Portal de Periódicos da CAPES

Proton-Pump Inhibitors, Clopidogrel, and Cardiovascular Adverse Events: Fact, Fiction, or Something in Between?

2009; Elsevier BV; Volume: 137; Issue: 3 Linguagem: Inglês

10.1053/j.gastro.2009.07.011

1528-0012

Mary K. Rude, William D. Chey,

Lipoproteins and Cardiovascular Health

Simon T, Verstuyft C, Mary-Krause M, et al. (Assistance Publique-Hopitaux de Paris, Hôpital Saint-Antoine, and Université Pierre et Marie Curie, Paris, France.) Genetic determinants of response to clopidogrel and cardiovascular events. N Engl J Med 2009;360:363–375. Simon et al recently evaluated whether certain known single nucleotide polymorphisms of genes involved in the absorption, metabolism, and biological activity of clopidogrel affect clinical outcomes in patients receiving clopidogrel after an acute myocardial infarction (MI). The investigators performed a post hoc analysis of patients enrolled in the French Registry of Acute ST-Elevation and Non-ST-Elevation Myocardial Infarction (FAST-MI). The analysis included patients who presented with acute MIs, defined as serum markers greater than twice the upper limit of normal, and either electrocardiographic changes or symptoms consistent with acute MI. The patients were subsequently admitted to intensive care units within 48 hours of onset of symptoms and received care "according to usual practice." Outcomes were evaluated for patients who subsequently received clopidogrel therapy. The primary outcome was measured as death from any cause, nonfatal MI, or stroke during 1 year of follow-up after admission for the initial event. Follow-up was available for 99.2% of the patient population. Clopidogrel is a prodrug that is intestinally absorbed and is then hepatically metabolized by the cytochrome P450 enzyme into its active form. The activated drug reduces platelet aggregation by irreversibly antagonizing the adenosine diphosphate platelet receptor, thereby inactivating the fibrinogen receptor. Simon et al performed genotyping for known single nucleotide polymorphisms with functional importance involved in this mechanism of action of clopidogrel. Certain polymorphisms in the genes related to clopidogrel absorption (ABCB1), metabolic activation (CYP3A5 and CYP2C19), and biological activity (P2RY12 and ITGB3) were identified in the study population, and then relationships between genotypes and primary clinical outcomes were evaluated. The FAST-MI study enrolled 3670 patients, of which 2430 patients provided a DNA sample. Of these patients, 2208 received clopidogrel therapy and were therefore included in this post hoc analysis. Of these 2208 patients, 294 experienced a primary outcome event as defined. Those patients who did experience a primary outcome on average were older, and more frequently had comorbidities such as hypertension, diabetes, previous MI, previous percutaneous coronary intervention (PCI), stroke, or heart failure. Patients who experienced primary outcome events less frequently underwent PCI or IV fibrinolysis at the time of their initial MI, and they were less likely to have been treated with statins, β-blockers, angiotensin-converting enzyme inhibitors, glycoprotein IIb/IIIa inhibitors, and heparin. The use of calcium channel blockers, aspirin, and proton pump inhibitors (PPIs) was similar between those who experienced a primary outcome and those who did not. The single nucleotide polymorphisms of the CYP3A5, P2RY12, or ITGB3 genes did not have a significant association with a primary outcome, as measured by a Cox proportional hazards model. However, there were significant increases in primary outcome events, even after controlling for comorbidities and previous treatments, in patients with 2 ABCB1 variant alleles compared with the wild type (15.5% vs 10.7%; adjusted hazard ratio, 1.72; 95% confidence interval [CI], 1.20–2.47), and in patients with 2 loss-of-function CYP2C19 variant alleles compared with those with wild-type alleles (21.5% vs 13.3%; adjusted hazard ratio, 1.98; 95% CI, 1.10–3.58). Of the 2208 patients, 1535 underwent PCI at the time of their initial event, and of these patients there was a significant increase in primary outcome events in those with 2 loss-of-function CYP2C19 variant alleles compared with those with wild-type alleles (adjusted hazard ratio, 3.58; 95% CI, 1.71–7.51). This study also investigated the effect of concomitant PPI use, including omeprazole, on clinical outcomes and found no significant increased rate of a primary outcome event. The authors concluded that, of