EMEA Issues Warning on Possible Clopidogrel-PPI Interaction, But Is There Really a Problem?

June 19, 2009

June 19, 2009 (London, United Kingdom) — The European Medicines Agency (EMEA) has now issued a public statement on a possible interaction between clopidogrel (Plavix, Sanofi-Aventis/Bristol-Myers Squibb) and proton-pump inhibitors (PPIs) and has recommended that the product information for all clopidogrel-containing medicines be amended to discourage concomitant use of PPIs unless absolutely necessary [1].

The UK medicines regulator, the Medicines and Healthcare Products Regulatory Agency (MHRA), has also issued advice to GPs that concomitant use of a PPI with clopidogrel is not recommended unless considered essential, urging a review of the prescribing of PPIs at the next appointment for patients taking clopidogrel.

This follows an "early communication" issued by the US FDA earlier this year, stating that PPIs might interfere with the effectiveness of clopidogrel and that clinicians should reevaluate starting or continuing treatment with a PPI in patients taking clopidogrel.

But several leading cardiologists have voiced concern that the studies on which these warnings are based have many limitations and that it is far from certain whether there really is an interaction between clopidogrel and PPIs.

The EMEA statement points out that, as heartburn and stomach ulcers can occur as side effects of clopidogrel, patients taking clopidogrel often take PPIs to prevent or ease these symptoms. Figures from the UK estimate that around 500 000 patients in that country are currently prescribed clopidogrel and around half are also prescribed PPIs. Many more may be buying omeprazole over the counter, as it is available without prescription. Other PPIs available on prescription include esomeprazole, lansoprazole, pantoprazole, and rabeprazole.

EMEA says the new concern "relates to several recently published studies examining clinical outcomes of clopidogrel users. Taken together, these studies suggest that a significant interaction might occur between clopidogrel and members of the PPI class of medicines, making clopidogrel less effective when given with these medicines." It adds: "One possible explanation for this observation is that some PPIs prevent the conversion of clopidogrel into its biologically active form in the body, reducing the effectiveness of clopidogrel and increasing the risk of heart attack or other conditions involving harmful clotting (eg, strokes). However, as different PPIs have different capacity to affect the metabolism of clopidogrel and as the outcome studies have not fully reflected the different effect of PPIs on activation of clopidogrel, there may be more than one explanation for the effect of this class of medicines on clopidogrel."

Is It Just Confounding?

But some cardiologists contacted by heartwire about this issue are concerned that the clopidogrel-PPI interaction has been given too high a profile, given that the clinical studies suggesting such problems are all observational.

Dr Peter Berger (Geisinger Health, Danville, PA) commented to heartwire : "There are lots of heavily confounded registry data circulating that have been widely reported and given more weight than they deserve."

He noted that these studies have shown that patients on clopidogrel and a PPI do worse than those on clopidogrel alone, but that there are three explanations for why that might be the case:

  • PPIs may interfere with the production of the clopidogrel active metabolite.

  • PIs may directly cause harm.

  • The effect may be due to confounding in the reported studies.

Berger says: "Everyone is assuming that the first explanation--that PPIs interfere with the metabolism of clopidogrel--is the reason, but I believe the third explanation is most likely correct. There is enormous confounding in these studies. Patients taking clopidogrel plus a PPI were much older and sicker than those taking clopidogrel alone. No amount of statistical adjustment can eliminate confounding when there are such large differences. That is always the problem with observational studies."

There are lots of heavily confounded registry data circulating that have been widely reported and given more weight than they deserve.

Berger also points out that there are two randomized databases that are not subject to confounding, and both suggest that there is no interaction between clopidogrel and PPIs. His group has performed an analysis of the CREDO trial, which was presented at the AHA 2008 meeting and showed the same reduction in risk with clopidogrel regardless of whether patients were or were not taking PPIs. And PPI use was independently associated with adverse cardiovascular events at 28 days and one year in the overall population.

The second randomized trial database is the TRITON trial, where, Berger notes, prasugrel showed the same relative benefit over clopidogrel in patients taking a PPI and those not taking a PPI. "Prasugrel is not believed to be affected by a PPI interaction, and if there were an interaction with clopidogrel and PPIs, you would expect to see a larger relative benefit of prasugrel over clopidogrel in patients taking PPIs than those not taking PPIs, but that was not seen," he added.

"In all randomized data so far, patients taking PPIs have more events--that is because they are older, with more comorbidities, and taking more medications. Patients on H2 blockers do worse as well. This tells us that patients who have stomach problems are older and sicker," Berger stated.


He notes that concerns about other drugs that were thought to interact with clopidogrel have been raised before but were subsequently dropped. "This happened with atorvastatin and calcium blockers, but these concerns were refuted in both cases when randomized studies were analyzed," he noted. "While no one knows for sure whether there is an interaction with PPIs or not, I believe that the burden of evidence at the moment supports the view that PPIs do not interact with clopidogrel. I believe the FDA and EMEA statements are premature," he added.

Issue Overpublicized?

Dr Shamir Mehta (McMaster University, Hamilton, ON) is another who believes the evidence for the interaction is shaky. He raised very similar points to Berger's. "The cohort studies suggesting an interaction are very interesting, but they are observational, and patients taking PPIs are fundamentally different from those not taking PPIs. There are probably many unmeasured variables that could account for the difference in events seen. And data from the randomized trials (CREDO and TRITON) do not support an interaction. The fact is that we simply don't know."

