P₂Y₁₂ inhibitors in neuroendovascular surgery: An opportunity for precision medicine

Abstract

Introduction

Dual antiplatelet therapy (DAPT), primarily the combination of aspirin with a P2Y12 inhibitor, in patients undergoing intravascular stent or flow diverter placement remains the primary strategy to reduce device-related thromboembolic complications. However, selection, timing, and dosing of DAPT is critical and can be challenging given the existing significant inter- and intraindividual response variations to P2Y12 inhibitors.

Methods

Assessment of indexed, peer-reviewed literature from 2000 to 2020 in interventional cardiology and neuroendovascular therapeutics with critical, peer-reviewed appraisal and extraction of evidence and strategies to utilize DAPT in cardio- and neurovascular patients with endoluminal devices.

Results

Both geno- and phenotyping for DAPT are rapidly and conveniently available as point-of-care testing at a favorable cost-benefit ratio. Furthermore, systematic inclusion of a quantifying clinical risk score combined with an operator-linked, technical risk assessment for potential adverse events allows a more precise and individualized approach to new P2Y12 inhibitor therapy.

Conclusions

The latest evidence, primarily obtained from cardiovascular intervention trials, supports that combining patient pharmacogenetics with drug response monitoring, as part of an individually tailored, precision medicine approach, is both predictive and cost-effective in achieving and maintaining individual target platelet inhibition levels. Indirect evidence supports that this gain in optimizing drug responses translates to reducing main adverse events and overall treatment costs in patients undergoing DAPT after intracranial stent or flow diverting treatment.

Introduction

Neuroendovascular interventions with the placement of intracranial stents and flow diverters are increasingly performed worldwide, a rise which is closely linked to increased utilization of brain imaging, improvements in interventional devices and techniques, and reduction in thromboembolic events (TEE) by combining a P2Y12 receptor inhibitor with aspirin, termed dual antiplatelet therapy (DAPT). With the growing complexity of neuroendovascular therapeutics, the risk profiles for adverse events have also changed and antiplatelet strategies driven by patient-specific characteristics have gained importance. Most of the evidence for DAPT effectiveness originates from extensive, randomized studies performed in high-risk cardiac patients which consistently showed substantial reductions of in-stent thrombosis and thromboembolic events employing various DAPT regimens, predominantly a combination of clopidogrel and aspirin. However, the newer oral P2Y12 inhibitors prasugrel and ticagrelor, together with the recent addition of the intravenous P2Y12 inhibitor cangrelor¹ and the anticipated availability of a specific reversal agent for ticagrelor,² provide increasing variety, and at times ambiguity, in addressing specific treatment scenarios.

The rationale to employ individualized P2Y12 inhibitor strategies is based on differences in genetic susceptibility, unpredictable pharmacokinetics, and clinical factors that introduce substantial inter- and intra-patient response variability in platelet reactivity (PR). For example, CYP gene polymorphisms cause significant variances in active clopidogrel metabolites, and genetic variation in high on-treatment platelet reactivity (HPR) is a consistent risk factor for stent thrombosis and TEE in both cardiac and neuroendovascular interventions. This complexity increases even further with the inclusion of clinical factors such as care setting (elective versus emergent), concomitant therapies, costs, side effects, medication adherence, and patient/physician preference. The integration of such factors with the ability to detect genetic polymorphism (genotype) and degree of PR (phenotype) at the bedside encourages a patient-specific, individualized therapeutic approach, also referred to as precision medicine. Precision medicine is a new frontier combining genomics, data analytics, and population health to develop “an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person.”³ The difference between traditional clinical assessments and precision medicine can be expressed as the degree of reliance on objective data to identify personalized treatments. As discussed in this review, adding genetic or molecular data to established laboratory and clinical risk profiles, allows neuroendovascular surgeons and interventionalists to tailor an antiplatelet regimen to an individual patient uniquely. Here, we integrate the pharmacogenetics of P2Y12 platelet inhibitors, neuroendovascular DAPT experience gained from the investigation of flow diverters, and pertinent cardiovascular evidence to conceptualize a precision medicine approach to DAPT.

Methods

We identified records through database searching, including MEDLINE (http://www.ncbi.nlm.nih.gov/pubmed/); EMBASE (Excerpta Medica Database; http://www.embase.com/); Cochrane CENTRAL (The Cochrane Central Register of Controlled Trials; http://uscc.cochrane.org/); AHRQ (Agency for Healthcare Research and Quality; http://www.ahrq.gov/); Lexicomp Clinical Drug Information; and MICROMEDEX Healthcare Series. When appropriate, Medical subject headings (MeSH) and keywords pertaining to neuroendovascular interventions, intracranial flow diversion, intracranial stent, DAPT, percutaneous coronary intervention; myocardial infarction, acute coronary syndrome; antithrombotic therapy; P2Y12 inhibitor, aspirin, clopidogrel, ticagrelor, prasugrel, ischemic stroke, hemorrhage, thrombosis, platelet reactivity, platelet function test, pharmacogenetic, polymorphism, outcome adverse events, and cost were used. Key publications were searched for additional citations. Identified publications were screened at the title and abstract level, and selected references independently evaluated by all authors for eligibility and discrepancies.

