PPIs and Cardiac Risk: Unveiling the Mask of Truth
PPIs and Cardiac Risk: Unveiling the Mask of Truth
Hello. I'm Dr David Johnson, professor of medicine and chief of gastroenterology at Eastern Virginia Medical School.
Proton pump inhibitors (PPIs) have gotten a lot of press recently. As you may recall, I posted a review on the scare that PPIs may be associated with myocardial infarction. I wanted to follow up on that discussion.
In that previous discussion, I cited the PLOS One article from the group at Stanford University and the Texas Heart Institute where they looked at the data extraction in patients and suggested that there was an increased risk for myocardial infarction with an odds ratio of 1.16 and increased myocardial death as it relates to the cardiac risk by a factor of 2. There was a lot of stir that PPIs may be associated with myocardial infarction and myocardial death.
The synopsis was that there was an animal study from their lab that suggested that the vascular flexibility changed through nitric oxide synthase release and that there was some potential risk that carried forward to a myocardial event and consequent mortality.
I noted that it is a retrospective study, a data extraction, a pharmacovigilant study that really could not co-adjust for the cardiac risk factors. I was very critical of their analysis because you couldn't co-adjust for other cardiac risk factors, such as obesity, diabetes, hypertension, and smoking. Again, these things would all be categoric, especially when your odds ratios are relatively low.
In addition, I noted that there was a significant variance in the animal study that suggested the endovascular change in the flexibility of blood vessels through nitric oxide synthase. The same authors had an article in press that showed in a prospective study that none of these changes were evident in humans.
So, again, it is more reassuring to us that PPIs, when taken appropriately, should be taken with confidence that there is not a conveyed myocardial risk.
A fly in the ointment—or a resolution to the story—is the most recent study that was published in Open Heart.
The authors evaluated the incidence of cardiovascular events and gastrointestinal (GI) bleeding in patients receiving clopidogrel with and without PPIs. It was an updated meta-analysis. The background is the whole story of PPIs potentially causing harm as it relates to clopidogrel, its interaction with the cytochrome P450 2C19 pathway, and the delayed biotransformation of clopidogrel, which is necessary for activation to be an antiplatelet agent.
This group did a phenomenal job. They systematically searched through the Cochrane Central Register of Controlled Trials and used all the rules of engagement for systematic reviews and meta-analyses following the appropriate gold standards. They applied a funnel plot analysis to look for publication bias. They ended up with 39 studies with a total of about 215,000 patients, of whom 74,000 (34%) received a combination of clopidogrel and a PPI.
What did they surmise? In a pooled analysis, all-cause mortality (ie, myocardial infarction, stent thrombosis, and cerebrovascular events) was actually more common in patients receiving both drugs.
So do we stop the presses? Do we say, "Yes, this corroborates what the Stanford group had shown us and that we had criticized?" No, not at all, because they then looked at the tests of heterogeneity. They found that there was significant heterogeneity, meaning differences of methodologies, across these studies.
When they called those out and looked at randomized controlled trials and propensity match studies, they came down to eight of those studies. This culled it to about 24,000 patients. They looked at all of these endpoints, and they found no significant cardiovascular risk. In fact, as it relates to revascularization, there was a protective factor, a decreased odds ratio, if you used a PPI. No adverse association was found across all of the other cardiovascular endpoints (ie, all-cause mortality, myocardial infarction, stent occlusion, and cerebrovascular accidents).
Again, these things all came down to potential stratification bias. These patients all had comorbid risks, so it wasn't that the PPI was causing this. It was just being used in patients with comorbid risks.
What does this tell us? Clearly it says that the patients who had cardiovascular risks also had other risks for comorbid diseases, which was the trump card for why they developed all of these endpoints of adverse events.
When they looked at the GI bleeding risk, there was a risk reduction of 60%. Again, this is something that plays out like what we've seen in trials, like the COGENT trial in the New England Journal of Medicine, where clopidogrel and omeprazole were randomized against clopidogrel alone. We do know that there was a risk protection as it relates to the overall risk for GI bleeding. Clearly, this was reinforced in this most recent meta-analysis.
I think we can say that PPIs are not an issue as it relates to cardiovascular risk.
As an aside issue in this particular study, they looked at five PPIs. Dexlansoprazole was not part of their analysis; but for omeprazole, esomeprazole, rabeprazole, pantoprazole, and lansoprazole, there was equal risk when they looked at the unbalanced cardiovascular endpoints of adverse events associated with PPI use. Co-adjusted, there was no relative risk when they looked in the randomized controlled trials.
What this tells us is that the patients receiving PPIs clearly have a higher burden of comorbidities. Thus, they are more likely to have an increased burden for adverse cardiovascular events. PPIs, however, have a significant potential to decrease the upper GI bleeding in patients taking clopidogrel.
I think this new analysis tells us that if you would appropriately co-adjust the risk for indications for PPI use in these patients, it's not a cardiovascular risk—it's a GI risk. They need these medications for prevention of an upper GI bleed. The cardiovascular risk should not be a concern as it relates to all of these particular studies.
I think we've unveiled the mask of truth on this. The evidence supports that when used appropriately in patients with a relative risk for upper GI bleeding, clopidogrel and PPIs can be done safely. Patients can be reassured, and I think we, as prescribers, can be reassured as well.
Hopefully this sets the record straight. The mask is off. I think this really should end the conversation as it relates to relative risk in this population. Certainly, PPIs are appropriate in the relatively defined high-risk patients.
I'm Dr David Johnson. I hope this provides some useful information for the next time you counsel your patients. I look forward to chatting with you again soon.
