Author + information
- Received May 22, 2017
- Revision received July 7, 2017
- Accepted July 11, 2017
- Published online January 15, 2018.
- Larisa H. Cavallari, PharmDa,∗ (, )
- Craig R. Lee, PharmD, PhDb,
- Amber L. Beitelshees, PharmD, MPHc,
- Rhonda M. Cooper-DeHoff, PharmD, MSa,d,
- Julio D. Duarte, PharmD, PhDe,∗,
- Deepak Voora, MDf,
- Stephen E. Kimmel, MDg,
- Caitrin W. McDonough, PhDa,
- Yan Gong, PhDa,
- Chintan V. Dave, PharmD, PhDh,
- Victoria M. Pratt, PhDi,
- Tameka D. Alestock, MSc,
- R. David Anderson, MDd,
- Jorge Alsip, MDj,
- Amer K. Ardati, MDk,
- Brigitta C. Brott, MDj,
- Lawrence Brown, MDl,
- Supatat Chumnumwat, PharmDe,
- Michael J. Clare-Salzler, MDm,
- James C. Coons, PharmDn,
- Joshua C. Denny, MD, MSo,
- Chrisly Dillon, MDp,
- Amanda R. Elsey, MHAa,
- Issam S. Hamadeh, PharmDa,†,
- Shuko Harada, MDq,
- William B. Hillegass, MDr,
- Lindsay Hines, PhDs,
- Richard B. Horenstein, MDc,‡,
- Lucius A. Howell, MDt,
- Linda J.B. Jeng, MD, PhDc,
- Mark D. Kelemen, MDc,
- Yee Ming Lee, PharmDe,§,
- Oyunbileg Magvanjav, MAa,
- May Montasser, PhDc,
- David R. Nelson, MDu,
- Edith A. Nutescu, PharmD, MSe,v,
- Devon C. Nwaba, MPHc,
- Ruth E. Pakyz, MSc,
- Kathleen Palmer, BSNc,
- Josh F. Peterson, MD, MPHo,
- Toni I. Pollin, MS, PhDc,
- Alison H. Quinn, PharmDe,
- Shawn W. Robinson, MDc,l,
- Jamie Schub, MDc,
- Todd C. Skaar, PhDw,
- D. Max Smith, PharmDa,
- Vindhya B. Sriramoju, MDt,
- Petr Starostik, MDm,
- Tomasz P. Stys, MDx,
- James M. Stevenson, PharmD, MSn,
- Nicholas Varunok, MSt,
- Mark R. Vesely, MDc,l,
- Dyson T. Wake, PharmDa,‖,
- Karen E. Weck, MDy,
- Kristin W. Weitzel, PharmDa,
- Russell A. Wilke, MDx,
- James Willig, MDj,
- Richard Y. Zhao, PhDz,
- Rolf P. Kreutz, MDw,
- George A. Stouffer, MDt,
- Philip E. Empey, PharmD, PhDn,
- Nita A. Limdi, PharmD, PhDaa,
- Alan R. Shuldiner, MDc,
- Almut G. Winterstein, PhDh,bb,
- Julie A. Johnson, PharmDa,d,
- on behalf of the IGNITE Network
- aDepartment of Pharmacotherapy and Translational Research, University of Florida, Gainesville, Florida
- bDivision of Pharmacotherapy and Experimental Therapeutics, Eshelman School of Pharmacy and McAllister Heart Institute, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
- cDepartment of Medicine, University of Maryland, Baltimore, Maryland
- dDepartment of Medicine, Division of Cardiovascular Medicine, University of Florida, Gainesville, Florida
- eDepartment of Pharmacy Practice, University of Illinois at Chicago College of Pharmacy, Chicago, Illinois
- fDepartment of Medicine, Center for Applied Genomics & Precision Medicine, Duke University, Durham, North Carolina
- gUniversity of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
- hDepartment of Pharmaceutical Outcomes and Policy, University of Florida, Gainesville, Florida
- iDepartment of Medical and Molecular Genetics, Indiana University School of Medicine, Indianapolis, Indiana
- jDivision of Cardiovascular Sciences, Department of Medicine, School of Medicine, University of Alabama at Birmingham, Birmingham, Alabama
- kDepartment of Medicine, University of Illinois at Chicago College of Medicine, Chicago, Illinois
- lVeterans Administration Medical Center, Baltimore, Maryland
- mDepartment of Pathology, Immunology and Laboratory Medicine, University of Florida, Gainesville, Florida
- nDepartment of Pharmacy and Therapeutics, Center for Clinical Pharmaceutical Sciences, University of Pittsburgh School of Pharmacy, Pittsburgh, Pennsylvania
- oDepartments of Biomedical Informatics and Medicine, Vanderbilt University Medical Center, Nashville, Tennessee
- pDepartment of Neurology, School of Medicine, University of Alabama at Birmingham, Birmingham, Alabama
- qDepartment of Pathology and Hugh Kaul Personalized Medicine Institute, School of Medicine, University of Alabama at Birmingham, Birmingham, Alabama
- rHeart South Cardiovascular Group, Department of Biostatistics, School of Public Health, University of Alabama at Birmingham, Birmingham, Alabama
- sDepartment of Neuropsychology, University of North Dakota, Fargo, North Dakota
- tDivision of Cardiology and McAllister Heart Institute, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
