https://orcid.org/0000-0003-3909-182X
Abstract
Carriers of cytochrome 2C19 (CYP2C19) loss-of-function (LoF) alleles treated with clopidogrel have impaired drug metabolism, resulting in reduced active metabolite levels, high platelet reactivity (HPR), and an increased risk of thrombotic events. Several alternative antiplatelet therapies have been proposed to overcome HPR in these patients, but their comparative effects remain poorly explored.
Randomized controlled trials (RCTs) comparing different oral antiplatelet therapies in carriers of CYP2C19 LoF alleles undergoing percutaneous coronary interventions (PCI) were included. A frequentist network meta-analysis was conducted to estimate mean difference (MD) or odds ratios and 95% confidence intervals (CI). The primary outcome was platelet reactivity assessed by VerifyNow and reported as P2Y12 reaction unit (PRU). The secondary outcome was the rate of HPR. Standard dose of clopidogrel (75 mg daily) was used as a reference treatment.
A total of 12 RCTs testing 6 alternative strategies (i.e. clopidogrel 150 mg, prasugrel 3.75 mg, 5 mg, and 10 mg, ticagrelor 90 mg bid, and adjunctive cilostazol 100 mg bid) were included in the network. Compared with standard-dose clopidogrel, the greatest reduction in PRU was observed with prasugrel 10 mg (MD −127.91; 95% CI −141.04; −114.78) and ticagrelor 90 mg bid (MD −124.91; 95% CI −161.78; −88.04), followed by prasugrel 5 mg (MD −76.33; 95% CI −98.01; −54.65) and prasugrel 3.75 mg (MD −73.00; 95% CI −100.28; −45.72). Among other strategies, adjunctive cilostazol (MD −42.64; 95% CI −64.72; −20.57) and high-dose clopidogrel (MD −32.11; 95% CI −51.33; −12.90) were associated with a modest reduction in PRU compared with standard-dose clopidogrel.
Among carriers of CYP2C19 LoF alleles undergoing PCI, standard-dose prasugrel or ticagrelor are most effective in reducing platelet reactivity, while double-dose clopidogrel and additional cilostazol showed modest effects. Reduced-dose of prasugrel may represent a balanced strategy to overcome HPR without a significant increase in bleeding. The clinical implications of these pharmacodynamic findings warrant further investigation.
Abbreviations
DAPT dual antiplatelet therapy
CYP Cytochrome P450
HPR high platelet reactivity
LoF loss-of-function
PCI percutaneous coronary intervention
PFT platelet function test
PD pharmacodynamic
Introduction
Dual antiplatelet therapy (DAPT) with aspirin and a P2Y12 inhibitor is the cornerstone of treatment in patients with acute coronary syndrome (ACS) or undergoing percutaneous coronary intervention (PCI).1 Clopidogrel is the most commonly used P2Y12 receptor inhibitor, but exhibits distinctive pharmacologic features leading to broad interindividual variability in its effects.2 In fact, clopidogrel is a pro-drug that requires a two-step hepatic conversion into its active metabolite by several cytochrome P450 (CYP) isoenzymes.3 Of these, the CYP2C19 isoforms are particularly important because they take part in both biotransformation oxidative processes required to generate clopidogrel's active metabolite.3,4 However, the genes that encode for the CYP2C19 enzymes are highly polymorphic, with some resulting in reduced [i.e. loss-of-function (LoF)] and others increased [i.e. gain-of-function (GoF)] clopidogrel metabolism.5 While the clinical implications of GoF alleles among clopidogrel treated patients remain controversial, it is well established that carriers of one (intermediate metabolizers, IMs) or two (poor metabolizers, PMs) LoF alleles exhibit reduced levels of clopidogrel active metabolite resulting in high platelet reactivity (HPR) and enhanced risk of thrombotic complications.5–8 Notably, carriers of LoF alleles are as frequent as 20–60% of treated patients, varying according to ancestry, underscoring its relevance in clinical practice.8
The association between CYP2C19 LoF alleles and increased rates of thrombotic events among clopidogrel-treated patients has led the Food and Drug Administration as well as other drug governing authorities to issue a cautionary advisory recommending the use of alternative P2Y12 inhibiting therapies in PMs.9,10 Similarly, the 2019 expert consensus statement on platelet function test (PFT) and genetic testing for patients undergoing PCI and the 2022 Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline and other expert consensus suggests the use of alternative P2Y12 inhibitors among carriers of CYP2C19 LoF alleles, particularly PMs.8,11–13 Although the use of ticagrelor or prasugrel is recommended in PM, a number of alternative antiplatelet treatment regimens have been proposed to reduce platelet reactivity and minimize the rate of HPR, but comprehensive studies exploring the comparative effects of different alternative strategies are lacking.14–22 Since levels of platelet reactivity are closely associated with adverse outcomes, pharmacodynamic studies comparing the effects of different alternative strategies on platelet function among CYP2C19 LoF carriers may provide important clinical information.7,23 On this background, we conducted a network meta-analysis to comprehensively assess the comparative effects of different alternative antiplatelet regimens among individuals carrying CYP2C19 LoF alleles, consolidating results from randomized controlled trials (RCTs).
