Introduction Variable response to clopidogrel can impact perioperative risk in elective endovascular intracranial aneurysm treatment. The present study aims to determine the efficacy of a two-test protocol in reaching in-range preoperative P2Y12 reaction units (PRU) of 60–240 and the rate of postoperative conversion to hyper-response.
Methods A 17-day two-test protocol (with tests on days 10 and 17) for patients starting clopidogrel in anticipation of elective endovascular intracranial aneurysm treatment was introduced in February 2013 at our institution. Records for patients started on this protocol through December 2014 were reviewed for preoperative and postoperative PRUs, patient and procedural data, and thromboembolic and hemorrhagic events within 30 days. Logistic regression analyses were performed to identify predictors of postoperative hyper-response (p<0.05 considered significant).
Results 103 patients (80 women) of mean age 57 years were included. 74 patients (71.8%) were in range at the first test and 92 patients (89.3%) were in range at the second test. A postoperative test was performed in 82 patients (79.6%) at a median of 9 days. 51 patients (62.2%) converted into hyper-responders. There were five non-disabling strokes and one intracranial hemorrhage within 30 days. There were no major strokes (modified Rankin Scale score >2) or deaths. There was no association between out-of-range PRU and thromboembolic or hemorrhagic neurological complications.
Conclusions The protocol achieves in-range preoperative PRU by the second test in almost nine of 10 patients. Nearly two-thirds of patients exhibited postoperative hyper-response to clopidogrel. Out-of-range PRU was not associated with thromboembolic or hemorrhagic neurological complications in this cohort of patients with actively managed P2Y12 inhibition.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Preoperative oral antiplatelet therapy has long been used to reduce the thromboembolic risk of neurointerventional procedures, particularly the elective embolization of intracranial aneurysms.1 In the setting of assistive devices like stents or flow diverters, dual antiplatelet therapy—usually in the form of aspirin and clopidogrel—is often recommended to prevent device thrombosis.2
A wide and dynamic variability in response to clopidogrel therapy among patients undergoing endovascular treatment of intracranial aneurysms has been shown, as measured by P2Y12 receptor inhibition testing (VerifyNow, Accriva Diagnostics, San Diego, California, USA).3 There may also be delayed conversion to clopidogrel hyper-response postoperatively. Additionally, it has been shown that the variable response to clopidogrel can impact thromboembolic and hemorrhagic perioperative risk.4 ,5
To accommodate the variable response to clopidogrel among our patients undergoing elective aneurysm embolization by allowing for preoperative dose adjustments, we instituted a two-test protocol using VerifyNow, which reports P2Y12 reaction units (PRU). Patients with wide-necked aneurysms, stents or flow diverters undergo postoperative P2Y12 inhibition testing as well.
The present study has two primary goals: (1) to determine the efficacy of this protocol in reaching in-range preoperative PRU by the second test; and (2) to determine the rate of postoperative conversion to clopidogrel hyper-response. A secondary goal is to determine if there is an association between out-of-range preoperative or postoperative PRU and thromboembolic or hemorrhagic neurological complications.
Institutional review board approval was obtained for this study, which was conducted in compliance with the Health Insurance Portability and Accountability Act. Informed consent was obtained for all procedures performed.
Beginning on 22 February 2013, we instituted a 17-day two-test protocol (described below) for patients starting clopidogrel as part of a dual antiplatelet regimen in anticipation of elective endovascular intracranial aneurysm treatment. Our prospectively maintained registry of endovascularly treated intracranial aneurysms was retrospectively reviewed for all patients who were started on this protocol through 1 December 2014. Of note, it is routine for all patients undergoing elective endovascular intracranial aneurysm treatment at our institution to be placed on dual antiplatelet therapy.
P2Y12 inhibition testing protocol
We started our patients on clopidogrel 75 mg once a day, 17 days in advance of the procedure. VerifyNow tests were performed on days 10 and 17. Patients outside our goal PRU range (60–240) underwent dose adjustment until a PRU within the target range was achieved. The clopidogrel dosing schedule used is shown in table 1 and our PRU-based dose adjustment schedule is shown in table 2.
