Abstract
Objective
This study measures effect of CYP2C19 genotype on ischemic stroke risk during clopidogrel therapy for asymptomatic, extracranial carotid stenosis patients.
Methods
Using deidentified electronic health records, patients were selected for retrospective cohort using administrative code for carotid stenosis, availability of CYP2C19 genotype result, clopidogrel exposure, and established patient care. Patients with intracranial atherosclerosis, aneurysm, arteriovenous malformation, prior ischemic stroke, or observation time <1 month were excluded. Dual-antiplatelet therapy patients were included. Patients with carotid endarterectomy or stenting were analyzed in a separate subgroup. Time-to-event analysis using Cox regression was conducted to model ischemic stroke events based on CYP2C19 loss-of-function allele and adjusted with the most predictive covariates from univariate analysis. Covariates included age, gender, race, length of aspirin, length of concurrent antiplatelet/anticoagulant treatment, diabetes, coagulopathy, hypertension, heart disease, atrial fibrillation, and lipid disorder.
Results
Total of 1110 patients met selection criteria for medical therapy cohort (median age 68 [Interquartile range (IQR) 60–75] years, 64.9% male, 91.9% Caucasian). Median study period was 2.8 [0.8–5.3] years. A total of 47 patients (4.2%) had an ischemic stroke event during study period. CYP2C19 loss-of-function allele was strongly associated with ischemic stroke events (One allele: HR 2.3, 95% CI 1.1–4.7, p=0.020; Two alleles: HR 10.2, 95% CI 2.8–36.8, p<0.001) after adjustment.
Conclusions
For asymptomatic carotid stenosis patients receiving clopidogrel to prevent ischemic stroke, CYP2C19 loss-of-function allele is associated with 2 to 10-fold increased risk of ischemic stroke. CYP2C19 genotype may be considered when selecting antiplatelet therapy for stroke prophylaxis in non-procedural, asymptomatic carotid stenosis.
Keywords: CYP2C19, Pharmacogenetics, Ischemic stroke, Carotid stenosis, Clopidogrel
Introduction
Stroke is a leading cause of long-term disability and death in the United States (US). Approximately 6 million Americans have had an ischemic stroke, with an estimated 692,000 acute cases annually [1]. Of these, carotid artery stenosis is attributable to 8–15% of ischemic strokes [2, 3]. Antiplatelet treatment, including clopidogrel, is a mainstay of stroke prophylaxis for patients with carotid stenosis. However, few trials have been able to demonstrate superiority of one antiplatelet agent over another [4].
Pharmacogenomics is an increasingly relevant tool to identify patients that are more likely to benefit from particular medications. CYP2C19 is a liver enzyme in the CYP450 pathway that plays a significant role in the conversion of clopidogrel to its active metabolite. Across different ethnicities, 18–45% of patients have a CYP2C19 loss-of-function (LOF) allele, while 2–35% of patients have a CYP2C19 gain-of-function (GOF) allele [5, 6]. It is well-established that LOF alleles impair the antiplatelet effects of clopidogrel and increase the risk of ischemic cardiovascular events during clopidogrel therapy [7]. The Clinical Pharmacogenetics Implementation Consortium (CPIC) currently recommends switching from clopidogrel following percutaneous coronary interventions [6]. Additionally, it has been reported that the GOF allele may be associated with an increased risk of bleeding during clopidogrel therapy due to overactivation of the antiplatelet effect [7].
In carotid stenosis, CYP2C19 LOF has been linked to a decreased laboratory response of clopidogrel function [8]. For secondary stroke prevention, CYP2C19 genotype has well-documented utility [9–12]. For primary stroke prevention, however, the effect of CYP2C19 genotype is still unclear. Given the considerable mortality and morbidity that develops from a patient’s first stroke, optimizing the efficacy of primary stroke prevention is an important clinical priority [1, 13].
For the primary prevention of ischemic stroke in asymptomatic, extracranial carotid stenosis, this study measures the effect of CYP2C19 genotype on ischemic stroke risk during clopidogrel therapy.
Materials and Methods
Data were obtained from the Synthetic Derivative, a deidentified database of approximately 2.8 million patients derived from our institution’s electronic medical record, for a retrospective cohort study [14]. Patient records from this database were selected if they had 1) Diagnostic code for extracranial carotid artery stenosis without cerebral infarction (ICD-9-CM 433.10; ICD-10-CM I65.21, I65.22, I65.23, I65.29) [15]; 2) CYP2C19 genotyping data from the PREDICT pharmacogenomics program or MEGA genotyping platform [14, 16]; 3) Clopidogrel exposure (two distinct mentions of clopidogrel via MedEx natural language processing) [17]; and 4) Prior, established patient care (patient visit within 1 year to 1 month before start of study period) (Figure 1). This study was approved by the Institutional Review Board (No. 190780) and formal patient consent was waived.
Figure 1:
Inclusion and exclusion criteria with sample sizes during development of medical therapy cohort and procedural cohort.
Patients were excluded if they had diagnostic code for intracranial atherosclerotic disease, aneurysm, or arteriovenous malformation at any point. Patients were additionally excluded if they had diagnostic code for ischemic stroke up to 2 weeks after start of study period or last mention of clopidogrel occurring <1 month after study start (in order to exclude patients with initial inpatient stay but no follow-up) (Figure 1). Dual antiplatelet therapy patients were included. During the study time, CYP2C19 genotype information was not available to the treating physicians during decision-making for carotid stenosis. This information was, however, available during decision-making for treatment of coronary artery disease (CAD).