Results

The need to improve on current antiplatelet therapy

Improved antiplatelet therapy is ever more pressing with the emergence of flow diversion as a first-line treatment for many intracranial aneurysms in the last decade. Flow diversion represents a conceptual shift in aneurysm treatment from packing the aneurysm to reconstructing the parent artery and the arising aneurysm using minimal porosity endoluminal devices. Two mechanisms primarily obtain occlusion: cutting off blood supply to the aneurysm which leads to thrombosis and ultimately, scar formation, providing a scaffold for neointimal neck overgrowth. Flow diverters, however, are thrombogenic, especially before completion of neointimal coverage. Thromboembolic rates with the Pipeline embolization device (PED) and DAPT have been reported to range from 8% to 12%,^4–6 and 6-month in-stent stenosis from 3.5% to 16%.^4–6 Small sample sizes explain this relatively wide distribution, retrospective study design, lack of anatomical stratification, and variation in DAPT between existing studies. Given these limitations, Saber et al.⁷ reported in their pooled analysis of 2,002 patients undergoing PED placement (2009 to 2017) an overall 7% (95% CI 6-9%) rate for thromboembolism and 5% (95% CI 4-6%) for hemorrhage For a concise overview, Table 1 summarizes ischemic and hemorrhagic event rates, as well as overall mortality rates for the respective DAPT regimens employed across six major prospective multicenter flow diversion trials for unruptured aneurysms.

Table 1.

Various DAPT regimens and reported ischemic and hemorrhagic complications of selected flow diversion publications.

	PITA68	PUFS69	PREMIER70	SAFE71	SCENT72	PARAT73	Adeeb et al.29	Atallah et al.74	Griessenauer et al.75	Moore et al.76	Soize et al.77
Study type	P / MC	P / MC	P / MC	P / MC	P / MC	P / RCT	R	R	R	R	R
Aneurysms	31	109	141	103	180	144	402	396	114	103	80
PreTx AntiplateletDosage	Not reported	ASA 325 mg/d x2 ds AND CLO 75 mg/d x7 dsOR CLO 600 mg x1d	ASA min 81 mg/d x7 dsANDCLO min 75 mg/d x7 ds	Not reported	ASA min 75 mg/d AND CLO 75 mg/d x5 ds	ASA 300 mg/d AND CLO 75 mg/d x3 ds	ASA 325 mg/d AND CLO 75 mg/d daily3 to 14 dsALT: CLO 600mgx1 or ticagrelor 90 mg/BID for nonresponders	ASA min 75 mg/d AND CLO 75 mg/d x10 ds ALT: ASA 325-650mg x1 AND CLO 600 mg x1 OR PRA 40–60 mg x1 OR TIC 90 mg x1	ASA min 81 mg/d AND CLO 75 mg/d min x10 ds	ASA 325 mg/d AND CLO 75 mg/d daily x14 ds OR Nonresponders TIC 180 mg x1 ∼3 h prior	ASA 325 mg/d AND CLO 75 mg/d x5 ds OR ASA 325 mg/d AND TIC 90 mg/BID x2ds
PostTx AntiplateletDosage	ASA 100 mg/d x180 dsCLO 75 mg/d x30 ds	ASA 325 mg d x 6 msCLO 75 mg/d X3 ms	ASA min 81 mg/dx6 msCLO 75 mg/d x3 ms	Not reported	ASA min 75 mg/d indefinitelyAND CLO 75 mg/d or equivalent x6 ms	<6 wks: ASA 300 mg/d AND CLO 75 mg/d; 6 wks to 3 ms: ASA 100 mg/d AND CLO 75 mg/d; then only ASA	DAPT x3 msthen only ASA	CLO 75 mg/d OR PRA 10 mg/d OR TIC 90 mg/BIDx6 to 12 ms	DAPT x3-6 ms	DAPT x3 msthen only ASA	DAPT x3 ms
PhenotypingTargets	None	None	VerifyNow PRU 60 to 200	None	Platelet aggregometry30% to 90%	None	Nonresponders if PRU >208	Nonresponders if PRU <30% inhibition	None	Aggregometry >50% with 5 μM ADP OR VerifyNow PRU >208	None
GenotypingAlternative	Not done	Not done	Not done	Not done	Not done	Not done	Not done	Not done	Not done	Not done	Not done
Follow-up	6 mo	6 ms and 5 yr	1 yr	14.7 ms	12 ms	6 ms	Not specified	15.8 ms	6.5 ms	7.2 ms	1 m
Thromboembolism	6.7%	2.8%	0.7%	6.8%	6.1%	Tubridge 10% SAC 10%	ASA CLO 5.6%ASA CLO plus BOOST 9.8%ASA TIC 2.7%	ASA CLO 7.5%ASA PRA OR TIC 4.5%	All: 8.5%	ASA CLO 6%ASA TIC 4.2%	ASA CLO 5%ASA TIC 12.5%
Hemorrhage	0%	1.9%	1.4%	1%	2.8%	Tubridge 6%SAC 3%	ASA CLO 3.2%ASA CLO plus BOOST 2%ASA TIC 0%	ASA CLO 5.6%ASA PRA OR TIC 0%	All: 1 patient	ASA CLO 1.9%ASA TIC 0%	ASA CLO 5%ASA TIC 5%