Patients not in range at the second test were generally rescheduled, particularly if a flow diverter was planned; this allowed for further dose adjustment and retesting every 7 days until the target range was achieved. For patients on steps 2 or 3 of the clopidogrel dosing schedule described in table 1 (ie, non-daily dosing), the testing day was independent of the dosing day.
Postoperative P2Y12 inhibition testing was performed on patients who underwent flow diversion, stent-assisted coiling or balloon-assisted coiling of a wide-necked aneurysm exposing a large surface area of coils to the parent artery circulation. This determination was made at the discretion of the operator. Postoperative testing was typically scheduled 7 days after the procedure. The management strategy for patients with out-of-range PRU was the same as the management preoperatively —that is, dose adjustments were made following tables 1 and 2 with a retest every 7 days until the target range was achieved. Dual antiplatelet therapy was continued for at least 5 months for flow diversion, 3 months for stent-assisted coiling, and 1 month for balloon-assisted coiling of a wide-necked aneurysm.
All procedures were performed in a dedicated biplane neuroangiographic suite (Axiom Artis, Siemens, Germany) by four interventional neuroradiologists with 30, 10, 7, and 6 years of experience. All procedures were performed under general endotracheal anesthesia using volatile inhalational agents. After induction of general anesthesia, transfemoral access was obtained. A 6 Fr guide catheter was placed in the internal carotid or vertebral artery, typically Neuron MAX (Penumbra, Alameda, California, USA). For flow diversion, Navien distal access catheter (Medtronic, Minneapolis, Minnesota, USA) was also used. The flow diverter, typically Pipeline Embolization Device (Medtronic), was deployed using a 027 microcatheter, usually Marksman (Medtronic). For stent-assisted coiling, typically LVIS Jr (MicroVention, Tustin, California, USA) would be deployed via Scepter XC 4×11 balloon (MicroVention) or Headway 17 microcatheter (MicroVention). For balloon-assisted coiling Scepter XC balloon was used.
Postoperatively, patients undergoing flow diversion were monitored in a dedicated neurological intensive care unit. All other patients were monitored on a dedicated neuroscience floor. Patients were typically discharged the following day unless a complication was encountered.
Patients were seen in clinic at 30 days and 6 months for a formal neurological examination. A modified Rankin Scale (mRS) was performed at each time point by an independent nurse practitioner with specific training in performing this assessment. If a patient was unable to come to clinic, the nurse practitioner obtained the mRS using a validated telephone questionnaire.6
In addition to all preoperative PRUs and the first postoperative PRU, the following patient data were recorded: age; sex; weight; body mass index; presence or absence of smoking, hypertension and diabetes; perioperative usage of medications that could interfere with clopidogrel metabolism including steroids, selective serotonin reuptake inhibitors, proton pump inhibitors, and statins; and preoperative laboratory values including creatinine, platelets, hemoglobin, and hematocrit.
The following procedural data were recorded: procedure date, assistive devices used, any intracranial hemorrhage within 30 days, any symptomatic thromboembolic event within 30 days (transient ischemic attack (TIA) or stroke referable to the arterial territory distal to the treated aneurysm). Strokes were classified as non-disabling if the mRS was ≤2 and major if the mRS was >2 at the time of the 6-month follow-up.
The rate of postoperative conversion to clopidogrel hyper-response and the degree of postoperative hyper-response were compared between initial hyper-responders and patients with in-range PRU at the first and second tests using Fisher's exact test and Student's t-test, respectively. Univariate analysis and multivariate logistic regression analysis were performed to identify predictors of postoperative hyper-response in our cohort. All calculations were performed using MedCalc V.16.1 (Ostend, Belgium). A p value of <0.05 was considered statistically significant.