- uCollege of Medicine, Division of Gastroenterology, Hepatology, and Nutrition, University of Florida, Gainesville, Florida
- vDepartment of Pharmacy Systems, Outcomes and Policy and Center for Pharmacoepidemiology and Pharmacoeconomic Research, University of Illinois at Chicago College of Pharmacy, Chicago, Illinois
- wDepartment of Medicine, Krannert Institute of Cardiology & Division of Clinical Pharmacology, Indiana University School of Medicine, Indianapolis, Indiana
- xDepartment of Medicine, University of South Dakota, Sanford School of Medicine, Sioux Falls, South Dakota
- yDepartment of Pathology and Laboratory Medicine, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
- zDepartment of Pathology, University of Maryland School of Medicine, Baltimore, Maryland
- aaDepartment of Neurology and Hugh Kaul Personalized Medicine Institute, School of Medicine, University of Alabama at Birmingham, Birmingham, Alabama
- bbDepartment of Epidemiology, Colleges of Medicine and Public Health and Health Professions, University of Florida, Gainesville, Florida
- ↵∗Address for correspondence:
Dr. Larisa H. Cavallari, Department of Pharmacotherapy and Translational Research, University of Florida, 1333 Center Drive, P.O. Box 100486, Gainesville, Florida 32610.
Objectives This multicenter pragmatic investigation assessed outcomes following clinical implementation of CYP2C19 genotype–guided antiplatelet therapy after percutaneous coronary intervention (PCI).
Background CYP2C19 loss-of-function alleles impair clopidogrel effectiveness after PCI.
Methods After clinical genotyping, each institution recommended alternative antiplatelet therapy (prasugrel, ticagrelor) in PCI patients with a loss-of-function allele. Major adverse cardiovascular events (defined as myocardial infarction, stroke, or death) within 12 months of PCI were compared between patients with a loss-of-function allele prescribed clopidogrel versus alternative therapy. Risk was also compared between patients without a loss-of-function allele and loss-of-function allele carriers prescribed alternative therapy. Cox regression was performed, adjusting for group differences with inverse probability of treatment weights.
Results Among 1,815 patients, 572 (31.5%) had a loss-of-function allele. The risk for major adverse cardiovascular events was significantly higher in patients with a loss-of-function allele prescribed clopidogrel versus alternative therapy (23.4 vs. 8.7 per 100 patient-years; adjusted hazard ratio: 2.26; 95% confidence interval: 1.18 to 4.32; p = 0.013). Similar results were observed among 1,210 patients with acute coronary syndromes at the time of PCI (adjusted hazard ratio: 2.87; 95% confidence interval: 1.35 to 6.09; p = 0.013). There was no difference in major adverse cardiovascular events between patients without a loss-of-function allele and loss-of-function allele carriers prescribed alternative therapy (adjusted hazard ratio: 1.14; 95% confidence interval: 0.69 to 1.88; p = 0.60).
Conclusions These data from real-world observations demonstrate a higher risk for cardiovascular events in patients with a CYP2C19 loss-of-function allele if clopidogrel versus alternative therapy is prescribed. A future randomized study of genotype-guided antiplatelet therapy may be of value.
- antiplatelet therapy
- cardiovascular events
- percutaneous coronary intervention
Treatment with a P2Y12 inhibitor plus aspirin is the standard of care following percutaneous coronary intervention (PCI) (1,2). The P2Y12 inhibitor clopidogrel is a prodrug requiring bioactivation by cytochrome P450 (CYP) 2C19. CYP2C19 loss-of-function (LOF) alleles lead to reduced or absent CYP2C19 activity, lower plasma concentrations of the clopidogrel active metabolite, and reduced inhibition of platelet aggregation during clopidogrel therapy (3,4). Retrospective analyses from randomized clinical trials and patient registries have demonstrated a higher risk for major adverse cardiovascular events (MACE) in clopidogrel-treated patients with versus without a CYP2C19 LOF allele, particularly after PCI (3,5–7).
