Proton Pump Inhibitors and Dementia: Cause for Concern?

David A. Johnson, MD


February 26, 2016

This feature requires the newest version of Flash. You can download it here.

Hello. I'm Dr David Johnson, professor of medicine and chief of gastroenterology at Eastern Virginia Medical School in Norfolk, Virginia. The news that's just out is that proton pump inhibitors (PPIs) might increase the risk for dementia.

A tremendous amount of attention has been given to this, considering the prevalence of dementia and its increasing recognition across the world. It's estimated that 35 million patients have a diagnosis of dementia, and this is thought to increase by more than 2.5 times by the year 2040.[1,2] It also has a major economic impact, with approximately $600 billion in worldwide costs in 2010.[3]

Could medications be a cause of this? Well, the report that was just published in JAMA Neurology suggests that there is a weak association between PPI use and dementia.[4] Let's look at the data.

This study was based on the follow-up of a smaller cohort study that had been reported in 2015, suggesting a similar association with a hazard ratio (HR) of about 1.4 for PPI use and dementia.[5] In the present study of more than 200,000 patients, approximately 29,000 received a diagnosis of dementia in the 11 years of follow-up. Dementia was categorized just as dementia, and only 2.7% of patients actually had Alzheimer disease.

The scientific hypothesis was that PPIs may change the development of beta-amyloid plaques, and that there is some potential, at least shown in mice, that the PPIs may alter the beta-secretase or gamma-secretase that lays down these plaques in the brain.[6] As you may or may not know, there is some hypothesis and demonstration that these amyloid plaques are seen in the Alzheimer variant of dementia. Now recalling how dementia was categorized in this study and that only 2.7% of patients had Alzheimer disease, it's important to put this percentage into perspective, given this hypothesis that PPIs may change the development of amyloid plaques.

There was a nominal difference between men and women, with an HR of 1.52 for men and an HR of 1.42 for women for the association between PPI use and dementia.

The authors adjusted for some variables, but certainly not all confounders. They adjusted for age, gender, polypharmacy, stroke history, depression, ischemic heart disease, and diabetes.

They didn't adjust for some very important risk factors for dementia, including alcohol use, family history of dementia, and hypertension. These are well-known risk factors for dementia, and for some reason, these weren't included in the analysis. Thus, we're missing a tremendous amount of potential risk factors that might account for the imbalance of incident dementia between PPI users and nonusers.

How Should We Interpret These Data?

Where do we go with this information?

We need to recognize that "dementia" was not well qualified. We have no way to know whether it was true dementia. Was it just cognitive impairment? There clearly are psychomotor and cognitive assessments that can differentiate cognitive impairment without dementia. We don't have any of that information.

As in a number of these recent pharmacovigilant studies, we've seen a nominal odds ratio association suggesting associated with a medication; in this case, it's a PPI. Interestingly, the same challenge was made with histamine-2 (H2) receptor antagonists. A study reported that H2 antagonist users had about 2.4 greater odds of cognitive impairment compared with nonusers.[7] This one is 1.4 times. In another study, H2 antagonist use was not associated with all-cause dementia or Alzheimer disease.[8]

There was also no risk-profile adjustment for dose. If you were going to report that this was a dose/response effect, you'd like to see some effect of higher doses or over longer periods of time.

In my mind, there are a lot of potential biases owing to the lack of statistical adjustment for key risk factors for dementia. We also don't have any knowledge about the qualitative assessments for the diagnosis of dementia in these patients, but we do know that there are mechanistic ways to do this using validated instruments, none of which were used in this particular study.

What Should We Tell Our Patients?

I think this study provides another healthy point of discussion with your patients, by evaluating whether they really need a PPI. If they need a PPI, they should take it. Certainly, there are benefits of PPIs, including treatment of severe reflux disease and prevention of bleeding and nonsteroidal anti-inflammatory drug (NSAID)-induced gastropathy. All of these benefits need to be kept in perspective. But if patients don't need the medication, this is a healthy platform to at least begin the discussion. Not because we really think they're going to develop dementia, but they just don't need the medication.

Keep things in perspective. There are a lot of issues with these data. Some of these pharmacovigilance studies with PPIs have created a lot of undue angst. When carefully analyzed, they really are not anything more than a little bit of noise right now. In my mind (no pun intended, and I do take a PPI), I'm very confident that right now you don't need to change therapies on the basis of concern about dementia.

I hope this puts this study in perspective when you discuss it with your patients. I'm Dr David Johnson. Thanks again for listening.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.