- Top of page
- Competing interests
Anti-platelet therapy is now widely used in patients with stroke, myocardial infarction or peripheral arterial disease. Clopidogrel and aspirin are the most important cornerstone agents in this therapy. To prevent stent thrombosis after implantation of bare metal and drug-eluting stents, dual anti-platelet therapy with aspirin and clopidogrel is now a standard therapy . However, such antiplatelet therapy increases the risk of gastrointestinal bleeding [2–7]. A proton pump inhibitor (PPI), therefore, is often prescribed with anti-platelet agents. In 2008, the American College of Cardiology Foundation (ACCF), the American College of Gastroenterology (ACG) and American Heart Association (AHA) published their statement on antiplatelet therapy which recommends the prescription of a PPI to patients with a risk of peptic ulcer and/or on two or more antiplatelet agents [8–10].
The drug–drug interaction between clopidogrel and a PPI has been receiving attention lately. Clopidogrel is metabolized by CYP2C19 to form its active metabolites. Therefore, the plasma concentration of the active metabolite of clopidogrel depends on the activity of CYP2C19 . CYP2C19 is also the main metabolizing enzyme of PPIs. Therefore, concomitant use of clopidogrel and a PPI induces a drug–drug interaction, resulting in the decreased activation of clopidogrel, which attenuates its the anti-platelet effect and could lead to an increased risk of re-infarction and/or stent thrombosis. PPIs per se are recognized as inhibitors of CYP2C19 , which would also contribute to the attenuation of activation of clopidogrel. Juurlink et al.  reported that concomitant therapy with a PPI other than pantoprazole was associated with an increased risk of re-infarction. Ho et al.  reported that concomitant use of clopidogrel and a PPI was associated with an increased risk of adverse outcomes compared with the use of clopidogrel alone. Toth et al.  reported that the efficacy of clopidogrel can also be reduced if patients are receiving concomitant therapy with a PPI such as omeprazole. However, Siller-Matula et al.  reported that the intake of pantoprazole or esomeprazole is not associated with an impaired response to clopidogrel. Thus, the influence of PPIs on clopidogrel efficacy seems controversial.
Interestingly, the influences of PPIs on CYP2C19 differ among different PPIs. Omeprazole is a substrate with strong affinity to CYP2C19. Lansoprazole is also a substrate for CYP2C19 but with a weak affinity in comparison with omeprazole. The metabolism of rabeprazole has been thought to be less associated with CYP2C19 . There have been several reports on drug–drug interactions between omeprazole and other drugs (e.g. warfarin, phenytoin, diazepam), while incidences of drug–drug interactions of rabeprazole and lansoprazole are not so many as observed with omeprazole . However, it has not been fully elucidated whether this difference in the affinity to CYP2C19 among different PPIs results in different effects on the anti-platelet function of clopidogrel.
Another problem with clopidogrel is interindividual difference in its anti-platelet activity among the different CYP2C19 genotype groups . Patients with intermediate or poor metabolizer genotype of CYP2C19 are at a higher risk of stent thrombosis and re-infarction because of decreased active metabolism of clopidogrel [18–22]. However, no study has evaluated the influence of CYP2C19 genotype status and different PPIs on the anti-platelet function of clopidogrel simultaneously.
Based on the backgrounds mentioned above, we prospectively examined the effect of three proton pump inhibitors, omeprazole, lansoprazole and rabeprazole, on the anti-platelet function of clopidogrel in relation to CYP2C19 genotype status. In addition, we examined whether the interaction between a PPI and clopidogrel could be avoided by administering the two drugs separately instead of taking them concomitantly.
- Top of page
- Competing interests
We evaluated the influences of PPIs on the effect of clopidogrel in relation to CYP2C19 genotypes and found that any of three PPIs (i.e. omeprazole 20 mg, lansoprazole 30 mg and rabeprazole 20 mg) could cause attenuation of the anti-platelet function of clopidogrel. We observed that the influence of a PPI on the effect of clopidogrel differed between the CYP2C19 genotype groups and that drug–drug interactions between clopidogrel and PPIs resulting in the decrease in IPA to the level of ‘low-responder’ were likely to occur in DMs of CYP2C19, who are carrying the decreased function allele of CYP2C19 (i.e. *2 and/or *3). We also observed that separate dosing of a PPI (i.e. clopidogrel in the morning and a PPI in the evening) did not prevent the drug–drug interaction between clopidogrel and a PPI. Therefore, we have to reconsider the risk and benefit balance of the concomitant use of clopidogrel and a PPI with reference to CYP2C19 genotype status.