the 2208 patients enrolled in the FAST-MI study who were subsequently treated with clopidogrel, those individuals with 2 loss-of-function alleles of the CYP2C19 gene had a significantly elevated rate of cardiovascular (CV) events 1 year after MI, as measured by death from any cause, nonfatal MI, or stroke. This rate of CV events was more pronounced in those who underwent PCI. The analysis also showed that PPIs as a class, specifically omeprazole, had no effect on the efficacy of clopidogrel in these patients. Recently, the issue of whether a drug–drug interaction exists between clopidogrel and PPIs has become a topic of interest. In 2008, the American College of Cardiology Foundation, the American College of Gastroenterology, and the American Heart Association recommended in their expert consensus document that all patients who are receiving nonsteroidal anti-inflammatory drugs (NSAIDs), aspirin, dual antiplatelet therapy, or concomitant anticoagulant therapy and who are at risk for gastrointestinal injury should receive prophylactic treatment with a PPI to reduce the risk of ulcer complications and gastrointestinal bleeding (J Am Coll Cardiol 2008;52:1502–1517). Because of the high prevalence of antiplatelet therapy in clinical practice, the possibility of decreased clopidogrel efficacy with PPI use has great clinical significance. There have been several recent publications that have investigated the clinical outcomes of patients concurrently taking clopidogrel and a PPI versus clopidogrel alone. Unfortunately, the results thus far have been conflicting. Furthermore, if an interaction does exist between clopidogrel and PPIs, it is unclear if any such interaction results in clinically meaningful adverse outcomes. Three post hoc analyses have recently been performed to address this issue. Two concluded that PPIs do not, whereas one concluded that PPIs do decrease the efficacy of clopidogrel. Zairis et al performed a post hoc analysis of the Combined Clopidogrel and Aspirin Resistance in Coronary Artery Stenting Study (3C study). The authors evaluated the 1-year incidence of death or MI of patients (n = 612) with stable or unstable coronary artery disease after elective coronary stenting and receiving clopidogrel. There was no difference in adverse CV outcomes in these patients when comparing those who were prescribed a PPI (n = 365) and those not prescribed a PPI (n = 247) with regard to death (95% CI, 0.7–1.4; P = .8) or MI (95% CI, 0.8–1.3; P = .9; Eur Heart J 2008;29[Suppl]:134). This study examined PPIs as a class rather than as individual PPIs and followed only patients who underwent elective coronary artery stenting. Another post hoc analysis was performed by Dunn et al of the Clopidogrel for Reduction of Events During Observation trial (CREDO). In this analysis, the authors evaluated both the 28-day incidence of death, MI, or urgent target vessel revascularization and the 1-year incidence of death, MI, or stroke for patients on clopidogrel with or without a PPI and placebo with or without a PPI. They found that baseline PPI use was independently associated with adverse CV events at 28 days and 1 year in the overall population (OR, 1.6; 95% CI, 1.1–2.5 at 28 days; OR, 1.5; 95% CI, 1.1–2.1 at 1 year; Circulation 2008;118:S815). The Simon study, discussed above, is a third post hoc analysis that analyzed the FAST-MI study (n = 2,208) and concluded that there was no difference in clinical outcomes at 1 year between patients with or without any PPI (P = .75), and also with or without omeprazole (P = .12; N Engl J Med 2009;360:363–375). There have also been several retrospective database analyses that have assessed a potential association between adverse CV events and the concurrent use of PPIs and clopidogrel. Acknowledging the weaknesses of such epidemiologic studies, the weight of this body of evidence suggests that clopidogrel is less efficacious when used with PPIs. Ho et al performed a retrospective cohort study using the Veterans Administration Hospital database to evaluate patients (n = 8,205) hospitalized for acute coronary syndrome (ACS) and subsequently prescribed clopidogrel at discharge. Of these patients, 63.9% were also prescribed a PPI (omeprazole, rabeprazole, lansoprazole, or pantoprazole). Patients were followed for a median of 521 days and monitored for a primary outcome of all-cause mortality or rehospitalization for ACS. Using adjusted ORs, the authors concluded that patients on clopidogrel with a PPI had an increased risk of death or rehospitalization