The need to further optimize current DAPT strategies is also identified in imaging patients undergoing neurointerventional procedures. In a prospective cohort study of 164 patients treated for unruptured aneurysms, new DWI-positive lesions on 3 T MRI appeared in 64% to 75% despite preprocedural 75 mg clopidogrel and 160 mg aspirin daily. Age and procedure duration were independent risk factors, and lesion diameter correlated with discharge Rankin scale.⁸ Yang et al. identified DWI-positive lesions in 54% of 193 aneurysm patients undergoing stent treatment while on daily 75 mg clopidogrel and 100 mg aspirin (ruptured aneurysms were preloaded with 300 mg of each clopidogrel and aspirin.) TEE incidence correlated with platelet ADP inhibition percentage, and 82% of events during an 8-month follow up were associated with clopidogrel resistance.⁹ A meta-analysis including 2,268 aneurysms identified that, depending on treatment type, 43% to 67% (95% CI 46-85%) had new DWI lesions that did not relate to rupture status, location, or size. Flow diversion as compared to other treatments showed a trend towards a higher lesion rate (67% vs. 45%).¹⁰ The neurocognitive impact of such brain lesions remains uncertain; however, one investigation assessed 51 patients undergoing flow diverter treatment for unruptured aneurysms and found no difference in Montreal Cognitive Assessment scoring between baseline and 1-month post-procedure.¹¹

Aspirin

Aspirin produces acetylation of the critical residue serine 529 of cyclooxygenase 1 (COX-1) leading to its inhibition and, consequently, reduced conversion of arachidonic acid to prostaglandin G2. Diminished prostaglandin substrate for thromboxane synthase in turn reduces thromboxane A2, which acts as a potent vasoconstrictor and platelet activator. As a result, both platelet aggregation and endothelial inflammatory factors are diminished. Low aspirin doses (75–300 mg) solely inhibit COX-1; nevertheless, PR can still be elevated due to competing for coactivating mechanisms. About 2–57% of aspirin users exhibit a suboptimal response, with some experiencing a cardiovascular event indicative of clinical aspirin failure. For others, in vitro platelet reactivity is not adequately blocked, even after aspirin addition (laboratory resistance). Pharmacokinetic resistance, or limited active drug at the target site, is commonly multifactorial and should be differentiated from pharmacodynamic resistance such as genetic polymorphism. While various laboratory assays quantify ranges for normal and therapeutic platelet reactivity or metabolites, these are mostly nonspecific and inconsistent, and assays are not considered to guide clinical decisions exclusively.¹² Genotyping aspirin resistance, however, has gained attention, i.e., for the genetic polymorphism of the PTGS1 gene that encodes COX-1, as specific short nucleotide polymorphisms and haplotypes dysregulate arachidonic acid-induced thromboxane production leading to inadequate aspirin responses.¹³

Aspirin HPR, a risk factor for major adverse cardiac events (MACE) after percutaneous coronary intervention (PCI),¹⁴ is observed in up to 15% of patients undergoing neuroendovascular interventions; however, in alignment with cardiovascular studies, a consistent association of aspirin PR and clinical outcomes is not evident. However, two recent meta-analyses suggest that in DAPT, low-dose (≤150 mg daily) as compared to high-dose aspirin given either before or after the neuroendovascular intervention increases late (>6 months) thromboembolism by about 2.5 times without differences in hemorrhagic events.^7,15 Furthermore, observational cardiac studies identified a roughly 4-fold increase in ischemic events in cardiovascular patients deemed resistant to aspirin.¹⁶ In an updated meta-analysis of 11,857 coronary patients, any aspirin resistance correlated with an increased risk of all-cause death (OR 2.42, 95% CI 1.86–3.15) and target vessel revascularization (OR 2.20, 95% CI 1.19–4.08).¹⁷ Low PR (LPR) on-aspirin-specific testing has been associated with an increased bleeding risk in one registry (HR 0.65, 95% CI 0.43–0.99) but not in another.¹⁸

P2Y12 inhibitors and genetic polymorphism

Clopidogrel and prasugrel are prodrugs biotransformed by multiple hepatic cytochrome P450 (CYP) enzymes to active metabolites that irreversibly inhibit the ADP P2Y12 platelet receptor. Of ingested clopidogrel, only 15% is transformed into active metabolites, an action predominantly promoted by the CYP2C19 subtype. Functional relevant polymorphisms of the CYP2C19 gene exist and expressions of different allelic combinations dictate active clopidogrel metabolite levels resulting in substantial inter-patient variabilities in platelet inhibition. However, the CYP2C19 genotype does not influence the clinical effectiveness of prasugrel and ticagrelor which exhibit more prompt, potent, and consistent platelet inhibition compared with clopidogrel. Prasugrel activation by CYP3A4 and CYP2B6, and to a lesser extent CYP2C19, is not affected by genetic polymorphisms. Ticagrelor, a reversible, non-competitive P2Y12 inhibitor, does not require activation and, moreover, 30% is metabolized by CYP3A4 into the equipotent antiplatelet metabolite AR-C124910XX.