Primary outcome was ischemic stroke event, defined as clinical or radiological diagnosis made by an attending physician. Potential events were identified using well-established administrative codes for ischemic stroke (ICD-9-CM 433.x1, 434.x1, 436; ICD-10 CM I63.x) and later confirmed or refuted using chart review [18]. The secondary outcome, intracranial hemorrhage event, was defined as clinical or radiological diagnosis of intraparenchymal, subarachnoid, or subdural hemorrhage made by an attending physician. Intraventricular hemorrhage was not included if presented as a solitary diagnosis without accompanying intraparenchymal, subarachnoid, or subdural hemorrhage. Microhemorrhages found on imaging and hemorrhagic conversion from acute ischemic infarct were excluded from this definition. Potential events were identified using well-established administrative codes for intraparenchymal (ICD-9-CM 431; ICD-10-CM I61.x), subarachnoid (ICD-9-CM 430; ICD-10-CM I60.x), and subdural (ICD-9-CM 432.1; ICD-10-CM I62.0x) hemorrhage and later confirmed or refuted using chart review [15, 18].
The primary variable was CYP2C19 genotype, converted into an indicator variable of four values: 1) Two LOF alleles, 2) One LOF allele, 3) Wild-type (control group), 4) GOF. LOF allele was defined as one of the following alleles: CYP2C19*2, *3, *4, *5, *6, *7, or *8. Wild-type was defined as the CYP2C19*1/*1 diplotype. GOF was defined as the diplotype of one GOF allele (*17) with wild-type allele (*1) or of two GOF alleles (i.e. CYP2C19 *1/*17 or CYP2C19*17/*17) [5]. Covariates included age, gender, race, length of aspirin use, length of warfarin use, length of statin use, diabetes mellitus, coagulopathy, hypertension, CAD, congestive heart failure (CHF), peripheral vascular disease, atrial fibrillation, and lipid disorder. Comorbidities were defined using presence of relevant ICD-9/10 codes within one year to one month before start of study period (Supplemental Table 1). The CHADS2 score was estimated for each patient using CHF, hypertension, age greater than or equal to 75 years, history of diabetes, and history of stroke. Additionally, the length of use of direct oral anticoagulant medications was calculated for apixaban, dabigatran, and rivaroxaban. Patient age was defined at start of study period. Race was categorized as Caucasian, African American, Asian, Hispanic, Native American, or other.
Two distinct cohorts were created: medical therapy cohort and procedural cohort. Patients with carotid artery procedure (endarterectomy or stenting) were placed in the procedural cohort, while patients treated with only medical therapy were placed in the medical therapy cohort (Supplemental Table 1). Characteristics of carotid stenosis (symptomatic vs. asymptomatic, stenosis grade) in the medical therapy cohort were estimated via chart review of a randomized sample of 100 patients. Chart review was performed on all procedural cohort patients to confirm asymptomatic characteristics of carotid stenosis, as well as define stenosis grades within population. Clopidogrel exposure time was defined as the time between first mention and 1 month after last mention of clopidogrel. For the medical therapy cohort, start of study period was defined as the first day when carotid stenosis diagnosis and clopidogrel exposure had been established. For the procedural cohort, start of study period was defined as the first day after carotid artery procedure when clopidogrel exposure began. For both cohorts, end of study period was either the day of ischemic stroke event or end of clopidogrel exposure, whichever occurred first. Patient loss to follow-up was defined as the absence of emergency, inpatient, or outpatient visit following the end of study period.
Medical Therapy Cohort Primary Outcome: Ischemic stroke
A time-to-event analysis utilizing multivariate Cox regression was performed to determine effect of CYP2C19 genotype on risk of ischemic stroke event during study period within the medical therapy cohort. The covariates with highest univariate association with ischemic stroke event were included, in order to achieve a minimum of 10 events per variable. As a sensitivity analysis, a second multivariate Cox regression model was created to include all covariates and ischemic stroke as the event. Kaplan-Meier curves representing ischemic stroke event-free rate were generated for each CYP2C19 genotype subset within the medical therapy cohort.
Medical Therapy Cohort Secondary Outcome: Intracranial hemorrhage
A time-to-event analysis utilizing Cox regression was performed to determine effect of CYP2C19 genotype on risk of intracranial hemorrhage events within the medical therapy cohort. Due to low number of events, CYP2C19 genotype was converted into an indicator variable with two possible values: 1) GOF, 2) LOF or Wild-type.
Procedural Cohort
Due to low number of ischemic stroke events in the procedural cohort, a descriptive analysis was performed to determine univariate effect of CYP2C19 genotype on risk of ischemic stroke events.
Statistical Software and Methodology
All analyses and plots were generated using R version 3.6 (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance for each test and model was defined as p<0.05. Univariate analysis methods were Kruskal-Wallis rank sum test and Pearson’s chi-square test. Median and interquartile range (IQR) were provided for variables that did not follow a normal distribution. Standard error was utilized as the measure of variance. Each Cox regression model was tested for proportional hazards assumption violation using Schoenfeld residuals against transformed time. If time dependent coefficients were identified, step functions using visually-identified time segments were utilized to adjust the model.
Results
Demographics
For the medical therapy cohort, a total of 1110 patients met selection criteria (median age 68 [60–75] years, 64.9% male, 91.9% Caucasian, 1.6% Two LOF alleles, 26.8% One LOF allele, 38.6% Wild-type, 33.1% GOF). A total of 47 patients (4.2%) had an ischemic stroke event and 16 patients (1.4%) had an intracranial hemorrhage event. Median study period was 2.8 [0.8–5.3] years with 309 (27.8%) patients lost to follow-up at the end of study period. A total of 16 patients (1.4%) had an ischemic stroke event within year 1, while 32 patients (2.9%) had an event within year 3.5 after study start. The high first-year incidence of ischemic stroke was largely driven by patients with one LOF (n=5; 31.3%) and two LOF (n=3; 18.8%) alleles. Median length of aspirin use was 100% [100–100] of study period. Median length of warfarin use was 0.0% [0.0–7.6] of study period. Median length of statin use was 100% [93–100] of study period (Table 1). Chart review of a randomized sample of 100 medical therapy cohort patients revealed 1/100 (1.0%) patient had symptomatic carotid stenosis of <50%, 6/100 (6.0%) patients had asymptomatic carotid stenosis of ≥60%, and 93/100 (93.0%) had asymptomatic carotid stenosis of <60%.