There appear to be differences in the effects on clopidogrel among the different PPIs. Juurlink et al.  have indicated that omeprazole, lansoprazole and rabeprazole, but not pantoprazole, attenuate the clinical effect of clopidogrel in patients with cardiovascular disorders. Ho et al.  have reported that omeprazole and rabeprazole attenuate the clinical effect of clopidogrel. Sibbing et al.  reported that omeprazole impaired the anti-platelet function of clopidogrel, but pantoprazole and esomeprazole did not. Gilard et al.  demonstrated that omeprazole impaired the anti-platelet function of clopidogrel [16, 27–29]. Small et al.  reported that lansoprazole decreased the anti-platelet function of clopidogrel. Siller-Matula et al.  reported that intake of pantoprazole or esomeprazole was not associated with impaired response to clopidogrel. On the other hand, Zuern et al.  reported that co-administration of PPIs including pantoprazole and esomeprazole significantly decreased the effect of clopidogrel on platelet aggregation. Together, all PPIs used in clinical practice appear to have the potential to attenuate the efficacy of clopidogrel. However, the recent report by O'Donoghue et al.  demonstrated that a clinically problematic interaction was not observed between clopidogrel and a PPI, although the ex vivo study demonstrated that PPI use attenuated the anti-platelet effect of clopidogrel. Rassen et al.  did not observe any conclusive evidence of a clopidogrel–PPI interaction of major clinical relevance. Therefore, a drug–drug interaction between clopidogrel and a PPI has been controversial.
In the present study, we tested omeprazole, lansoprazole and rabeprazole and found that the influence of three PPIs on the anti-platelet function of clopidogrel differed among them. However, the levels of attenuation of clopidogrel by PPIs depended on CYP2C19 genotype status. Although the efficacy of clopidogrel was decreased by a PPI in RMs of CYP2C19, the levels of anti-platelet function of clopidogrel after attenuation by a PPI in this group were mostly not problematic (i.e. rarely decreased to the levels of ‘low-responder’). Of 15 RMs, only one case became ‘low-responder’ after concomitant use of rabeprazole 20 mg. On the other hand, the efficacy of clopidogrel appeared to be unstable with concomitant use of any of the PPIs used in the present study especially in the DM group. Six of 16 DMs who were judged as ‘responders’ during the dosing with clopidogrel alone became ‘low responders’ when clopidogrel was dosed with either of three PPIs, suggesting that DMs could be at a higher risk of becoming a ‘low-responder’ to clopidogrel and could easily become ‘low responders’ when clopidogrel was dosed with a PPI. Therefore, we have to be careful with the concomitant use of clopidogrel and a PPI in DMs of CYP2C19 in particular.
The drug–drug interactions of PPIs with other drugs through cytochrome P450s have not been considered so problematic. Several drug–drug interactions have been reported with omeprazole, but such reports on lansoprazole and rabeprazole have been rare . However, recent reports and our results have demonstrated that omeprazole as well as lansoprazole and rabeprazole attenuate the efficacy of clopidogrel. The following explanations could be offered as the reason why the PPIs that had few reports on drug–drug interactions with many drugs influenced clopidogrel as shown in this study. The major metabolic route of clopidogrel is its metabolism by esterase to the inactive metabolite (SR26334) and the active metabolism of clopidogrel by CYP2C19 is a minor route . Moreover, CYP2C19 catalyzes the two step activation of clopidogrel requires metabolism by CYP2C19 twice . Therefore, in the subjects with decreased activity of CYP2C19, such as DMs, active metabolism of clopidogrel seemed to be easily impaired by concomitant use of a PPI. Moreover, plasma concentrations of PPIs in DMs of CYP2C19 are higher and sustained longer in comparison with RMs [34–36], which could further attenuate the active metabolism of clopidogrel.
However, in our study, some DMs, who were judged to be ‘low responders’, converted to ‘responders’ when a PPI was dosed. There is wide variation in the activity of CYP2C19 in DMs of CYP2C19. The individuals who were judged as the ‘low responders’ might have a lower activity of CYP2C19. However, omeprazole is known to induce CYP1A2 in individuals with lower activity of CYP2C19 . Therefore, the concomitant use of a PPI might induce CYP1A2, which metabolized clopidogrel to the active metabolites, resulting in conversion of some ‘low-responders’ to ‘responders’. Together, concomitant use of a PPI with clopidogrel makes the effect of clopidogrel unstable.