Clopidogrel metabolism can be categorized by its underlying, genetically determined CYP metabolism using fast and reliable bedside assays, as well as laboratory-based methods, to detect CYP polymorphism and guide P2Y12 inhibitor therapy (Table 2). Based on the specific CYP2C19 allelic combination, five activity phenotypes with resultant clopidogrel metabolic responses are clinically recognized: ultrarapid, rapid, normal, intermediate (IM), and poor metabolizers (PM) (Table 3).¹⁹ Poor metabolizers have homozygous polymorphism CYP2C19 no-function (NF; formerly loss-of-function, LOF), both CYP2C19 wild-type (*1) alleles are replaced, for example, by *2, *3, or other combinations of no function alleles. The absence of any active clopidogrel metabolites is seen in about 5% of the US and 12% of Asian populations, though on average approximately 30% of the populations throughout the US, Africa, Middle East, and Europe are IMs/PMs; this prevalence increases to about 57% in Asian and 94% in Oceanian populations.^20,21 Intermediate metabolizers have reduced amounts of active clopidogrel. They are either heterozygous with one normal *1 and one NF allele leading to *1/*2, *1/*3, or show mixed *2/*17, *3/*17 polymorphism by replacing both *1 alleles with the combination of a single no function and gain-of-function allele (*17). In contrast, CYP2C19*17 is a gain-of-function (GoF) allelic variant that increases clopidogrel metabolic activity. Individuals with one or two of the GoF allelic variant *17 have rapid and ultrarapid clopidogrel metabolisms, respectively. Increased bleeding risk due to GoF has been reported in some,²² but not all, cardiovascular studies. Additional genetic variants with insignificant impact on clopidogrel metabolism have been described,²³ and there is no data to support genotyping in prasugrel- or ticagrelor-treated patients. Overall, *17 polymorphism outcome data are inconsistent and insufficient to support its role in treatment decisions. As discussed further below, reproducible correlations do exist between CYP2C19 *2 and *3 NF, reduced capacity for clopidogrel bioactivation, impaired platelet inhibition, and significantly higher risks of TEE substantiate CYP genotyping to assess inter-patient variability.

Table 2.

Pharmacology of oral P2Y12 inhibitors and expected clopidogrel bioactivation frequencies within the U.S. population.

	Clopidogrel (Plavix, Bristol-Myers Squibb/Sanofi)	Prasugrel (Effient, Eli Lilly)	Ticagrelor (Brilinta, AstraZeneca)	Cangrelor (Kengreal, Chiesi)
Drug class	Thienopyridine	Thienopyridine	Cyclopentyl-triazolopyrimidine	Adenosine triphosphate analogue
Receptor blockade	Irreversible	Irreversible	Reversible	Reversible
Prodrug	Yes	Yes	No but active metabolite	No
Half-life parent drug	≈6 h	<5 min	6–12 h	3–6 min
Half-life active metabolite	30 mins	Distribution 30-60 mins Elimination 2–15 h	8–12 h	1-2 mins
Binding site1	ADP	ADP	Allosteric, non-ADP	site uncertain
Route, dosage, frequency	oral, 75- and 300-mg tablets, once daily	oral, 5- and 10-mg tablets, once daily	oral, 60- and 90-mg tablets, twice daily	10 mL vial containing 50 mg
Loading dose	300 mg or 600 mg	60 mg	180 mg	30 mcg/kg IV bolus
Maintenance dose	75 mg daily	10 mg daily(5 mg if <60 kg)	90 mg twice daily	mcg/kg/min IV for ≥2 h
Onset2-offset of action	2–8 h; 5–10 d	0.5–4 h; 7–10 d	0.5–4 h; 3–5 d	≈2 min; <1 h
CYP drug interaction3	CYP2C19	no	CYP3A	no
Use in renal/hepatic impairment	no dose adjustment	caution in moderate-to-severe renal or severe hepatic impairment	caution in moderate and avoid in severe hepatic impairment	no dose adjustment
Nonbleeding side effects	none	none	dyspnea; elevated uric acid; elevated creatinine	dyspnea
Average wholesale price 30-day supply at maintenance dose	$232 (brand) $184 (lowest generic)	$551 (brand) $495 (lowest generic)	$236 (brand) no generic	$899 (brand) for 50 mg solutionno generic

¹Bind to ADP P2Y12 receptor preventing ADP from binding and activating GPIIb/IIIa complex.

²After loading dose-bolus.

³Indicates clinically significant drug interactions.

ADP = adenosine diphosphate; CYP = cytochrome P450; IV = intravenous.

Table 3.

Five activity phenotypes of clopidogrel metabolic responses.

Metabolizer Phenotype	Genotype with Examples	Frequency	Antiplatelet Response
Poor	2 LOF alleles(*2*2, *2/*3, *3/*3)	1%–5%Highest in Asia: 12.2%	Effectively absent
Intermediate	1 LOF/NF allele (*1/*2, *1/*3, *2/*17, *3/*17)	20%–30%Highest in Asia: 45.5%	Reduced
Normal	2 NF alleles (*1/*1)	35%–50%	Normal
Rapid	1 GOF allele (*1/*17)	20%–30%Highest in Middle East: 32%	Normal to increased
Ultrarapid	2 GOF alleles (*17/*17)	1%–5%Highest in Americas: 4.2%	Normal to increased

LOF/NF = Loss of/No Function; GOF = Gain of Function; NF = Normal Function.

