July 16th, 2012

Does PPI + DAPT = MI?


Dual antiplatelet therapy (DAPT) with aspirin and clopidogrel is the standard of care for prevention of coronary stent thrombosis. Unfortunately, DAPT is associated with an increased risk for gastrointestinal hemorrhage. Although this risk can be reduced with the concomitant administration of a proton pump inhibitor (PPI), initial observational studies suggested that PPI use with DAPT is associated with an increased incidence of myocardial infarction (MI), which was attributed to PPI inhibition of the CYP2C19 cytochrome pathway necessary for the activation of clopidogrel. However, the authors of a study released last week concluded that “the interaction between proton pump inhibitors and clopidogrel is clinically unimportant.”

Why don’t the studies agree?  

Observational studies — like those first reporting an adverse interaction of PPI and DAPT — attempt to control for confounding variables (i.e., using proportional hazards models), but they can’t adjust for factors that weren’t measured or those associated with PPI use that may independently increase the risk for MI. In other words, differences between patients who are and are not prescribed PPIs could account for the harmful effects observed.

To address this issue, Douglas and colleagues performed two different observational studies on the same dataset. The first was designed as a traditional cohort study, and the second was designed as a self-controlled case series (in which subjects function as their own controls, eliminating confounding between subjects). As in previous cohort studies, the traditional analysis showed that use of PPIs with DAPT was associated with a significantly increased risk for MI (hazard ratio, 1.35; 95% confidence interval, 1.26–1.45). However, the within-subject analysis on the same dataset found no association between PPI use and MI (hazard ratio, 0.75; 95% CI, 0.55–1.01). To confirm their findings, the investigators performed the same analyses for other known CYP2C19 inhibitors and again found no associations with MI in the within-person analyses despite significant associations with MI in the proportional-hazards analyses.

So which is it. . .are PPIs safe or harmful when used with DAPT?

When residual confounding is minimized, no association is demonstrable between PPI use and MI in patients receiving DAPT.

So, PPI or not?

We agree with the the author of the accompanying editorial that clinicians should strongly consider prescribing a PPI to all patients on DAPT who are considered to be at increased risk for gastrointestinal complications, including those over 60 years of age, those with medical comorbidities, and those taking nonsteroidal anti-inflammatory drugs or anticoagulants.

Are you convinced it’s safe to administer PPIs to your patients on DAPT?

5 Responses to “Does PPI + DAPT = MI?”

  1. aziz ismail, MD says:

    Yes, weighing all the factors i still consider it is safe.

  2. It might be more safe with pantoprazol than omeprazol. Is this the way out? (… according to some data … the weight of evidence I cannot remember)

  3. Jeff Dickey, MD says:

    why would pantoprazole be better?

  4. All PPIs are metabolized by CYP2C19, but to a varying degree. Thus interaction with clopidogrel. Pantoprazol may be better in this aspect. Yet the data are not consistent. And what I asked myself and found no answer from colleagues is the weight of evidence. Just some links to the topic. „The interaction between clopidogrel and proton pump inhibitors (PPI): is there any clinical relevance?“ http://www.ncbi.nlm.nih.gov/pubmed/22291500
    Differential clinical outcomes are important, I think, because “Platelet function studies do not demonstrate a clear or consistent interaction between clopidogrel and PPIs.“ http://www.ncbi.nlm.nih.gov/pubmed/22204719
    And some more links …

  5. Diego Fernandez Redondo, Medical Doctor, Cardiology. says:

    I think they are safe. We don´t forget that this kind of patients have a greater risk of a new event than general population. The warning starts with an observational article with few patients. Since no other evidences were found. If a really clinical interaction exists, I´ll hope see more evidence publicated.
    Kind regards