Table 1.
Baseline demographics for medical therapy cohort, segmented by CYP2C19 genotype
| Overall | GOF allele | Wild-type | One LOF allele | Two LOF alleles | p | |
|---|---|---|---|---|---|---|
| n | 1110 | 367 | 428 | 297 | 18 | |
| Age (years) (median [IQR]) | 68.00 [60.00, 75.00] | 68.00 [59.00, 73.50] | 68.00 [60.00, 75.00] | 68.00 [59.00, 75.00] | 65.00 [57.50, 74.25] | 0.764 |
| Male (%) | 720 (64.9) | 248 (67.6) | 263 (61.4) | 199 (67.0) | 10 (55.6) | 0.199 |
| Race (%) | NaN | |||||
| Caucasian | 1015 (91.9) | 329 (90.4) | 394 (92.5) | 276 (93.2) | 16 (88.9) | |
| African American | 74 (6.7) | 30 (8.2) | 23 (5.4) | 19 (6.4) | 2 (11.1) | |
| Asian | 5 (0.5) | 3 (0.8) | 2 (0.5) | 0 (0.0) | 0 (0.0) | |
| Hispanic | 7 (0.6) | 2 (0.5) | 5 (1.2) | 0 (0.0) | 0 (0.0) | |
| Native American | 3 (0.3) | 0 (0.0) | 2 (0.5) | 1 (0.3) | 0 (0.0) | |
| other | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Time of study period (years) (median [IQR]) | 2.76 [0.83, 5.25] | 2.87 [0.78, 5.27] | 2.41 [0.90, 5.15] | 3.12 [1.02, 5.47] | 1.11 [0.54, 2.33] | 0.035 |
| Loss to follow-up (%) | 309 (27.8) | 113 (30.8) | 112 (26.2) | 80 (26.9) | 4 (22.2) | 0.459 |
| Ischemic stroke event (%) | 47 (4.2) | 11 (3.0) | 12 (2.8) | 21 (7.1) | 3 (16.7) | 0.001 |
| Ischemic stroke event by Year 1 (%) | 16 (1.4) | 6 (1.6) | 2 (0.5) | 5 (1.7) | 3 (16.7) | <0.001 |
| Ischemic stroke event by Year 3.5 (%) | 32 (2.9) | 8 (2.2) | 7 (1.6) | 14 (4.7) | 3 (16.7) | <0.001 |
| Intracranial hemorrhage event (%) | 16 (1.4) | 7 (1.9) | 5 (1.2) | 4 (1.3) | 0 (0.0) | 0.785 |
| Aspirin: percent of study period (median [IQR]) | 100.00 [100.00, 100.00] | 100.00 [100.00, 100.00] | 100.00 [100.00, 100.00] | 100.00 [100.00, 100.00] | 100.00 [100.00, 100.00] | 0.413 |
| Warfarin: percent of study period (median [IQR]) | 0.00 [0.00, 7.55] | 0.00 [0.00, 11.17] | 0.00 [0.00, 2.65] | 0.00 [0.00, 34.74] | 0.00 [0.00, 2.01] | 0.160 |
| Apixaban: percent of study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.446 |
| Dabigatran: percent of study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.545 |
| Rivaroxaban: percent of study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.381 |
| Statin: percent of study period (median [IQR]) | 100.00 [93.03, 100.00] | 100.00 [93.91, 100.00] | 100.00 [91.65, 100.00] | 100.00 [92.22, 100.00] | 100.00 [100.00, 100.00] | 0.364 |
| CHADS2 score (median [IQR]) | 2.00 [1.00, 3.00] | 2.00 [1.00, 3.00] | 2.00 [1.00, 2.00] | 2.00 [1.00, 3.00] | 2.00 [1.25, 2.00] | 0.862 |
| Diabetes mellitus (%) | 449 (40.5) | 147 (40.1) | 183 (42.8) | 111 (37.4) | 8 (44.4) | 0.521 |
| Coagulation disorder (%) | 100 (9.0) | 41 (11.2) | 34 (7.9) | 22 (7.4) | 3 (16.7) | 0.179 |
| Hypertension (%) | 845 (76.1) | 280 (76.3) | 328 (76.6) | 222 (74.7) | 15 (83.3) | 0.827 |
| Coronary artery disease (%) | 957 (86.2) | 316 (86.1) | 369 (86.2) | 256 (86.2) | 16 (88.9) | 0.990 |
| Peripheral vascular disease (%) | 390 (35.1) | 121 (33.0) | 156 (36.4) | 109 (36.7) | 4 (22.2) | 0.438 |
| Congestive heart failure (%) | 421 (37.9) | 146 (39.8) | 159 (37.1) | 111 (37.4) | 5 (27.8) | 0.689 |
| Atrial fibrillation (%) | 233 (21.0) | 78 (21.3) | 81 (18.9) | 70 (23.6) | 4 (22.2) | 0.508 |
| Lipid disorder (%) | 836 (75.3) | 271 (73.8) | 332 (77.6) | 219 (73.7) | 14 (77.8) | 0.561 |
Abbreviations: GOF=Gain-of-function; LOF=Loss-of-function; IQR=Interquartile range. Continuous values reported as median [IQR]. Categorical values reported as number (percent). Statistical tests are Kruskal-Wallis Rank Sum Test and Pearson’s chi-square. Loss-of-function allele was defined as one of the following alleles: CYP2C19*2;*3;*4;*5;*6;*7;or *8.Wild-type was defined as CYP2C19*1/*1. Gain-of-function was defined as the diplotype of one GOF allele (*17) with wild-type allele (*1) or of two GOF alleles (i.e. CYP2C19 *1/*17 or CYP2C19*17/*17).