Determining platelet reactivity

Assessment of platelet function with reliable and reproducible tests is crucial to identify patients not or over-responding to ongoing antiplatelet treatment. Because platelets react to a plethora of stimuli, different methodologies were developed to investigate certain platelet function aspects. Commonly employed test batteries include assessing platelet adhesion and aggregation, exposure of platelets to shear conditions, analysis of clot properties or measuring platelet compounds after activation. Light transmission platelet aggregometry (LTA) studies the changes of light penetration through an optically dense sample of plasma-rich platelets in response to agonists such as collagen, ADP, thrombin, ristocetin, epinephrine, and arachidonic acid. The photometer results are graphically depicted as the ratio of patient-to-standard (platelet-poor plasma or transparent) transmission during aggregation and expressed as the velocity of aggregation (slope of the curve), the maximal extent of aggregation (%), and latency time (lag phase). It is considered the gold standard to evaluate platelet function; however, limitations include aggregation testing without the presence of other blood cells, the need for trained operators, and sensitivity to pre-analytical (hemolysis, low platelet count, anticoagulant presence, etc.) and procedural conditions (reagent preparation, variations in agonists, etc.). In contrast, VerifyNow measures platelet aggregation in whole blood by a turbidimetric-based optical detection using a cartridge containing fibrinogen-coated beads and platelet agonists. Agonists, such as arachidonic acid, ADP, prostaglandin E1, or thrombin receptor activating peptide (iso-TRAP) aggregated platelets, and changes in optical transparency are quantified in the whole blood specimen. This primarily evaluates arachidonic acid resistance, with results expressed as Aspirin Reaction Units (ARU) and resistance to ADP/PGE1 P2Y12 inhibitors as P2Y12 Reaction Units (PRU). VerifyNow is widely used as a point-of-care device; its advantages and limitations are outlined in Figure 1, and platelet function tests in clinical use.

Figure 1.

Platelet reactivity and therapeutic range of P2Y12 inhibition.

Phenotyping and clinical outcome

Two neurointerventional studies applied receiver operating characteristic analyses to identify the best on-clopidogrel PR cutoffs for predicting adverse events. Kang et al.²⁴ derived from an elective cohort with 209 aneurysms a PRU ≥295 as the strongest significant cutoff to predict TEE (sensitivity 75%, specificity 57%). Nishi et al.²⁵ reported intracranial bleeding in 5.26% vs. 2.56% of 174 patients undergoing elective coil embolization applying a PRU cutoff ≤175. Smaller, retrospective studies reported PRU values of <70 and >150,⁴ >208,²⁶ <60 and >240²⁷ for an increase in hemorrhagic and thromboembolic rates, respectively (Figure 1). Evidence indeed exists to support a modified treatment strategy based on PRU testing. Using the ARU and PRU cutoff of 550 and 213, respectively, to define HTPR, Hwang et al.²⁸ prospectively identified among 228 patients undergoing elective coil embolization 126 with and 102 without HTPR. Patients with HTPR were randomized to either standard (100 mg ASA and 75 mg clopidogrel daily) or modified regimen (additional 300 mg aspirin if ARU ≥550 and 200 mg cilostazol if PRU >213). All non-HTPR patients also received standard treatment. Thromboembolism was defined as TIA or stroke within seven days and bleeding within 30 days after coiling. The TEE rate in HPR patients receiving standard treatment was significantly worse than those without HTPR, 16.3% and 1%, respectively. However, patients with HTPR receiving the modified regimen had a significant decrease in TEE (1.6% vs. 11.1%) compared to HTPR patients on a standard regimen without a detectable change in bleeding rate among groups. Further, Kang et al.²⁴ demonstrated a positive relationship between increasing PRU quartiles with 17% TEE and 21% procedure-related adverse events in the 4th quartile. Adeeb et al.²⁹ retrospectively evaluated 414 PED procedures and reported significantly lower TEE between clopidogrel responders and non-responders (5.6% vs. 17.4%), and among non-responders switched to ticagrelor rather than remaining on clopidogrel (2.7% vs. 24.4%). Podlasek et al.,³⁰ in a meta-analysis of five retrospective studies with control comparison groups, included 1,005 patients on DAPT undergoing flow diversion treatment of an unruptured or ruptured aneurysm. In addition to aspirin, 832 (82.8%) of all analyzed patients received clopidogrel while the ticagrelor and prasugrel comparison groups represented only 137 (13.6%) and 36 (3.6%) patients, respectively. The authors summarized a significant mortality risk reduction with ticagrelor or prasugrel (4.57, CI 1.23-16.99), similar TEE prevention rates, and in 2 studies, an insignificant increase in intracerebral hemorrhage among patients receiving ticagrelor.^31,29