Patients and devices
Between 22 February 2013 and 1 December 2014, 103 patients (80 women) of mean age 57 years (range 32–81) were started on dual antiplatelet therapy in anticipation of elective endovascular intracranial aneurysm embolization. Five of the women were counted twice due to starting dual antiplatelet therapy on two separate occasions for two different procedures. As such, there were 97 distinct individuals.
Within our cohort of 103 patients there were 26 who underwent placement of a flow diverter, 21 who underwent stent-assisted coil embolization, and 56 who did not receive a device in the parent artery (ie, balloon-assisted or primary coiling).
The first VerifyNow test was performed after a mean, median and mode of 10 daily doses (range 6–17) of clopidogrel 75 mg. Overall, PRU values ranged from 0 to 315, mean (SD) 118 (70). Seventy-four patients (71.8%) were within the target range (PRU 60–240). Of the 29 patients (28.2%) who were outside the target range, 26 (25.2%) were below the target range (PRU <60) and 3 (2.9%) were above the target range (PRU >240).
Of the 26 patients below the target range, dosing was decreased by one step in 11 patients, two steps in nine patients, and three steps in six patients, according to tables 1 and 2. Dosing was increased by one step in all three patients above the target range.
The second VerifyNow test was performed after a mean, median and mode of 7 days (range 3–10) after the first test. Overall, PRU values ranged from 41 to 261, mean (SD) 148 (50). Sixty-five of the 74 patients (87.8%) who were in range at the first test remained in range at the second test. Of the nine patients (12.2%) who were initially in range but no longer in range at the second test, seven (77.8%) became hyper-responders (PRU <60; highest first PRU among these was 106) and two (22.2%) became hypo-responders (PRU >240; first PRUs in these patients were 225 and 235).
Of the 29 patients who were initially out of range, 27 patients (93.1%) were in range at the second test after one dose change, 24 of 26 patients (92.3%) who had been below target PRU and all three patients who had been above target PRU. Overall, 92 patients (89.3%) were in range at the end of the two tests.
Of the 11 patients out of range at the end of the two tests (9 below target PRU and 2 above target PRU), dosing was decreased by one step in six patients, two steps in one patient, and three steps in two patients, according to tables 1 and 2. Dosing was increased by one step in both patients above the target range.
Of the 11 patients (10.7%) who were out of range at the second test, five had their procedures delayed due to hyper-response. These PRUs were 1, 2, 10, 51, and 52. The two patients with the lowest PRUs (1 and 2) started out as hyper-responders (PRUs 20 and 0 at the first test, respectively). The other three patients were initially in range (PRUs 81, 87, and 87). All three patients made it into the target range by the third test. The longest time from the start of dual antiplatelet therapy to the procedure was 33 days. The other six patients who were out of range at the second test (4 hyper-responders and 2 hypo-responders) did not have their procedures delayed; none were flow diversion procedures.
A postoperative test was performed in 82 of 103 patients (79.6%) at a mean of 11 days after the procedure (median 9 days, range 1–34 days). Fifty-one of the 82 patients (62.2%) converted into hyper-responders (figure 1) and three of the 82 patients (3.7%) converted into hypo-responders. Twenty-eight patients (34.1%) remained in range.
Of the 51 postoperative hyper-responders, 37 (72.5%) went on to have changes in their regimen per our protocol with additional P2Y12 inhibition testing. Of the three postoperative hypo-responders, two (66.7%) had changes in their regimen with additional P2Y12 inhibition testing. Seven of the 39 patients (17.9%) who were out of range with additional testing remained out of range after two additional postoperative tests (5 hyper-responders and 2 hypo-responders).
Relationship between preoperative and postoperative clopidogrel hyper-response
Of the 26 patients with initial hyper-response (PRU <60 on the starting dose of clopidogrel 75 mg daily), 19 (73.1%) had a postoperative PRU available. Twelve of the 19 patients (63.2%) had postoperative conversion to hyper-response after being in range preoperatively. Of the 65 patients with in-range PRU of 60–240 after both 10-day and 17-day tests, 57 patients (87.7%) had a postoperative PRU available. Thirty-five of the 57 patients (61.4%) had a postoperative conversion to hyper-response. Initial hyper-responders were not significantly more likely than patients who were in range at both 10 and 17 days to have a postoperative conversion to hyper-response (p>0.05).