For the procedural cohort, a total of 84 patients met selection criteria (median age 66 [62–71] years, 69.0% male, 95.2% Caucasian, 2.4% Two LOF alleles, 19.0% One LOF allele, 50.0% Wild-type, 28.6% GOF). A total of 8 patients (9.5%) had an ischemic stroke event and 0 patients had an intracranial hemorrhage event during the study period. Median study period was 2.8 [0.8–5.2] years with 22 (26.2%) patients lost to follow-up at the end of the study period. Median length of aspirin use was 100% [100–100] of study period. Median length of warfarin use was 0.0% [0.0–0.0] of study period. Median length of statin use was 100% [22–100] of study period (Table 2). Chart review of procedural cohort showed that 84/84 (100%) of the population had asymptomatic carotid stenosis. In addition, a total of 81/84 (96.4%) patients had carotid artery stenosis ≥60%, while 3/84 (3.6%) patients had clinically-significant progression of ≥50% stenosis as described by attending physician.
Table 2.
Baseline demographics for procedural cohort, segmented by CYP2C19 genotype
| Overall | GOF allele | Wild-type | One LOF allele | Two LOF alleles | p | |
|---|---|---|---|---|---|---|
| n | 84 | 24 | 42 | 16 | 2 | |
| Age (years) (median [IQR]) | 66.00 [61.75, 71.00] | 64.00 [61.00, 68.00] | 66.50 [62.25, 71.00] | 68.00 [61.75, 71.25] | 74.00 [72.00, 76.00] | 0.246 |
| Male (%) | 58 (69.0) | 17 (70.8) | 28 (66.7) | 11 (68.8) | 2 (100.0) | 0.791 |
| Race (%) | NaN | |||||
| Caucasian | 80 (95.2) | 23 (95.8) | 39 (92.9) | 16 (100.0) | 2 (100.0) | |
| African American | 4 (4.8) | 1 (4.2) | 3 (7.1) | 0 (0.0) | 0 (0.0) | |
| Asian | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Hispanic | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Native American | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| other | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Time of study period (years) (median [IQR]) | 2.84 [0.79, 5.21] | 3.07 [1.46, 5.10] | 3.13 [0.51, 5.62] | 2.45 [1.72, 4.02] | 1.95 [1.03, 2.87] | 0.748 |
| Loss to follow-up (%) | 22 (26.2) | 7 (29.2) | 9 (21.4) | 4 (25.0) | 2 (100.0) | 0.100 |
| Ischemic stroke event (%) | 8 (9.5) | 3 (12.5) | 3 (7.1) | 2 (12.5) | 0 (0.0) | 0.826 |
| Ischemic stroke event by Year 1 (%) | 4 (4.8) | 1 (4.2) | 2 (4.8) | 1 (6.2) | 0 (0.0) | 0.978 |
| Ischemic stroke event by Year 3.5 (%) | 6 (7.1) | 3 (12.5) | 2 (4.8) | 1 (6.2) | 0 (0.0) | 0.666 |
| Intracranial hemorrhage event (%) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | NaN |
| Aspirin: percent of study period (median [IQR]) | 100.00 [100.00, 100.00] | 100.00 [100.00, 100.00] | 100.00 [100.00, 100.00] | 100.00 [100.00, 100.00] | 100.00 [100.00, 100.00] | 0.897 |
| Warfarin: percent of study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 41.82] | 0.00 [0.00, 0.00] | 0.497 |
| Apixaban: percent of study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.715 |
| Dabigatran: percent of study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.801 |
| Rivaroxaban: percent of study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.801 |
| Statin: percent of study period (median [IQR]) | 100.00 [21.57, 100.00] | 97.94 [0.00, 100.00] | 100.00 [99.41, 100.00] | 100.00 [6.93, 100.00] | 0.00 [0.00, 0.00] | 0.025 |
| CHADS2 score (median [IQR]) | 2.00 [1.00, 2.00] | 1.50 [1.00, 2.00] | 2.00 [1.00, 2.00] | 2.00 [1.00, 2.00] | 2.00 [2.00, 2.00] | 0.682 |
| Diabetes mellitus (%) | 32 (38.1) | 8 (33.3) | 18 (42.9) | 5 (31.2) | 1 (50.0) | 0.784 |
| Coagulation disorder (%) | 10 (11.9) | 2 (8.3) | 7 (16.7) | 1 (6.2) | 0 (0.0) | 0.581 |
| Hypertension (%) | 73 (86.9) | 21 (87.5) | 37 (88.1) | 13 (81.2) | 2 (100.0) | 0.847 |
| Coronary artery disease (%) | 72 (85.7) | 19 (79.2) | 37 (88.1) | 14 (87.5) | 2 (100.0) | 0.703 |
| Peripheral vascular disease (%) | 44 (52.4) | 15 (62.5) | 19 (45.2) | 8 (50.0) | 2 (100.0) | 0.296 |
| Congestive heart failure (%) | 24 (28.6) | 6 (25.0) | 12 (28.6) | 6 (37.5) | 0 (0.0) | 0.665 |
| Atrial fibrillation (%) | 17 (20.2) | 5 (20.8) | 8 (19.0) | 4 (25.0) | 0 (0.0) | 0.856 |
| Lipid disorder (%) | 62 (73.8) | 20 (83.3) | 32 (76.2) | 9 (56.2) | 1 (50.0) | 0.223 |
Medical Therapy Cohort Primary Outcome: Ischemic Stroke
Univariate analysis showed significant differences for occurrence of ischemic stroke event (Chi-square=16.3, p<0.001, df=3) and events per year (Kruskal-Wallis chi-square=16.7, p<0.001, df=3) among the four CYP2C19 groups, which strongly inferred an increased occurrence of ischemic stroke events among CYP2C19 loss-of-function groups (One LOF and Two LOF alleles) (Table 1 and Table 3). Multivariate Cox regression showed that CYP2C19 LOF (One allele: HR 2.3, 95% CI 1.1–4.7, p=0.020; Two alleles: HR 10.2, 95% CI 2.8–36.8, p<0.001) was associated with increased risk of ischemic stroke after adjustment (Table 4). The sensitivity analysis model which included all possible covariates yielded nearly identical hazard ratio for the one LOF allele group, and slightly increased hazard ratio for the two LOF group (HR: 11.9, 95% CI 3.2–44.0, p<0.001).
