Clopidogrel: What Is a Cardiologist to Do?

The FDA required a black box warning for clopidogrel based on the assumption that it was a CYP2C19 substrate. Therefore, those patients with defective CYP2C19 alleles would not be as able to effectively metabolize clopidogrel to the active metabolite. The boxed warning informed the physician that “tests are available to identify a patient's CYP2C19 genotype and can be used as an aid in determining therapeutic strategy.” Thus, in some hospitals, as well as the cost, PCIs could be delayed, until such genotyping was completed, as the hospital might be subject to liability, if there was a bad outcome in a patient, whose CYP2C19 genotype was unknown. Unfortunately, because FDA did not require the patients to be genotyped, most insurers did not reimburse additional costs incurred by genotyping. Those patients needing a percutaneous coronary intervention (PCI) are at high risk and FDA acted appropriately in placing this warning. There were, though, already some concerns, as to whether clopidogrel was indeed a CYP2C19 substrate, as it had previously considered to be a CYP3A substrate. Now there is serious doubt. FDA should review their position on the clopidogrel black box and ensure that the label is still “truthful and not misleading.” While most physicians welcome prompt action by FDA when a risk with a drug is identified, others have noted FDA's inconsistencies in their updating of drug labels. Seminerio and Ratain1 have suggested that FDA adopt a more stringent protocol for such updating. Based on the amount of usage, clopidogrel should surely qualify for such an updating. The metabolism of clopidogrel was studied initially at Sanofi and the structure of the active metabolite was determined, except for the stereochemistry at C4.2 An article by Savi et al3 defined a chemically reasonable process to the active metabolite, namely a single oxidative step to 2-oxo-clopidogrel and hydrolysis to the active thiol acid. The characterization of clopidogrel, as a CYP2C19 substrate, stimulated a lot of interest amongst geneticists. Gilard et al4 carried out a study in 124 patients, who had had a coronary artery stent placed. They were randomized post-procedure to placebo or omeprazole 20 mg/day and treated with aspirin and clopidogrel. Platelet aggregability was measured by the VASP method. It was shown that omeprazole antagonized effects of clopidogrel on platelet aggregability. As part of the clinical effort by Lilly and Sankyo to develop prasugrel, multiple articles appeared to support the role of CYP2C19 in the metabolism of clopidogrel. However, the majority of such articles used platelet aggregability, as a surrogate for a clinical outcome. This issue of the Journal includes another such article, this time from the FDA Clinical Pharmacology Division.5 This is a small study of 18 subjects (6 EMs, 6 IMs, and 6 PMs) who were treated with clopidogrel in a multi-dose study, three-period, fixed-sequence crossover. They concluded that increased doses of clopidogrel might be necessary to overcome the effects of poor CYP2C19 metabolism. The study seems well conducted. Metabolic processes in man, though, still observe the basic laws of chemistry. There is only one oxidation step in the sequence from clopidogrel to the active thiol acid, as Savi pointed out. The intermediate 2-oxo-clopidogrel can open hydrolytically. Mansuy and Dansette, who derived the present metabolic scheme, postulated two oxidation steps and a final reduction step.6 As the postulated intermediates would be easily transformed by lipases to the inactive carboxylic acids, this scheme seems improbable. Also metabolism involving multiple CYPs and pathways is hard to block with a single inhibitor. The underlying problem is that when the initial in vitro metabolic work was carried out in human hepatosomes, it showed clopidogrel to be a CYP3A substrate. These findings were confirmed in the early work at Sankyo by Kurihara et al7 using human hepatosomes. Subsequently they were ignored in favor of the data derived from recombinant CYPs. It seems, there was no work done to cross validate the two methods, as was done with CYP2C9 by McGinnity et al.8 He recognized that false negatives can occur with recombinant CYPs. Clopidogrel is thought to undergo rapid first pass metabolism by CYP3A within the enterocyte of the small bowel. Human hepatosomes can replicate the intracellular milieu of the enterocyte, whereas the recombinant CYP3A cannot. Many studies have appeared in an attempt to support the revision of the metabolic pathway. One was a placebo controlled crossover interaction study of omeprazole and placebo in healthy volunteers treated with clopidogrel.9 The volunteers were initially given 300 mg of