Prasugrel and ticagrelor are alternative P2Y12 inhibitors, shown to be superior to clopidogrel in preventing cardiovascular events in patients with acute coronary syndrome (ACS) in the TRITON–TIMI 38 (Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel–Thrombolysis in Myocardial Infarction 38) and PLATO (Platelet Inhibition and Patient Outcomes) trials, respectively (8,9). Post hoc genetic analyses from these trials showed no effect of CYP2C19 genotype on outcomes with either prasugrel or ticagrelor (7,10). However, both drugs are more expensive than clopidogrel, which is available generically, and are associated with an increased bleeding risk. Ticagrelor is also associated with more frequent discontinuation because of side effects compared with clopidogrel (8).
Although several institutions have implemented clinical CYP2C19 genotyping to guide antiplatelet therapy selection after PCI (11–14), the effect of this strategy on clinical outcomes is not well defined. Therefore, among patients who underwent PCI and clinical CYP2C19 genotyping, we compared the risk for cardiovascular events between patients with a CYP2C19 LOF allele prescribed clopidogrel 75 mg/day and those with a CYP2C19 LOF allele prescribed alternative antiplatelet therapy. We also compared MACE risk between those with an LOF allele prescribed alternative therapy and those without a CYP2C19 LOF allele treated with any P2Y12 inhibitor.
This was a multicenter investigation of clinical CYP2C19 genotype–guided antiplatelet therapy post-PCI. The study design was pragmatic, on the basis of delivery of the genotype intervention as part of clinical care, the ultimate decision to order genetic testing and choice of drug therapy left to the discretion of the physician, unobtrusive collection of data from the electronic health record (EHR), and the focus on an objectively measured and clinically meaningful outcome (15–17). Seven institutions (the University of Florida; the University of North Carolina, Chapel Hill; the University of Maryland, Baltimore; the University of Alabama, Birmingham; the University of Illinois, Chicago; the University of Pittsburgh; and Indiana University) implemented clinical CYP2C19 genotyping, with results returned via the EHR for consideration during antiplatelet therapy prescribing. All sites participated in the National Institutes of Health–funded Implementing Genomics in Practice Network Pharmacogenetics Working Group and contributed data (18). All patients from each site ≥18 years of age who underwent PCI and CYP2C19 genotyping (per the strategy described in Online Table 1) and received a P2Y12 inhibitor after PCI were included, regardless of length of follow-up. A total of 1,815 patients across the 7 institutions met these criteria and were included in the analysis.
CYP2C19 genotyping and phenotyping
Genotyping was performed at each institution in a Clinical Laboratory Improvement Amendments–licensed laboratory, with the test ordered prior to or at the time of PCI. All sites genotyped for the LOF CYP2C19*2 and CYP2C19*3 alleles, with additional rare alleles genotyped at 5 institutions (Online Table 1). CYP2C19 LOF allele status was defined by presence of at least 1 LOF allele.
CYP2C19 phenotype was assigned similarly across sites on the basis of standardized definitions (19). Patients with 1 or 2 LOF alleles were assigned the intermediate metabolizer (IM) or poor metabolizer (PM) phenotype, respectively. Alternative antiplatelet therapy, consisting of prasugrel or ticagrelor in the absence of contraindications, was recommended for IMs and PMs, according to Clinical Pharmacogenetics Implementation Consortium guidelines (20). Ultimate antiplatelet therapy selection was left to the discretion of the prescriber.
Data abstraction procedures were approved by the institutional review board at each institution. Clinical data at baseline and for up to 12 months following the index PCI, defined as the PCI performed in association with genotyping, were manually abstracted from the EHR through review of patient encounters, including the index PCI hospitalization and subsequent hospitalizations and outpatient visits, using a common data collection form (18). The occurrence of clinical outcomes of interest, including death, myocardial infarction (ST-segment or non–ST-segment elevation myocardial infarction), ischemic stroke, stent thrombosis, and unstable angina occurring over the 12-month period after the index PCI, was determined. The date range for data abstraction was June 2012 through April 2016.