Table 3.
Univariate associations with occurrence of ischemic stroke and intracranial hemorrhage within medical therapy cohort
| No ischemic stroke | Ischemic stroke | p | No intracranial hemorrhage | Intracranial hemorrhage | p | |
|---|---|---|---|---|---|---|
| n | 1063 | 47 | 1094 | 16 | ||
| Age (years) (median [IQR]) | 68.00 [60.00, 75.00] | 64.00 [56.50, 71.00] | 0.051 | 68.00 [60.00, 75.00] | 66.00 [59.00, 72.00] | 0.364 |
| Male (%) | 691 (65.0) | 29 (61.7) | 0.758 | 710 (64.9) | 10 (62.5) | 1.000 |
| Race (%) | NaN | NaN | ||||
| Caucasian | 972 (92.0) | 43 (91.5) | 1001 (92.0) | 14 (87.5) | ||
| African American | 70 (6.6) | 4 (8.5) | 72 (6.6) | 2 (12.5) | ||
| Asian | 5 (0.5) | 0 (0.0) | 5 (0.5) | 0 (0.0) | ||
| Hispanic | 7 (0.7) | 0 (0.0) | 7 (0.6) | 0 (0.0) | ||
| Native American | 3 (0.3) | 0 (0.0) | 3 (0.3) | 0 (0.0) | ||
| other | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | ||
| CYP2C19 (%) | 0.001 | 0.785 | ||||
| GOF | 356 (33.5) | 11 (23.4) | 360 (32.9) | 7 (43.8) | ||
| Wild-type | 416 (39.1) | 12 (25.5) | 423 (38.7) | 5 (31.2) | ||
| One LOF allele | 276 (26.0) | 21 (44.7) | 293 (26.8) | 4 (25.0) | ||
| Two LOF alleles | 15 (1.4) | 3 (6.4) | 18 (1.6) | 0 (0.0) | ||
| Time of study period (years) (median [IQR]) | 2.80 [0.84, 5.26] | 2.14 [0.65, 4.59] | 0.328 | 2.74 [0.83, 5.20] | 4.44 [1.79, 8.76] | 0.098 |
| Aspirin: total years within study period (median [IQR]) | 2.61 [0.73, 5.15] | 2.12 [0.37, 4.54] | 0.369 | 2.45 [0.71, 5.10] | 4.44 [0.71, 8.76] | 0.165 |
| Warfarin: total years within study period (median [IQR]) | 0.00 [0.00, 0.12] | 0.00 [0.00, 0.32] | 0.023 | 0.00 [0.00, 0.12] | 0.25 [0.00, 4.17] | 0.005 |
| Apixaban: total years within study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.674 | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.461 |
| Dabigatran: total years within study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.269 | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.525 |
| Rivaroxaban: total years within study period (median [IQR]) | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.355 | 0.00 [0.00, 0.00] | 0.00 [0.00, 0.00] | 0.156 |
| Statin: total years within study period (median [IQR]) | 1.80 [0.32, 4.68] | 2.12 [0.32, 4.32] | 0.855 | 1.80 [0.32, 4.66] | 2.96 [0.41, 7.63] | 0.225 |
| CHADS2 score (median [IQR]) | 2.00 [1.00, 3.00] | 2.00 [1.00, 2.00] | 0.576 | 2.00 [1.00, 3.00] | 1.50 [1.00, 2.00] | 0.373 |
| Diabetes mellitus (%) | 432 (40.6) | 17 (36.2) | 0.646 | 444 (40.6) | 5 (31.2) | 0.618 |
| Coagulation disorder (%) | 98 (9.2) | 2 (4.3) | 0.367 | 97 (8.9) | 3 (18.8) | 0.352 |
| Hypertension (%) | 809 (76.1) | 36 (76.6) | 1 | 834 (76.2) | 11 (68.8) | 0.688 |
| Coronary artery disease (%) | 917 (86.3) | 40 (85.1) | 0.993 | 944 (86.3) | 13 (81.2) | 0.830 |
| Peripheral vascular disease (%) | 376 (35.4) | 14 (29.8) | 0.53 | 386 (35.3) | 4 (25.0) | 0.554 |
| Congestive heart failure (%) | 403 (37.9) | 18 (38.3) | 1 | 414 (37.8) | 7 (43.8) | 0.823 |
| Atrial fibrillation (%) | 224 (21.1) | 9 (19.1) | 0.894 | 232 (21.2) | 1 (6.2) | 0.250 |
| Lipid disorder (%) | 795 (74.8) | 41 (87.2) | 0.078 | 823 (75.2) | 13 (81.2) | 0.793 |
Table 4.
Cox regression model predicting occurrence of ischemic stroke event during clopidogrel treatment of carotid artery stenosis, within medical therapy cohort
| Variable | Hazard ratio | 95% CI | p |
|---|---|---|---|
| One LOF allele | 2.3 | 1.1–4.7 | 0.020 |
| Two LOF alleles | 10.2 | 2.8–36.8 | <0.001 |
| GOF allele | 0.9 | 0.4–2.2 | 0.898 |
| Warfarin use: total years within study period | 0.9 | 0.8–1.1 | 0.271 |
Concordance=0.598 (se=0.053)
Likelihood ratio test=15.79 on 4 df, p=0.003
Wald test=19.7 on 4 df, p<0.001
Score(logrank) test=25.56 on 4 df, p<0.001”
Medical Therapy Cohort Secondary Outcome: Intracranial Hemorrhage
Univariate analysis showed no significant differences for occurrence of intracranial hemorrhage events among the four CYP2C19 groups (Chi-square=1.07, p=0.785, df=3) (Table 3). Univariate Cox regression failed to establish association of intracranial hemorrhage events with CYP2C19 GOF (HR 1.6, 95% CI 0.6–4.2, p=0.386).