In acute coronary syndromes, DAPT with aspirin and a P2Y12 receptor inhibitor significantly reduces the RR of MACE by about 20% at 1 year, and an additional 20% when prasugrel or ticagrelor are used instead. Depending on the clinical setting, DAPT duration commonly ranges between 1 month to >1 year. While on DAPT, approximately 10% experience thrombotic and 2% hemorrhagic events within one year. Current cardiology guidelines define LPR, optimal PR (OPR), and HPR categories as <95, 95–208, and >208 PRU for VerifyNow testing.³² Similar to aspirin treatment,³³ the presence of risk factors seems a significant contributor to the clinical relevance of HPR and outcome.³⁴ A meta-analysis of 6,000 individual on-clopidogrel patient data³⁵ showed that HPR was prognostic for MACE in a dose-dependent fashion. Low-risk cardiovascular patients had a 2% yearly risk of TEE; however, with ≥2 risk factors (including age >75, hypertension, and diabetes) the annual TEE increased from 2.5% to 8% in those with lower and higher PR, respectively. A meta-analysis including individual data from >20,000 on-clopidogrel patients associated LPR with a 1.74 RR (95% CI 1.47–2.06) of bleeding. Also constrained by statistical heterogeneity, a recent meta-analysis of 13 RCTs including >7,000 patients showed PFT-driven therapy decreases the incidence of cardiovascular events without increasing the risk of bleeding.³⁶

Genotyping and clinical outcome

Studies evaluating the role of genotyping in neuroendovascular patients are few. Ge et al.³⁷ analyzed 215 patient undergoing stent-assisted coiling for intracranial aneurysms and treated with 75 mg clopidogrel and 100 mg aspirin daily. In multivariate analysis carriage of CYP2C19 NF alleles and clopidogrel resistance were predictors of TEE, while normal (formerly referred to as ‘extensive') metabolizers were significantly associated with bleeding at 3-month follow up. Limited by small and heterogeneous subgroup sizes, Lin et al. suggested, in contrast to previously reported findings, an increase in TEE in 108 neuroendovascular patients with the allele CYP2C19*17, which did not correlate to the level of PR.³⁸

A meta-analysis supported by clinical trials showed that poor and intermediate clopidogrel responders have an increased HR of 1.57 (95% CI, 1.13–2.16) for MACE and 2.81 (95% CI, 1.81–4.37) for stent thrombosis.^39,40 Notably, increasing clopidogrel dosages did not reliably overcome the incomplete platelet inhibition seen in IM and PMs.⁴¹ For example, in patients with acute CAD, the presence of *2 polymorphism has been shown to increase the risks for MACE and stent thrombosis despite high-dose 600/900 mg clopidogrel loading followed by 150 mg daily.⁴² In clopidogrel IMs/PMs, prasugrel or ticagrelor (both unaffected by CYP2C19 genotype) significantly reduce MACE with a HR of 0.98 (95% CI, 0.80–1.20) and 0.77 (95% CI, 0.60–0.99), respectively.^43,44 The majority of studies using a genotype-targeted switch from clopidogrel to prasugrel or ticagrelor alternatives identified a significant reduction in MACE without a change in bleeding risks. The largest benefit of genotype-guided therapy is seen in PCI, in which IMs/PMs with clopidogrel increased risks of MACE 2-3 times compared to prasugrel or ticagrelor.⁴⁵ In the genotype-driven de-escalation study, POPULAR-GENETICS, 2,488 ST-elevation myocardial infarction (STEMI) patients were randomized after PCI either to a group receiving standard prasugrel/ticagrelor or a group guided by genotype such that *2 carriers received either prasugrel or ticagrelor while non-carriers received clopidogrel instead. Using the composite ischemic and hemorrhagic outcome, genotyping was not inferior and displayed fewer bleeding events (9.8% vs. 12.5%; HR = 0.78, 95% CI 0.61–0.98, p = 0.04). Furthermore, the PHARMCLO trial screened for three CYP2C19 and ABCB1 gene polymorphisms and included clinical parameters to select the P2Y12 inhibitor after PCI. The composite outcome occurred in 15.9% of patients with genetic testing and 25.9% with standard care (HR = 0.58, 95% CI 0.43–0.78, p < 0.001).⁴⁶ However, a subpart of TROPICAL-ACS study inferred no added benefits from using genotyping to predict ischemic and bleeding risks among patients who had undergone a de-escalation treatment guided by a PFT.⁴⁷ Recently, the TAILOR-PCI followed genotype-guided, oral P2Y12 inhibitor selection in ACS and stable CAD patients undergoing PCI, where a CYP2C19 LOF (NF) finding triggered in ticagrelor instead of clopidogrel therapy.⁴⁸ The authors concluded that there was no significant difference between groups for the composite end point of cardiovascular death, MI, stroke, stent thrombosis, and severe recurrent ischemia at 12 months. However, 15% of CYP2C19 LOF carriers still received clopidogrel; since the study enrollment in 2012 coronary stents and delivery techniques have improved; and genotype-guided therapy had a 34% lower occurrence of major cardiovascular events as well the inclusion of multiple end points significantly favored genotype-based therapy (HR, 0.60 [95% CI, 0.41-0.89]; P = .01). Furthermore, consistent with other studies,⁴⁹ post hoc analysis showed an about 80% reduction with genotype guidance during the first 3 months (HR, 0.21 [95% CI, 0.08-0.54]; P = .001).