Relationship between preoperative hyper-response and degree of postoperative hyper-response
In those who converted to hyper-response (51 of 82 patients (62.2%) with postoperative testing), the mean postoperative PRU was 14 among initial hyper-responders (n=12, 23.5%) and 18 among those with in-range PRU at 10 and 17 days (n=35, 68.6%). This difference was not significant (p>0.05).
Identifying predictors of postoperative conversion to hyper-response
In univariate analysis and a multivariate logistic regression analysis of all patient data recorded (age; sex; weight; body mass index; presence or absence of smoking, hypertension and diabetes; perioperative usage of medications that could interfere with clopidogrel metabolism including steroids, selective serotonin reuptake inhibitors, proton pump inhibitors and statins; preoperative laboratory values including creatinine, platelets, hemoglobin and hematocrit) there were no independent predictors of postoperative conversion to hyper-response.
Of the five patients who were counted twice for two separate procedures, only one converted to hyper-response both times, three converted once only, and one had a postoperative PRU for only one procedure.
Thromboembolic and hemorrhagic neurological complications
There were five non-disabling strokes (mRS ≤2) in the 30-day postoperative period (4.9%). Additionally, there were two TIAs in the 30-day postoperative period. There were no major strokes (mRS >2). All of the thromboembolic complications were in patients with in-range preoperative PRUs. Two of the non-disabling strokes and one TIA occurred in patients among the 26 initial hyper-responders (3/26 11.5% thromboembolic event rate). Three of the non-disabling strokes and one TIA occurred in patients among the 74 patients who were in range at the first test (4/74, 5.4% thromboembolic event rate). This difference was not significantly different (p=0.372). Of the five patients who had a stroke, four of them actually had postoperative hyper-response and one was in range. There were no thromboembolic complications related to postoperative hypo-response.
There was one intracranial hemorrhage, a delayed ipsilateral intraparenchymal hematoma resulting in an mRS 2, in the 30-day postoperative period. This was in a patient who was in range preoperatively (PRU 211) on every-other-day clopidogrel dosing, but received an extra dose of clopidogrel outside the protocol on the first postoperative day resulting in hyper-response at the time of hemorrhage 2 days after the procedure (PRU 26). There was one case of a retinal hemorrhage in a hyper-responder at the preoperative test. There were no other hemorrhagic complications including systemic complications such as gastrointestinal bleeding or major groin hematoma requiring transfusion. There were no deaths.
Our data demonstrate the efficacy of the presented 17-day two-test protocol in reducing the number of patients outside the preoperative target PRU range. The data also confirm the previously described wide and dynamic variability in response to clopidogrel therapy.3 Initial PRU values in this study ranged from 0 to 315, with 28.2% of patients outside the target PRU range of 60–240. The number of patients outside the target range dropped to 10.3% by the time of their procedure following the protocol, meaning that almost nine in 10 patients do not require rescheduling.
After the procedure, nearly two-thirds of patients converted to hyper-response (PRU <60) in our population. Having out-of-range preoperative or postoperative PRU was not associated with a higher complication rate. We hypothesize that this is due to the active management of the PRU limiting the amount of time a patient is exposed to out-of-range platelet inhibition.
Multiple studies have shown an association between antiplatelet resistance and thromboembolic complications in the setting of coronary and carotid stenting.7–12 Specifically in patients undergoing neurointerventional procedures, hypo-response and hyper-response to clopidogrel have been associated with thromboembolic and hemorrhagic complications, respectively.3–5 ,13 A recent randomized trial showed that a modified antiplatelet regimen for patients with high on-treatment platelet reactivity reduced the thromboembolic event rate in coiling unruptured aneurysms from 11.1% to 1.6% without increasing bleeding.14
We perform preoperative P2Y12 inhibition testing using the VerifyNow assay on every patient undergoing elective aneurysm embolization. Our practice evolved to the rescheduling of most patients outside the target PRU range of 60–240 in order to minimize the likelihood of a thromboembolic or hemorrhagic complication. As such, having an effective testing protocol that reduces the disruption caused by the rescheduling of cases is important.