Methods
The study protocol adhered to the recommendations of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) for network meta-analyses (PRISMA-NMA) of RCTs (Supplementary material online, Table S1) and Cochrane Collaboration.24,25 The study protocol was registered in the Prospective Register of Systematic Reviews (PROSPERO ID: CRD42024509513). As the present research was a meta-analysis of published studies, the requirement to obtain approval from an ethics committee was waived.
Eligibility criteria
Studies were considered for inclusion in the network meta-analysis when the following criteria were satisfied: (i) patients were carriers of CYP2C19 LoF alleles (e.g. CYP2C19*2, CYP2C19*3) diagnosed with atherosclerotic cardiovascular disease and undergoing PCI; (ii) random allocation to at least two different antiplatelet strategies; (iii) pharmacodynamic (PD) assessment (i.e. platelet reactivity) with at least one PFT.
Search, data extraction, and qualitative assessment
A systematic digital search using MEDLINE with PubMed interface, Cochrane Library, and Web of Science databases was carried out from database inception to 17 January 2024. Major scientific websites as well as abstracts and presentations from major cardiovascular meetings were also screened. The electronic search process was integrated by exploration of the reference lists of each eligible study (i.e. backward and forward snowballing). The full research strategy is reported in Supplementary material online, Table S2. Literature search terms were reviewed by an experienced medical librarian. No language restrictions or filters were applied.
Following the execution of search queries, two reviewers (R.L., G.O.) independently screened titles and abstracts to identify studies that met the inclusion criteria. Duplicate entries identified across various electronic databases were removed. The remaining reports underwent independent full-text screening by the same reviewers to confirm compliance with the eligibility criteria. Disagreements were solved by consensus under the supervision of the lead investigator (M.G.). Data pertaining to the outcomes of interest, as well as the primary clinical and procedural characteristics, design features, and endpoints definitions were extracted at study level and incorporated into dedicated electronic spreadsheets. Before running the statistical analyses, the reviewers collegially assessed the quality of each trial by using the Cochrane's Risk of Bias (RoB) 2 tool.
Inconsistency was evaluated by means of net heat plots, and node-splitting analysis. Finally, the impact of small-study effects and publication bias was estimated by visual inspection of comparison-adjusted funnel plots, and formally assessed by means of Egger's regression test.
Study endpoints
The pre-specified primary endpoint was the level of platelet reactivity between treatment arms as assessed by the VerifyNow P2Y12 assay and measured as P2Y12 reaction unit (PRU). The secondary endpoint was trial-defined rates of HPR. PD assessments obtained while on maintenance dose therapy were chosen over those following a loading dose. All study outcomes were extracted at the longest available follow-up.
Statistical analysis
A frequentist network meta-analysis was conducted to estimate the posterior mean effect, reported as odds ratio (OR) or mean difference (MD), as appropriate, and the 95% confidence interval (CI), for all the outcomes of interest. An estimated effect was considered significant when the upper or lower CI did not include the unity (i.e. null value). The analysis was conducted using the adjusted random effects weight, calculating the pooled effect size using the inverse variance method.
The network of evidence was visually and numerically assessed in terms of weights, comparisons, and individual trial influence for each outcome. Consistency between direct and indirect comparisons was examined by node splitting and visual inspection of net-heat plots. Results were displayed using forest plots illustrating the relative contribution of individual trials. The standard dose of clopidogrel (75 mg daily) was used as a reference treatment. Antiplatelet strategies were ranked according to P-scores, ranging between 0 (poor performance) and 1 (good performance). Between-study heterogeneity (i.e. statistical heterogeneity) was assessed according to the Cochran's Q test, Higgins and Thompson's I2 statistics (with an I2 value >50% being considered the result of severe heterogeneity); the heterogeneity variance τ2 and standard deviation τ. Statistical analysis was conducted using R version 3.6 (R Foundation for Statistical Computing, Vienna) by using the packages ‘netmeta’ and ‘gemtc’.
Pre-specified sensitivity analyses were conducted to explore the possible impact of clinical presentation [i.e. ACS vs. chronic coronary syndrome (CCS)] or ethnicity (i.e. East Asian vs. non-East Asian patients) on each endpoint.
Results
Studies selection, network of evidence, and baseline characteristics
After screening, a total of 12 RCTs were included in the network meta-analysis (Figure 1).14,26–36 The network of evidence for each outcome is illustrated in Figure 2. Seven antiplatelet strategies were compared: standard-dose (i.e. 75 mg od) clopidogrel (reference strategy), high-dose clopidogrel (150 mg od), adjunctive cilostazol (100 mg bid), standard-dose ticagrelor (90 mg bid), standard-dose prasugrel (10 mg od), low-dose prasugrel (5 mg od), and very-low-dose prasugrel (3.75 mg od).29 Two loops were closed, made by the comparisons of standard-dose clopidogrel, high-dose clopidogrel, and standard-dose prasugrel for the first loop, and standard-dose clopidogrel, high-dose clopidogrel, and cilostazol for the second loop. For these strategies, both direct and indirect evidence were available. All other comparisons were made by indirect comparisons.
Systematic review flow-chart. The figure illustrates the level at which each study was excluded during the review process. CYP, Cytochrome P450.
Network of the evidence. Direct comparisons are depicted as thick lines and indirect comparisons as dotted lines. Numbers indicate how many randomized controlled trials were included for each comparison.