clopidogrel and the subsequently 75 mg per day. Both Day 1 and Day 5 data showed a statistical reduction in the amount of active metabolite during the omeprazole treatment. The amount of the thiol acid, the active metabolite, was decreased by 40%, but this is hardly the magnitude of effect that one would expect to be clinically meaningful. There were problems with the study. In an effort to ensure statistical significance, large group sizes (∼65) had to be employed and there was no measurement of 2-oxo-clopidogrel levels. The 2-oxo-clopidogrel intermediate was purported to undergo further oxidation by CYP3A, to yield the active metabolite.10 This is, though, a rather implausible hypothesis as thiolactones open hydrolytically not oxidatively. An additional problem with the two step oxidation mechanism is that while a CYP2C19 substrate, omeprazole, undergoes rapid hydroxylation to 5-hydroxy omeprazole, neither clopidogrel or 2-oxo-clopidogrel undergo oxidation under such conditions.11 Recently, a drug interaction cross-over study of clopidogrel with grapefruit juice versus water has been published.12 There was no need for large group sizes. Fourteen healthy volunteers were adequate to demonstrate a profound reduction in the active metabolite AUC. The AUC0–inf fell from 167 ng h/mL in the water treatment to 27 ng h/mL with grapefruit juice. Grapefruit juice is an established CYP3A inhibitor. The effects on platelet aggregability have served to provide support for the clinical consequences of CYP2C19 loss-of-function alleles. The CYP2C19 hypothesis has been supported by a large number of clinical studies in both ACS and PCI patients that used platelet aggregation as a surrogate endpoint for clinical outcomes. There were doubts raised about the validity of this assumption. One study that compared the various platelet aggregation methods concluded that light transmittance aggregometry, Verify Now, Plateletworks, and Innovance PFA P2Ygave results that were associated with the primary end-point. However, the tests were only modestly predictive and were unable to even provide prognostic information regarding the risk of bleeding.13 There were further concerns. One large clinical study with clopidogrel, CURE, was reinvestigated.14 Many of the acute coronary syndrome (ACS) patients in this study were located and offered CYP2C19 genotyping. From the CURE study, 5,059 of the original 12,562 patients agreed. They were well distributed between clopidogrel (2,549) and placebo (2,510). The effect of clopidogrel as compared to placebo was consistent and independent of CYP2C19 loss-of-function carrier status. At the AHA Meeting in 2012, Dr. Gurbel and Dr. Montalescot published their large prospective sub-studies from TRILOGY-ACS15 and ARCTIC.16 It was concluded in TRILOGY-ACS that there was “no significant association existed between platelet activity and occurrence of ischemic outcomes” and in ARCTIC that “This study showed no significant improvements in clinical outcomes with platelet function monitoring and treatment adjustment for coronary stenting, as compared with standard antiplatelet therapy without monitoring.” If clopidogrel were a CYP2C19 substrate, then one would expect to see its effects impaired by concomitant administration of a CYP2C19 inhibitor, such as omeprazole. There was one prospective study of clopidogrel that was randomized to placebo or omeprazole.17 It was not completed due to the bankruptcy of the sponsor. Omeprazole had a significant and positive effect on the incidence of bleeding. There was no trend towards antagonizing the effect of clopidogrel on major adverse cardiovascular events. There have been many meta-analyses looking at the effects of omeprazole and other proton pump inhibitors (PPIs) that inhibit CYP2C19. One such study attracted attention.18 Holmes reported that from a study of 42,016 patients involving 3545 cardiovascular events “there was no significant association of the CYP2C19 genotype with cardiovascular events” There are many other meta-analyses that support this conclusion, for example, Kwok and Loke.19 The clinical studies of clopidogrel do present a confusing picture. Some meta-analyses of interest on initial review of the data can show an effect of a CYP2C19 defective allele. When the data are re-reviewed for bias and adjusted for quality modifiers, such effects can disappear.20 However, the metabolism of clopidogrel does not present a confused picture. It is primarily a CYP3A substrate. The TV commercials that tell one not to take clopidogrel with Nexium® are not supported by the data. There is also need for clarification by FDA regarding the use of clopidogrel, so it may be used in procedures by cardiologists, without the suggested need for CYP2C19 genotyping.