There were wide variation in the IPA in the DMs of CYP2C19, as demonstrated in our previous report . We cannot offer any appropriate explanation for the wide variation in clopidogrel efficacy in DMs. However, we are tempted to hypothesize that most of the DMs are heterozygous for a wild type allele and a mutated allele. Which of the two alleles, a mutated allele or a wild type allele, is dominant, may differ among different individuals. DMs in whom a mutated allele is dominant may show the lower efficacy of clopidogrel and appear ‘low-responders’. On the other hand, DMs in whom the wild type allele is dominant, where clopidogrel is extensively metabolized to the active metabolites, appear ‘responders’.
In the present study, we tested whether the separate dosing of a PPI (i.e. clopidogrel in the morning and omeprazole in the evening) could prevent the drug–drug interaction between clopidogrel and a PPI. However, our study results indicated that separate dosing could not prevent the drug–drug interaction between clopidogrel and a PPI. The results in the RM group suggested that the evening dosing of omeprazole could decrease the influence of omeprazole on the efficacy of clopidogrel dosed in the morning. However, as shown in the present study, the RMs rarely became non-responders even if clopidogrel was dosed concomitantly with a PPI. Therefore, whichever dosing scheme, concomitant or separate dosing of clopidogrel and a PPI, does not appear to be a problem in RMs.
On the other hand, in the DM group, the separate dosing of clopidogrel and omeprazole could not prevent the attenuation of clopidogrel efficacy by omeprazole. Because the plasma half life and Cmax of omeprazole in the DM group are increased in comparison with RMs  and because activity of CYP2C19 to metabolize clopidogrel to its active metabolite is decreased in comparison with RMs, omeprazole dosed in the evening could last for a long in the systemic circulation and interfere with the efficacy of clopidogrel dosed in the morning in DMs.
The prophylactic use of a PPI for GI bleeding in patients undergoing anti-platelet therapy has been recommended [8–10]. However, our results indicate that prophylactic use of a PPI should be based on CYP2C19 genotype status. In the RMs, clopidogrel is extensively metabolized to its active metabolites, which indicates an increased risk of GI bleeding with clopidogrel. Therefore a PPI should be used, because the effect of a PPI on clopidogrel efficacy in this group may be not so problematic. On the other hand, in the DMs of CYP2C19, efficacy of clopidogrel is decreased in comparison with RMs. Moreover, a PPI has a higher risk of conversion from ‘responder’ to ‘low-responder’ in this group. Therefore, we assume that the CYP2C19 genotyping test could be useful for the optimal prophylactic treatment for patients undergoing anti-platelet therapy.
Recently, a histamine H2-receptor antagonist (H2RA) was reported to be effective for the prevention of aspirin-induced gastric injury . Yasuda et al.  demonstrated that an H2RA as well as a PPI completely prevented the GI bleeding and that an H2RA did not increase the risk of stent thrombosis or new lesions in coronary arteries in patients taking dual antiplatelet therapy after coronary stenting, suggesting that an H2RA could replace a PPI as the prophylactic agent in patients taking dual antiplatelet therapy. However, this study is retrospective and no prospective study has been performed. Therefore, which anti-secretory agent, a PPI or an H2RA, is better as the prophylactic agent in dual anti-platelet therapy must be verified from the points of view of safety and efficacy in a future prospective study.
Lastly, our results must be interpreted within the limitations of the study. First of all, our study subjects were all healthy volunteers, not patients. Second, the sample size was small. However, we could not estimate the appropriate sample size because of lack of a preliminary study. Third, we did not use aspirin, although many patients were treated with dual anti-platelet agents such as clopidogrel and aspirin. Therefore, our study results should be considered as preliminary. However, we would like to emphasize that the risk of attenuation of clopidogrel by a PPI depends on CYP2C19 genotype status. Therefore, we believe that the testing of CYP2C19 genotype could contribute to the development of tailored optimal anti-platelet therapy. However, a further large scale study is necessary to verify the clinical usefulness of CYP2C19 genotyping in anti-platelet therapy.