Table 1 summarizes the key clinical characteristics of the included trials. The number of included patients from randomized trials reporting data on carriers of CYP2C19 LoF ranged from 24 to 485. Women were under-represented, with a percentage ranging between 9.2 and 32.8%. Four studies were conducted exclusively in East Asia, and among those three out of four focused exclusively on patients with ACS.30–32 The proportion of PM or IM was not reported in 6 out of the 12 included trials, but only represented a minority in the overall population in the other 6 RCTs, ranging from 5 to 20%. Three studies were conducted exclusively in patients with CCS.14,29,35,37 Importantly, all studies assessed platelet reactivity reported as PRU using VerifyNow-P2Y12 Assay. A detailed description of the assays used for genotyping and PFT across trials is reported in Supplementary material online, Table S3. HPR definitions across trials are reported in Supplementary material online, Table S4. The overall and individual risk for bias of included trials is shown in Supplementary material online, Figure S1. Some concerns arose in six studies,26,29,30,32,34 three of them due to RoB arising from the randomization process,30,32 and three of them due to missing outcome data.26,29,34 Although Egger's regression test suggested absence of publication bias, the inspection of the funnel plot for visual asymmetry revealed slight asymmetry for both platelet reactivity and HPR (Supplementary material online, Figure S2).
Key characteristics of randomized trials included in the systematic review and meta-analysis
| Study, Year | LoF carriers | Percentage of poor metabolizers n (%) | Age | Female (%) | Race (%) | Genotype | Diabetes (%) | Scenario (%) | Interventional Group | Control Group |
|---|---|---|---|---|---|---|---|---|---|---|
| Varenhorst et al. (2009)35 | 52 | NA | 63.8 | 9.2 | § Caucasian: 100% | Multiple, Heterozygous | 19.4% | § CCS: 100% | Prasugrel 60 mg LD/10 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Alexopoulos et al. (2011)26 | 21 | NA | 65.0 | 11.0 | NA | Multiple, Heterozygous | 36.0% | § ACS: 70.3% | Clopidogrel 150 mg/Prasugrel 10mg | Prasugrel 10 mg/Clopidogrel 150 mg |
| Kim In-Suk et al. (2011)31 | 77 | 19 (15) | 61.4 | 24.7 | § Asian: 100% | Multiple, Heterozygous | 28.6% | § ACS: 100% | Adjunctive cilostazol 100mg | Clopidogrel 150 mg |
| Mega et al. (2011)14 | 86 | 5 (5.81) | 60.2 | 25.2 | § White: 88.0% § Asian: 2.4% § Black: 9.0% | Heterozygous or Homozygous | 35.4% | § CCS: 100% | Clopidogrel 150 mg | Clopidogrel 75 mg |
| Park et al. (2011)36 | 283 | 52 (10.7) | 63.4 | 30.1 | § Asians: 100% | Multiple, Heterozygous | 36.7 | § Referred for PCI:100% | Adjunctive Cilostazol 200 mg LD/100 mg MD | Clopidogrel 600 mg LD/75mg |
| Roberts et al. (2012)33 | 46 | 7 (15.2) | 60.2 | 22 | NA | Multiple, Heterozygous | 22.5% | § ACS: 37.4% § CCS: 62.6% | Clopidogrel 600 mg LD/Prasugrel 10 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Braun et al. (2013)27 | 49 | NA | 61.6 | 18.9 | § Caucasian: 90.3% | Multiple, Heterozygous | 25.7% | NA | Prasugrel 60 mg LD/10 mg MD | Clopidogrel 75 mg MD |
| Gurbel et al. (2014)29 | 41 | NA | 65.5 | 32.8 | § Caucasian: 93.9% | Multiple, Heterozygous | NA | § CCS: 100% | Prasugrel 10 mg MD Prasugrel 5 mg MD | Clopidogrel 75 mg MD Clopidogrel 75 mg MD |
| Rossi et al. (2014)34 | 50 | NA | 63.3 | 29.2 | NA | Heterozygous or Homozygous | 42.1% | NA | Clopidogrel 150mg | Clopidogrel 75 mg |
| Ogawa et al. (2016)32 | 485 | 154 (19.9) | 64.3 | 18.9 | § Asian: 100% | Heterozygous or Homozygous | 35.7% | § ACS: 100% | Prasugrel 20 mg LD/3.75 mg MD | Clopidogrel 300 mg LD/75 mg MD |
| Franchi et al. (2020)28 | 65 | 8 (12) | 59.0 | 21.5 | § White: 67.7% | Multiple, Heterozygous | 37.0% | § ACS: 54.9% § CCS: 45.1% | Prasugrel 60 mg LD/10 mg MD | Ticagrelor 180 mg LD/90 mg MD |