Procedural Cohort Outcome: Ischemic Stroke
Univariate analysis showed no significant differences for occurrence of ischemic stroke events among the four CYP2C19 groups (Chi-square=0.887, p=0.826, df=3) (Table 2).
Discussion
CYP2C19 loss-of-function allele was associated with 2 to 10-fold increased risk (2-fold for one allele, 10-fold for two alleles) of first-time ischemic stroke during clopidogrel treatment of non-surgical, asymptomatic carotid artery stenosis (Figure 2). This increased risk of ischemic stroke was not found among patients procedurally-treated (carotid endarterectomy, carotid stenting) for asymptomatic carotid stenosis with subsequent clopidogrel therapy. Further, no association was found between intracranial hemorrhage events and CYP2C19 genotype among patients on clopidogrel for asymptomatic carotid stenosis.
Figure 2:
Kaplan-Meier curve illustrating first-time ischemic stroke event-free rate for carotid stenosis patients undergoing clopidogrel therapy in medical therapy cohort, segmented by CYP2C19 genotype.
While CYP2C19 LOF has previously been found to be useful for secondary stroke prevention [9–12], this is the first study that shows CYP2C19 utility in primary stroke prevention. A patient’s first stroke is associated with 40% 5-year mortality rate, 65% rate of severe disability, and chronic complications that include neurocognitive decline, seizures, depression, pain syndromes, gait instability, and falls [1, 13]. For patients with carotid stenosis, using CYP2C19 genotype to select the most efficacious antiplatelet medication is therefore critical to prevent this stroke-related mortality and morbidity.
Previous studies have established a synergistic antiplatelet effect between clopidogrel and aspirin [19]. The majority of the medical therapy cohort was undergoing dual antiplatelet therapy, likely due to high rates of concurrent CAD. This suggests that the synergistic effect of aspirin was not sufficient to overcome the lack of clopidogrel efficacy associated with CYP2C19 LOF for the primary prevention of ischemic stroke.
A 2016 study found that LOF patients on clopidogrel therapy for carotid artery stenting after ischemic stroke were at increased risk for recurrent ischemic stroke [20]. In contrast to the previous measurements of recurrent stroke risk, our study investigated the significance of LOF for first-time stroke prevention within the procedural cohort. However, this analysis failed to show such an association. Notably, the procedural cohort was characterized by higher stenotic grades compared to the medical therapy cohort. This is likely the cause of increased ischemic stroke in the procedural cohort compared to the medical therapy cohort, and may also entail that CYP2C19 genotype is less predictive for patients with higher stenotic grades. However, the low occurrence of ischemic stroke events (n=8) in this cohort necessitates future studies with increased sample size to reduce risk of type II error.
Our analysis was unable to find that CYP2C19 genotype was associated with an increased risk of intracranial hemorrhage. Physiologically, it may be expected that CYP2C19 GOF predisposes to adverse bleeding effects from over-activation of clopidogrel, as previously demonstrated for coronary stent placement [21]. Within the percutaneous neurointervention setting, mixed findings showing both increased bleeding events and ischemic events have been reported [22, 23]. The lack of association in this study may represent type II error: there were only 16 intracranial hemorrhage events in our population, inviting a future study with larger sample size.
While the utility of CYP2C19 genotyping is controversial [24], CYP2C19 genotype data is becoming more readily-available in electronic medical records owing to its utilization in CAD [25, 26]. With accessible CYP2C19 data, clinical support tools are being developed. The PREDICT pharmacogenomics program was a successful mechanism to guide genotype-specific antiplatelet recommendations for CAD [16]. Similar processes have been employed at other institutions showing both clinical feasibility and fidelity [26, 27]. For primary ischemic stroke prevention in asymptomatic carotid stenosis, medical centers should utilize CYP2C19 clinical support tools to guide antiplatelet selection.
While our study found an important longitudinal association between CYP2C19 and primary ischemic stroke risk, there are several limitations. Loss to follow-up is a significant concern, and we attempted to minimize this problem by requiring a prior visit within one year before study start and excluding patients with clopidogrel exposure time <1 month. A related limitation is the possibility of an ischemic stroke or intracranial hemorrhage event occurring at an outside institution. In this case, a positive event may have been missed in our analysis. Confounding medications and medical conditions are additional limitations in retrospective clinical studies. We attempted to exclude major intracranial vascular conditions and adjust for medical confounders in the analysis, but it is important to note the possibility of missed features that may influence results. Notably, patients who had previous remote history of transient ischemic attack (TIA) were not systematically excluded from this study. However, 100% of the procedural cohort and an estimated 99% of the medical therapy cohort were asymptomatic at the time of carotid stenosis diagnosis. In addition, due to the relatively low number of events identified (47 ischemic stroke events) in our analysis, the total number of confounders able to be added as covariates in the primary model was limited to 1. However, a sensitivity analysis model which included all clinically-significant confounders yielded similar associations for the primary variables. Further, the high rates of dual antiplatelet therapy and CAD comorbidities in this population suggests that these medications may have been prescribed for previous coronary stent or bypass procedures. While the clopidogrel prescription for these patients may not have been initially intended to prevent stroke, the study window was designed to analyze exposure to this medication after carotid stenosis diagnosis. Nonetheless, the generalizability of these results may be limited to patients with higher overall cardiovascular risk. Dose data were missing in this analysis due to unreliable charting in the data source. However, as clopidogrel’s dosing guidelines are largely limited to 75 mg per day, there is likely minimal dosing variation [4]. Furthermore, administrative data are imperfect, creating risk of misclassifications, missing data, and/or false positive events. For these reasons, our inclusion/exclusion criteria were chosen to be specific, rather than sensitive, and our event identification methodology included a manual chart review step to ensure validity. The procedural cohort and intracranial hemorrhage analyses were limited by low sample size, and may benefit from replication with larger study cohorts. Additionally, the study population had a limited sample of patients with two LOF alleles (n=18, with three ischemic stroke events), was predominantly Caucasian (91%), and was based in the Southeastern US, which limits the generalizability of these results. These analyses should therefore be replicated with other ethnic groups at multiple institutions with an increased number of patients with two LOF alleles.