Costs of genotyping

Over 3 million people annually in the United States receive prescriptions for clopidogrel after PCI, and in 2010 clopidogrel sales of US$9.4 billion were the second highest selling medication worldwide.^50,51 Cost prediction remains challenging, and cost-effectiveness models suffer from price changes in generic and branded medication and associated test kits. These models must also consider diverse estimates of adverse risk and target populations (i.e., ethnicity, age, principal diagnosis, and comorbidities). For example, cost-effectiveness changes rapidly with the incidence of cerebral ischemia and hemorrhage due to high management cost, significant socioeconomic long-term impact, and a drastic reduction in the patient’s quality of life. While the incidences for various treatment and medication regimens are relatively well-defined in patients with acute coronary syndromes, data from neuroendovascular trials are sparse. Cost-effectiveness studies provide insight into the tradeoffs and consequences of individual management choices and are helpful for clinical guidelines, but they are not well suited for individual decision-making, incorporation into institutional values, or for the complex determination of resource allocation.⁵²

Data on medication cost-effectiveness in neuroendovascular patients are lacking. Kim et al.⁵³ compared the cost-effectiveness of both genotype- and phenotype-guided strategies for selecting P2Y12 inhibitors in patients with ACS. Incremental cost-effectiveness ratios (ICER), or the ratio of extra cost per extra unit of health benefit, were most acceptable for a clopidogrel + phenotype and a ticagrelor + genotype with $12,119 per quality-adjusted life-year (QALY) and $29,412/QALY, respectively; the ICER of universal ticagrelor was cost-prohibitive $1,42,456/QALY. Another decision model in which intermediate metabolizers (those with persistent HTPR despite high-dose clopidogrel at 225 mg daily) were switched to an alternative inhibitor, found that combining both genotype and phenotype strategies was less costly and more effective.⁵⁴ Limdi et al.⁵⁵ used data derived from trials, observations, US life tables, Medicare claims, and guidelines to estimate the incidence of MACE events in ACS patients and, compared with universal clopidogrel, both universal ticagrelor and genotype-guided strategy were superior in QALY. However, only the genotype strategy was cost-effective at US$42,365/QALY. Additionally, AlMukdad et al.⁵¹ concluded in a summary analysis of 13 published studies of genotype-guided versus universal antiplatelet therapy in PCI patients that genotype-guided escalation was cost-effective. Johnson et al.⁵⁶ reasoned in 2015 that changing from to an all-genotyping approach in 1,000 ACS patients leads to savings approaching US$5,00,000 per year. Taken together, recent evidence identifies genotyping as a cost-effective strategy to determine DAPT therapy for ACS. Single gene testing, however, is falling out of favor as the costs for single gene vs. multi-gene (average price of US$250, range US$146-386, for a 4-gene panel) are comparable and multi-gene diagnostics will add lifelong value to the patient’s care with predicted cost-savings.⁵⁷

Bergmeijer et al.⁵⁸ studied the clinical feasibility of two genotyping devices, TaqMan StepOnePlus and Spartan RX POC. Test results of the Spartan RX POC, processing a single patient at a time, were available within 1 h after collecting a buccal swab allowing quick genotyping. Together with its POC feature, this is appealing for same-day surgery, including interventional or advanced care settings. Despite these benefits, the difficulties presented with capturing POC data in the medical record may make this option less desirable to some institutions due to the lifelong value of germline genetic data. TaqMan genotyping relies on blood sampling and trained laboratory personnel, which increase turnaround times. However, in addition to the basic to CYP2C19*2, *3, and *17 genotyping obtainable from Spartan RX POC, TaqMan types more polymorphisms allowing simultaneous batch-testing, which may present a cost advantage. Cost data on laboratory-based CYP2C19 genotyping vary and are estimated at US$100 (range US$50-250) per test, including reagent and personnel costs but excluding equipment. For example, utilizing the Spartan RX CYP2C19 system costs about $200 per test kit and approximately US$15,000 for the POC device.

Discussion

Antiplatelet therapy is the foundation of medical treatment preventing TEE and hemorrhagic events in neuroendovascular patients. Current antiplatelet treatment strategies rely on evidence-based medicine primarily obtained from cardiovascular RCT. Similar, level 1 and 2 neuroendovascular-specific results are currently unavailable due to sample size limitations, anatomical and clinical heterogeneity, device advancements, among others. The lack of large-scale neuroendovascular data and the unavailability of statistical assurance to select the optimal antiplatelet therapy in specific treatment scenarios encourages the application of precision medicine concepts. Recent and expected developments increasingly allow concurrent integration of pharmacodynamic and genomic data, promptly monitored platelet reactivity, and clinical factors (i.e. employing risk scores) to facilitate individualized risk mitigation.