| Jin et al. (2020)30 | 70 | NA | 63.0 | 29.6 | § Asian: 100% | Multiple, Heterozygous | 31.7% | § ACS: 100% | Prasugrel 30 mg LD/5 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Study, Year | LoF carriers | Percentage of poor metabolizers n (%) | Age | Female (%) | Race (%) | Genotype | Diabetes (%) | Scenario (%) | Interventional Group | Control Group |
|---|---|---|---|---|---|---|---|---|---|---|
| Varenhorst et al. (2009)35 | 52 | NA | 63.8 | 9.2 | § Caucasian: 100% | Multiple, Heterozygous | 19.4% | § CCS: 100% | Prasugrel 60 mg LD/10 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Alexopoulos et al. (2011)26 | 21 | NA | 65.0 | 11.0 | NA | Multiple, Heterozygous | 36.0% | § ACS: 70.3% | Clopidogrel 150 mg/Prasugrel 10mg | Prasugrel 10 mg/Clopidogrel 150 mg |
| Kim In-Suk et al. (2011)31 | 77 | 19 (15) | 61.4 | 24.7 | § Asian: 100% | Multiple, Heterozygous | 28.6% | § ACS: 100% | Adjunctive cilostazol 100mg | Clopidogrel 150 mg |
| Mega et al. (2011)14 | 86 | 5 (5.81) | 60.2 | 25.2 | § White: 88.0% § Asian: 2.4% § Black: 9.0% | Heterozygous or Homozygous | 35.4% | § CCS: 100% | Clopidogrel 150 mg | Clopidogrel 75 mg |
| Park et al. (2011)36 | 283 | 52 (10.7) | 63.4 | 30.1 | § Asians: 100% | Multiple, Heterozygous | 36.7 | § Referred for PCI:100% | Adjunctive Cilostazol 200 mg LD/100 mg MD | Clopidogrel 600 mg LD/75mg |
| Roberts et al. (2012)33 | 46 | 7 (15.2) | 60.2 | 22 | NA | Multiple, Heterozygous | 22.5% | § ACS: 37.4% § CCS: 62.6% | Clopidogrel 600 mg LD/Prasugrel 10 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Braun et al. (2013)27 | 49 | NA | 61.6 | 18.9 | § Caucasian: 90.3% | Multiple, Heterozygous | 25.7% | NA | Prasugrel 60 mg LD/10 mg MD | Clopidogrel 75 mg MD |
| Gurbel et al. (2014)29 | 41 | NA | 65.5 | 32.8 | § Caucasian: 93.9% | Multiple, Heterozygous | NA | § CCS: 100% | Prasugrel 10 mg MD Prasugrel 5 mg MD | Clopidogrel 75 mg MD Clopidogrel 75 mg MD |
| Rossi et al. (2014)34 | 50 | NA | 63.3 | 29.2 | NA | Heterozygous or Homozygous | 42.1% | NA | Clopidogrel 150mg | Clopidogrel 75 mg |
| Ogawa et al. (2016)32 | 485 | 154 (19.9) | 64.3 | 18.9 | § Asian: 100% | Heterozygous or Homozygous | 35.7% | § ACS: 100% | Prasugrel 20 mg LD/3.75 mg MD | Clopidogrel 300 mg LD/75 mg MD |
| Franchi et al. (2020)28 | 65 | 8 (12) | 59.0 | 21.5 | § White: 67.7% | Multiple, Heterozygous | 37.0% | § ACS: 54.9% § CCS: 45.1% | Prasugrel 60 mg LD/10 mg MD | Ticagrelor 180 mg LD/90 mg MD |
| Jin et al. (2020)30 | 70 | NA | 63.0 | 29.6 | § Asian: 100% | Multiple, Heterozygous | 31.7% | § ACS: 100% | Prasugrel 30 mg LD/5 mg MD | Clopidogrel 600 mg LD/75 mg MD |
ACS, acute coronary syndrome; CCS, chronic coronary syndrome; n, number; NA, not available; LD, loading dose; MD, maintenance dose; LoF, loss of function.
Key characteristics of randomized trials included in the systematic review and meta-analysis
| Study, Year | LoF carriers | Percentage of poor metabolizers n (%) | Age | Female (%) | Race (%) | Genotype | Diabetes (%) | Scenario (%) | Interventional Group | Control Group |
|---|---|---|---|---|---|---|---|---|---|---|
| Varenhorst et al. (2009)35 | 52 | NA | 63.8 | 9.2 | § Caucasian: 100% | Multiple, Heterozygous | 19.4% | § CCS: 100% | Prasugrel 60 mg LD/10 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Alexopoulos et al. (2011)26 | 21 | NA | 65.0 | 11.0 | NA | Multiple, Heterozygous | 36.0% | § ACS: 70.3% | Clopidogrel 150 mg/Prasugrel 10mg | Prasugrel 10 mg/Clopidogrel 150 mg |
| Kim In-Suk et al. (2011)31 | 77 | 19 (15) | 61.4 | 24.7 | § Asian: 100% | Multiple, Heterozygous | 28.6% | § ACS: 100% | Adjunctive cilostazol 100mg | Clopidogrel 150 mg |
| Mega et al. (2011)14 | 86 | 5 (5.81) | 60.2 | 25.2 | § White: 88.0% § Asian: 2.4% § Black: 9.0% | Heterozygous or Homozygous | 35.4% | § CCS: 100% | Clopidogrel 150 mg | Clopidogrel 75 mg |
| Park et al. (2011)36 | 283 | 52 (10.7) | 63.4 | 30.1 | § Asians: 100% | Multiple, Heterozygous | 36.7 | § Referred for PCI:100% | Adjunctive Cilostazol 200 mg LD/100 mg MD | Clopidogrel 600 mg LD/75mg |