PPIs and Cardiovascular Risk: A Follow-up
Hello. I'm Dr David Johnson, professor of medicine and chief of gastroenterology at Eastern Virginia Medical School.
Proton pump inhibitors (PPIs) have gotten a lot of press recently. As you may recall, I posted a review on the scare that PPIs may be associated with myocardial infarction. I wanted to follow up on that discussion.
In that previous discussion, I cited the PLOS One article from the group at Stanford University and the Texas Heart Institute where they looked at the data extraction in patients and suggested that there was an increased risk for myocardial infarction with an odds ratio of 1.16 and increased myocardial death as it relates to the cardiac risk by a factor of 2. There was a lot of stir that PPIs may be associated with myocardial infarction and myocardial death.
The synopsis was that there was an animal study from their lab that suggested that the vascular flexibility changed through nitric oxide synthase release and that there was some potential risk that carried forward to a myocardial event and consequent mortality.
I noted that it is a retrospective study, a data extraction, a pharmacovigilant study that really could not co-adjust for the cardiac risk factors. I was very critical of their analysis because you couldn't co-adjust for other cardiac risk factors, such as obesity, diabetes, hypertension, and smoking. Again, these things would all be categoric, especially when your odds ratios are relatively low.
In addition, I noted that there was a significant variance in the animal study that suggested the endovascular change in the flexibility of blood vessels through nitric oxide synthase. The same authors had an article in press that showed in a prospective study that none of these changes were evident in humans.
So, again, it is more reassuring to us that PPIs, when taken appropriately, should be taken with confidence that there is not a conveyed myocardial risk.
A Resolution to the Story?
A fly in the ointment—or a resolution to the story—is the most recent study that was published in Open Heart.
The authors evaluated the incidence of cardiovascular events and gastrointestinal (GI) bleeding in patients receiving clopidogrel with and without PPIs. It was an updated meta-analysis. The background is the whole story of PPIs potentially causing harm as it relates to clopidogrel, its interaction with the cytochrome P450 2C19 pathway, and the delayed biotransformation of clopidogrel, which is necessary for activation to be an antiplatelet agent.
This group did a phenomenal job. They systematically searched through the Cochrane Central Register of Controlled Trials and used all the rules of engagement for systematic reviews and meta-analyses following the appropriate gold standards. They applied a funnel plot analysis to look for publication bias. They ended up with 39 studies with a total of about 215,000 patients, of whom 74,000 (34%) received a combination of clopidogrel and a PPI.
What did they surmise? In a pooled analysis, all-cause mortality (ie, myocardial infarction, stent thrombosis, and cerebrovascular events) was actually more common in patients receiving both drugs.
So do we stop the presses? Do we say, "Yes, this corroborates what the Stanford group had shown us and that we had criticized?" No, not at all, because they then looked at the tests of heterogeneity. They found that there was significant heterogeneity, meaning differences of methodologies, across these studies.
When they called those out and looked at randomized controlled trials and propensity match studies, they came down to eight of those studies. This culled it to about 24,000 patients. They looked at all of these endpoints, and they found no significant cardiovascular risk. In fact, as it relates to revascularization, there was a protective factor, a decreased odds ratio, if you used a PPI. No adverse association was found across all of the other cardiovascular endpoints (ie, all-cause mortality, myocardial infarction, stent occlusion, and cerebrovascular accidents).
Again, these things all came down to potential stratification bias. These patients all had comorbid risks, so it wasn't that the PPI was causing this. It was just being used in patients with comorbid risks.
Where Does This Leave Us?
What does this tell us? Clearly it says that the patients who had cardiovascular risks also had other risks for comorbid diseases, which was the trump card for why they developed all of these endpoints of adverse events.
When they looked at the GI bleeding risk, there was a risk reduction of 60%. Again, this is something that plays out like what we've seen in trials, like the COGENT trial in the New England Journal of Medicine, where clopidogrel and omeprazole were randomized against clopidogrel alone. We do know that there was a risk protection as it relates to the overall risk for GI bleeding. Clearly, this was reinforced in this most recent meta-analysis.
I think we can say that PPIs are not an issue as it relates to cardiovascular risk.
As an aside issue in this particular study, they looked at five PPIs. Dexlansoprazole was not part of their analysis; but for omeprazole, esomeprazole, rabeprazole, pantoprazole, and lansoprazole, there was equal risk when they looked at the unbalanced cardiovascular endpoints of adverse events associated with PPI use. Co-adjusted, there was no relative risk when they looked in the randomized controlled trials.
What this tells us is that the patients receiving PPIs clearly have a higher burden of comorbidities. Thus, they are more likely to have an increased burden for adverse cardiovascular events. PPIs, however, have a significant potential to decrease the upper GI bleeding in patients taking clopidogrel.
I think this new analysis tells us that if you would appropriately co-adjust the risk for indications for PPI use in these patients, it's not a cardiovascular risk—it's a GI risk. They need these medications for prevention of an upper GI bleed. The cardiovascular risk should not be a concern as it relates to all of these particular studies.
I think we've unveiled the mask of truth on this. The evidence supports that when used appropriately in patients with a relative risk for upper GI bleeding, clopidogrel and PPIs can be done safely. Patients can be reassured, and I think we, as prescribers, can be reassured as well.
Hopefully this sets the record straight. The mask is off. I think this really should end the conversation as it relates to relative risk in this population. Certainly, PPIs are appropriate in the relatively defined high-risk patients.
I'm Dr David Johnson. I hope this provides some useful information for the next time you counsel your patients. I look forward to chatting with you again soon.