In the traditional treatment paradigm, an antiplatelet agent is prescribed without the weighted integration of all available data. An ineffective antiplatelet agent is switched only in response to an adverse event. This is the cornerstone of reactive medicine, an approach intrinsically inefficient concerning safety, outcomes, and costs. The process can repeat itself several times over without statistical assurance of its effectiveness. The critical point of precision medicine in antiplatelet strategies is to increase the likelihood of prescribing the individually efficient drug right from the beginning. Employing global risk algorithms for personalized antiplatelet therapy that integrate clinical, biological, and genetic data are increasingly utilized in cardiac patients, but not yet neuroendovascular patients. For example, Angiolillo et al. recently introduced the ABCD-GENE hybrid score incorporating age, BMI, renal impairment, DM, and *2 genotyping. With a c-statistic of 0.66, higher than seen in VerifyNow or LTA testing, good discrimination and calibration in identifying HPR status and predicting adverse ischemic events following PCI was demonstrated. However, an essential step towards individualized neurointerventional decision-making could be done by integrating univariate clinical, biological, and genetic associations, nowadays quickly available at the bedside. Furthermore, genetic point-of-care testing is increasingly available, reducing the clinical uncertainties in P2Y12 therapy both at therapy initiation and hypo- or hyper-responders. The RAPID GENE study, which randomized verified hypometabolizers to prasugrel instead of clopidogrel post PCI, represents one example of successful integration of genetic point-of-care genetic testing, as one week post-procedure the genetic testing group was significantly less likely to have HTPR compared to no genetic testing.⁵⁹

CYP2C19 genotyping tests are available and results have become as fast as PR levels. However, genotype determination is not a perfect surrogate for the phenotype. In contrast to homozygous NF allele carriage, heterozygotes express a more intermediate and variable response to clopidogrel, although still significantly worse than that observed in wild-type patients.⁶⁰ At the same time, carriers of *2/*17 and *3/*17 have largely unpredictable reactivities. Hence, ascertainment of the individual gene-dose effect, that is, correlation of geno- and phenotype, remains vital as identified in the ELEVATE-TIMI 56 trial. Here, tripling of clopidogrel maintenance dose in 1-NF allele patients achieved levels of platelet reactivity similar to those seen with the standard 75-mg dosing in non-carriers; however, in 2-NF allele patients with doses as high as 300 mg daily, this did not result in comparable PR.⁶¹

It is essential to realize that antiplatelet therapy failure is commonly multifactorial and only targeted testing may provide real-time, objective data. The prevalence of cardiovascular medication non-adherence in high-income countries is estimated at 50%, with rates greater still among marginalized groups.⁶² Recently, Fanaroff et al. presented a post hoc analysis of the cluster randomized clinical trial ARTEMIS where 8,373 patients with MI were followed up for 1 year at 287 US hospitals with measurement of P2Y12 inhibitor presence. Medication compliance rates were 48% using pharmacy fills compared with 15% patient self-reported, and agreement between P2Y12 inhibitor drug levels and patient-reported or fill-based persistence was low. Patients who were nonpersistent by both pharmacy fills and self-report had the highest 1-year adverse cardiac event rates.⁶³ To that effect, a promising intervention was described by Griessenauer et al.⁶⁴ demonstrating that a pharmacy-supervised antiplatelet management protocol with closed-loop feedback for patients undergoing PED implantation is safe and efficacious.

Nevertheless, platelet function results are not the holy grail for thromboembolic and hemorrhagic risk assessments, and results must be weighed and examined within the clinical setting (Figure 2). High PR cutoff values have high negative predicative value for TEE, but their positive predictive value is low,⁶⁵ the latter caused by very low event rates; clinical event rates are dependent on multiple factors in addition to HPR. HPR is a significant and modifiable, though not stand-alone, risk factor for TEE, but rather should be part of a risk algorithm and other data points. Multiple risk factors and covariates such as diabetes, old age, previous myocardial infarction, renal insufficiency, and smoking status influence platelet physiology. Kirtane et al. analyzing PR data from an observational study of 8,582 acute coronary patients (ADAPT-DES) found that a monotonic increase in PRU was independently associated with cardiac stent thrombosis within the highest PRU quintile (median 317), and a threshold set at >208 PRUs showed a 2.3-fold increase of 2-year stent thrombosis. In contrast, clinically relevant bleeding was independently associated with the lowest PRU quintile (median 57) without an identifiable threshold.⁶⁶

Figure 2.

Perioperative on-treatment platelet reactivity-based de-escalation and escalation strategies.

Lastly, it is essential to appreciate that CYP genotyping results may affect other treatments beyond antiplatelet therapy. The CYP2C19 enzyme is also involved in the metabolism of selective serotonin reuptake inhibitors, tricyclic antidepressants, proton pump inhibitors, and voriconazole.⁶⁷ To further enhance overall patient care, predefined protocols to regulate access of obtained genotyping results, i.e. to the clinical pharmacist, are available. Genotyping results should also be shared with the local pharmacy and general practitioner.

Conclusion

In conclusion, both geno- and phenotyping are rapidly and conveniently available as point-of-care testing at a favorable cost-benefit ratio allowing detailed, real-time assessment of expected and actual platelet reactivity in response to antiplatelet agents. When taken together with a risk factors assessment for thrombotic or hemorrhagic adverse events, most reliably in the context of a quantifying clinical risk score and a technical risk assessment for potential adverse events, a more precise and individualized approach to modern-day P2Y12 inhibitor therapy can be achieved.