Reports

This report is based on medical evidence presented at sanctioned medical congress, from peer reviewed literature or opinion provided by a qualified healthcare practitioner. The consumption of the information contained within this report is intended for qualified Canadian healthcare practitioners only.

PRIORITY PRESS - 58th Annual Scientific Session of the American College of Cardiology

Orlando, Florida / March 29-31, 2009

The antiplatelet clopidogrel has a pivotal role in reducing the risk of thrombotic events in a variety of populations, including those who have had a previous myocardial infarction (MI). Proton pump inhibitors (PPIs), which reduce gastric acid secretion to prevent acid-related diseases such as reflux esophagitis, are among the most widely prescribed agents in the world. Concern that a drug-drug interaction between these agents leads to diminished clopidogrel activity has generated a recent series of publications yielding conflicting data. However, the most recent studies observing the relative interactions within the PPI class provide an explanation: the specific pathways of PPI metabolism are important to risk, and the interaction is not a class effect.

Observational Study

“We have demonstrated that PPIs do not have the same effect on platelet response to clopidogrel. This is consistent with our own clinical observations, and it explains why there has been inconsistency in the retrospective studies,” reported Dr. Dirk Sibbing, Deutsches Herzzentrum München, Germany. He suggested that those studies which have associated PPIs with adverse outcomes were likely to have had a high representation of patients taking a PPI that interacts with clopidogrel, whereas studies unable to identify an association were more likely to have had a high representation of PPIs with a low risk of an interaction.

In the cross-sectional observational study, 1000 patients who had undergone a percutaneous intervention (PCI) for coronary artery disease (CAD) were enrolled. All patients were treated with dual antiplatelet therapy consisting of 75 mg clopidogrel and ASA. Platelet function testing was performed at a follow-up visit by standardized measurement of adenosine diphosphate (ADP)-induced platelet aggregation with multiple electrode platelet aggregometry (MEA). Of the 1000 patients evaluated, 268 were on concomitant PPI therapy and 732 were not taking a PPI.

Platelet aggregation values differed markedly. MEA expressed as area under the curve of arbitrary units (AU*min) averaged 220 AU*min in those not taking a PPI. In those taking esomeprazole, aggregation was slightly less (209 AU*min). In those taking pantoprazole, aggregation was slightly more (226 AU*min). Neither measure was significantly different than that of non-PPI users. However, the average MEA aggregation among those taking omeprazole was 295.5 AU*min, which was significantly greater than PPI non-users (P=0.001).

“In a previous study when we defined a low clopidogrel response as the upper quintile of MEA from a large set of prospectively collected data, a low response was significantly associated with an increased risk of thrombotic events,” Dr. Sibbing reported. In this study, using the same value for low response, there were no significant differences between esomeprazole, pantoprazole and no PPI-use, but “the low response rates were greater on omeprazole vs. non-PPI use, and the difference was highly significant [32.8% vs. 19.1%; P=0.008].”

Corroborative Findings

The same findings were generated by a similar study published electronically shortly before the ACC (Siller-Matula et al. Am Heart J 2009;157:148:e1-148.e5). In this study, 300 patients with CAD undergoing PCI were evaluated with the same ADP-induced platelet aggregation used in the ACC study and by vasodilator-stimulated phosphoprotein phosphorylation (VASP), which generated a platelet reactivity index (PRI). The study only compared PPI non-users to those on esomeprazole or pantoprazole, but there were no significant differences. This result was considered to be consistent with what is currently known about the risk of interaction.

The Metabolic Pathway CYP2C19

Although all PPIs are metabolized to a varying degree by the P450 cytochrome system, not all PPIs are dependent on CYP2C19. The authors observed that neither esomeprazole nor pantoprazole are dependent on the CYP2C19 isoenzyme, concluding that drug interactions observed between PPIs and clopidogrel should not be regarded as a class effect.

The evidence for the clinical significance of an interaction between PPIs and clopidogrel remains mixed. In new data presented here at the ACC, 535 patients enrolled in the National Heart, Lung, and Blood Institute (NHLBI) Dynamic Registry at the University of Pittsburgh, Pennsylvania, were evaluated for events after one year on clopidogrel. Of the patients, 138 patients (26%) were also taking PPIs and 397 (74%) were not. The two groups were comparable for age, prevalence of comorbidities such as diabetes or renal impairment, smoking habits, or procedural success.

At the end of one year, there were no differences in adverse outcomes between the two groups; rather, absolute rates of many of the major outcomes were slightly lower in the group taking a PPI. This included MI (3.7% vs. 4.2%; P=0.83), coronary artery bypass grafting (3.1% vs. 4.1%; P=0.53), death (3.0% vs. 5.9%; P=0.18) and death plus MI (6.7% vs. 9.6%). Repeat PCI (13.4% vs. 10.1%; P=0.23) and repeat revascularization (15.8% vs. 14.2%; P=0.65) were slightly higher in the group receiving PPIs, but again, the differences were not significant.

“Our data do not support an increased risk of thrombotic events or other adverse outcomes in patients on clopidogrel who are also taking a PPI,” stated Dr. José F. Ramírez, Cardiovascular Institute, University of Pittsburgh. He concluded that PPI treatment in patients on clopidogrel after PCI “appears to be safe.”

Differentiating Between PPIs

However, based on the pharmacology of PPIs and the apparent mechanism of drug-drug interactions, differentiation of PPIs may be essential. Population-based studies are hard to interpret when the proportional representation of specific PPIs is unknown, particularly because formulary inclusions and exclusions may allow one or two PPIs to dominate in a given study. In fact, the findings of the Pittsburgh study are contradicted by one from Denver published just weeks earlier (Ho et al. JAMA 2009;301:937-44). In that study of 8205 patients taking clopidogrel after hospital discharge for an acute coronary syndrome (ACS), PPI use was associated with a 25% increased risk of death or rehospitalization for ACS (HR 1.25; 95% CI, 1.11-1.41) over follow-up ranging from two to five years. Neither the Pittsburgh nor Denver studies provided data on the rate of specific PPI use, which may explain these differences.

The importance of differentiating PPIs was a basis for the study led by Dr. Sibbing. He noted that concern about an increase in thrombotic events developed when physicians at his institution became aware that patients with a discharge prescription for a preferred PPI were returning at follow-up visits on omeprazole. At that point, there was already concern about the risk for drug-drug interactions, so platelet function was tested in these individuals based on type of PPI they were currently taking. Only three PPIs were represented in the study because only three had been offered to the participating patients.

“We found that many patients were switched automatically to generic omeprazole after discharge as a cost-saving measure,” Dr. Sibbing told delegates. Now that he has obtained objective evidence that there are differences between omeprazole, esomeprazole and pantoprazole in platelet aggregation, Dr. Sibbing stated that steps have been taken to prevent future generic substitution, including educating patients about the potential differences between PPIs.

Summary

The clinical significance of the potential drug-drug interaction between PPI therapy and clopidogrel remains unclear because of conflicting data, but new evidence suggests that the risk of an interaction at the level of platelet function is not shared equally by agents within the PPI class. Consistent with the mechanism of metabolism, PPIs dependent on the CYP2C19 pathway, such as omeprazole, appear to pose the greatest risk of diminishing the antiplatelet effect of clopidogrel, which uses the same metabolic pathway. Esomeprazole and pantoprazole, which employ a broader array of metabolic pathways, have not been associated with any impact on the ability of clopidogrel to inhibit platelet activity. These findings may be important to drug selection in the large population who require both clopidogrel and a PPI.