Primary and secondary outcomes
The primary outcome was a major adverse cardiovascular event (MACE), defined as the composite of first occurrence of myocardial infarction, ischemic stroke, or death within 12 months following the index PCI (5). Secondary outcomes were the composite of a MACE plus stent thrombosis and unstable angina and individual cardiovascular events within the MACE definition. Outcomes were identified on the basis of physician-reported diagnoses abstracted from the cardiac catheterization laboratory report, hospital discharge summary notes, or clinical notes in the event of death. Antiplatelet therapy was assessed at the time of event or last follow-up in which P2Y12 inhibitor treatment was documented. The number of days between the index PCI and initiation of alternative antiplatelet therapy was determined in patients with an LOF allele and was available for all but 4 patients.
Data were curated and aggregated at the University of Florida. The time of index PCI was considered time 0. Patients who did not experience a MACE during the 12 months post-PCI were censored at the time of the last EHR-documented follow-up in which treatment with a P2Y12 inhibitor was documented. Event rates were calculated as the number of events divided by follow-up time (time to event or censoring) and expressed as events per 100 patient-years.
Patient characteristics and PCI features at the time of the index PCI were compared between patients with an LOF allele prescribed either clopidogrel 75 mg/day (LOF-clopidogrel group) or alternative antiplatelet therapy (LOF-alternative group) using the Student unpaired t-test, chi-square analysis, or the Fisher exact test as appropriate. Additional comparisons were made between non-LOF patients prescribed clopidogrel or alternative antiplatelet therapy (non-LOF group) and LOF-alternative patients and between clopidogrel-treated versus alternatively treated patients in the non-LOF group. All patients were included in the MACE outcome analyses. A pre-specified secondary analysis limited to patients with ACS indications (ST-segment elevation or non–ST-segment elevation myocardial infarction or unstable angina) at the index PCI was also conducted.
Kaplan-Meier plots were generated to estimate the cumulative risk for an event comparing patients in the LOF-clopidogrel versus LOF-alternative groups and also patients in the non-LOF versus LOF-alternative groups. To adjust for differences between groups, we used logistic regression to estimate the probability (propensity score) of receiving clopidogrel versus alternative antiplatelet therapy using previously reported risk factors for cardiovascular events and study site (2). Propensity scores were estimated separately for each comparison. Stabilized inverse probability of treatment weights (IPTWs) were calculated using the estimated propensity score (21). Covariate balance between groups was assessed by examining the magnitude of any residual differences between groups after applying the weights. Differences were quantified as the weighted standardized differences, for which a threshold of 10% was used to signify a meaningful difference in covariates (21). We also examined the weights themselves to ensure that no observations were overly influential. To compare risk for primary and secondary outcomes, we constructed cause-specific Cox proportional hazard models, weighted by IPTWs. As a sensitivity analysis, we also estimated cluster robust standard errors by accounting for clustering by study site. All statistical analyses were performed in SAS version 9.4 (SAS Institute, Cary, North Carolina).
A total sample size of 1,815 patients, with at least 30% having an LOF allele and 60% of LOF allele carriers receiving alternative therapy, provided >90% power with an alpha level of 0.05 to detect a hazard ratio (HR) of 2.0 for the occurrence of a MACE between the LOF-clopidogrel and -alternative groups.
Patient characteristics are summarized in Table 1. Among 1,815 patients, 1,210 (66.7%) presented with ACS, and 1,794 (98.8%) had stents placed at the time of PCI. The majority of patients received drug-eluting stents (83.6%) and were prescribed aspirin (98.2%) in addition to a P2Y12 inhibitor.
Genotypes and associated phenotypes are shown in Online Table 2. The median time from index PCI to available genotype result was 1 day (interquartile range: 1 to 3 days). LOF alleles were present in 572 patients (31.5%) (Figure 1, Online Table 2); 518 patients (28.5%) were IMs, and 54 patients (3%) were PMs. Alternative antiplatelet therapy was prescribed to a higher proportion of patients with an LOF allele (60.5%) compared with patients without an LOF allele (15.5%; p < 0.0001) (Figure 1). A total of 58% of IMs and 87% of PMs were prescribed alternative antiplatelet therapy (Online Figure 1). Among patients with an LOF allele, the median time from genotype result to initiation of alternative antiplatelet therapy was 1 day (interquartile range: 1 to 6 days).
There were differences between the LOF-clopidogrel and -alternative groups in age, prevalence of diabetes, stroke, peripheral vascular disease, and use of oral anticoagulation (Table 1). Age and prevalence of hypertension and diabetes differed between the non-LOF and LOF-alternative groups. These imbalances were negligible after adjustment with propensity score-derived IPTWs (Table 2).