Conclusions
Among patients undergoing non-surgical treatment for asymptomatic, extracranial carotid artery stenosis with clopidogrel, the CYP2C19 loss-of-function allele is associated with 2 to 10-fold increased risk of suffering a first-time ischemic stroke event. No association was found among carotid endarterectomy or carotid artery stenting patients exposed to clopidogrel. Intracranial hemorrhage was not found to be associated with CYP2C19 genotype. Future prospective studies and clinical trials on asymptomatic carotid stenosis management may be designed to include CYP2C19 genotype. As the push for personalized medicine continues, it will be important for clinicians to capture the full value of genomic data, particularly for patients who have been previously genotyped. While randomized trials are necessary to ascertain the value of genomic testing, the results of this study indicate that CYP2C19 genotyping data may be considered when selecting an antiplatelet medication for prevention of first-time ischemic stroke in asymptomatic carotid artery stenosis patients.
Supplementary Material
Funding:
Xinnan Niu and Joshua Denny were supported by NIH U01HG010232, which partially supported this project.
MEGA Genotyping platform, MEGA dataset(s) used for the analyses described were obtained from Vanderbilt University Medical Center’s BioVU which is supported by numerous sources: institutional funding, private agencies, and federal grants. These include the NIH funded Shared Instrumentation Grant S10RR025141; and CTSA grants UL1TR002243, UL1TR000445, and UL1RR024975. Genomic data are also supported by investigator-led projects that include U01HG004798, R01NS032830, RC2GM092618, P50GM115305, U01HG006378, U19HL065962, R01HD074711; and additional funding sources listed at https://victr.vanderbilt.edu/pub/biovu/.
Footnotes
Declarations
Ethics approval: Study approved by Vanderbilt Institutional Review Board #190780
Consent to participate: Waived due to retrospective design.
Consent for publication: Waived due to retrospective design.
Code availability: N/A
Conflicts of interest/Competing interests: Dr. Rohan V. Chitale receives research grants from Medtronic and Cerenovus. Dr. Josh F. Peterson consults for Color Genomics.
Availability of data and material:
Data available upon request.
References
- 1.Benjamin EJ, Muntner P, Alonso A, Bittencourt MS, Callaway CW, Carson AP et al. Heart Disease and Stroke Statistics-2019 Update: A Report From the American Heart Association. Circulation. 2019;139(10):e56–e528. doi: 10.1161/cir.0000000000000659. [DOI] [PubMed] [Google Scholar]
- 2.Kolominsky-Rabas PL, Weber M, Gefeller O, Neundoerfer B, Heuschmann PU. Epidemiology of ischemic stroke subtypes according to TOAST criteria: incidence, recurrence, and long-term survival in ischemic stroke subtypes: a population-based study. Stroke. 2001;32(12):2735–40. doi: 10.1161/hs1201.100209. [DOI] [PubMed] [Google Scholar]
- 3.Flaherty ML, Kissela B, Khoury JC, Alwell K, Moomaw CJ, Woo D et al. Carotid artery stenosis as a cause of stroke. Neuroepidemiology. 2013;40(1):36–41. doi: 10.1159/000341410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wabnitz AM, Turan TN. Symptomatic Carotid Artery Stenosis: Surgery, Stenting, or Medical Therapy? Curr Treat Options Cardiovasc Med. 2017;19(8):62-. doi: 10.1007/s11936-017-0564-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Dean L. Clopidogrel therapy and CYP2C19 genotype. Medical Genetics Summaries [Internet]. National Center for Biotechnology Information (US); 2018.
