• Aneurysm
• arteriovenous malformation
• brain
• spine
• stroke
• artery
• angiography
• angioplasty
• atherosclerosis
• fistula
• stenosis
• stent
• subarachnoid
• thrombectomy

Two recent papers have evaluated the safety and efficacy of prasugrel in suspected clopidogrel non-responders in the neurointerventional population.1 ,2 In both papers the platelet response to clopidogrel was assessed preprocedurally and prasugrel was administered to patients meeting the criteria for inadequate platelet inhibition. The first found a statistical trend towards increased hemorrhagic complications with prasugrel while the second reported similar complication rates between prasugrel and clopidogrel. We commend the authors for these valuable contributions to the literature, highlighting the delicate balance between thrombosis and hemorrhage that must be maintained in the care of neurointerventional patients.

We wish to advance the following comments.

Since the FDA warning in March 2010 about the diminished effectiveness of clopidogrel in patients unable to fully convert the drug into its active form, there has been growing interest in the topic of clopidogrel response and routine point-of-care patient testing. There is good data to suggest that the variability in clopidogrel response has important clinical sequelae, with increased rates of cardiovascular morbidity and mortality in poor responders.3 ,4 However, no standard method exists to measure the platelet effects of clopidogrel, which leads to significant intra- and inter-individual variability in this continuous parameter.5 Such variability makes dichotomisation into clopidogrel responders and non-responders especially problematic.

Both studies use the VerifyNow technology (Accumetrics Inc, San Diego, California, USA) which uses an adenosine diphosphate agonist to induce platelet activation and ascertains the percentage by which these medications impair platelet reactivity. This is one of a variety of assays available, and although frequently used, it remains a means of assessing the effect of antiplatelet therapy that is yet to be truly validated. In addition, there are a number of commercially available genetic test kits that identify the presence of the CYP2C19 alleles associated with impaired clopidogrel activation. However, these tests are expensive, not routinely covered by most insurance policies and, importantly, address only one of multiple causes of clopidogrel non-response.

Furthermore, in spite of the burgeoning cardiovascular literature on this topic, there remains no universal or validated definition of ‘poor responder’. Although most current definitions of clopidogrel responsiveness are formulated according to the degree of inhibition of platelet aggregation, a recent cardiovascular meta-analysis including 3960 patients across 15 studies found a range of 32–70% for the aggregation cut-off used to define clopidogrel non-response.6 Indeed, in the two papers discussed here, responders are defined as >40% platelet P2Y12 receptor inhibition by Akbari et al and >20% by Stetler et al.

With these limitations in mind, no prospective studies exist that clearly demonstrate that the routine use of these tests coupled with modification of antiplatelet therapy improves clinical outcomes or reduces subsequent clinical events. It remains a matter of debate whether testing for completeness of antiplatelet inhibition following a prasugrel load or evaluating the antiplatelet effect of aspirin (as performed by Stetler et al but not by Akbari et al) is clinically important. Additionally, much of the work in this realm has been reported in the cardiology literature, raising the question of whether ‘adequate’ antiplatelet action in the coronary arteries is sufficient in the cerebral circulation or whether high-risk groups in cardiology mirror those in the neurointerventional population.

Does the use of prasugrel in clopidogrel non-responders reduce ischemic complications during neurointerventional procedures? Neither study was designed to evaluate clopidogrel non-responders who remained on dual therapy with clopidogrel and aspirin. Akbari et al found no difference in ischemic complications between their clopidogrel and prasugrel groups. Stetler et al did not detect any ischemic complications in their prasugrel cohort, but do not mention the occurrence of ischemia in their clopidogrel cohort. They also use a prasugrel dose of 5 mg or 10 mg daily, but details are lacking as to who and how many patients received the lower dose. (Might this be a reason for the lower rates of bleeding complications when compared with the 10 mg prasugrel regime used by Akbari et al?) Small sample sizes additionally limit our ability to draw any firm conclusions.

Akbari et al highlight the risk of hemorrhage, both spontaneously (eg, epistaxis, gastrointestinal bleeding) and peri-procedurally (following microwire perforations or other complications), that is likely to be a major hurdle to the widespread adoption of these agents by our practice. We reported our initial experience with prasugrel in a neurointerventional patient who developed a large retroperitoneal hematoma.7

The 2011 ACCF/AHA guidelines for the management of patients with unstable angina or non-ST-elevation myocardial infarction incorporate a new recommendation (class IIb, level of evidence B) that cautiously suggests that an approach of assessing platelet reactivity in combination with other procedural and patient characteristics may be considered to aid in patient management decisions.8 While we may progress to this within our field too, we are not there yet.

We expect that many similar studies will be published in the near future. However, until we standardise our methods for platelet inhibition testing and apply uniform definitions, we will be unable to effectively compare findings across studies or to assess whether they can be applied to our own practice. In addition, this problem will limit our ability to evaluate newer antiplatelet medications. P2Y12 receptor antagonists such as the reversible ticagrelor have been tested in cardiac patients and appear to be superior to clopidogrel.9 Ticagrelor is not a prodrug like clopidogrel and prasugrel and thus does not require bioactivation. However, its role in neurointervention remains to be explored. The twice daily dosage compared with the once daily dosing of prasugrel may also raise issues for longer term compliance. In addition, triple antiplatelet therapy (we currently use this in the form of rescue therapy through abciximab or epitifibatide infusions in patients already on dual antiplatelet agents who experience procedural thrombosis) may have a role in certain cardiac patients,10 and potentially by extension in neurointerventional patients. These questions become increasingly important as flow diverters and other complex constructs enter the market. Establishing industry standards for definitions of clopidogrel non-response and a uniform method of testing for it will be absolutely essential for future research and clinical practice. The JNIS has begun developing practice standards and guidelines11 and we anticipate that, as more data is published to guide policy, this topic will lend itself well to a societal guideline.

At this time, given the degree of uncertainty, our practice continues to use clopidogrel and aspirin as our antiplatelet agents of choice and to reserve prasugrel use (and testing for clopidogrel non-response) for cases of clinical antiplatelet failure. We join Akbari et al in their call for further evaluation and reporting of outcomes in patients treated with alternative antiplatelet therapies. We greatly look forward to contributing to and reading further articles that will help establish protocols and standards unique to the field of neurointervention.