Mehta is not critical of the regulatory authorities' statements on the possible interaction, saying, "They are just being cautious." But he expressed surprise about the amount of publicity that this issue has attracted. "There are observational analyses published all the time suggesting different things, but this PPI-clopidogrel interaction has sparked a huge amount of interest. Why this has been singled out I don't know."

There are lots of heavily confounded registry data circulating that have been widely reported and given more weight than they deserve.

The fact is that we simply don't know.

Another to voice a similar view is Dr Gabriel Steg (Centre Hospitalier Bichat-Claude Bernard, Paris, France). He commented to heartwire : "I think the recommendation that clopidogrel and PPIs should not be used together is too strong. Yes, there are good pharmacokinetic data showing an interaction between clopidogrel and omeprazole, but there are other studies suggesting that the clinical impact of the pharmacokinetic interaction may be limited." Steg also points out the hazards of observational studies: "We just need to remember the many observational studies that pointed to the major cardiovascular benefit of HRT in postmenopausal women." He adds: "More important, patients receiving clopidogrel often receive aspirin and other antithrombotics, and the potential for GI bleeds in this population is also real. So it is important to look at the big picture of overall clinical benefit of prevention of GI bleed vis-a-vis the potential for reduced effectiveness. In my view, the data are not definitive yet."

Other experts were more cautious. Dr Shaun Goodman (St Michael's Hospital, Toronto, ON) said: "We’ve been 'fooled' by observational studies before. But there is some biological plausibility based on the genetic-polymorphism data and some platelet-inhibition studies that warrant caution." Dr Robert Harrington (Duke Clinical Research Institute, Durham, NC) pointed out that in some of the observational data sets there was inadequate/incomplete information on concurrent aspirin use and that the higher event rate in the PPI patients may have been simply due to lack of aspirin in patients at risk for GI bleeding. But he added, "Nonetheless, there is a consistency across the observational data that should cause us to take pause."

Is There a Difference Between PPIs?

Another point of uncertainty is whether there may be a difference between individual PPIs, with some pharmacodynamic studies suggesting an interaction with omeprazole but not with pantoprazole. Dr Dirk Sibbing (Deutsches Herzzentrum, Munich, Germany), who conducted one such study, says he believes there is a difference between the various PPIs. "From my point of view, pantoprazole is safe to use with clopidogrel. The clinical evidence, however, is conflicting. But there has been one clinical trial from Canada suggesting an interaction with omeprazole but not with pantoprazole, which fits with our results," he said. He added that from a mechanistic view it is known that omeprazole is metabolized by the CYP219 enzyme, which converts clopidogrel into its active metabolite. And while pantoprazole can also be metabolized by this enzyme, it also uses other routes.

So What Are Doctors Supposed to Do?

All the experts contacted by heartwire agreed that although the data are not definitive, doctors should still be cautious, and PPIs should be prescribed to patients taking clopidogrel only if they are having stomach problems that are not controlled with H2 antagonists. There has been a recommendation that PPIs be given as blanket gastric protection to patients at risk of gastric problems taking dual antiplatelet therapy, but everyone heartwire spoke to said this was no longer advisable, given the possibility of an interaction. Goodman said he didn’t think the data were strong enough to support routine use of a PPI with dual antiplatelet therapy even before this whole interaction issue came to the forefront.

Remember the many observational studies that pointed to the major cardiovascular benefit of HRT in postmenopausal women.

Berger said: "Obviously, given the uncertainty, caution is appropriate, but in patients who we know need a PPI, I would not change or stop therapy at this point just because they are taking clopidogrel."

Sibbing commented: "In our clinic, we try not to use PPIs in patients taking clopidogrel, but if they really need one, we would give pantoprazole." But Berger says he would not differentiate between any of the PPIs or switch a patient off omeprazole and onto a different PPI. "Okay, some pharmacodynamic studies may have suggested a problem with omeprazole but not with the others, but I do not believe it is appropriate to guide therapy based on ex vivo studies of platelet function quite yet," he argued.

Harrington believes that the decision whether to prescribe a PPI to a patient on clopidogrel must be made on an individual patient basis. "A blanket statement to use or not use these drugs in combination seems to be overstepping the available evidence. Caution should be used regarding combining the drugs, selecting types of stents (ie, drug-eluting) that necessitate more prolonged antiplatelet therapies in patients with an increased risk for GI bleeding, etc," he commented. Dr Deepak Bhatt (VA Boston Healthcare System, MA) said: "There is a great deal of confusion. For the time being, it seems reasonable to make sure that patients who are prescribed PPIs really have a good indication for them. Beyond that, it seems premature to change clinical practice, unless new, more compelling data become available."

More Data on the Horizon

All the experts called for a randomized trial to investigate this question. But the only one that was under way--COGENT-1--was stopped prematurely by the sponsor due to financial issues. But other ongoing trials of clopidogrel have been tracking PPI use, including two large studies being reported at the forthcoming European Society of Cardiology meeting--OASIS 7/CURRENT and PLATO--and Berger has also requested permission from Bristol-Myers Squibb/Sanofi-Aventis to analyze the CAPRIE database regarding PPI usage. So more data should be available soon.