464/491). I (and I suspect most clinicians) remain uncertain about the clinical relevance of biomarker elevations. It is possible that by not measuring post-PCI biomarkers routinely, we remain blissfully unaware of the potential benefit of these new and ‘expensive’ therapeutics.

    I am curious to find out the extent of biomarker elevations (>3, >5, >10xULN) as I did not see the results in the main paper or the supplement.

    Another observation is that 0.6% difference in Stent Thrombosis did not translate into a difference in mortality or Q-MIs. Stent thrombosis are typically associated with large thrombi and often result in large MIs and/or deaths.

    • Great points.

      1. CHAMPION PHOENIX did not compare cangrelor against clopidogrel started upstream, so practices that start clopidogrel in the emergency department, for example, may not feel these results apply to their practices. I agree that the data for clopidogrel pretreatment are not as ironclad as is widely believed. I have participated in observational analyses that show an association between clopidogrel pretreatment and lower rates of ischemic complications, but my thinking has evolved in the face of negative trials like PRAGUE-8 and (though not yet published) ACCOAST, which used the more potent prasugrel but still found no apparent benefit, but did find bleeding – and was therefore terminated by the DSMB.

      2. Subsequent presentations/papers will provide further data regarding MI, but as I mentioned during the late-breaking clinical trial Q+A, the primary endpoint does remain statistically significant even when one uses the current universal definition of MI (which uses 5xULN, though coupled with clinical events). I agree that there is controversy regarding silent periprocedural MI and its true importance – my own thinking on the matter has evolved with the data. I don’t think many interventionalists will argue with the reduction in stent thrombosis though. Q-wave MI was numerically lower, but it is true this was not statistically significant, likely due to statistical power (despite this being a large trial, we are talking about low frequency, albeit important, events). Regarding case fatality of stent thrombosis, it does depend on the setting and in-hospital stent thrombosis is less likely to result in mortality, assuming it is promptly recognized and treated. We are working on providing more data from CHAMPION PHOENIX, and indeed the entire CHAMPION program, as quickly as we can. It may take a while for these additional analyses to work their way through peer review, but we will strive to get the data out to answer these and other excellent lingering questions.

  3. Alain Efstratiou, MD says:

    I do not see, if one decides to use an intravenous agent, why not stick with a IIbIIIa that inhibits platelet aggregation every single way as opposed to cangrelor that blocks only ADP mediated activation. Kastrati’s work has validated a single bolus approach that I find very convenient and efficacious (and probably cheaper).

    • The ADP receptor is a well-validated target of antiplatelet therapy. ADP receptor antagonism targets both platelet activation and aggregation, while GP IIb/IIIa inhibition only targets aggregation. There are theoretical reasons why preventing initial activation might be useful. The intravenous GP IIb/IIIa inhibitors are not as quickly reversible as cangrelor. This would be a potential advantage of cangrelor if emergent surgery were needed. Perhaps that explains why in many studies they were associated with an increase in major bleeding. The pivotal trials that led to the approval of glycoprotein IIb/IIIa inhibitors all used an infusion after the initial bolus, though it is true there are some data supporting use of a bolus only strategy.