The median follow-up from index PCI to MACE or censoring was 4.8 months (interquartile range: 0.6 to 9.9 months). The occurrence of a MACE was documented in 108 patients (5.95%) over the follow-up period (event rate of 13.5 per 100 patient-years). There was a higher rate of MACE in the LOF-clopidogrel group (n = 18 events, event rate 23.4 per 100 patient-years) compared with the LOF-alternative group (n = 16 events, event rate 8.7 per 100 patient-years, log-rank p = 0.016) (Table 3, Figure 2). After propensity score adjustment, the risk for a MACE remained significantly higher in the LOF-clopidogrel versus LOF-alternative group (adjusted HR: 2.26; 95% confidence interval [CI]: 1.18 to 4.32; p = 0.013) (Table 3). There was no difference in event rates between the non-LOF and LOF-alternative groups (13.7 vs. 8.7 per 100 patient-years; log-rank p = 0.15; propensity score–adjusted HR: 1.14; 95% CI: 0.69 to 1.88; p = 0.60) (Table 3). Similarly, within the non-LOF group, there was no difference in MACE rates between non-LOF patients treated with clopidogrel versus non-LOF patients treated with alternative therapy after adjusting for clinical differences between groups (adjusted HR: 1.01; 95% CI: 0.52 to 1.94; p = 0.98) (Online Tables 3 and 4). Accounting for within-site clustering did not change the estimates for the primary outcome (Online Table 5).
In the 518 patients with the IM phenotype only, the risk for a MACE was significantly higher with clopidogrel versus alternative antiplatelet therapy (log-rank p = 0.003) (Online Figure 2), with event rates of 24.0 versus 6.7 per 100 patient-years, respectively. Only 7 of 54 PMs (13%) were treated with clopidogrel, precluding analysis of outcomes in this group.
Secondary outcomes are shown in Table 3. The risk for a MACE plus other ischemic events (stent thrombosis and unstable angina) was higher in the LOF-clopidogrel versus LOF-alternative group (adjusted HR: 1.82; 95% CI: 1.07 to 3.12; p = 0.027). There was no difference in the occurrence of a MACE plus ischemic events between non-LOF and LOF-alternative groups (adjusted HR: 1.09; 95% CI: 0.72 to 1.63; p = 0.69) (Table 3) or between non-LOF patients treated with clopidogrel versus alternative therapy (adjusted HR: 1.02; 95% CI: 0.59 to 1.78) (Online Table 4). Patients presenting with ACS at the time of index PCI (1,210 of 1,815 patients) contributed the majority of events to the analysis, including 86 of 108 events in the composite MACE outcome (80%) and 48 of 58 myocardial infarctions (83%). Consistent with the analysis of the overall study population, among patients with ACS, the LOF-clopidogrel group had a higher rate of a MACE compared with the LOF-alternative group (39.0 per 100 patient-years vs. 8.9 per 100 patient-years, respectively; adjusted HR: 2.87; 95% CI: 1.35 to 6.09; p = 0.013) (Online Figure 3, Online Table 6). The risk for a MACE plus other ischemic events (adjusted HR: 2.10; 95% CI: 1.12 to 3.90; p = 0.019) was also significantly elevated. No differences in outcomes were observed between non-LOF and LOF-alternative groups in the ACS subset. Although not an outcome of the study, moderate and severe or life-threatening bleeding events, defined according to the GUSTO (Global Utilization of t-PA and Streptokinase for Occluded Coronary Arteries) criteria, were observed in 2.3% of patients in the overall study population and were similar across groups (22).
On the basis of the prevalence of patients with LOF alleles (31.5%), the number needed to genotype to identify an LOF allele carrier for whom alternative therapy would be recommended was 3.2. With an absolute difference in the proportion of patients who had a MACE in the LOF-clopidogrel (8.0%) and LOF-alternative (4.6%) groups (Table 3), the number of patients with LOF alleles needed to treat with alternative antiplatelet therapy to prevent 1 event was 29 (1/0.034). Therefore, the number of patients needed to genotype, with alternative antiplatelet therapy prescribed for all patients with LOF alleles, to prevent 1 cardiovascular event was 93 (29 × 3.2).
Our study is the first large multicenter study to examine outcomes after clinical implementation of CYP2C19 genotype–guided antiplatelet therapy. We demonstrate the feasibility of genotype-guided antiplatelet therapy after PCI across multiple institutions, with efficient return of genotype results and high uptake of alternative antiplatelet therapy in patients with an LOF allele. More important, our results show a higher risk for a MACE in patients with CYP2C19 LOF alleles who are treated with clopidogrel versus alternative antiplatelet therapy. Most events occurred in patients with ACS indications at the index PCI, in whom the risk for a MACE was higher in LOF-clopidogrel versus LOF-alternative patients.