for ACS (adjusted OR, 1.25; 95% CI, 1.11–1.41) compared with patients on clopidogrel alone (JAMA 2009;301:937–944). Juurlink et al also performed a retrospective database analysis from the Ontario Public Drug Program that included patients (n = 13,636) who filled a prescription for clopidogrel within 3 days of being discharged for an acute MI and were monitored for readmission for an acute MI within 90 days. The authors concluded that there was an increased risk of readmission for an acute MI in patients taking clopidogrel and with exposure to a PPI within 30 days (adjusted OR, 1.27; 95% CI, 1.03–1.57) compared with those on clopidogrel alone. Furthermore, stratification of the results showed that current use of omeprazole, lansoprazole, and rabeprazole was associated with readmission (adjusted OR, 1.40; 95% CI, 1.10–1.77), although this was not true for pantoprazole (adjusted OR, 1.02; 95% CI, 0.70–1.47; CMAJ 2009;180:713–718). In a 3rd study, Aubert et al performed a retrospective cohort study of the National MedCo Integrated Database that included patients (n = 14,383) who underwent coronary artery stent placement and were subsequently started on clopidogrel. Of these patients, 68.6% were on clopidogrel alone and 31.4% took clopidogrel and a PPI. These patients were followed for 1 year for hospitalization for stroke, MI, angina, or coronary artery bypass grafting. The authors concluded that patients on clopidogrel and a PPI were more likely to have an adverse event compared with patients on clopidogrel alone. This was true in both patients with no previous CV events (adjusted OR, 1.79; 95% CI, 1.62–1.97) and in those with a previous CV event (adjusted OR, 1.86; 95% CI, 1.63–2.12; Circulation 2008;118:S815). Based on the conflicting results yielded by post hoc analyses of previously performed randomized trials and retrospective database studies, it remains unclear whether PPIs interfere with the cardioprotective benefits of clopidogrel. If there is indeed an association between PPIs, clopidogrel, and an increased risk of adverse cardiac events, how might such an association be explained? It is possible that PPIs in and of themselves increase CV events irrespective of clopidogrel use. In a statement from December 2007, this issue was addressed by the US Food and Drug Administration (FDA), which performed a detailed CV safety analysis using data from 2 long-term studies in patients with severe gastroesophageal reflux disease who were treated with either omeprazole or esomeprazole as well as data from an additional 14 comparative studies of omeprazole, 4 of which were placebo controlled. The FDA's statement was careful to point out that these studies were not specifically designed to assess the risk of adverse cardiovascular events and that patient follow-up was incomplete. Acknowledging these weaknesses, the FDA concluded that the available data "do not suggest an increased risk of heart problems with the use of Prilosec or its newer formulation Nexium" (available: http://www.fda.gov/bbs/topics/NEWS/2007/NEW01754.html). It is also possible that a true interaction exists between PPIs and clopidogrel. Given our understanding of PPI and clopidogrel metabolism, this possibility is biologically plausible (J Am Coll Cardiol 2008;51:256–260). As mentioned, the activation and metabolism of both clopidogrel and PPIs rely on CYP2C19; therefore, co-administration could theoretically lead to changes in the pharmacodynamics and, ultimately, the clinical effects of these drugs. It remains unclear whether any potential interaction between PPIs and clopidogrel is limited to specific PPIs or represents a class effect (Am Heart J 2009;157:148.e1–5; Eur Heart J 2008;29[Suppl]:832). As Simon et al suggest, genetic polymorphisms of CYP2C19 are quite likely to influence the efficacy of clopidogrel (N Engl J Med 2009;360:363–375). If this is true, it is possible that individuals with certain genetic polymorphisms may be more prone to an interaction between PPIs and clopidogrel. A 3rd potential explanation for an apparent association between PPIs and clopidogrel is "channeling bias." Channeling bias refers to the tendency of physicians to prescribe certain medications to patients with more severe underlying illness/es. In this way, the use of certain medications does not truly predispose individuals to develop specific diseases, but rather falsely gives the appearance of an association because persons with certain diseases are more likely to be taking the medication in question. Put simply, channeling