High frequency of CYP2C19*2 carriers in PCI-treated patients switched over from clopidogrel to prasugrel based on platelet function monitoring

To the editor

Among patients undergoing percutaneous coronary intervention (PCI), subjects with high on-clopidogrel treatment platelet reactivity (HCPR) exhibit a high risk for post-procedural thrombotic events including stent thrombosis [1]. Numerous non-genetic and genetic variables were discovered as major factors causing HCPR. Among the genetic variables, the common loss-of-function CYP2C19*2 polymorphism was found to be associated with an attenuated response to clopidogrel [2] and a higher risk for stent thrombosis [3]. However, prior studies suggested that – based on statistical models – this genetic marker may only account for about 5–12% of the observed variability in clopidogrel response [4], [5]. This suggests that a single genetic variant such as CYP2C19*2 may only have little impact on the overall response to clopidogrel treatment and genotyping of this variant may be of limited value in routine clinical practice. The prevalence of CYP2C19*2 as a causative factor for high levels of on-clopidogrel treatment platelet reactivity was never investigated in a clinical scenario where patients are switched over from clopidogrel to prasugrel treatment due to clopidogrel treatment failure. To close this gap of knowledge, the aim of this study was to determine CYP2C19*2 allele carriage in PCI-treated patients initially treated with clopidogrel and switched over to prasugrel therapy due to high levels of platelet reactivity on clopidogrel treatment in a setting of routine platelet function testing.

The prevalence of the CYP2C19*2 allele was compared between two consecutively recruited PCI cohorts: The first consecutive (= control) cohort of patients (n = 1524) stems from a study that enrolled clopidogrel-treated patients undergoing PCI and that aimed to determine the prognostic value of HCPR [6] and genetic markers [7]. As a consecutive cohort, this study included patients with and without HCPR but without any switch of antiplatelet treatment based on obtained platelet function testing results. The second consecutive cohort of PCI-treated patients (n = 143) was recruited between 2009 and 2011 [8] and within this cohort all patients were switched over from clopidogrel to prasugrel treatment based on detected high levels of platelet reactivity on clopidogrel treatment. Blood for genotyping was available in 124 (87%) of the 143 patients who constitute the study group (= cases) for this specific analysis. Genotypes in both cohorts were determined by a TaqMan assay as previously described [7] and were compared between groups by chi-square test. In both cohorts, the adenosine diphosphate (ADP)-induced platelet aggregation (in AU·min; AU; aggregation unit) was tested on a Multiplate analyzer (Verum Diagnostica, Munich, Germany). Patients in the second cohort showed either HCPR (n = 108, defined per consensus definition [1]: ≥468 AU·min) or high levels of platelet reactivity on clopidogrel treatment (n = 16, range 402–467 AU·min) and were switched from clopidogrel over to prasugrel in a setting of routine platelet function monitoring. The time point of switching was an individual decision of the attending physician and was based on obtained platelet function testing results measured after single or repeated clopidogrel loading doses (LDs). Up to three clopidogrel LDs (600 mg) were given to patients before a switch to prasugrel treatment and LD administration (high LD of 60 mg prasugrel in all except for three patients who received a low LD). On-treatment platelet aggregation measurements were not normally distributed, are presented as median with interquartile range, and were compared between groups with two-sided unpaired or paired Wilcoxon test.