Retrospective genetic substudies of large clinical trials suggested worse outcomes in patients with CYP2C19 LOF alleles treated with clopidogrel (3,5–7), but there is a paucity of data from large prospective clinical trials on outcomes with CYP2C19 genotype–guided antiplatelet therapy. This has hindered inclusion of CYP2C19 genotyping in clinical practice and PCI practice guidelines. Guidelines state that routine genetic testing is not recommended but might be considered in high-risk patients (Class IIb, Level of Evidence: C) (1,2). A randomized controlled trial assessing the efficacy of genotype-guided antiplatelet therapy in a target population of more than 5,000 patients began in 2013 but is not expected to be completed until 2020 (NCT01742117). Given the magnitude, time, and expense of conducting traditional randomized controlled trials, other approaches are needed to generate an evidence base quantifying clinical outcomes related to pharmacogenetic-tailored therapy. Our pragmatic approach to assessing outcomes following clinical implementation of CYP2C19 genotyping with real-world evidence is one such approach.
Our findings are in line with recent data from the Netherlands and Spain (23,24). The Dutch study focused on patients undergoing elective PCI, who were clinically genotyped, with prasugrel recommended for PMs (23). Fewer adverse cardiovascular events were observed in PMs treated with prasugrel versus clopidogrel. The Spanish study included patients undergoing elective or emergent PCI, and recommendations for alternative therapy were made for both PMs and IMs (24). Compared with historical control subjects without genotyping, fewer adverse events were observed in patients receiving genotype-guided therapy. Our data are also consistent with those from 2 trials of Chinese patients undergoing PCI and randomized to either clopidogrel 75 mg/day or genotype-guided antiplatelet therapy, consisting of high-dose clopidogrel (150 mg/day) for IMs and high-dose clopidogrel plus cilostazol for PMs in 1 trial and ticagrelor for PMs in the other (25,26). Both studies observed significant reductions in cardiovascular events in the genotype-guided arm.
The U.S. Food and Drug Administration–approved clopidogrel label warns of reduced clopidogrel effectiveness in PMs and recommends consideration of alternative antiplatelet therapy in these patients, but is silent regarding risk and recommendations in IMs (27). In our study, conducted in the context of routine clinical care, the majority of PMs (87%) were treated with alternative antiplatelet therapy, consistent with labeling recommendations, but a lower proportion of IMs were treated with alternative antiplatelet therapy (58%). This observation suggests that an IM test result is weighed less heavily than a PM result in the post-PCI antiplatelet therapy prescribing decision. However, when limiting our analysis to IMs, we observed a significantly higher rate of cardiovascular events with clopidogrel versus alternative antiplatelet therapy. These data indicate that IMs, like PMs, are at higher risk for adverse cardiovascular outcomes if treated with clopidogrel and suggest that alternative antiplatelet therapy should be considered in both IMs and PMs.
Our real-world data corroborate those from retrospective analyses of clinical trials showing a higher risk for cardiovascular events among clopidogrel-treated patients with versus without an LOF allele (3,5–7). In the absence of an LOF allele, data from the TRITON–TIMI 38 trial also suggest that the risk for adverse events may be comparable between clopidogrel and prasugrel, with a genetic substudy of the trial showing a relative risk for a MACE of 0.98 (95% CI: 0.80 to 1.20) with prasugrel compared with clopidogrel in patients without an LOF allele (28). The association between genotype and outcomes with clopidogrel appears to be indication specific, with strong and consistent associations in patients undergoing PCI, but not among lower-risk patients, such as those with atrial fibrillation or ACS managed medically (5,7,29,30). This is most clearly demonstrated in a meta-analysis showing an association between CYP2C19 LOF genotype and risk for a MACE when analyzing studies of clopidogrel-treated patients undergoing PCI, but not when analyzing those without PCI (6).
Across institutions in our study, clopidogrel was the most commonly prescribed P2Y12 inhibitor among patients without an CYP2C19 LOF allele. This is consistent with other data suggesting that clopidogrel remains the predominant antiplatelet therapy in the United States, prescribed 60% to 70% of the time according to published data through the first half of 2013 (31,32). To provide more contemporary use data, we interrogated practice patterns at Vanderbilt University Medical Center in the southeastern United States and Sanford Medical Center, which draws patients from 6 states in the Midwest. Current practice patterns in these geographic locales demonstrate that clopidogrel remains commonly prescribed following PCI, regardless of clinical context. Specifically, among 10,115 patients who underwent PCI (41% with ACS) at Vanderbilt University Medical Center from 2010 to 2015, clopidogrel use fell minimally, from 93% to 72%, during that time (personal communication from Josh F. Peterson, 2017). Similarly, among 1,260 patients from Sanford Medical Center who underwent PCI at 15 catheterization laboratories during the first 9 months of 2016, 53% with ACS and 77% without ACS were discharged on clopidogrel (personal communication from Tomasz P. Stys 2017).