- 6.Scott SA, Sangkuhl K, Stein CM, Hulot JS, Mega JL, Roden DM et al. Clinical Pharmacogenetics Implementation Consortium guidelines for CYP2C19 genotype and clopidogrel therapy: 2013 update. Clinical pharmacology and therapeutics. 2013;94(3):317–23. doi: 10.1038/clpt.2013.105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Klein MD, Lee CR, Stouffer GA. Clinical outcomes of CYP2C19 genotype-guided antiplatelet therapy: existing evidence and future directions. Pharmacogenomics. 2018;19(13):1039–46. doi: 10.2217/pgs-2018-0072. [DOI] [PubMed] [Google Scholar]
- 8.Bačković D, Ignjatović S, Rakićević L, Kusić-Tišma J, Radojković D, Čalija B et al. Influence of Cyp2c19*2 Gene Variant on Therapeutic Response During Clopidogrel Treatment in Patients with Carotid Artery Stenosis. Journal of medical biochemistry. 2016;35(1):26–33. doi: 10.1515/jomb-2015-0009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang Y, Zhao X, Lin J, Li H, Johnston SC, Lin Y et al. Association Between CYP2C19 Loss-of-Function Allele Status and Efficacy of Clopidogrel for Risk Reduction Among Patients With Minor Stroke or Transient Ischemic Attack. Jama. 2016;316(1):70–8. doi: 10.1001/jama.2016.8662. [DOI] [PubMed] [Google Scholar]
- 10.Pan Y, Chen W, Xu Y, Yi X, Han Y, Yang Q et al. Genetic Polymorphisms and Clopidogrel Efficacy for Acute Ischemic Stroke or Transient Ischemic Attack: A Systematic Review and Meta-Analysis. Circulation. 2017;135(1):21–33. doi: 10.1161/circulationaha.116.024913. [DOI] [PubMed] [Google Scholar]
- 11.Tornio A, Flynn R, Morant S, Velten E, Palmer CNA, MacDonald TM et al. Investigating Real-World Clopidogrel Pharmacogenetics in Stroke Using a Bioresource Linked to Electronic Medical Records. Clin Pharmacol Ther. 2018;103(2):281–6. doi: 10.1002/cpt.780. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Yi X, Lin J, Zhou J, Wang Y, Huang R, Wang C. The secondary prevention of stroke according to cytochrome P450 2C19 genotype in patients with acute large-artery atherosclerosis stroke. Oncotarget. 2018;9(25):17725–34. doi: 10.18632/oncotarget.24877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Adamson J, Beswick A, Ebrahim S. Is stroke the most common cause of disability? Journal of Stroke and Cerebrovascular Diseases. 2004;13(4):171–7. doi: 10.1016/j.jstrokecerebrovasdis.2004.06.003. [DOI] [PubMed] [Google Scholar]
- 14.Roden D, Pulley J, Basford M, Bernard G, Clayton E, Balser J et al. Development of a Large-Scale De-Identified DNA Biobank to Enable Personalized Medicine. Clinical Pharmacology & Therapeutics. 2008;84(3):362–9. doi: 10.1038/clpt.2008.89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Andrade SE, Harrold LR, Tjia J, Cutrona SL, Saczynski JS, Dodd KS et al. A systematic review of validated methods for identifying cerebrovascular accident or transient ischemic attack using administrative data. Pharmacoepidemiology and drug safety. 2012;21 Suppl 1(Suppl 1):100–28. doi: 10.1002/pds.2312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Pulley JM, Denny JC, Peterson JF, Bernard GR, Vnencak-Jones CL, Ramirez AH et al. Operational Implementation of Prospective Genotyping for Personalized Medicine: The Design of the Vanderbilt PREDICT Project. Clinical Pharmacology & Therapeutics. 2012;92(1):87–95. doi: 10.1038/clpt.2011.371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Xu H, Stenner S, Doan S, Johnson K, Waitman L, Denny J. MedEx: a medication information extraction system for clinical narratives. J Am Med Inform Assoc. 2010. Jan-Feb;17(1):19–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.McCormick N, Bhole V, Lacaille D, Avina-Zubieta JA. Validity of Diagnostic Codes for Acute Stroke in Administrative Databases: A Systematic Review. PLoS One. 2015;10(8):e0135834. doi: 10.1371/journal.pone.0135834. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Johnston SC, Easton JD, Farrant M, Barsan W, Conwit RA, Elm JJ et al. Clopidogrel and Aspirin in Acute Ischemic Stroke and High-Risk TIA. New England Journal of Medicine. 2018;379(3):215–25. doi: 10.1056/NEJMoa1800410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zhu W-Y, Zhao T, Xiong X-Y, Li J, Wang L, Zhou Y et al. Association of CYP2C19 Polymorphisms with the Clinical Efficacy of Clopidogrel Therapy in Patients Undergoing Carotid Artery Stenting in Asia. Scientific Reports. 2016;6:25478. doi: 10.1038/srep25478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Umemura k, Furuta t, Kondo k. The common gene variants of CYP2C19 affect pharmacokinetics and pharmacodynamics in an active metabolite of clopidogrel in healthy subjects. Journal of Thrombosis and Haemostasis. 2008;6(8):1439–41. doi: 10.1111/j.1538-7836.2008.03050.x. [DOI] [PubMed] [Google Scholar]
- 22.Saiz-Rodríguez M, Romero-Palacián D, Villalobos-Vilda C, Caniego JL, Belmonte C, Koller D et al. Influence of CYP2C19 Phenotype on the Effect of Clopidogrel in Patients Undergoing a Percutaneous Neurointervention Procedure. Clinical Pharmacology & Therapeutics. 2019;105(3):661–71. doi: 10.1002/cpt.1067. [DOI] [PubMed] [Google Scholar]
- 23.Lin M, Todaro M, Chan J, Churilov L, Zhu WS, Ramdave S et al. Association between CYP2C19 Polymorphisms and Outcomes in Cerebral Endovascular Therapy. AJNR American journal of neuroradiology. 2016;37(1):108–13. doi: 10.3174/ajnr.A4481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Fuster V, Sweeny JM. Clopidogrel and the reduced-function CYP2C19 genetic variant: a limited piece of the overall therapeutic puzzle. Jama. 2010;304(16):1839–40. doi: 10.1001/jama.2010.1566. [DOI] [PubMed] [Google Scholar]
- 25.Cavallari LH, Lee CR, Beitelshees AL, Cooper-DeHoff RM, Duarte JD, Voora D et al. Multisite Investigation of Outcomes With Implementation of CYP2C19 Genotype-Guided Antiplatelet Therapy After Percutaneous Coronary Intervention. JACC Cardiovascular interventions. 2018;11(2):181–91. doi: 10.1016/j.jcin.2017.07.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lee CR, Sriramoju VB, Cervantes A, Howell LA, Varunok N, Madan S et al. Clinical Outcomes and Sustainability of Using CYP2C19 Genotype-Guided Antiplatelet Therapy After Percutaneous Coronary Intervention. Circ Genom Precis Med. 2018;11(4):e002069-e. doi: 10.1161/CIRCGEN.117.002069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Cavallari LH, Franchi F, Rollini F, Been L, Rivas A, Agarwal M et al. Clinical implementation of rapid CYP2C19 genotyping to guide antiplatelet therapy after percutaneous coronary intervention. Journal of Translational Medicine. 2018;16(1):92. doi: 10.1186/s12967-018-1469-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data available upon request.