| Roberts et al. (2012)33 | 46 | 7 (15.2) | 60.2 | 22 | NA | Multiple, Heterozygous | 22.5% | § ACS: 37.4% § CCS: 62.6% | Clopidogrel 600 mg LD/Prasugrel 10 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Braun et al. (2013)27 | 49 | NA | 61.6 | 18.9 | § Caucasian: 90.3% | Multiple, Heterozygous | 25.7% | NA | Prasugrel 60 mg LD/10 mg MD | Clopidogrel 75 mg MD |
| Gurbel et al. (2014)29 | 41 | NA | 65.5 | 32.8 | § Caucasian: 93.9% | Multiple, Heterozygous | NA | § CCS: 100% | Prasugrel 10 mg MD Prasugrel 5 mg MD | Clopidogrel 75 mg MD Clopidogrel 75 mg MD |
| Rossi et al. (2014)34 | 50 | NA | 63.3 | 29.2 | NA | Heterozygous or Homozygous | 42.1% | NA | Clopidogrel 150mg | Clopidogrel 75 mg |
| Ogawa et al. (2016)32 | 485 | 154 (19.9) | 64.3 | 18.9 | § Asian: 100% | Heterozygous or Homozygous | 35.7% | § ACS: 100% | Prasugrel 20 mg LD/3.75 mg MD | Clopidogrel 300 mg LD/75 mg MD |
| Franchi et al. (2020)28 | 65 | 8 (12) | 59.0 | 21.5 | § White: 67.7% | Multiple, Heterozygous | 37.0% | § ACS: 54.9% § CCS: 45.1% | Prasugrel 60 mg LD/10 mg MD | Ticagrelor 180 mg LD/90 mg MD |
| Jin et al. (2020)30 | 70 | NA | 63.0 | 29.6 | § Asian: 100% | Multiple, Heterozygous | 31.7% | § ACS: 100% | Prasugrel 30 mg LD/5 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Study, Year | LoF carriers | Percentage of poor metabolizers n (%) | Age | Female (%) | Race (%) | Genotype | Diabetes (%) | Scenario (%) | Interventional Group | Control Group |
|---|---|---|---|---|---|---|---|---|---|---|
| Varenhorst et al. (2009)35 | 52 | NA | 63.8 | 9.2 | § Caucasian: 100% | Multiple, Heterozygous | 19.4% | § CCS: 100% | Prasugrel 60 mg LD/10 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Alexopoulos et al. (2011)26 | 21 | NA | 65.0 | 11.0 | NA | Multiple, Heterozygous | 36.0% | § ACS: 70.3% | Clopidogrel 150 mg/Prasugrel 10mg | Prasugrel 10 mg/Clopidogrel 150 mg |
| Kim In-Suk et al. (2011)31 | 77 | 19 (15) | 61.4 | 24.7 | § Asian: 100% | Multiple, Heterozygous | 28.6% | § ACS: 100% | Adjunctive cilostazol 100mg | Clopidogrel 150 mg |
| Mega et al. (2011)14 | 86 | 5 (5.81) | 60.2 | 25.2 | § White: 88.0% § Asian: 2.4% § Black: 9.0% | Heterozygous or Homozygous | 35.4% | § CCS: 100% | Clopidogrel 150 mg | Clopidogrel 75 mg |
| Park et al. (2011)36 | 283 | 52 (10.7) | 63.4 | 30.1 | § Asians: 100% | Multiple, Heterozygous | 36.7 | § Referred for PCI:100% | Adjunctive Cilostazol 200 mg LD/100 mg MD | Clopidogrel 600 mg LD/75mg |
| Roberts et al. (2012)33 | 46 | 7 (15.2) | 60.2 | 22 | NA | Multiple, Heterozygous | 22.5% | § ACS: 37.4% § CCS: 62.6% | Clopidogrel 600 mg LD/Prasugrel 10 mg MD | Clopidogrel 600 mg LD/75 mg MD |
| Braun et al. (2013)27 | 49 | NA | 61.6 | 18.9 | § Caucasian: 90.3% | Multiple, Heterozygous | 25.7% | NA | Prasugrel 60 mg LD/10 mg MD | Clopidogrel 75 mg MD |
| Gurbel et al. (2014)29 | 41 | NA | 65.5 | 32.8 | § Caucasian: 93.9% | Multiple, Heterozygous | NA | § CCS: 100% | Prasugrel 10 mg MD Prasugrel 5 mg MD | Clopidogrel 75 mg MD Clopidogrel 75 mg MD |
| Rossi et al. (2014)34 | 50 | NA | 63.3 | 29.2 | NA | Heterozygous or Homozygous | 42.1% | NA | Clopidogrel 150mg | Clopidogrel 75 mg |
| Ogawa et al. (2016)32 | 485 | 154 (19.9) | 64.3 | 18.9 | § Asian: 100% | Heterozygous or Homozygous | 35.7% | § ACS: 100% | Prasugrel 20 mg LD/3.75 mg MD | Clopidogrel 300 mg LD/75 mg MD |
| Franchi et al. (2020)28 | 65 | 8 (12) | 59.0 | 21.5 | § White: 67.7% | Multiple, Heterozygous | 37.0% | § ACS: 54.9% § CCS: 45.1% | Prasugrel 60 mg LD/10 mg MD | Ticagrelor 180 mg LD/90 mg MD |
| Jin et al. (2020)30 | 70 | NA | 63.0 | 29.6 | § Asian: 100% | Multiple, Heterozygous | 31.7% | § ACS: 100% | Prasugrel 30 mg LD/5 mg MD | Clopidogrel 600 mg LD/75 mg MD |
ACS, acute coronary syndrome; CCS, chronic coronary syndrome; n, number; NA, not available; LD, loading dose; MD, maintenance dose; LoF, loss of function.