Per PCI guidelines, the use of alternative antiplatelet agents in preference to clopidogrel after ACS and PCI is a Class IIa recommendation, on the basis of a moderate quality of evidence (1). In patients started on alternative P2Y12 inhibitors according to PCI guidelines, CYP2C19 genotype may still have an important role in informing therapy, especially after the 30-day post-PCI period when risk for events is highest (1). In patients with a high risk for bleeding or difficulty affording or tolerating newer agents, knowledge that a patient does not carry an LOF allele may give physicians increased confidence when considering switching the patient to the more affordable clopidogrel.
We did not directly assess why physicians chose to start some patients with an LOF allele on alternative therapy but not others. However, some speculation can be made on the basis of prescribing patterns and characteristics of treatment groups. First, the higher use of alternative therapy in PMs versus IMs suggests that physicians may heed the boxed warning in the clopidogrel labeling and place a stronger emphasis on using alternative therapy in PMs, the focus of the boxed warning. Second, the higher prevalence of stroke or transient ischemic attack history and concurrent anticoagulant agent use in the LOF-clopidogrel group versus the LOF-alternative group suggests that patient bleeding risk influenced preference for antiplatelet therapy.
First, there was no control group of patients who did not undergo genotyping. However, some comparisons can be made with published event rates. In particular, in a genetic substudy of the TRITON–TIMI 38 trial of patients with ACS and PCI, 8% of clopidogrel-treated patients without an LOF allele and 12% with an LOF allele had a MACE (3). Rates in our study were comparable, with a MACE occurring in 7% of patients with ACS without an LOF allele (most of whom received clopidogrel) and 11% of clopidogrel-treated patients with an LOF allele.
Second, genotype-guided therapy was not randomized, because of our pragmatic design, and antiplatelet therapy selection was left to physician discretion. Thus, propensity scoring methods were used to mitigate the potential confounding effects related to differences across groups, and balance was achieved across comparison groups after weighting with IPTWs. However, residual confounding regarding the choice of antiplatelet therapy may remain.
Third, because ascertainment of outcomes was confined to the EHR without event adjudication, clinical events, including deaths, may have been missed.
Fourth, length of follow-up was variable, as is typical for the clinical setting. As such, event rates were reported.
Fifth, bleeding events were not objectively and systematically collected, given known difficulties in accurate bleeding assessment and the focus of data collection for this analysis on ischemic outcomes (33).
Finally, because of the limited number of patients who underwent elective PCI, we did not examine outcomes separately in this group.
We demonstrate that implementation of clinical CYP2C19 genotyping to guide post-PCI antiplatelet therapy is feasible across multiple institutions. Our data also demonstrate that cardiovascular outcomes were worse when clopidogrel versus alternative antiplatelet therapy was prescribed after PCI in patients with an LOF allele. The higher risk for a MACE in LOF carriers prescribed clopidogrel was also evident when analyses were confined to patients with ACS and, separately, to IMs (with a single LOF allele). Our data suggest that obtaining genotype data early after PCI allows the identification of patients with a CYP2C19 LOF allele in whom alternative antiplatelet therapy would reduce risk for events. A future randomized study of genotype-guided antiplatelet therapy may be of value.
WHAT IS KNOWN? CYP2C19 LOF alleles impair clopidogrel activation and effectiveness after PCI. However, the impact of genotype-guided antiplatelet therapy on clinical outcomes is not well defined.
WHAT IS NEW? This study demonstrates that patients with 1 or 2 CYP2C19 LOF alleles prescribed an alternative P2Y12 inhibitor after PCI exhibit a lower risk for cardiovascular events compared with patients with 1 or 2 CYP2C19 LOF alleles prescribed clopidogrel.
WHAT IS NEXT? Strategies to more broadly incorporate genotyping into clinical care to inform antiplatelet therapy prescribing decisions after PCI, and to evaluate the impact on health care costs, warrant further investigation.
↵∗ Present address: Department of Pharmacotherapy and Translational Research, University of Florida, Gainesville, Florida.
↵† Present address: Carolinas HealthCare System, Charlotte, North Carolina.
↵§ Present address: University of Chicago Medical Center, Chicago, Illinois.