bias leads to "guilt by association" rather than a true "cause and effect" relationship. In the case of PPIs and clopidogrel, it is quite possible that patients with more severe CV disease or a greater number of comorbid illnesses are more likely to receive PPI therapy than individuals with less severe CV disease or fewer comorbid illnesses. In fact, individuals with multiple comorbid illnesses are more likely to develop ulcer bleeding when taking aspirin and/or NSAIDs and, as such, most management guidelines suggest that these individuals receive co-therapy with a PPI to reduce the risk of adverse GI outcomes (Am J Gastroenterol 2008;103:2908–2918). There are recent data from retrospective, post hoc analyses that address the possibility of channeling bias. Using data from the CREDO study, Dunn et al concluded that at 1 year, baseline PPI use was associated with an increase in CV events in patients receiving clopidogrel (OR, 1.63; 95% CI, 1.02–2.63), and in those taking a PPI but not clopidogrel (OR, 1.55; 95% CI, 1.03–2.34; Circulation 2008;118:S815). On the other hand, Ho et al concluded that the use of a PPI alone (without clopidogrel) was not associated with an increased risk of death or rehospitalization for ACS (adjusted OR, 0.98; 95% CI, 0.85–1.13; JAMA 2009;301:937–944. Given all of the controversy and conflicting information, how should clinicians handle this contentious issue while awaiting the arrival of more definitive data? The FDA is currently engaged in an ongoing safety review of this topic. In January 2009, the FDA issued an early communication on their website notifying healthcare professionals that a safety review of this topic with cooperation from the manufacturers of clopidogrel remains ongoing. In their statement from January, the FDA recommended the following: "(1) Healthcare providers should continue to prescribe and patients should continue to take clopidogrel as directed, because clopidogrel has demonstrated benefits in preventing blood clots that could lead to a heart attack or stroke. (2) Healthcare providers should re-evaluate the need for starting or continuing treatment with a PPI, including Prilosec OTC, in patients taking clopidogrel. (3) Patients taking clopidogrel should consult with their healthcare provider if they are currently taking or considering taking a PPI, including Prilosec OTC" (available: http://www.fda.gov/cder/drug/early_comm/clopidogrel_bisulfate.htm). A practical if untested strategy to overcome this possible interaction is to dose clopidogrel and PPIs at different times during the day. Evidence suggests that the active metabolite of clopidogrel has a very short half-life and that plasma concentrations are undetectable approximately 2 hours after administration (CMAJ 2009;180:713–718). Similarly, the elimination half-life of PPIs ranges from 1 to 2 hours (Pharmacol Ther 2006;23:2–8). Therefore, dosing these medications at different times of the day could theoretically overcome any competitive drug–drug interaction that exists. There are also some data to suggest that the interaction between PPIs and clopidogrel may not be relevant to all of the individual PPIs. As mentioned, the Juurlink et al study suggested that the proposed interaction between PPIs and clopidogrel may not be relevant to pantoprazole. It is unclear as to why this might be the case, because pantoprazole is extensively metabolized in the liver through the cytochrome P450 system, and the main metabolic pathway involves CYP2C19, as do all other PPIs (CMAJ 2009;180:713–718). Nonetheless, these data raise questions about whether an interaction with clopidogrel represents a PPI class effect or is only applicable to specific PPIs. Finally, it is also possible that newer antiplatelet agents, such as prasugrel, which are less dependent on activation by CYP2C19, may be less subject to an interaction with PPIs. Ultimately, it is likely that large, prospective trials will be necessary to more definitively settle this issue. In the meantime, it is important for consultants in gastroenterology to weigh the available evidence and to consider risk management strategies that minimize the potential impact of any interaction that might exist between PPIs and clopidogrel. At the same time, it will be important for gastroenterologists to continue to advocate for gastroprotective strategies in patients taking antiplatelet therapy and/or an NSAID who are at increased risk of developing an adverse GI outcome. Let's not "throw the baby out with the bathwater"—at least not yet.