Network meta-analysis
Compared with standard-dose clopidogrel, the primary endpoint of PRU was reduced with all alternative antiplatelet strategies, except for adjunctive cilostazol. Specifically, the reduction was greatest with standard-dose prasugrel (MD −127.91; 95% CI −141.04; −114.78) and standard-dose ticagrelor (MD −124.91; 95% CI −161.78; −88.04), followed by low-dose prasugrel (MD −76.33; 95% CI −98.01; −54.65), and very-low dose prasugrel (MD −73.00; 95% CI −100.28; −45.72) (Figure 3). High-dose clopidogrel and adjunctive cilostazol were associated with a modest reduction of PRU compared with standard-dose clopidogrel (MD −29.49; 95% CI −46.41; −12.58 and −42.64; 95% CI −64.72.; −20.57, respectively) (Figure 3).
Forest plot for the primary endpoint (i.e. platelet reactivity). The figure illustrates treatment estimate among different antiplatelet strategies for the primary endpoint. Standard-dose (75 mg od) clopidogrel was used as reference strategy. The primary endpoint was the level of platelet reactivity between treatment arms as assessed by the VerifyNow P2Y12 assay and reported as P2Y12 reaction unit (PRU). For each comparison, treatment estimates and respective 95% CI are illustrated by one circle and line, respectively, along with text highlighting the specific mean difference and 95% CI. BID, bis in die; CI, confidence interval; MD, mean difference; OD, once daily.
The node split analysis for the comparisons reporting both direct and indirect evidence (standard-dose prasugrel, standard-dose clopidogrel, and high-dose clopidogrel) for the primary endpoint, did not reveal inconsistency, with results of the network mostly laying on direct comparisons (Supplementary material online, Figure S3). The net heat plot revealed the consistency between direct and indirect evidence (Supplementary material online, Figure S4),
The secondary outcome of HPR rates was not available for the treatment arms of standard-dose ticagrelor and very low-dose prasugrel. All the alternative antiplatelet strategies were associated with reduced HPR, but the reduction was statistically significant only with standard dose of prasugrel (OR 0.03; 95% CI 0.00; 0.19) (Figure 4).
Forest plot for the secondary endpoint (i.e. high-platelet reactivity rates). The figure illustrates the treatment estimate among different antiplatelet strategies for the secondary endpoint of trial-defined high-platelet reactivity rates. Standard-dose (75 mg od) clopidogrel was used as reference strategy. For each comparison, treatment estimates and respective 95% CI are illustrated by one circle and line, respectively, along with text highlighting the specific odds ratio and 95% CI. BID, bis in die; CI, confidence interval; OD, once daily; OR, odds ratio.
Node split analysis for the comparisons reporting both direct and indirect evidence (standard-dose prasugrel, standard-dose clopidogrel, and high-dose clopidogrel for the first loop, and standard-dose clopidogrel, high-dose clopidogrel, and cilostazol for the second loop) did not reveal any inconsistency, with results of the network laying mostly on direct comparisons (Supplementary material online, Figure S5). Net heat plot revealed strong consistency between direct and indirect evidence (Supplementary material online, Figure S6).
Ranking of treatments
Ranking of treatments according to P-scores for primary and secondary outcomes is displayed in the central figure and in Supplementary material online, Figures S7 and S8. With regard to the primary endpoint, standard-dose prasugrel ranked as the best treatment (P-score = 0.928), followed by ticagrelor (P-score = 0.901), and low-dose prasugrel (P-score = 0.595), very-low dose prasugrel (P-score = 0.5654), cilostazol (P-score = 0.318), high-dose clopidogrel (P-score = 0.193). Rankograms for the secondary endpoint should be interpreted in light of the fact that rates of HPR were not available for the treatment arms of standard-dose ticagrelor and very low-dose prasugrel.
Sensitivity analyses
The replication of analyses after the exclusion of studies selectively including patients with ACS (Supplementary material online, Figure S9) or CCS (Supplementary material online, Figure S10) showed consistent results with the main analysis. Similarly, the analysis run after the exclusion of studies selectively including Caucasian (Supplementary material online, Figure S11) or East Asian patients (Supplementary material online, Figure S12) showed results similar to the main analysis.