↵‖ Present address: NorthShore University HealthSystem, Evanston, Illinois.
This work was supported by National Institutes of Health (NIH) grants U01 HG007269 (to Drs. Johnson, Cavallari, Weitzel, Elsey, McDonough, Gong, Cooper-DeHoff, Nelson, Starostik, Smith, Wake, and Skaar), U01 HG007775 (to Drs. Pollin, Horenstein, Jeng, Nelson, Palmer, and Shuldiner), U01 HG007253 (to Drs. Peterson, Wilke, and Denny), and U01 HG007762 (to Dr. Skaar, Pratt, and Kreutz) as part of the NIH IGNITE (Implementing Genomics in Practice) network. Additional support was provided by NIH grants U01 GM074492 and U01 HL105198 (both part of the NIH Pharmacogenomics Research Network); by substantial institutional support from the University of Florida and its Clinical Translational Science Institute (grants UL1 TR000064 and UL1 TR001427 to Drs. Cavallari, Weitzel, Nelson, Clare-Salzler, Wake, Smith, and Johnson); NIH grant U01 HL105198; support from the University of Maryland Medical Center and University of Maryland School of Medicine Program for Personalized and Genomic Medicine (to Drs. Beitelshees, Alestock, Jeng, Brown, Horenstein, Kelemen, Montasser, Nelson, Pakyz, Palmer, Pollin, Robinson, Schub, Vesely, Zhao, and Shuldiner); the University of Illinois at Chicago Offices of the Vice President for Health Affairs and Vice Chancellor for Research (to Drs. Nutescu and Duarte); NIH grant K23 HL112908 (to Dr. Nutescu); NIH grant K23 GM112014 (to Dr. Duarte); the American Society of Health System Pharmacists (to Drs. Empey and Coons); NIH grant UL1TR0000005 (to Dr. Empey); an anonymous donor (to Drs. Empey, Coons, and Stevenson); the Penn Center for Precision Medicine at the Perelman School of Medicine at the University of Pennsylvania to S.T.; NIH grants R01HL092173 and 1K24HL133373, University of Alabama, Birmingham Health Service Foundations’ General Endowment Fund, and NIH grant UL1TR000165 (to Dr. Limdi); and the Indiana University Health – Indiana University School of Medicine Strategic Research Initiative (to Drs. Kreutz and Pratt). Dr. Kimmel has served as a consultant for Pfizer and Bayer, neither related to the content of the study. Dr. Kreutz has received consulting fees and research funding from Haemonetics. Reagents were provided by Nanospere to the University of Maryland. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute coronary syndrome(s)
- confidence interval
- cytochrome P450
- electronic health record
- hazard ratio
- intermediate metabolizer
- inverse probability of treatment weight
- major adverse cardiovascular event(s)
- percutaneous coronary intervention
- poor metabolizer
- Received May 22, 2017.
- Revision received July 7, 2017.
- Accepted July 11, 2017.
- 2018 American College of Cardiology Foundation
- Levine G.N.,
- Bates E.R.,
- Bittl J.A.,
- et al.
- Levine G.N.,
- Bates E.R.,
- Blankenship J.C.,
- et al.
- Sorich M.J.,
- Rowland A.,
- McKinnon R.A.,
- Wiese M.D.
- Mega J.L.,
- Close S.L.,
- Wiviott S.D.,
- et al.
- Peterson J.F.,
- Field J.R.,
- Unertl K.M.,
- et al.
- Ford I.,
- Norrie J.
- Loudon K.,
- Treweek S.,
- Sullivan F.,
- Donnan P.,
- Thorpe K.E.,
- Zwarenstein M.
- Sox H.C.,
- Lewis R.J.
- Cavallari L.H.,
- Beitelshees A.L.,
- Blake K.V.,
- et al.
- Mehran R.,
- Rao S.V.,
- Bhatt D.L.,
- et al.
- Deiman B.A.,
- Tonino P.A.,
- Kouhestani K.,
- et al.
- Sanchez-Ramos J.,
- Davila-Fajardo C.L.,
- Toledo Frias P.,
- et al.
- Shen D.L.,
- Wang B.,
- Bai J.,
- et al.
- Holmes D.R. Jr..,
- Dehmer G.J.,
- Kaul S.,
- Leifer D.,
- O’Gara P.T.,
- Stein C.M.
- Doll J.A.,
- Neely M.L.,
- Roe M.T.,
- et al.
- Sherwood M.W.,
- Wiviott S.D.,
- Peng S.A.,
- et al.
- Fan W.,
- Plent S.,
- Prats J.,
- Deliargyris E.N.