Baseline characteristics (see below, first vs. second cohort) for both cohorts have been reported previously [7], [8]: In brief, age (67.4 vs. 67.5, p = 0.89) and the proportion of women (23% vs. 23%, p = 0.99) were comparable between both cohorts but other variables such as diabetes mellitus (28% vs. 44%; p < 0.001) and the proportion of ACS patients (11% vs. 53%; p < 0.0001) differed significantly. In the second cohort, repeated clopidogrel LDs were given in the majority of patients (single LD in 124 patients, two LDs in 76 patients, and three LDs in 7 patients) before switching over to prasugrel treatment. The ADP-induced platelet aggregation was significantly lower after prasugrel LD administration as compared to values obtained after the clopidogrel LD administration proximate (LD₁ or LD₂ or LD₃) to prasugrel switch (140 [84–242] vs. 557 [470–704] AU·min; p < 0.0001), showing that prasugrel was highly effective in this cohort in terms of lowering high on-treatment platelet reactivity values. The CYP2C19*2 genotype distribution for both study cohorts (cases vs. controls) is shown in Figure 1(A) and differed significantly between groups with a higher proportion of both wt/*2 (wt, wild type) and *2/*2 patients in the second cohort of patients with switch of treatment. Patients with a switch of treatment due to high levels of platelet reactivity on clopidogrel treatment were significantly more often CYP2C19*2 allele carriers as compared to patients without antiplatelet treatment switch (43.5% vs. 24.7%, respectively; odds ratio 2.4, 95% CI 1.6–3.4, and p < 0.0001). Individual aggregation measurements following prasugrel loading for CYP2C19*2 non-carriers vs. carriers are shown in Figure 1(B). The ADP-induced platelet aggregation was similar in CYP2C19*2 carriers (n = 54, wt/*2 n = 48, *2/*2 n = 6) vs. non-carriers (wt/wt, n = 70) with aggregation values of 148 [68–261] vs. 137 [88–236] AU·min; p = 0.94.

Figure 1. CYP2C19*2 genotypes and influence of *2 allele carriage on prasugrel treatment: (A) CYP2C19*2 genotype frequency in the cohort without (in blue) vs. with (in red) treatment adjustment (p-value was calculated across genotypes) and (B) platelet aggregation measurements following prasugrel LD for patients without (n = 70) and with (n = 54) CYP2C19*2 carriage. Notes: The blue lines represent median values per group. Red dots denote individual platelet aggregation measurements after prasugrel LD administration. The dotted line in black denotes the HCPR consensus cut-off value (≥468 AU·min). ADP, adenosine diphosphate; AU, aggregation units; LD, loading dose; wt, wild type.

The key findings of this study are that (a) carriage of the CYP2C19*2 allele is very frequent in patients with clopidogrel treatment failure who are switched over to prasugrel in a setting of routine platelet function monitoring and (b) that in this specific cohort of high-risk patients, the antiplatelet response to prasugrel is not influenced by CYP2C19*2. Numerous studies as well as their meta-analyses have shown that CYP2C19*2 is associated with an attenuated response to clopidogrel and an increased risk for stent thrombosis [3], [9–12]. However, the contribution of the *2 allele on clopidogrel response variability was suggested to be rather modest, accounting for about 5–12% [4], [5] of the overall observed variability in drug response. In contrast to this, our data suggest that in a clinical setting of routine platelet function monitoring and tailored antiplatelet treatment for patients with high levels of platelet reactivity on clopidogrel treatment, the prevalence of the CYP2C19*2 allelic variant is very high and must therefore be considered as a major genetic risk factor for clopidogrel treatment failure. Of note, a selection bias that may have influenced results seems very unlikely as in both cohorts [7], [8], genotyping data were generated post hoc and phenotyping by platelet function testing was the only variable that determined a switch of treatment in the second study cohort. As a further important finding, an influence of CYP2C19*2 was not observed for on-prasugrel treatment aggregation values. This observation is in line with prior studies that also failed to show an influence of CYP2C19*2 on platelet response to prasugrel [13] or the clinical outcome of prasugrel-treated patients [14]. Thus, our study expands prior findings to a high-risk cohort (high proportion of diabetes and ACS) of PCI-treated patients showing a clopidogrel treatment failure.

In conclusion, the CYP2C19*2 allele is very frequent in patients with clopidogrel treatment failure and a consecutive switch over to prasugrel in a setting of individualized and platelet function testing guided antiplatelet treatment. On-prasugrel treatment platelet reactivity is not influenced by this genetic variant. The clinical impact of these findings warrants further investigation, as well as the value of an individualized antiplatelet treatment approach based on genotyping or phenotyping by platelet function testing.

Declaration of interest: Dr Sibbing reported receiving fees for advisory board activities from Verum Diagnostica, Eli Lilly and Company and Daiichi Sankyo.

Notes

Both Isabell Bernlochner and Katharina Mayer contributed equally to this study.