Discussion
The results of this network meta-analysis assessing the comparative effects of different alternative antiplatelet regimens by consolidating results from RCTs among individuals carrying CYP2C19 LoF alleles can be summarized as follows: (i) all alternative antiplatelet strategies were associated with reduced platelet reactivity compared with standard-dose clopidogrel, the magnitude of which, however, varied widely across different strategies; (ii) standard-dose prasugrel or ticagrelor are associated with the greatest reduction in platelet reactivity, followed by reduced- and very low-dose of prasugrel, compared with standard-dose clopidogrel; (iii) double-dose clopidogrel and adjunctive cilostazol were associated with a modest reduction in platelet reactivity, compared with standard-dose clopidogrel; (iv) compared with standard-dose clopidogrel, there was a reduction in HPR with all treatments, that was statistically significant with standard-dose prasugrel.
Platelet function may be assessed by different assays, allowing for the identification of patients at HPR, low platelet reactivity (LPR), or optimal range of platelet reactivity (OPR).8 Point-of-care assays have the advantage of providing rapid bedside results by non-expert personnel in a timely fashion compared with laboratory-based assays. Among these, the VerifyNow P2Y12 assay is one of the most commonly used.8,38 HPR among clopidogrel-treated patients undergoing PCI has been consistently shown to be associated with increased major adverse cardiovascular events (MACE), including death and stent thrombosis, and that LPR is not associated with a further reduction of MACE, but increased bleeding, as compared with OPR patients.7,23 This has represented the rationale for proposing a potential ‘therapeutic window’ of platelet inhibition in PCI patients.8,39,40
The understanding that the interindividual variability in clopidogrel response lies, at least in part, in the fact that 20–60% of treated patients are carriers of CYP2C19 LoF alleles (e.g. CYP2C19*2, CYP2C19*3), the prevalence of which is largely influenced by ethnicity, being more common in East Asian than European and African ancestry populations.4,8 The fact that heterozygote patients are IMs associated with reduced clopidogrel active metabolite, and homozygote patients are PMs associated with nearly no generation of clopidogrel active metabolite, has stimulated the interest in the use of genetic testing to identify patients non-responding to clopidogrel.4 Indeed, genetic testing for CYP2C19 LoF alleles have the key advantage that the genetic makeup of an individual remains unvaried over time and does not require the patient to be on treatment with clopidogrel.4,8 Studies have shown that IM, and particularly PM phenotypes for CYP2C19 LoF alleles, are associated with reduced levels of clopidogrel active metabolite, diminished platelet inhibition, and high rates of HPR resulting in increased rates MACE and stent thrombosis following PCI.41 On this background, several drug governing authorities have issued a cautionary advisory recommending the use of alternative P2Y12 inhibiting therapies in carriers of CYP2C19 LoF alleles, particularly for PM, a recommendation also supported by expert consensus documents.8–11
On this background, several studies have explored the clinical impact of using alternative oral antiplatelet therapies among carriers of CYP2C19 LoF alleles treated with clopidogrel.14–22 Early strategies tested in RCTs first included increased dose of clopidogrel (i.e. 150 mg od) or adjunctive cilostazol on top of DAPT.14–20 Subsequent studies have tested the use of newer generation P2Y12 inhibitors (i.e. prasugrel or ticagrelor).21,22 Although the results of these studies were not univocal due to a number of pitfalls in their design, the overall clinical evidence suggest that the use of prasugrel or ticagrelor as alternative therapy to clopidogrel could reduce MACE in CYP2C19 LoF carriers but with increased bleeding without any additional reduction in ischaemic events in patients responding to clopidogrel.41 Importantly, the use of PFT and genetic testing has been associated with improved outcomes compared with standard antiplatelet therapy in both ACS and CCS patients undergoing PCI.42,43
The clinical impact of alternative antiplatelet therapy in CYP2C19 LoF allele carriers stems from their pharmacologic characteristics. Indeed, ticagrelor does not require hepatic activation by the CYP2C19 enzyme, while the role of this enzyme in the conversion of prasugrel into its active metabolite is limited.32 Moreover, cilostazol is a selective phosphodiesterase-3 inhibitor that inhibits platelet aggregation by increasing cyclic adenosine monophosphate in platelets and is not affected by the CYP2C19 enzyme.44 However, the comparative effects of these alternative antiplatelet agents among CYP2C19 LoF carriers has never been tested. Because platelet reactivity assessed by PFT is closely associated with adverse clinical outcomes, we performed a network meta-analysis of RCTs comparing different alternative antiplatelet agents among CYP2C19 LoF carriers and reporting measures deriving from PFT.7 We found that all strategies were associated with reduced rates of HPR, but that the reduction in platelet reactivity varied widely across the different alternative regimens, being maximum with standard dose of the newer generation P2Y12 inhibitors (prasugrel 10 mg od or ticagrelor 90 mg bid), and progressively decreasing with the reduction of its dose (i.e. prasugrel 5 and 3.75 mg). There is no data with reduced dose of ticagrelor (e.g. 60 mg bid). Notably, double-dose clopidogrel and additional cilostazol had a modest reduction in platelet reactivity, compared with standard-dose clopidogrel. These findings are consistent with previous studies, but add important insights on how reduced doses of prasugrel may perform in CYP2C19 LoF carriers compared with other doses and other alternative strategies. In fact, a dose reduction of prasugrel after a short course of standard DAPT represents one of the possible de-escalation strategies after ACS/PCI, although the clinical evidence in support of this strategy is scarce and comparative effects of different de-escalation antiplatelet strategies are lacking.45–48
Moreover, our results support previous findings suggesting that adjunctive cilostazol on top of DAPT could enhance the inhibition of P2Y12 signalling, reducing platelet reactivity and improving outcomes in patients undergoing PCI.49 However, it should be acknowledged that VerifyNow is a test primary designed to assess the effects of P2Y12 inhibitors, and therefore it is inevitably less sensitive to assess the antiplatelet effects of cilostazol.
