A recent report suggested an association between the use of proton pump inhibitors (PPI) and myocardial infarction (MI). This study has generated a great deal of attention in the lay press with coverage in The New York Times, Washington Post and many others.
What Did the Study Find and What Does it Mean?
The risk of MI on PPI was increased by 16%. While this sounds impressive, the low rate of MI would require treatment of 4000 patients with PPI to produce one additional MI. Even that finding is questionable. This is a data-mining pharmacovigilance study of 2 large electronic record databases. These studies take advantage of “big data” sets where a huge number of patients are thrown together and computer power is used to suggest an association. The term “suggest” is most important since studies of this type cannot prove causality and certainly do not define a mechanism.
An Example of the Potential Pitfalls of “Big Data”
The nature of this type study prevents matching. So the PPI patients included in the analysis may well have been sicker in some way or have other risk factors for MI leading to substantial stratification bias. Importantly, there have been several other large, prospective trials that have not demonstrated this connection. The authors speculate on a potential, yet unproven, mechanism. However, it is more likely that the PPI-treated patients were in some way different from the untreated patients. It is also possible that PPI therapy might allow patients to overeat, or eat less healthy foods, and might thereby increase risk factors for MI.
Practice Implications and Key Take-Aways for Patients
I would not change my practice based on this study, and believe the following messages should be emphasized to our colleagues and patients:
There is nothing from this study or any other to date that would push patients who truly need a PPI away from those medications.
- On the other hand, patients who do not really need a PPI (or any other drug) should not take them.
- The surgical and endoscopic therapies for GERD are appropriate for selected patients, but have a morbidity and mortality risk far in excess of risks quoted in this, or any other, study.
- Patients on PPI, and all others, should pay attention to diet, exercise, blood pressure, lipid profile and other activities that may actually alter their risk of MI.
Important Note about the PLOS One Model
Readers should also understand the publication model of PLOS One. PLOS One is peer-reviewed, but has an acceptance rate of 69%. According to the journal’s website, all papers that are technically sound will be published, and thus may not be subject to the same level of methodological scrutiny as papers published in other leading journals. PLOS One does not judge the importance of papers and leaves that to post-publication discussion, which as of today has been minimal for this article.
Kenneth R. DeVault, MD, FACG
About Dr. DeVault
Dr. DeVault is on the Board of Trustees and President-Elect of the American College of Gastroenterology. He is Professor of Medicine and Chair, Department of Medicine, at the Mayo Clinic in Jacksonville, FL.