As our experience with P2Y12 inhibition testing has grown, we now strive to reach the optimal ADAPT-DES10 range of 95–207 up to postoperative day 30 in order to minimize the risk of perioperative complications (table 3). Planned flow diversion and stent-assisted coiling procedures are rescheduled if the preoperative PRU is >207 or <60. While the present study has shown that strokes can still occur despite aggressive management of antiplatelet therapy, there were no major strokes in this series and controlling clopidogrel dosing to maintain a defined target PRU range may minimize stroke severity.
Both hemorrhagic complications in the present study occurred in patients with clopidogrel hyper-response at the time of hemorrhage. Although this relationship was not statistically significant, being able to predict who will convert to clopidogrel hyper-response would be useful for the management of antiplatelet dosing in an effort to reduce the risk of hemorrhagic complications. To this end, a multiple logistic regression analysis was performed in an attempt to identify predictors of postoperative hyper-response; however, none was found.
Undetected postoperative conversion to hyper-response is likely a contributing factor to the known risk of delayed parenchymal hemorrhages seen in flow diversion and stent-assisted coil embolization cases.15 We hypothesize that postoperative hyper-response may be induced, possibly related to a postoperative change in hepatic clopidogrel metabolism related to general anesthesia or the physiological stress of the procedure. Hepatic enzymatic changes would explain why the effect is somewhat delayed, long after general anesthetic agents would typically clear. However, further research on this topic is necessary.
It is important to note that many neurointerventionalists do not place patients in whom a primary coiling is planned without an adjunctive device on dual antiplatelet therapy, as we have. This is part of our practice as a way to be prepared for unforeseen challenges during a case that might necessitate a stent—for example, coil prolapse into the parent artery.
Limitations of this study include those inherent in a single-center retrospective review, the limited study population, and the fact that one in five patients did not have a postoperative VerifyNow test. In addition, the testing day was independent of the dosing day for patients not on daily dosing, which may have resulted in peak PRU values in some patients and trough PRU values in others.
A variable response to clopidogrel can impact thromboembolic and hemorrhagic perioperative risk in elective endovascular intracranial aneurysm treatment. To accommodate for the variable response to clopidogrel among our patients, we instituted a two-test protocol using VerifyNow that allows for dose adjustments in order to reach in-range PRU of 60–240 by the second test. This was achieved in nearly nine out of 10 patients, limiting the rate of potentially rescheduled procedures. Nonetheless, nearly two-thirds of the patients who were in range pre-procedure exhibited hyper-response to clopidogrel in post-procedure VerifyNow testing. Out-of-range PRU was not associated with thromboembolic or hemorrhagic neurological complications in this cohort of patients with actively managed P2Y12 inhibition.
The authors acknowledge Sandee K Verootis and Kira Tran, Neurointerventional Radiology, Abbott Northwestern Hospital, for their contributions to data collection for this manuscript.
Twitter Follow Yasha Kayan at @yashakayan
Contributors YK: study design, data collection and analysis, statistical analysis, manuscript preparation, manuscript editing, guarantor of the study. JEDA: study design, data collection and analysis, statistical analysis, manuscript editing. JLF: data collection and analysis, statistical analysis, manuscript editing. AMM and JMS: data collection, manuscript editing. MM: data analysis, manuscript editing.
Competing interests YK is a consultant for Medtronic, Microvention, and Penumbra. JEDA is a consultant for Accriva, Medtronic, Microvention, Penumbra, and Sequent. JLF, AMM, JMS, and MM have nothing to disclose.
Ethics approval Ethics approval was obtained from the Abbott Northwestern Hospital Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.