Given the importance of OPR in optimizing the balance between ischaemic and bleeding risks in ACS/PCI patients, our findings suggest that a reduced-dose of prasugrel may provide an adequate balance between efficacy, by reducing the rates of HPR in carriers of CYP2C19 LoF, and safety, by reducing the rates of patients expected to be at LPR compared to other alternative strategies (i.e. standard-dose ticagrelor or prasugrel).39 These findings are hypothesis generating and should be confirmed in dedicated clinical studies.
Limitations
This analysis has several limitations. First, this is a study-level rather than a patient-level meta-analysis. Therefore, it was not possible to stratify the analysis according to IM vs PM. In fact, the proportion of IM or PM was not reported in 6 out of the 12 included trials, but only represented a minority in the population of the other 6 RCTs. Although the pre-specified sensitivity analysis according to ethnicity found consistent results after excluding East Asian patients—that are those characterized by the higher prevalence of PM—further studies specifically focusing on IM are needed to support the validity of our findings in these patients. Second, we focused on PD measures rather than clinical outcomes as our endpoints. This was due to the fact that RCTs reporting clinical outcomes in this setting are scarce and do not allow for the implementation of a network to compare different alternative antiplatelet strategies. Therefore, whether these findings could translate into similar clinical outcomes remains to be explored. Third, although we performed a sensitivity analysis according to ethnicity, it should be acknowledged that RCTs reporting outcomes of low-dose prasugrel were exclusively conducted in East Asians, thus limiting the generalizability of its findings to other ethnicities.50 Forth, the secondary endpoint result of HPR should be interpreted with caution, given that trial-defined HPR definitions were used, the limited statistical power and the fact that this outcome was not available for all included regimens. Fifth, a limitation in several of the provided effect estimates is represented by the presence of ‘open-loops’, thus mainly relying on indirect comparisons. Nevertheless, the results obtained through the node-split analysis for the closed loops, showed consistency between direct, indirect, and network evidence for both the primary and secondary endpoints and provide an important focus on the reliability of standard-dose prasugrel when compared to standard- and high-dose clopidogrel. Sixth, the population included in our analysis encompassed both patients with ACS and CCS. Nevertheless, we conducted a pre-specified sensitivity analysis to explore the influence of clinical presentation among outcomes. Seventh, the timing in the evaluation of platelet reactivity differed across studies, but the optimal timing for PFT to be performed remains debated.8
Conclusions
Our findings show that all alternative antiplatelet strategies reduced platelet reactivity compared to standard-dose clopidogrel among carriers of CYP2C19 LoF alleles, but the reduction varies widely across strategies. Standard-dose prasugrel or ticagrelor are associated with the greatest reduction in platelet reactivity, followed by low- and very low-dose of prasugrel. Double-dose clopidogrel and adjunctive cilostazol showed modest effects. Reduced-dose of prasugrel may represent a balanced strategy to overcome high-platelet reactivity without a significant increase in bleeding. The clinical implications of these pharmacodynamic findings warrant further investigation.
Funding
The study was not supported.
Disclosures
M.G. declares that he has received consulting fees by Terumo, outside the present work. G.B.Z. has consulted for Amarin, Balmed, Cardionovum, Crannmedical, Endocore Lab, Eukon, Guidotti, Innovheart, Meditrial, Microport, Opsens Medical, Terumo, and Translumina, outside the present work. F.F. declares that he has received payment as an individual for consulting fee or honoraria from AstraZeneca, outside the present work; and institutional payments for grants from PLx Pharma and The Scott R. MacKenzie Foundation. G.B.Z declares that he has consulted for Amarin, Balmed, Cardionovum, Crannmedical, Endocore Lab, Eukon, Guidotti, Innovheart, Meditrial, Microport, Opsens Medical, Terumo, and Translumina, outside the present work. D.J.A. declares that he has received consulting fees or honoraria from Abbott, Amgen, AstraZeneca, Bayer, Biosensors, Boehringer Ingelheim, Bristol-Myers Squibb, Chiesi, CSL-Behring, Daiichi-Sankyo, Eli Lilly, Faraday, Haemonetics, Janssen, Merck, Novartis, Novo Nordisk, PhaseBio, PLx Pharma, Pfizer, Sanofi, and Vectura, outside the present work; D.J.A. also declares that his institution has received research grants from Amgen, AstraZeneca, Bayer, Biosensors, CeloNova, CSL Behring, Daiichi-Sankyo, Eisai, Eli Lilly, Faraday, Gilead, Janssen, Matsutani Chemical Industry Co., Merck, Novartis, Osprey Medical, Renal Guard Solutions and Scott R. MacKenzie Foundation. All other authors have no disclosures to declare.
Data availability
This study did not combine individual patient data. However, trial-level data used across analyses are available upon reasonable request.
References
Author notes
MG and GO contributed equally.
